1 MECHANISTIC LINKAGE S BETWEEN WATER SATURATION AND ORGANO METAL INTERACTIONS IN SANDY COASTAL PLAIN SOIL S By CHUMKI BANIK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2 Â© 2014 Chumki Banik
3 To my pa rents and my husband
4 ACKNOWLEDGMENTS It is an immense gratification to me that I am getting the opportunity to thank everyone, whoever is being with me during this long and hard journey and make this an easy and wonderful one. The first person I would thank from bottom of my heart is my advisor, Dr. Willie G. Harris. He is one of the best human being I have ever seen. He guided me with his scientific vision, power of imagination, intellectual thoughts and huge experience. He always had faith on my abilities and that was the strength to complete my PhD work. I am thankful to him for his advice and supp ort outside my academic world to make me a better person. I would also like to express my sincere thanks to my advisory committee members: Dr. Vimala D. Nair, Dr. Andy Ogram, Dr. Matt Cohen for their valuable advice and assistance to achieve this academi c goal. Our lab assistant Mr. Kafui Awuma deserves special thanks from me for all his encouragement. I deeply appreciate his generous help and assistance during my research. I am deeply indebted to my husband Dr. Santanu Bakshi for the support and sincere faith which gave me the strength and patience to finish my PhD work. I thank all my professors and friends in USA and India for supporting me during the crisis period of my life. I would never be able to be here or achieve anything in life without the su pport of my parents, my in laws and my other family members. I would love to dedicate my research to my beloved parents and husband who helped me to understand my responsibilities.
5 Having faith on the almighty is a wonderful thing. I am grateful to him for his blessings and providing me the mental strength I required.
6 T ABLE OF CONTENTS Page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 Observations ................................ ................................ ................................ ........... 15 Overarching Question ................................ ................................ ............................. 16 Hypotheses ................................ ................................ ................................ ............. 1 6 Research Objectives ................................ ................................ ............................... 17 2 LITERATURE REVIEW ................................ ................................ .......................... 19 Distribution and Theories of Spodosol Formation ................................ ................... 19 Processes of Podzolization ................................ ................................ ..................... 21 Hydrologic Factors in Podzolization ................................ ................................ ........ 24 Minerals Associated with Spodosols ................................ ................................ ....... 25 Stripping Process to form Eluvial Horizon ................................ ............................... 25 Forms of Aluminum Complexed with Spodic Carbon ................................ .............. 27 Metal Complexing Low Molecular Weight Organic Acids ................................ ........ 28 Microbial Effects on Carbon Fluxes ................................ ................................ ........ 29 3 RELATIONS OF IRON, ALUMINUM, AND CARBON ALONG TRANSITIONS FROM UDULTS TO AQUODS ................................ ................................ ............... 31 Materials and Methods ................................ ................................ ............................ 32 Site Selec tion ................................ ................................ ................................ ... 32 Well Construction and Deployment ................................ ................................ .. 33 Soil Description, Sampling and Analysis ................................ .......................... 34 Statistical Analysis ................................ ................................ ............................ 35 Results and Discussion ................................ ................................ ........................... 36 Water Table in Relation to Morphology, Classification, and pH ........................ 36 Relations of C, Al, and Fe along the Drainage Gradient ................................ ... 37 Mechanistic Implcations ................................ ................................ ................... 39 Summary ................................ ................................ ................................ ................ 40
7 4 CARBON AND METAL RESPONSES TO CONTROLLED WATER TABLE FLUCTUATIONS IN SANDY SOIL MATERIAL ................................ ...................... 57 Materials and Methods ................................ ................................ ............................ 59 Site and Soil Selection ................................ ................................ ..................... 59 Sample C haracterization ................................ ................................ .................. 60 Column Experimental Design ................................ ................................ ........... 60 Statistical Analysis ................................ ................................ ............................ 63 Results and Discussion ................................ ................................ ........................... 63 Sample Pro perties ................................ ................................ ............................ 63 Depth of Water Table in Relation to CO 2 Flux ................................ .................. 63 Water Table Depth and Soil Pore Water Characteristics ................................ .. 64 Possible Decomposition Pathways ................................ ................................ ... 65 Treatment Effects on Soil C, Metals and Finer (<50 Âµm) Material Distributi on ................................ ................................ ................................ .... 66 Summary ................................ ................................ ................................ ................ 66 5 FACTORS RELATED TO ACID INDUCED IRON AND ALUMINUM RELEASE FROM SANDY COASTAL PLAIN SOILS ................................ ............................... 79 Materials and Methods ................................ ................................ ............................ 80 Sites and Soils ................................ ................................ ................................ .. 80 Basic Soil Characterization ................................ ................................ ............... 81 Extractions ................................ ................................ ................................ ........ 82 Chemical extractants ................................ ................................ ................. 82 Natural extractants ................................ ................................ ..................... 82 Extracting Procedure ................................ ................................ ........................ 83 Statistical Analysis ................................ ................................ ............................ 83 Results and Discussion ................................ ................................ ........................... 83 Metal Extr action ................................ ................................ ................................ 83 DOC Adsorption and Metals Release ................................ ............................... 84 Interaction Time Fe Reduction ................................ ................................ ...... 84 Possible Adsorption Release Mechanism ................................ ....................... 85 Summary ................................ ................................ ................................ ................ 86 6 SUMMARY AND CONCLUSIONS ................................ ................................ .......... 96 LIST OF REFERENCES ................................ ................................ ............................. 104 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 111
8 LIST OF TABLES Table page 3 1 Examples of vegetative, soil descriptive, and metals data for Udult and Aquod profiles along transects 1 (LFR location) and 3 (ACMF location). ........... 42 3 2 Correlation coefficients and P values for relations be tween upper Bh horizon properties and mean water table depth (MWT), lowest (deepest) water table depth (LWT) and highest (shallowest) water table depth (HWT). ....................... 44 3 3 Comparison for selected properties between weak and well expressed upper Bh horizons. ................................ ................................ ................................ ........ 45 3 4 Comparison for selected properties between A horizons and weak Bh horizons of driest side. ................................ ................................ ........................ 46 4 1 Soil profiles from which A and E h orizon materials were collected for this study, along with selected soil physica l and chemical characteristics. . .............. 68 4 2 Comparison o f <50 Âµm material percentages before and after treatments, ................................ ................................ ..... 69 5 1 Description of soils and their physicochemical characteristics . ........................... 87 5 2 Chemical characterization data for extractants used ................................ .......... 88 5 3 Data for Fe and Al extracted by 1 mmol HCl, 1 mmol oxalic acid (ox), and soil derived dissolved organic matter (DOM) ................................ ...................... 89
9 LIST OF FIGURES Figure page 3 1 Locations of transects where wells were i nstalled and soil profiles were sampled. ................................ ................................ ................................ ............. 47 3 2 Relation between depth to water table and distance between wells for each transect ................................ ................................ ................................ ............... 48 3 3 Transect 1 as an example of the drainage gradient and morphological trends that were common to all transects ................................ ................................ ...... 49 3 4 Photograph of a trench at Austin Cary Memorial Forest, 1.2 km S of transect 3, showing upward fading of Bh from wetter (back ground) to drier (foreground) part of the landscape ................................ ................................ ..... 50 3 5 Comparison of well expressed (left) and weakly expressed (right) Bh hori zons along transect 4 (wells 4 4 and 4 1, respectively), as displayed at proportionate depth in a tray from auger excavation ................................ .......... 51 3 6 Relation between depth to upper Bh (n = 48) boundary and A) acid ammonium oxalate extractable Fe (Fe aao ) and B)Al (Al aao ). ................................ 52 3 7 Comparison between pyrophosphate extractable Fe (Fe p ) and Al (Al p ) for eight weakly expressed Bh and fourteen strongly express upper Bh ................. 53 3 8 Comparison between A (n = 25) and Bh (n = 48) horizons with respect to relations between total C (C t ) and the variables pyrophosphate C (C p ) and pyrophosphate Al (Al p ). ................................ ................................ ...................... 54 3 9 Comparison between mean ratios of pyrophosphate extractable Al and C (Al p/Ct) for A versus Bh horizons, by transect. ................................ ....................... 55 3 10 Evidence for C , metals and clay accumulation in Bh with respect to E for each transect.. ................................ ................................ ................................ .... 56 4 1 Schematic diagram of a landscape ecological transition from better drained sandhill (left) to poorly drained flatwoods (right), showing a change in vegetation, water table position and an upward fading Bh.. ............................... 70 4 2 Comparison between treatments of soil C loss as CO 2 , showing A) average loss per specified time interval for all columns within treatments and B) average loss by each set of columns over the 45 days of incubation.. ............... 71 4 3 Cumulative C loss as CO 2 over the 45 days of incubation, by treatment effect.. ................................ ................................ ................................ ................. 72
10 4 4 Averag e C loss per specified time interval for all columns within treatments over a 45 day study period. ................................ ................................ ................ 73 4 5 Comparison between treatments (SWT = shallow water table; DWT = deep water table) of average Fe release from columns, based on 18 days of leachate collection. ................................ ................................ ............................. 74 4 6 Comparison between C released as DOC into soil solution under treatments of SWT (shallow water table) and DWT (deep water table). ............................... 75 4 7 Relations between C and metal released into pore water, by SWT (shallow water table) and DWT (deep water table) treatments. ................................ ........ 76 4 8 Relations between Fe and Al released into pore water soil solution, by treatments SWT (shallow water table) and DWT (deep water table) .................. 77 4 9 Comparison between treatments effect on total C loss from each set of columns ................................ ................................ ................................ .............. 78 5 1 Locations of s amples collected for the study in Florida, USA. Bronson, NATL, and Ordway samples were in Levy, Alachua, and Putnam Counties, respectively. ................................ ................................ ................................ ........ 90 5 2 Relation between Fe and Al extracted by 1mmol oxalic acid, by locations. Excluding t he circled data point R 2 = 0.73, and P < 0.0001. ............................... 91 5 3 Relation between DOC adsorption and non crystalline Al and crystalline Fe concentrations durin g 1 h extraction by sample site ................................ ........... 92 5 4 DOM extractable Al and Fe, as released over time ................................ ............ 93 5 5 Relation between DOM extractable Fe and Al re leased after7 d incubation of all samples Ordway samples and Bronson NATL samples. ............................... 94 5 6 Relation between acid ammonium oxalate (Al aao ) and DOM extractable Al (Al DOM7 ) after 7 d incubation. ................................ ................................ ............. 95 6 1 Schematic example of a landscape ecological transition from better drained sandhill (left) to poorly drained flatwoods (right), showing a soil Fe depletion gradient, water table position and a downward strengthening Bh .................... 103
11 LIST OF ABBREVIATIONS aao Acid ammonium oxalate cdb Citrate dithionate bicarbonate DOC Dissolve organic carbon DOM Dissolve organic matter HWT Highest water table LWT Lowest water table LMW Low molecular weight MWT Mean water table OC Organic carbon OM Organic matter p Pyrophosphate (e.g., Alp : pyrophosphate extractable Al) SWT Shallow water table WT Water table
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MECHANISTIC LINKAGE BETWEEN WATER SATURATION AND ORGANO METAL INTERACTIONS IN SANDY COASTAL PLAIN SOILS By Chumki Banik August 2014 Chair: Willie G. Harris Major: Soil and Water Science Southern Spodosols occur on poorly landscapes. Their Bh horizons constitute a globally significant pool of subsurface carbon (C) associated with metals but the role of hydrology in their formation is uncertain. This research addressed relations between water table fluctuation and fate of C, iron (Fe) and alumi num (Al). Carbon, Al, and Fe distributions were determined along four monitored drainage transitions from better drained Paleudults to poorly drained Alaquods . The Bh horizons faded in expression and became shallower with diminishing frequency and duratio n of near surface saturation. Iron declined from weak Bh of drier soil to well expressed Bh of Alaquods. Weak, shallow Bh horizons were distinguished from A horizons by having higher Al/C molar ratios. Carbon atmospheric flux and solution metal C relation s were compared under controlled shallow and deep fluctuating water tables . Net C loss was less for shallow fluctuation beca use of lower CO 2 evolution . However , shallow fluctuation showed greater vertical movement of dissolved organic C (DOC) and more C m etal interactions. Magnitude and direction of C flux were influenced by depth of water table fluctuation.
13 An inverse relation between Fe and Bh expression suggested that Fe could inhibit podzolic organo Al interactions. A batch study of 18 E horizons was conducted to evaluate Fe effects on Al release by organic and inorganic acids. Releases of Fe and Al to soil DOC were delayed and correlated. Organic acid extraction efficacy was much greater than that of hydrochloric acid , confirming significance of organo complexation. Overall results elucidate highly specific hydrologic control of C and metal dynamics in Alaquods, and implicate Fe as a hydrologi c linkage. A broader implication is that climate driven hydrologic changes can affect soil C dynamics at a global scale, at depths much greater than surface horizons.
14 CHAPTER 1 INTRODUCTION Spodosols are soils with black to brownish or reddish illuvial s ubsurface horizons (Bh and Bhs; spodic diagnostic horizons) that are rich in organic C (McKeague et al., 1983; Petersen, 1984; Schaetzl and Harris, 2011) . These horizons can be quite thick (Daniels et al., 1975) and constitute a significant pool of global C (Stone et al., 1993) . They commonly underlie strongly expressed eluvial (E) horizons that have been stripped of metals. Different suborders of Spodosols (Aquods, Orthods, Humods, and Cryods ) are found in many places on earth under environmental conditions conducive to their formation. These conditions generally include coarse textured parent material; cool, moist climate; and coniferous vegetation. Spodosols occur extensively in the lower co with which they are associated elsewhere. These coastal plain Spodosols (mostly Aquods) are strongly associated with zones of fluctuating water tables (Tan et al., 19 99) whereas northern Spodosols are commonly well drained and presumed to form via mechanisms (discussed below) not involving water table influences . The proposed research focuses on mechanisms that would help to explain this water table linkage. It is p ertinent to processes of C sequestration in soils as related to hydrologic conditions. Organo metal complexes are thought to be important ingredients in Spodosol formation (DeConinck, 1980; McKeague et al., 1983) . These complexes form a black to brown or reddish horizon within or sometimes deeper tha n 2m from the soil surface. DeConinck (1980) proposed that metals are mobilized via complexation with relatively hydrophilic organic acids near the soil surface (E horizon f ormation) and are subsequently immobilized as a consequence of increased metal loading (spodic horizon
15 formation). Metal loading, by his theory, promotes immobilization of the complex by reduction of surface charge, increase of mass (van der Waals effect), and metal bridging. The mechanism proposed by DeConinck is consistent with the morphology and chemistry of Florida Aquods but it provides no explanation for the linkage with flu ctuating water table. One study showed that artificial spodic horizon forma tion a near surface water table be maintained for some minimal time (Harris et al., 1995) . The process involved dissolution of sand gr ain coatings cements by oxalic acid (a complexing acid surrogate used in the experiments), subsequent release and colloidal mobilization of all coating materials, and accumulation of these materials in Bh like (artificial) layers that darkened over time. R esults of that study still did not resolve the question as to why a fluctuating water table is necessary for the Al dissolution and organo Al complex formation. Observations 1. Spodosol occurrence is associated with fluctuating water tables that are seasonall y shallow in the SE coastal plain of the US but in northern climates Spodosols commonly occur on well to excessively drained landscapes. 2. Spodosol formation involves mostly mobilization via organo Al complexation in the SE USA, but Fe concentrations are hi gher in Spodosols of northern latitudes . That is likely because well drained northern Spodosols have not been depleted of Fe via redox effects. 3. The Bh horizon of Spodosols have lower Fe concentration than subsurface horizons of be tter drained associated so ils. Al uminum mobilization does NOT significantly occur in better drained sandy soils of the SE USA but DOES occur in well drained northern soils in Spodosol formation.
16 Overarching Question How is seasonal saturation promoting Al mobilization in the formation of southern Spodosols? Hypotheses The research hypotheses are as give below: 1. Removal of Fe by near surface saturation induced reducing conditions fosters mobilization and redistribution of Al that serves as a binding agent for sand grain coatings. 2. Mobilization of Al via complexation with organic acids results in redistribution of sand grain coatings bound by Al prior to Al complexation. Mobilization and depletion of Fe via chemical redu ction under conditions of near surface water saturation leaves behind only Al as an agent to bind grain coatings. Al uminum must bear increasing pressure of complexation as Fe becomes more thoroughly removed toward the wetter part of the landscape. Hence, Bh horizons will have a deeper upper boundary and be more prominent and thick in poorly drained flatwoods soils with fluctuating water tables because these hydrological conditions foster Fe removal. 3. Near surface saturation favors a higher rate of dissolv ed organic carbon (DOC) production relative to C mineralization, resulting in greater organo Al complexation. A reduced rate of C mineralization under shallow water table conditions could result from saturation and anaerobic conditions that favor C dissol ution as organic species (DOC) over its oxidative gaseous loss as CO 2 . Thus, saturation might indirectly favor Al mobilization via organic acid complexation. In the case of deep water table conditions (lower moisture content near the surface) , more C could be lost to the
17 atmosphere via mineralization or microbial respiration and less entrained as DOC in the soil solution. 4. Vertical movement of C is greater under shallow than under deep fluctuating water table conditions because there is less DOC production and mobilization for deep water tables. The combination of shallow water table and presence of complete or partially decomposed organic matter produce more DOC. This DOC moves vertically with downward moving water table and is trapped deeper into the soil profile. As the water table gets deeper more C is lost by mineralization and hence there is less chance of DOC contributing to Bh formation. 5. Total C loss from sandy soils , not accounting for primary production, is greater under conditions of deep tha n sha llow water table fluctuation . This is based on the expectation of higher loss as CO 2 due to faster mineralization under aer obic than anaerobic conditions. 6. Aluminum release into soil solution inversely relates to crystalline Fe content because surfaces of these oxides mitigate organo Al complexing potential. 7. Iron release, as by biochemical reduction, will be accompanied by Al release due to loss of Fe oxide surfaces that can otherwise compete with Al oxides for ligand reactions. Hypotheses 1 and 2 were tested with a field study ; hypothesis 3, 4 and 5 with a laboratory based column study ; and hypothesis 6 and 7 with a batch study. Research Objectives 1. Determine distributional relations between C, Al, and Fe in A and Bh horizons in soil s along a hydrologic transition and evaluate their mechanistic implications. (Chapter 3) 2. Monitor change in carbon fluxes under different fluctuating water table depths in laboratory based column study. (Chapter 4) 3. Evaluate C metal relations in leachates and from soil surface from columns with different water table treatments. (Chapter 4)
18 4. Compare proportion of Al extracted by different extractant s from soils with different crystalline Fe concentrations. (Chapter 5)
19 CHAPTER 2 LITERATURE REVIEW The process of Spodosol formation entails a combination of different soil biogeochemical processes. To provide background for exploring the underlying mechanisms explaining the restriction of southern Spodosols to landscapes of shallow water table fluctuations, this literature review focu ses mainly on processes and theories related to podzolization (Spodosol formation ). However, r esearch articles pertaining to hydrologic effects on soil physic al and biogeochemical processes are also reviewed since soil hydrology is a variable of interest f or research conducted in this dissertation. Carbon is one of the major element s involved in Spod osol formation . Hence research findings related to C dynamics as influenced by hydrology related microbial processes are pertinent to the understanding of Spodosol formation. F ormation of low molecular weight ( LMW ) acids is related to microbial processes (as described in the literature) and these acids are powerful metal (Fe and Al) chelating agents that contribut e to Spodic horizon formation. Distribution a nd Theories of Spodosol Formation Spodosols are generally formed in cool climate s , and a t sites with sandy parent material. Spodosol formation has been thoug ht to involve formation and eluviation of organo metallic complexes (Wright and Schnitzer, 1963) , in some cases accompanied by eluviation of silt and clay (Guillet et al., 1975; Harris and Hollien, 1999; Lin et al., 2002; Ugolini et al., 1977) . Wright and Foss . , ( 1968) s tudied Spodosols for many years and proposed various mechanisms to explain their occurrence. Spodosols are significant C sink and hence the processes that control their formation are pertinent to global C dynamics (Stone et al., 1993) . For example, Florida Spodosols to a depth of 2
20 m have been estimated to contain about 0.025 to 0.05% of the global organic carbon (OC) pool (Stone et al., 1993) . Di stribution of Spodosols is worldwide; they can be found in Africa, Asia, Australia, Europe, New Zealand, South America, and other regions of North America (McKeague et al., 1983; Petersen, 1984; Schaetzl and Harris, 2011) . In the coastal plain of the SE US they are mainly associated with fluctuating water tables and classified as Aquods (Brasfield et al., 1973; Collins, 1990) . They occur predominantly in flatwoods ecosystem s which are characterized by pine forests and nearly level topography. Orthods also occur in the SE US coastal plain; Orthods are better drained than Aquods b ut are likely still influenced by fluctuating water table (Harris, 2001) . Many studies have shown that the major components involved in the process are organic C, Al, and Fe (Lundstrom et al., 2000; McKeague et al., 1983; Petersen, 1976; Petersen, 1984) . Howev er, some studies have shown that Al and Fe can be transported as inorganic short range order silicic acid complexes (Farm er et al., 1980) . This inorganic variation of Spodosol genesis, known as imogolite proto imogolite theory, involves the complextion of Al with silicic acid at pH<5, to form hydroxy alumin um orthosilicate which is transported downward to ultimately combi ne with eluviated organic C to form spodic horizons. Another theory of Spodosol genesis combines the processes of complexation of polyvalent cations, mostly Fe and Al, and their downward transport as mobile complexes to ultimately become immobilized in the formation of the spodic horizon (DeConinck, 1980) . This immobilization is theoriz ed to occur as follows. Organic acids bind Al/Fe and help in the process of mobilization, migration and deposition
21 respectively, depending upon their charge satisfaction. The hydrophilic groups ( COOH , OH , NH 2 ) in the organic compounds partially dissociate and their negative charges get neutralized by the positively charged metal ions. This interaction results in the formation of organo metallic complexes. If the diffuse double layer (Olphen, 1963) is thick then the discrete organo metal complexes repel each other and remain mobile. However, with increasing sorption (loading) of metal ions the double layers of organo metal collo ids or species get compressed resulting in their flocculation or precipitation and their accumulation to form a spodic horizon. Morphological and chemical characteristics both are important for the identification of spodic horizons according to th e DeConin ck theory. Organic substances play a major role. Processes of Podzolization Podzolization can be described as a pedological redistribution of components within soil, resulting in depletion of metals in one zone and enrichment of metals and C in an underlyi ng horizon (Franzmeier et al., 1965) . The morphology of many Spodosols dramatically reflects this redistribution. White to light gray E (albic) horizons represent the zone of depletion; the zone of accumulation (illuviation) is expressed by the darker colored B (spodic) horizon. To indicate the prevalence of C (Bh), metals (Bs), or both (Bhs), the spodic B horizon is appended with appropriate subscripts. The morphology and components of flatwood Spod osols are generally consistent with the formation via classical concepts. A study showed that there is a direct relationship between organic acids released and Al /Fe dissolution in soil and that about 1 6% of the DOC is comprised of low molecular weight o rganic acids (Pohlman and McColl, 1988) . Hence, organic acids can affect metal solubility and speciation as they form stable organo metal complexes.
22 Spodic horizons can be distinguished from buried A horizons by morphological features, the nature of organic matter present, and selective dissolution results (Yuan, 1966) . Thickness of Bh horizon depend upon the water movement (Daniels et al., 1975) and Bh horizons are thicker when water movement is mainly vertical. Organic matter along with metals may be carried downward by moving water. From these works it can be inferred that po dzolization is a biogeochemical process. Vegetation and drainage (topography and microclimate) are major factors of Aquod formation. Typical plant species in Aquods/flatwoods ecosystems include longleaf and slash pine ( Pinus palustrus and elliottii ), saw p almetto ( Seronoa repens ), gallberry ( Ilex gabra ) and various herbaceous and grass like plants (Abrahamson and Hartnett, 1990) . The Bh horizon of Spodosols have lower Fe concentration th an subsurface horizons of better drained associated soils (Harris et al., 1995) . The strong empirical linkage between podzolization and shallow fluctuating water table distinguishes southern from northern Spodosols since the latter most commonly occur on well drained landscapes. The role of the water table in the formation of Florida Aquods is not completely understood. This proposed research would test ideas related to how near surface saturation may foster podzolization in southern Spodosols. The ratio of organic C to metals is high in spodic horizons of Florida as revealed by various selective dissolution procedures (Yuan, 1966) . Hence they are designated s show inverse depth trends for Al and C, i. e, with typically an increase in Al with depth but with the greater C concentration being at the top of the Bh (Bardy et al., 2007) . Shallow spodic horizons are exceptions of this trend, and could be explained by morphological and hydrological anomalies. Also
23 bisequel flatwoods Spodosols, with d eeper Bh horizons, can show an increase in C and decrease in Al with depth (Harris, 2001) . Ultrafiltration and size exclusion chrom atography were used (Van Hees et al., 2000) to determine the fraction of Al in soil solution bound to low molecular weight organic compounds (Al LMW ). A positive r elation was found between the quantity of organic acid and Al LMW . In this case Al LMW was most abundant in the O i horizon and declined with depth. The quantity of Al LMW consistently declined/increased with decreasing/increasing levels of low molecular wei ght organic acids (Van Hees et al., 2000) . A distinguishing feature of podzolization is the lightening of the E horizon result from exposure of sand gain surfaces by the stripping effect of organic acids. Quartzos sandy paren t materials of SE US Spodosols are nearly devoid of weatherable primary minerals but secondary minerals persist as sand grain coatings which are subject to acid stripping (Brasfield et al., 1973) . Aquods have diverse age; some are much older than glacially influenced landscapes (Holzhey et al., 1975) . Aquods formation of the SE US involves mostly mobilization of Al via organo Al complexation since Fe is consistently low in concentration based on Florida soil characterization data ( http://flsoils.ifas.ufl.edu/ ) . This is another distinction between southern and northern Spodosols since Fe is commonly more abundant in the latter (Schaetzl and Harri s, 2011) . That is likely because well drained northern Spodosols have not been depleted of Fe via redox effect and can have appreciable amount of crystalline and noncrystalline forms of Fe (Franzmei er et al., 1965) . This lack of Fe in Aquods may indicate a redox related predisposition of soils to form Aquods by virtue of Fe depletion (Harris et al., 1995) . During the podzolization process the agents that binds the clay and silt size d mineral particles to sand grains
24 are apparently dissolved (i.e. by organic acids), resulting in stripping of grains and formation of a light colored E horizon (Harris, 2001) . These coatings in Florida sandy soil horizons are commonly dominated by hydroxyl interlayered vermiculite , kaolinite, gibbsite, and quartz (Harris et al., 1987a) . Hydrologic Factors in Podzolization Specific hydrologic conditions such as frequency and duration of saturation are associated with podzolization (Tan et al., 1999) ; these researchers compared the hydrologic responses of h orizons to rainfall along with Aquod Udult transition. Piezometric heads were consistently higher in Aquods than in Udults. In Udults the period of saturation was less in comparison to Aquods after a short period of rain. Flatwoods landscapes may appear fl at to a casual observer, but subtle topographic changes are discernible by the distinct difference in hydrology, vegetation, and soils. Water table fluctuations can be large in some areas of flatwoods. In dry periods of summer, the water table can drop to a greater relative extent than in depressions. Though the selected site was at an intermediate relative elevation on the landscape, and the gradient was minimal, the upper part of the Udult a rgillic horizon did not have any redoximorphic features but that part of the Aquods a rgillic did have redox concentration s and depletion s . Hence, the absence of a podzolic E Bh horizon sequence in the Udult may relate to its better drained condition . Harris et al., ( 1987a) inferred that the water table is likely a factor behind the redistribution of phyllosilicates in Alaquods. Their study of clay mineralogy of coastal plain soils with sandy epipedons was based on X ray diffraction peak height ratios. They found mineralogica l consistency of coated sand materials regardless of sand and silt content but that the mineralogy of stripped sands differed markedly from those with coatings.
25 Minerals Associated with Spodosols All coated /slightly coated sand materials were dominated b y hydroxy interlayered minerals (HIM), gibbsite, kaolinite, and quartz while stripped sands contained large amounts of quartz, variable smectite, little or no HIM, and no gibbsite (Harris et al., 1987a) . Similar m ineralogical distinctions were made in a comparison between a Quartzipsamment and Alaquod soil profile (Harris and Hollien, 2000) . This redistribution of HIM is not only seen in Alaquods via podzolization but also reducing environment in the deeper gleyed horizons with loamy textures (Btg and Cg) may experience a similar coating destabilization effect. M ineralogical differences were shown between clay coatings separated from Bh horizon sand grains by ultrasound and the remainder of the s oil matrix, with the coatings being enriched in hydroxyl interlayerd minerals (HIM) relative to the clay in the matrix (Harris et al., 1987b) . According to the authors, partitioning of fines across the E/Bh boundary can be explained by podzolization. Removal of Fe and Al cementing agents from the E horizon in the podzolization process would strip any coatings and liberate the incorporated minerals. Some of the minerals liberated and not degra ded by the acidic leachate may eluviate along with chelates and become entrapped within organo metal coatings that form on sand grains in the Bh Stripping Process to form Eluvial Horizon The stripping of grain coatings during artificial formatio by microscopic analysis (Harris et al., 1995) . Accumulation of these materials in an for podzolization. An initial hypothesis of the study was that Aquod like features would
26 form in the sands taken from the toeslopes but not in those taken from summits because the grain coatings of the toeslope sands would be predisposed for destabilization due to redox depl etion of Fe oxides which help to bind coating components to grains. The chemical reduction of Fe was supported by the presence of (Harris et al., 1995) . Th is study verified that a water table was necessary, in addition to predisposition of sand, but the experiment provided no specific means to explain the water table effect on Al mobilization . The presence of organic acids in soil causes the lowering of pH and as a consequence Fe/Al dissolution may increase in the presence of those organic acids. So both biological and chemical pressures are required for the dissolution of inorganic components for Spodosol formation. The dominance of quartzose sand (devoid of weatherable minerals) as the parent material of southern Aquods means that the g r ain coatings may be a major source of metals in the podzolization process for these soils. During organo metal complexation, the binding agents Fe or Al become less efficie nt in holding the other coating materials. This results in downward colloidal translocation of significant amounts of the fine mineral coating materials and their accumulation to form a finer texured zone (Harris and Hollien, 1999) . This zone can retain more water than the overlying stripped zone, thus organo metal complexes get sufficient time and encounter favorable conditions for precipitation or adsorption as the water table recedes . In effect, the water retention characteristics of the Bh is a relevant factor in that this fine matrix may act as a template for Bh development. This would be a physical explanation for the typical abrupt upper boundary of the Bh even when the boundary
27 topography is wavy or convoluted . Convolution of upper illuv ial boundaries can be arise from the legacy effects of root growth and other soil disturbances on water flow and colloidal migration (Schaetzl and Harris, 2011) . Sharp color chang es across the E Bh boundary in artificially formed Aquods were found to be related to mineralogical differences between the two horizons (Harris and Hollien, 2000) . Quartz was the dominant mineral in the white E, where ph y llosilicates we re barely detectable. However, the subjacent dark colored Bh contained appreciable amounts of phyllosilicates. The accumulation of phyllosilicates was due to the redistribution of minerals from E to Bh. Carbon also accumul ated in the Bh. The boundary betwe en E and Bh was abrupt even when topography was irregular. The Bh horizon was darkest at its upper boundary. Sand grains were stripped above, but not below, the E Bh boundary. Forms of Aluminum Complexed with Spodic Carbon Dissolution and complexation reaction s of Al in Spodosols are dependent on the concentrations of organic acids present in the soil system, with a threshold ratio of Al/C in pyrophosphate extract (Al p /C p ) of less than about 0.1 according to Tipping et al., ( 1995) . They found that the presence of a reactive Al(OH) 3 phase regulates the activity of Al +3 in Spodosols at pH >4.1 but this Al phase decreases with increasing str ong acid concentrations at pH < 4.1. From their results they concluded that organo Al complexation might be controlling the Al +3 activities depends on a pH . They proposed a model describing various pools of Al contributing Al +3 solubility. The model reveal s soil organic matter to be the only rapidly reacting solid phase controlling Al +3 solubility. Their data indicated that simple cation exchange did not explain solution activity of Al +3 . An
28 e xperiment by Coelho et al . , ( 2010) revealed that formation and maintenance of low crystallinity Al minerals was inhibited by Al sorption to soil organic matter. Metal Complexing L ow Molecular Weight Organic Acids Decomposition of litter and dead roots by fungi and bacteria produces low molecular weight organic acids, having c oncentration in the range of 1 1000 ÂµM (Fox and Comerford, 1990) . These acids form strong complexes with Al and promote its eluviation. A range of low molecular weight acids ha s been found in soils, to include oxalic, citric, formic, acetic, malic, succinic, malonic, maleic, lactic, aconitic, fumeric, gallic, vanillic, benzoic, and p hydroxybenzoic acid. However, the type of organic acids varies with vegetation and climate. Oxal ic (Fox and Comerford, 1990) and citric acids (Wickman , 1996) have been found to be prevalent in soils. Formic acid is found in anaerobic systems as a decomposition product of organic matter (Stevenson, 1967) ; fungi and bacteria synthesize this acid (Hodgkinson, 1977) . Formic acid is produced in stoichiometric quantity through oxalate decarbox y lation by the decarboxylase enzyme. Several species of wood decomposing fungi produce this enzyme, in acid environments (Hodgkinson, 1977) , with optimum production occurring between pH 3.0 and 5.2. Oxalic, formic, and citric acids can all form complexes with metals but the stability of these comp lexes differs. Oxalic and citric acid form more stable complexes than formic acid. The stability of these organo metal complexes relates to equilibrium constants and kinetics of their reactions of formation. Rates of formation can be important in comparing the acids with respect to like ly occurrence in soils. In one kinetics study of HCOOAl +2 (Knoche and LÃ³pez Quintela, 1983) , it was found that the rate of formation of this complex is related to the electrical conductivity o f the solution. Overall conductivity is influenced by the stability of the inner sphere
29 complex. When the Al ion forms a complex with formic acid, a proton gets releas ed with the formation of alumin um formate, and due to the high mobility of protons conduc tivity of the solution increases. Carbon mineralization rate from low molecular weight organic acids like citric acid can drop significantly by organo metal complex formation or by their adsorption on metal surfaces as the metal surfaces have a protective effect . Boudot . , ( 1992) p roposed bound in insoluble organo Al compl e xes >>alophanes > imogolite . The rate of mineralization of citric acid also depends on nutrient releasing capacity of the soil as found by (Boudot et al., 1986) , where high P retention capacity of a high Al containing Andosol cause d reduction of P release in soil solution . T hus decomposition of citric acid dropped as a consequence of low microbial activity in response to low nutrient availability . Microbial Effects on C arbon Fluxes The relationship between C mineralization and DOC production, the latter being a cr itical factor in podzolization potential, depend upon temperature and water content of the system (Chow et al., 2006) . Low upward C f lux can be favorable to increase total C content in a soil if the primary production is high . M icrobial processes are important an determinant to upward C fluxes. Downward flow of DOC can result in its loss of C from the soil system as leachates (Buckau et al., 2000) or its accumulation in the soil subsurface as in the case of podzolization. R esearchers have found significantly lower evolution of CO 2 C from S podosol Bh horizons relative to a soil with low Al concentration; one explanation is Al related OM resistance to microbial decomposition , result ing in higher OM accumulation (Aran et al.,
30 2001) . Boudot et al., ( 1989) found that a high Fe and Al to C ratio can reduce the decomposition rate of organic matter. According to this study, as the charge of complexing functional groups beco me satisfied by metal adsorption the organic colloidal particles start to flocculate. The rate of flocculation is much higher for humic like substances which had the effect of reducing the C mineralization rate. On treatment with Al, structural changes in the microbial community were observed for both surface organic and subsurface mineral horizons in a study conducted on forest soils (Joner et al., 2005) , showing that the soil microbial community can b e a sensitive indicator of soil chemical compositional changes. W et and dry cycles , such as occur seasonally in Aquods, have an impact on microbial communities via factors such as oxidation reduction potential (Fierer et al., 2003 b) . In effect, t he change in microbial community can alter C mineralization rates (Fierer et al., 2003a) ,which could in turn influence potential for vertical C flux leading to podzolization . Prevalence of different microbial populations is dependent upon oxidation reduction potential. A stud y of a Louisiana coastal swamp revealed that fungi produce more CO 2 than bacteria under a moderately reducing condition (Eh > +250 mV) but bacteria contribute more in total CO 2 production under a highly reducing (Seo and deLaune, 2010) . Fungi, by growing mycelium , explore soil space; this growth can promote buildup of soil organic matter. A batch experiment on agricultural peat soil showed that under floode d condition DOC concentration was higher than initial conditions for surface soil and C mineralization rate was highest at day 3 but decreased gradually. Under isothermal condition (20 0 C) dry and wet cycles were even more effective in production of DOC tha n dry period or flooded condition.
31 CHAPTER 3 RELATIONS OF IRON, ALUMINUM, AND CARBON ALONG TRANSITIONS FROM UDULTS TO AQUODS Spodosols are a large sink of C (Johnson et al., 2011; Stone et al., 1993) as well as potential C source if conditions change such that C forms become unstable (Brady et al., 2011) . Classical theories of Spodosol formation (Schaetzl and Harris, 2011) invoke soil organic acid production (Fox and Comerford, 1990) and organo metal complexation (Van Hees et al., 2000; Wright and Schnitzer, 1963) that result in the redistribution of metals and subsurface C accumulation. Florida Spodosols to a depth of 2 m have been estimated to contain about 0.025 to 0.05% of the global organic carbon (OC) pool (Stone et al., 1993) . However, this 2 m assessment is a significant underestimation because Bh horizons in some regions can extend well below 2 m (Daniels et al., 1975; Harris et al., 2004; Thompson, 1992) . Spodosols of the SE USA coastal plain are associated with shallow fluctuating water table s and commonly classify as Aquods (Brasfield et al., 1973; Garman et al., 1981; Harris, 2001; Tan et al., 1999; Watts and Collins, 2008) . Hence, hydrology is an apparent factor in the accumula tion of a significant pool of C and Al , with minimum Fe concentration , but mechanistic details of the hydrologic linkage are still uncertain. Clues to the water table linkage can be found along transitions between Aquods and better drained soils such as Psamments and Udults. Studies of water table dynamics along these transitions document close correspondence between degree of podzolic E Bh development and water table depth (Garman et al., 1981; Tan et al., 1999) . Specifically, the Bh weakens in morp hological expression (becomes thinner and lighter in color) and approaches the soil surface (E horizon becomes thinner) with diminishing near surface water table incidents , e.g., shorter periods of near surface
32 saturation, greater depths to saturation, etc . Divergence of seasonal high water saturation and depth of Bh along the hydrologic transition shows that despite the strong association between fluctuating water table and podzolization in these soils, the depth of Bh horizon is not an indicator of season al high water table depth, per se. However , t h e presence and relative depth of the Bh does provide information about the hydrologic trend of a landscape. Th e rise in Bh depth with deepening of seasonal high water tables also confirms, in conjunction with w ell expressed overlying podzolic E horizons, that Bh horizons in these settings do not form simply by C precipitation with groundwater entrained metals. Rather, there is a redistribution of metals, primarily Al in the case of regional Aquods. The followin g questions arise from the trends described above: How does seasonal water saturation promote the mobilization of Al and its subsequent accumulation in association with C in the Bh? Why is Fe concentration consistently lower than Al in Aquod Bh horizons? Hy potheses related to this study are (H1) Removal of Fe by near surface saturation induced reducing conditions fosters mobilization and redistribution of Al that serves as a binding agent for sand grain coatings and (H2) Mobilization of Al via complexation with organic acids results in redistribution of sand grain coatings bound by Al prior to Al complexation. Objectives of this study were to determine distributional relations between C, Al, and Fe in A , E and Bh horizons along these soil and hydrologic tr ansitions and to evaluate their mechanistic implications. Materials and Methods Site Selection A major selection criterion was that the sites constitute a hydrologic transition from poorly drained flatwoods ecosystems dominated by Aquods to better drained
33 ecosystems with little or no expression of podzolization. This inferred drainage gradient, based initially on soil morphological indicators of wetness, topography, and plant communities, was later documented by water table monitoring well data (see below) . Details of these data will be reported elsewhere (Harris et al., 2013) . Four transect sites were selected along hydrologic transition s f rom poorly drained Aquods to better drained Udults . T wo transects were located at the Longleaf Flatwoods Reserve (LFR) managed by the St. Johns River Water Management District and two at the Austin Cary Forest (ACF) managed by the School of Forest Resources and Conservation at the University of Florida (Figure 3 1). All four sites had vegetation typif ies such transitions: pine and dense understory of saw palmetto ( Serenoa repens ) in the flatwoods (Abrahamson and Hartnett, 1990) and mixed oak and pine with less dense understory on the drier end of the trans ects . All transitions showed typical soil morphology in which the Bh horizon becomes lighter in color and shallower or disappears altogether as the drier end of the transition is approached (Garman et al., 1981; T an et al., 1999) . Slopes along transects ranged from <1 to 2%. A survey level transit and level rod were used to determine ground elevations relative to a previously surveyed benchmark. The survey loop closed with a vertical elevation error of 0.6 cm. All elevations determined are therefore accurate to +/ 1 cm. W ell Construction and Deployment Wells were constructed of screened PVC pipe (7.5 cm internal diameter) with fabric wrapped around the perforations to prevent clogging. They were installed to 2 m depth , backfill ed with sand and seal ed at the top with bentonite. Risers were also installed to increase visibility such that they would be less likely to be damaged by
34 land management practices (e.g., controlled burns). A rain gauge was available at bot h the LFR and ACMF locations. Six wells were installed at the first LFR transect (1 1 to 1 6) and at both ACMF transec ts (3 1 to 3 6 and 4 1 to 4 6). Five wells were installed along the second LFR transect (2 1 to 2 5). Depth to water table was measured weekly for approximately 2.5 years (with few exc eptions) using a contact meter. Rainfall was monitored via rain gauges deployed at ACMF and LFR. There was a period of drought between November, 2011 and March, 2012, when ~30 mm of rainfall was recorded (50 year annual average is >1250 mm), and during which water tables were commonly below the bottom of all wells. There were 98, 245, 217 and 280 days when all wells were dry for transect 1, transect 2, transect 3 and transect 4, respectively. Following the dro ught several high rainfall events resulted in all well s having water, sometime above the soil surface for brief periods at the wettest ends of the transects. Distance between wells ranged from approximately 10 to 30 m. Wells were located so as to capture t he soil morphological range of the transition and to accommodate some local constraints (a road and fire trails ). Total transect lengths were approximately 120 m (transect 1), 80 m (transect 2) , 99 m (transect 3) and 84 m (transect 4) . Soil Description, S ampling and Analysis Soils were described and sampled by horizon after well construction at all well installation points (Table 3 1) using an auger within a radius of 1 2 m of each well. Soils at the driest end of these hydrologic transects are classified as loamy, siliceous, subactive, hyperthermic Grossarenic Paleudults (activity class was based on family of soils mapped at the site), and at the wettest sites, either as sandy, siliceous, hyperthermic Ultic Alaquods or Aeric Alaquods. Samples were air drie d, sieved to < 2
35 mm and stored in sealed plastic bags. Electrical conductivity (EC) was determined by standard conductivity meter after equilibrating at 1:2 (mass:mass) soil to doub le de ionized water ratio. Sample pH was measured using a standard pH meter with silver/silver chloride reference electrode, after equilibrating at 1:1 (mass:mass) soil to double de ionized water. Particle size of E and Bh horizons was determined by pipette method (Soil Survey Staff, 2004 ) . Total C was determined by flash combustion using a C/N analyzer. Sodium pyrophosphate (Soil Survey Staff, 2004) which is thought to be selective for organically complexed metals, was used to extract Al, Fe and C. Non crystalline secondary oxides of Fe and Al were extracted by a cid (pH 3) ammonium oxalate (aao) in the dark (McKeague and Day, 1966b) . Total secondary Fe oxides were extracted by sodium citrate dithionite bicarbonate (cdb) (Mehra and Jackson, 1960) . All metals were determined by inductively coupled plasma (ICP) spectrophotometry. Pyrophosphate extractable C (C p ) was determined by flash combustion using a liquid CO 2 analyzer. Numerical averages of water table elevations and depths were calculated for days when water was in all wells along a given transect to maintain relativity within a specified range of conditions (i.e., when depth to saturation was known for the entire range). Otherwise relativity would be distorted because lower depth of drier posi tions would not be accounted for below the well depth. The number of days for which the averages were calculated 26, 23, 31 and 16 days for transect 1, transect 2, transect 3, and transect 4 re s pectively . Statistical Analysis A s lope test of two regressions regarding C vs Al plots for surface A and subsurface Bh horizon w as conducted to determine whether they are significantly
36 different ( P <0.05). Relation between depth to upper Bh boundary and Fe concentrations was determined by non linear regress ion. The effect of water table depth on selected soil parameters were analyzed using one way ANOVA. All statistical analysis was done using JMP software (version 10, 2013, SAS Institute, Cary, NC). Results and Discussion Water Table in Relation to Morpholo gy, Classification, and pH Each transect of this study captured a drainage gradient (F igure 3 2) of moderately well drained Udults to poorly drained Aquods with fluctuating water table conditions across the gradient . Water tables fluctuated significantly over a period of approximately 2.5 years ; in the Aquod zone of transects they ranged in depth from 0 to >1.5 m. However, the drainage gradient (depth from soil surface to saturation) was maintained over the study period. There were minimal topographic clue s to the soil drainage transitions since slopes along transects were nearly imperceptible. However, the transition was reflected in vegetative changes and soil morphology (Table 3 1; Figures 3 3 and 3 5). An inverse relation was found between depth to upp er Bh boundary and all shallowest water table depth measured at each sampling point during the study), (ii) me when water was measurable in all wells ( above the bottom depth of well installation ), and mean water table (MWT; average water table depth when water was in all wel ls of a transect) (Figure 3 3; Table 3 2). The Bh horizons also became deeper and more strongly expressed (Table 3 3) moving from drier to wetter condi tions along the transects (Figures 3 4 and 3 5). These transitional hydrologic podzolic trends are consistent with
37 those reported by others for soils of the region (Garman et al., 1981; Tan et al., 1999) . Drier sites at each transect (1 2, 2 1, 3 1, and 4 1) had Bh horizons that were too weakly expressed to meet the criteria for Spodic horizon. Hence, all of the soils at the driest wells still classif ied as Udults. based on proximity to the drier end of the transect, relatively shallow depth (not thickness), and slightly higher color value (10YR 4/2 or 4/4) rather than diagnostic spodic material (Figur e3 5; Table 3 3). Weak Bh horizons had less C but comparable Al p to strongly expressed Bh horizons. There were 8 Udult profiles with weak Bh horizons. All soils were acidic (pH from 3.5 to 5.5), with A horizons being more acidic tha n Bh horizons ( P < 0.000 1). A significant inverse relation was found between pH of upper Bh and depth of its upper boundary (R 2 =0.51; P < 0.001), with well expressed Alaquods being significantly more acidic than the better drained soils. The indicators of relative water table dept hs (MWT, LWT and HWT) showed a significant positive relation with pH (Table 3 2). There was no significant relation between pH and Fe p or Al p . However, pH decreased significantly with increasing C t ( P < 0.0001) and C p ( P < 0.0001) for all horizons, suggesting that acidity is controlled more by activity of organic acids than by buffering from metal hydrolysis for these soils. Relations of C, Al, and Fe along the Drainage Gradient Concentrations of Fe p in Bh horizons change d systematically along the transitions. Metals extracted by acid ammonium oxalate are only slightly higher in concentration than pyrophosphate extractable metals, and they were highly correlated (for Al, R 2 =0.96; P <0 .0001 ; slope=1.33 and Fe; R 2 = 0.87; P <0. 0001 ; slope= 1.33 ) for all horizons, suggesting that most of the non crystalline Fe and Al present in these soils
38 are organically bound. Metals extracted by these two procedures yielded equivalent interpretations. Extractable Fe concentrations of Bh correl ated positively with depth to HWT, LWT, and MWT (Table 3 2) and inversely with depth to upper Bh boundary (Figure 3 6). These trends of Fe p amount to diminishing Fe abundance with strengthening of Bh (Table 3 3) and wetter conditions (Table 3 2). Crystalli ne Fe was so low (highest concentration was 3 mmol kg 1 ) in the samples analyzed that no trends were evident within the drainage context of this study and so those data are not reported . No significant relation was found between HWT, MWT or LWT and Bh horizon Al p but Bh horizon C p was related to MWT (Table 3 2). Aluminum concentration of the Bh did not significantly increase with depth to the upper Bh boundary, suggesting that some Al is lost (could be along with C) from the soil zone during the podzol ic redistribution. The slopes of C p versus C t , and Al p versus C t , are significantly higher ( P <0.05) for Bh than A horizons considering the pooled data from all transects (Fig ure 3 8) as well as the d ata for separate transects (Figure 3 9). The consistent ly higher Al p to C t ratio for Bh than A horizons for each transect (Figure 3 9) and in each profile (Table 3 4) reflects the higher proportion of metal associated C for Bh relative to A horizons, the latter being dominated by decomposing plant detritus. Hence, this ratio, as well as the Al p to C p ratio, could be used as an indicator to distinguish between Bh and surface horizon material, even in the case of weakly expressed Bh horizons (Table 3 4). Increases of up to 88% in Ct were observed for Aquods re lative to Udults when C t was calculated per m 2 on a comparable depth basis (data not shown). Hence
39 conversion of Udults to Aquods by rising water tables would constitute considerable subsurface C sequestration. Mechanistic Implcations The inverse relation between depths of upper Bh boundary and relative WT depths is consistent with the idea that frequency and duration of near surface saturation is a driver of podzolization for Spodosols of this region. More precisely, the rate and extent of podzolization ma y be largely controlled by the frequency at which near surface saturation occurs at a sufficient duration to reach certain moisture driven biogeochemical thresholds. Possible thresholds are biogeochemical reduction of Fe and the speciation and activity of organic acids but it did show declining Fe concentration with strengthening of Bh expression toward the wetter end of the transect. It also revealed that Fe and Al are significantly related for weakly exp ressed Bh horizons but not f or well expressed ones (Figure 3 7). The latter finding indicates that there is a zone along the drier end of the drainage transition, where weak Bh horizons tend to occur, in which Fe is being vertically redistributed in tandem with Al. Beyond that zone, where the Bh deepens and becomes darker, Fe is essentially depleted such that it shows no relation with Al or C. Redox depletion of Fe is a possible predisposing factor for podzolization in southern coastal plain soils where S podosols are mainly restricted to wet sites with fluctuating water table conditions. In effect, Fe may serve as an inhibitor of the process by reducing the organo complexing pressures born increasingly by Al as Fe is progressively depleted toward the wette r part of the landscape. A previous study (Harris et al., 1995; Harris and Hollien, 2000) showed that while Fe concentration was significantly lower on toeslopes than summits, Al did not differ significantly until the
40 position where podzolization became evident. Sam ples with higher Fe content were shown to be less inclined to form experimental Aquod like features. These experimental results are consistent with Fe serving as an inhibitor. An alternative explanation is that preemptive redox depletion of Fe as the wette r part of the landscape is approached would reduce the pool of Fe available for mobilization via organo complexation even if Fe were not an inhibitor of podzolization. Either of these scenarios is consistent with the dominance of Al in Bh horizons of the r egion and with the positive relation between Al p and C p ( P <0.0001 ) for Bh horizons observed in the present study; Fe p and C p were not related. However, Fe is not thoroughly depleted in the zones of marginal wetness as inferred from its presence in weak Bh horizons (Table 3 3). The podzolic eluviation which resulted in the strongly expressed E horizon of the Aquods can be inferred to have resulted from stripping of metals along with other components of sand grain coatings that were still retained on the Udu lt side of the transects. The data for all transects confirm consistent eluvial illuvial relations between the E and Bh not only fo r metals but also for clay (Figure 3 10). The metal relation is in accordance with classical podzolization theories. Clay ill uviation is not commonly invoked in these theories. However, others have also found evidence of clay translocation associated with podzolization (Guillet et al., 1975; Harris and Hollien, 1999; Lin et al., 2002; Ugolini et al., 1977) or with podzolic features produced experimentally in soil col umns (Harris et al., 1995; Harris and Hollien, 2000) . Summary The morphological transition from Udults to Aquods parallels hydrological (drier to wetter, with fluctuating water tables), chemical (e.g., pH), and compositional (Fe, Al, C and clay) changes. Implications of these changes are that Al and C distributions along
41 the transition are mainly controlled by podzo lization whereas Fe concentrations diminish with progressive wetness and podzolization. These results implicate Fe as a possible inhibitor of the process for southern Spodosols where podzolization is mainly restricted to poorly drained landscapes. However, its higher concentration and correlation with Al in weak Bh horizons indicate limited podzolic redistribution of Fe in marginally wet zones. Near surface wetness effect on microbial processes influencing organic acid production is an additional factor tha t should be investigated in explaining the hydrologic link. Results of this study have implications for the consequences of changing hydrology on direction and magnitude of C flux as well as fate of metals on coastal plain landscapes of the SE USA.
42 Table 3 1. Examples of vegetative, soil descriptive, and metals data for Udult and Aquod profiles along transects 1 (LFR location) and 3 (ACMF location). All A and E horizons were sands with very friable to lo o se consistence. The Bt and Btg horizons were loamy (sandy loam to sandy clay loam) with friable consistence. These soils are representative of morphological trends along all 4 transects. [ C p = pyrophosphate extractable C; Al p = pyrophosphate extractable Al ; Fe p = pyrophosphate extractable Fe ] Transect Profile and Vegetation Horizon Lower depth (cm) Dominant color Fe p Al p C p ----------mmol kg 1 ----------1 * 1 1 Loamy, siliceous, subactive, hyperthermic Grossarenic Paleudult Hardwoods and pines with sparse understory A 30 10YR 3/1 2.21 12.06 156 EA 68 10YR 4/2 1.8 10.4 59 E1 105 10YR 7/2 0.65 5.79 12 E2 149 10YR 8/1 0.46 3.83 0 Bt 168 10YR 5/8 Btg 187+ 10YR 5/2 1 5 Sandy, siliceous, hyperthermic, Ultic or Aeric Alaquod Pine with dense understory of mainly saw palmetto (S erenoa repen s) A 35 10YR 5/1 0.11 1.29 92 E 71 10YR 6/1 0.02 0.13 86 Bh1 80 10YR 3/2 0.16 16.41 366 Bh2 90 10YR 3/3 0.11 11.28 255 Bh3 106 10YR 4/3 0.12 5.23 109 E1 115 10YR 6/3 0.06 4.85 100 E2 134 10YR 5/3 & 4/2 0.13 4.51 95 E3 175+ 10YR 6/2 0.04 2.74 63 * Transect# Well#
43 Table 3 1. Continued. Transect Profile and Vegetation Horizon Lower depth (cm) Dominant color Fe p Al p C p ----------mmol kg 1 ----------3 3 1 Loamy, siliceous, subactive, hyperthermic Grossarenic Paleudult Pine and scattered hardwoods with sparse understory A 5 10YR 4/1 1.39 8.26 306 Bh 17 10YR 4/2 3.79 18.91 149 EBh 82 10YR 5/3 3.30 8.97 6 E1 101 10YR 7/2 1.07 3.02 0 E2 153 10YR 8/1 0.44 1.77 0 Bt 200+ 10YR 5/6 3 6 Sandy, siliceous, hyperthermic, Ultic Alaquod Pine with dense understory of mainly saw palmetto (S erenoa repen s) A1 13 10YR 2/1 0.97 2.92 553 A2 23 10YR 3/1 0.38 0.76 142 E 43 10YR 5/1 0.31 0.21 36 Bh1 58 10YR 2/2 1.26 24.79 698 Bh2 80 10YR 3/2 0.46 34.99 558 Bt 102 10YR 5/3 Btg1 124 10YR 6/2 Btg2 151+ 10YR 6/1
44 Table 3 2. Correlation coefficients and P values for relations between upper Bh horizon properties and mean water table depth (MWT), lowest (deepest) water table depth (LWT) and highest (shallowest) water table depth (HWT). Water table measurements apply to conditions when water was in all wells along transects. C p = pyrophosphate ex tractable C; Al p = pyrophosphate extractable Al; Fe p = pyrophosphate extractable Fe Soil properties (upper Bh) MWT (cm) LWT (cm) HWT (cm) r P r P r P Depth to Bh (cm) 0.71 0.0003 0.82 <0.0001 0.57 0.0039 C p mmol kg 1 0.47 0.028 0.41 0.062 0.22 0.32 Al p mmol kg 1 0.09 0.687 0.02 0.93 +0.09 0.70 Fe p mmol kg 1 +0.71 0.0003 +0.72 0.0002 +0.73 0.0002 pH +0.68 0.0008 +0.63 0.002 +0.48 0.03
45 Table 3 3. Comparison for selected properties between weak and well expressed upper Bh horizons. Designation of slightly higher color value. C p = pyrophosphate extractable C; Al p = pyrophosphate extractable Al; Fe p = pyrophosphate extractable Fe; C t = total C Transect Well # Bh development Bh upper depth (cm) Color pH Al p Fe p * C p * C t * ---------mmol kg 1 ---------1 2 weak 32 10YR 4/2 4.6 13.2 0.8 100.7 305.4 5 strong 71 10YR 3/2 4.1 16.4 0.2 366.2 521.6 2 1 weak 12 10YR 4/2 4.4 18.0 2.4 144.1 513.0 4 strong 56 10YR 2/2 4.2 15.1 0.1 341.1 587.3 3 1 weak 5 10YR 4/2 4.7 18.9 3.8 148.8 536.8 4 strong 35 10YR 2/2 4.2 35.4 0.5 563.5 1137.5 4 1 weak 16 10YR 4/4 4.9 25.5 5.6 127.6 260.9 4 strong 65 10YR 2/1 3.8 47.3 0.3 1091.6 1693.8 *Indicates statistically significant difference between means of weak and strong Bh.
46 Table 3 4 . Comparison for selected properties between A horizons and weak Bh horizons of driest side . C p = pyrophosphate extractable C; Al p = pyrophosphate extractable Al; Fe p = pyrophosphate extractable Fe; C t = total C Transect Well # Horizon Lower depth (cm) Color pH Al p * Fe p C p C t * Al p /C p * Al p /C t * ---------mmol kg 1 ---------1 2 A 32 10YR 4/1 4.1 3.7 0.9 101 482.9 0.04 0.008 Bh 69 10YR 4/2 4.6 13.2 0.8 101 305.4 0.13 0.04 2 1 A 12 10YR 4/1 4.0 8.9 1.7 222 1143.3 0.04 0.008 Bh 27 10YR 4/2 4.4 18.0 2.4 144 513.0 0.13 0.04 3 1 A 5 10YR 4/1 4.4 8.3 1.4 306 2485.8 0.03 0.003 Bh 17 10YR 4/2 4.6 18.9 3.8 149 536.8 0.13 0.04 4 1 A 16 10YR 4/1 4.4 5.7 2.3 149 776.5 0.04 0.007 Bh 30 10YR 4/4 4.9 25.5 5.6 128 260.9 0.19 0.10 *Indicates statistically significant difference between means of A and weak Bh.
47 Figure 3 1. Locations of transects where wells were installed and soil profiles were sampled. Each transect followed a drainage gradient from better drained Udults to poorly drained Aquods
48 Distance between wells (m) Figure 3 2 . Relation between depth to water table and distance b etween wells for each transect A ) transect 1, B ) transect 2, C ) transect 3 and D ) transect 4. The distance from left to right is moving from better drained Udults to poorly drained Aquod s. Mean water table (MWT), highest (shallowest) water table (HWT) and lowest (deepest) water table (LWT) were calculated considering all days (26 days, 23 days, 31 days, and 16 days for transect 1, transect 2, transect 3 and transect 4 respectively) when w ater was present for all six wells for the entire study
49 NAD88* = North American Vertical Datum 1988 Figure 3 3 . Transect 1 as an example of the drainage gradient and morphological trends that were common to all transects. Mean water table (MWT) was calculated considering all days when water tables were present for all six wells. Highest (shallowest) water table (HWT ) was based on the highest water level measure during the study for this transect. The dark region depicts the Bh horizon along the transect. Note a vertical exaggeration of approximately 20
50 Figure 3 4. Photograph of a trench at Austin Cary Memorial Forest, 1.2 km S of transect 3, showing upward fading of Bh from wetter (background) to drier (for eground) part of the landscape. The trench extends about 100 m from the edge of a wetland (background) through a zone of poor (moving toward foreground; dense understory of Serenoa repens ) to a better vegetation extending behind the photographer) Picture taken by Geof Locuta
51 Figure 3 5. Comparison of wel l expressed (left) and weakly expressed (right) Bh horizons along transect 4 (wells 4 4 and 4 1, respectively), as displayed at proportionate depth in a tray from auger excavation. Mean water table depth increased from left to right, inversely with Bh dept h and degree of expression
52 Figure 3 6 . Relation between depth to upper Bh (n = 48) boundary and A ) acid ammonium oxalate extractable Fe (Fe aao ) and B )Al (Al aao ). A similar trend with Bh depth was observed for pyrophosphate extractable Fe (Fe p ) and Al (Al p )
53 Figure 3 7. Comparison between pyrophosphate extractable Fe (Fe p ) and Al (Al p ) for eight weakly expressed Bh and fourteen strongly express upper Bh
54 Figure 3 8 . Comparison between A (n = 25) and Bh (n = 48) horizons with respect to relations between total C (C t ) and the variables A ) pyrophosphate C (C p ) and B ) pyrophosphate Al (Al p ). The two circled data points were statistically verified to be outliers; R 2 = 0.91 excluding outliers
55 Figure 3 9 . Comparison between mean ratios of pyrophosphate extractable Al and C (Al p/Ct) for A versus Bh horizons, by transect. Error bars represent standard error. Differences between means within transects were statistically significant ( P <0.05; t test) for eac h transect.
56 Figure 3 10 . Evidence for C A), metals B) and clay C) accumulation in Bh with r espect to E for each transect. Differences between means within transects were statistically significant ( P <0.05; t test) for each transect.
57 CHAPTER 4 CARBON AND METAL RESPONSES TO CONTROLLED WATER TABLE FLUCTUATIONS IN SANDY SOIL MATERIAL Increase in concentrations of atmospheric greenhouse gases has generated much concern. Carbon is a major element participating in global warming as carbon dioxide ( CO 2 ) or methane ( CH 4 ) . Soil C can be released either as gaseous or dissolved form , with different consequences with respect to its environmental distribution . For example, release as CO 2 constitutes an immediate release to the atmosphere whereas dissolved org anic C (DOC) has potential for transport, transformation, and persistence in soils and aquatic environments. The present study focuses on the depth of controlled water table fluctuation as a determinant of the direction (upward gaseous versus downward DOC) and magnitude of C flux from soil columns. S tudies focused on wetlands soils and sediments have shown that C distribution can be influenced by parameters such as temperature (Vicca et al., 2009) , moisture (Lamparter et al., 2009) , drainage (Salm et al., 2009) water table effect (Chivers et al., 2009; Olivas et al., 2010) , management practices and tidal effect (Yamamoto et al., 2009) . Photosynthesis fixes atmospheric CO 2 and promotes a build up in the soil C pool; soil microorganisms complete the C cycle by returning the soil C to the atmosphere via ae robic and anaerobic processes. Changes in soil parameters can affect how these microbial communities (Fierer et al., 2003b; Rousk et al., 2010) influence greenhouse gas fluxes (Singh et al., 2010) . Wetland and peat land soils are mainly C sinks (Maynard et al., 2011) but they could also be affected b y environmental changes and act as greenhouse gas sources (Gorham, 1991; Qualls and Richardson, 2008) . Downward flow of dissolved organic C (DOC) can result in its loss from the soil system as leachates (Buckau et al., 2000) or its a ccumulation in the soil subsurface as
58 in the case of podzolization. Laboratory incubation studies have shown that the mineralization rate of C or DOC production and its movement depend on factors like temperatur e, pH (Andersson et al., 2000) , and types of vegetation (Hansson et al., 2009) . The dissolved forms of C are significant as they are not only a source of nutrients but also a facilitator of the transport of hydrophobic organic solutes (Rav Acha and Rebhun, 1992) and metals (Chen et al., 2008) to water bodies. Downward movement of C can, under certain conditions, result in its subsurface accumulat ion in association with metals within spodic (Bh and Bhs ) horizons of Spodosols. These horizons constitute a significant pool of soil C (Johnson et al., 2 011; Stone et al., 1993) . There is e vidence that clay material s (e.g., silicates) are also transported and eventually accumulated along with metal s and C in the formation of spodic horizons (Banik et al., 2014; Guillet et al., 1975; Harris and Hollien, 1999; Ugolini et al., 1977) . The occurrence of southern Spodosols (e.g., the SE USA coastal plain) is strongly linked to fluctuating water tables (Brasfield et al., 1973; Garman et al., 1981; Harris and Hollien, 2000; Tan et al., 1999; Thompson, 1992; Watts and Collins, 2008) but mechanisms of the hydrological effect are uncertain (Harris et al., 1995) . The C in Bh horizons of southern Aquods was found to differ significantly in characteristics from that of overlying surface horizons (Banik et al., 2014) . This study was c onducted to determine the influence of water table depth on flux of C, metals, and < 50 Âµm ( finer fractions as silt and clay size s ) components in soils columns. My overarching hypothesis is that the upper depth of seasonal high saturation influences the di rection and magnitude of C fluxes in imperfectly drained sandy coastal plain soils such that a change in hydrology would change the C dynamics and potenti al
59 of subsurface C and metal accumulation. The following specific hypotheses were addressed in this st udy: (a) near surface hydrologic regime favors a higher rate of DOC production relative to C mineralization, resulting in greater organo metal complexation (H1), (b) vertical movement of C is greater under shallow than under deeper fluctuating water table conditions because there is less DOC production and mobilization for deeper water tables (H2), and (c) total C loss (sum of mineralization and leaching ) is less for shallow water table conditions (H3). These hypotheses were tested using soil columns with fluctuating water tables established at two different upper depth limits. Materials and Methods Site and Soil Selection The soil materials used in this study were collected from the A and upper E horizons of five somewhat poorly drained soils (loamy, siliceous, hypert hermic Grossarenic Paleudults) located at the margins of poorly drained flatwoods and transitioned to Alaquo ds (Table 4 1). S oils at such locations have been shown to be predisposed to podzolic like changes under artificial water table regimens (Harris et al., 1995; Harris and Hollien, 2000; Harris and Rischar, 2012) . They were the driest soils in fi ve transects sampled as part of a study focused on podzolization along hydrologic transitions of sandy coastal plain soils (Figure 4 1 ) (Banik et al., 2014) . All five soils (Cox et al., 2004) of oaks ( Quercus sp ecies ) and pine ( Pinus palustrus ; Pinus taeda ) with sparse understory. The adjacent flatwoods had similar tree species but greater dominance of pine and denser understory dominated by saw palmetto ( Serenoa repens ) and gallberry ( Ilex glabra ) (Abrahamson and Hartnett, 1990 ) , a plant community consistent with podzolization.
60 Sample Characterization Soil s amples collected as described above were sieved through 2mm sieve and stored at the same moisture content as sampled until used for preparing columns. Moisture content wa s determined gravimetrically after oven drying at 105 0 C. Organic matter content was measured by loss on ignition after heating at 450 0 C overnight (Nelson and Somm ers, 1996) . A few hundred grams of these samples were air dried for physical and chemical analyses. A pH meter with silver/silver chloride reference electrode was used to measure sample pH in double de ionized (DDI) water (equilibrated at 1:1, mass: mass) and electr ical conductivity (EC) was measured by standard conductivity meter after equilibrating with DDI water at 1:2 (mass: mass). Column Experimental Design Ten acrylic tubes (2 to represent each of the 5 sites) of 2.5 cm diameter each and of 90 cm height were p acked to a thickness of 60 cm with E horizon sand. Above this E horizon material was packed 15 cm of A horizon material from same soil. Soil materials were packed under field moist condition to a bulk density of approximately 1.5 (on dry weight basis) for each column and for each horizon. Five of these tubes received fluctuating shallow water table (SWT) treatments whereby the column water table (WT) was maintained at the soil surface for a period of time (specified below), followed by drainage. The other 5 tubes simulated fluctuating deeper water table (DWT) conditions which entailed saturation at a depth of approximately 25 cm for the same time periods as the SWT columns, followed by drainage. Glass wool was used in the bottom of tubes to retain the soil i n the columns during drainage. A stopper with small glass tube was attached in the bottom to introduce water from below and to serve as an outlet. Clear flexible tubing was fitted over the glass tubes to enable control of water
61 flow in and out of columns; water was introduced into the columns from the bottom in order to minimize air entrapment. A constant water flux was maintained for all columns throughout the study period, with water table depth and duration being the independe nt variables of the experime nt. The top of each column was fitted with a CO 2 absorber (1M NaOH) to monitor the amount CO 2 evolved (Anderson, 1994) in each column. A blank column contain ing only DDI water was monitored in the same way. Columns were initially filled to the soil surface with DDI water. Water was maintained at the surface for SWT columns while for the DWT columns water was dra ined within 1 h to a 25 cm target water table. S aturation was maintained at these depths for 3 days followed by drainage and the cycle was repeated 8 times. Following these treatments the same saturation regimen was performed of all columns 3 more times but with 7 day incubations . The incubation regimen of 3 days followed by 7 days was established based on providing sufficient magnification time to monitor upward C flux from SWT columns while also achieving representative ranges of saturation duration and redox potentials. The experiment was designed suc h that the total water flux was the same for all columns for both treatments. After each treatment period each NaOH trap was replaced with a fresh NaOH solution and the old one was back titrated with HCl (1M). To back titrate the NaOH barium chloride was a dded followed by Phenolphthalein indicator addition. Column le achates were collected for 1 h and preserved in freezer (< 0 0 C) to analyze Fe, Al and DOC concentration later. After 45 days, pore water was withdrawn from each column after allowing them to f ill to 3 5 cm above the soil surface and before draining the DWT columns to their targeted depth . These solutions were analyzed for
62 DOC, Al, and Fe concentrations and the degree of condensation of aromatic C network by determining E 4 /E 6 ratios (Kononova, 1966; Tan, 2003) . The process of pore water collection was repeated 4 m ore times at 14 day intervals. The oxidation reduction potential (ORP) was measured in two sets of columns at the end of the experiment to confirm that the SWT condition induced a lower potential than the DWT. The process of ORP measurement was repeated three more times to verify ORP differences statistically, by different WT conditions. A Pt wire was inserted at a depth of 7cm in the SWT columns and at depths of 7 and 38 cm in DWT columns. The Pt wire was sealed at the point of insertion with Si gel. A A g/AgCl electrode was used as a reference electrode and hydroquinone at pH 4 and pH 7 buffer solutions were used to standardize the set up before taking each reading. The change in electron volts was measured for 1 hr time period using a voltmeter. At the e nd of the column experiment the columns were allowed to drain for 3 days. Then samples from each column were collected at 5 6 cm increments. This was done by sliding a thin plastic tube (marked at 5 cm) of slightly smaller diameter than the column downward along the column interior to the prescribed increment, inserting a rubber stopper at the top end of the plastic tube to form a air lock to retain the soil material, and removing the tube with retained sample. These samples as well as soils before treatmen t were dried and analyzed for organic C, metal (Fe and Al), total C, and <50 Âµm material content to evaluate treatment effect on distribution of these components. Total C and N were determined by flash combustion using a C/N analyzer. Metals and C were ex tracted using sodium pyrophosphate (Soil Survey Staff, 2004) , which is thought to be selective for organically complexed metals. The concentration of <50 Âµm materials was determined by ultrasonic dispersion and
63 separation of the silt and clay sized materials from sand size particles (dispersed until the solution looks clear) using water as solvent , via elutriation and weight difference between initial material and final sand materials . Statistical Analysis The effect of water table depth effect on metal and C released were analyzed using one way ANOVA. All pair Tukey test was used to test difference between statistical significa nce for this study. Statistical analysis was done using JMP software (version 10, 2013, SAS Institute, Cary, NC). Results and Discussion Sample Properties The soil materials used in the experiment were acidic, sandy, and relatively low in organic matter co ntent (Tables 4 1 and 4 2). Hence they are representative of the sandy coastal plain soil materials targeted in this study. Depth of Water Table in Relation to CO 2 Flux The atmospheric CO 2 flux was consistently higher for DWT treatments than for SWT, for all days of measuremen t and all set comparisons (Figure 4 2 ). No statistically significant difference in CO 2 flux was found among the 5 sites. The mineralization rate decreased slightly with time as inferred from the cumulative CO 2 release (Figure 4 3 ). T his likely resulted from a decline in readily decomposable material over the course of the experiment. This study verified that a lower WT depth resulted in increased C mineralization rate and emission of CO 2 to the atmosphere and reduced potential for C s torage in soil systems.
64 Water Table Depth in Relation to Leachates Characteristics In leachates, C released from columns as DOC was related to WT position. Total DOC concentrations were relatively low (maximum 0.15 mg L 1 ) but the SWT condition re leased s ignificantly ( P < 0.005, t test) more C than DWT (Figure 4 4 ). The net C retention was much higher for SWT than DWT columns despite the slightly greater leaching loss from the former due to the much higher CO 2 loss from the latter. Carbon release diminished over time for all columns and both treatments. Iron was higher in leachates for SWT treatment than DW T during the first 18 days (Figure 4 5 ), after which Fe was below the detection limit. Lower ORP under SWT might initially promote Fe reduction and release in leachates. However, other results (see section 3.4 below) confirmed that Fe was not completely depleted from the column. Aluminum concentrations were generally low or below detection limit, with an avera ge of 2 mg L 1 for SWT and 0.5 mg L 1 for DWT; no significant difference in Al concentrations in leachates over time was found between treatments. Aluminum is not ORP sensitive and Water Table Depth and Soil Pore Water Characteristics Release of DOC in pore water was significantly higher (3 5 times) under SWT than DWT ( P < 0.05) (Figure 4 6 ) while average SWT CO 2 loss by mineralization wa s about 1/4 that of DWT (Figures 4 2 A and 4 3 ). Thus the depth o f WT fluctuation profoundly influenced the path (dissolved or gaseous efflux) of C transformations over the study period. The E 4 65 /E 6 65 ratio, which relates to degree of condensation of aromatic constituents and relative molecular weights of soil organic compounds (Kononova, 1966; Tan, 2003) was in range of humic substances (3 4) for DOC produced by the
65 SWT system b ut DWT its E 465 /E 665 data . Metal concentrations were related posit ively ( P <0.0001) to DOC C concentrations in SWT soil pore water (Figure 4 7). These data suggest that SWT treatments promoted functional groups condensation and DOC metal interactions. Complexation of DOC with Fe and Al under SWT treatment could possibly r educe su rface charge of DOC molecules, making them less hydrophilic and more prone to precipitation and immobilization such as may occur in podzolization (DeConinck, 1980) . The eventual result of such DOC metal interactions would be a decrease in DOC and metal concentrations in leachates, as was observed in this study (Figures 4 4 and 4 5). The positive significant relation between Fe and Al in pore water for SWT but not DWT columns (Figure 4 8) suggest s that release of Al is related to Fe release due to drop in redox potentials . The pH of pore water was consistently higher (by1 2 units ) for SWT than DWT treatments. All these processes had the effect of diminishing the total C loss from SWT systems (Figures 4 9 ). The mass loss of C as CO 2 over the course of the study ranged from 6 10% of the original C content of column materials for DWT treatment and from 1 3% for SWT treatment. Possible Decomposition Pathways Soil organic matter is essentially the only nutrient source for the sandy soil used in this study. As soil microbial processes are sensitive to changes in soil environmental conditions, maintaining different water table positions indirectly affected the rate of microbial decomposition processes. Low oxygen content of the SWT treatme nt would tend to limit soil aerobic microbial respiration over time. Soluble DOC released by the systems would readily decompose at the earlier stages of incubation, during which higher microbial activity under DWT treatment was more effective in releasing CO 2 than
66 SWT. A drop in redox potential with longer incubation can raise pH which can promote hydrolysis of complex polymers of soil organic matter and release more DOC into soil solution. Eventually the fractions released could become a more recalcitrant and less nutrient rich substrates for microbes. These conditions might have resulted in slower respiration rate and hence explain the leveling off of cumulative CO 2 release for the 7 day incubation period even for the DWT treatment. The SWT condition cou ld provide enough time for DOC to stabilize as well as for the E 4 /E 6 ratio to increase. Under the DWT conditions the low release of DOC may relate to Eh/pH conditions less favorable for hydrolysis of organic matter (OM). Treatment Effects on Soil C, Meta ls and Finer (<50 Âµm) Material Distribution Before an d after percentage comparisons of < 50 Âµm materials (Table 4 2) showed evidence of significant decline in the upper 5 cm of columns but of grains in the subjacent zones for the SWT treatment, suggesting some downward redistribution of fine particles during the treatment. A significant relation ( P < 0.001) between pyrophosphate extractable C and < 50Âµm materials for both treatments suggests some sorbed or metal complexed C has accumulated in the SWT column s, consistent with less C release to the at mosphere (Figures 4 2 and 4 3 ) and l ess total C release (Figures 4 9 ) relative to the DWT columns. These observations again support our hypothesis of C stability under SWT, but in addition suggest that the stabili ty might be the result of both of chemical (complex formation with metals) and physical (downward movement of finer fractions by WT fluctuation) processes. Summary Soil C flux is dependent to a large extent on moisture driven biogeochemical processes. Thi s study experimentally documented that atmospheric and net C loss is
67 significantly less when the upper saturation limit is at the soil surface than when it is deeper. The upward (atmospheric) flux of C as CO 2 , favored by deeper saturation, constitutes imm ediate C loss from soil whereas downward C flux as DOC, favored by shallower saturation, amounts to greater duration of C retention in the soil. The vertical movement of DOC could result in its eventual loss to streams via groundwater flow or its long term retention in subsurface (Bh) horizons under shallow water table conditions such as prevail in flatwoods ecosystems of the SE USA coastal plain. An implication is that changes in water table depth can alter not only net soil C storage but also the proporti on of C converted to CO 2 versus DOC. Differences in the proportion would have consequences for global C dynamics.
Table 4 1 . S oil profiles from which A and E horizon materials were collected for this study, along with selected soil physical and chemical characteristics. TC = total C; TN = total N; C p , Fe p , and Al p = pyrophosphate extractable C, Fe, and Al, respectively. Transect Horizon Color pH (H 2 O) EC 1 ) Moisture content (g g 1 ) Organic matter ** TC TN C p Fe p Al p ----------% ----------mmol kg 1 1 A 10YR 3/1 4.2 29.1Â±0.52 * 0.07 2.9 1.31 0.03 58.6 1.1 5.9 E 10YR 5/1 4.8 12.6Â±0.32 0.06 N.D 0.48 0.03 32.1 1.9 16.9 2 A 10YR 3/1 4.2 29.5Â±0.32 0.07 2.6 1.22 0.04 58.4 1.7 7.7 E 10YR 5/3 4.8 11.7Â±0.17 0.05 N.D 0.44 0.01 22.7 2.8 14.0 3 A 10YR 3/2 4.4 32.8Â±0.47 0.06 2.4 1.32 0.04 53.3 2.0 9.1 E 10YR 5/2 4.9 11.9Â±0.26 0.07 N.D 0.4 0.01 26.3 3.6 20.9 4 A 10YR 3/1 4.5 23.4Â±0.15 0.06 2.7 1.16 0.04 56.8 1.8 5.1 E 10YR 5/2 4.9 10.6Â±0.2 0.05 N.D 0.34 0.01 24 4.1 16.7 5 A 10YR 3/1 4.3 40.8Â±0.64 0.07 1.5 0.85 0.04 31.9 1.1 3.0 E 10YR 6/4 5.0 11.5Â±0.1 0.05 N.D 0.62 0.02 18.5 2.8 15.3 *MeanÂ±standard error (n=3) ; ** Using loss on ignition ; N.D = Not determined
69 Table 4 2. Comparison of <50 Âµm material percentages before and after treatments, showing statistically significant loss Sample ID % <50 Âµm material before treatment % <50 Âµm material after SWT treatment (upper 5 cm ) % <50 Âµm material after DWT treatment (upper 5 cm) *P value comparison initial to final SWT (<50Âµm) materials *P value comparison initial to final DWT (<50Âµm) materials 1 A 5.39 3.44 4.82 0.01 0.08 2 A 5.01 5.05 5.95 3 A 6.47 5.08 5.08 4 A 6.24 5.36 4.56 5 A 4.33 2.88 3.08 1 E 2.77 3.53 3.18 0.002 0.1 2 E 3.18 3.94 3.29 3 E 3.52 4.05 3.74 4 E 4.1 4.49 4.24 5 E 2.9 3.2 2.72
70 Figure 4 1 . Schematic diagram of a landscape ecological transition from better drained sandhill (left) to poorly drained flatwoods (right), showing a change in vegetation, water table position and an upward fading Bh. The relative rise in water table (moving towards the surface) and in creased understory towards flatwoods results in subsurface C sequestration in the form of Bh horizons which tend to be best expressed under poorly drained conditions.
71 Figure 4 2. Comparison between treatments of soil C loss as CO 2 , showing A ) average loss per specified time interval for all c olumns within treatments and B ) average loss by each set of columns over the 45 days of incubation. All treatment mean differences for data shown were statistically significant. Error bars indicating standard erro r.
72 Figure 4 3. Cumulative C loss as CO 2 over the 45 days of incubation, by treatment effect. Where n (n = 1, 2, 3, 4, 5) designates a particular location from which the column material was collected and S and D stand for shallow and deep water table tre atments , respectively .
73 Figure 4 4. Average C loss per specified time interval for all columns within treatments over a 45 day study period. Differences between means for SWT (shallow water table) and DWT (deep water table) treatments were statistic ally significant ( P <0.05; t test).
74 Figure 4 5. Comparison between treatments ( SWT = shallow water table ; DWT = deep water table) of average Fe release from columns, based on 18 days of leachate collection. Differences between means were statistically significant ( P <0.05; t test).
75 Figure 4 6 . Comparison between C released as DOC into soil solution under treatments of SWT (shallow water table) and DWT (deep water table) .
76 Figure 4 7 . Relations between C and metal released into pore water , by SWT (shallow water table) and DWT (deep water table) treatments.
77 Figure 4 8 . Relations between Fe and Al released into pore water soil solution, by treatments SWT (shallow water table) and DWT (deep water table) .
78 Figure 4 9 . Comparison between treatments effect on total C loss from each set of columns. A significantly higher loss of C was observed for DWT (deep water table) than SWT (shallow water table) t reatments .
79 C HAPTER 5 FACT ORS REL A TED TO ACID INDUCED IRON AND ALUMINUM RELEASE FROM SANDY COASTAL PLAIN SOILS Stability of iron (Fe) oxides in soil is highly variable and chemistry behind the dissolution is complex, as related to their high sensitivity to different soil bio geo chemical changes. Soil redox oscillation can cause the ir dissolut ion /re precipitat ion and transform their surface properties ( Thompson et al., 2006) , thus chang ing reactivity. Iron and aluminum (Al) can each substitute in the crystalline and noncrystalline oxides of the other such that their fate and influence in soils tend to be interrelated (Desphande et al., 1968; Ekstrom et al., 2010) . D issolution of Fe oxides can release not only Al but a number of other soil metals , confirming their substitution in the soil Fe oxide phase (Singh and Gilkes, 1992) . The dissolution mechanism of Fe oxides involves adsorption of the ligands on their surface followed by Fe release into solution, depending on the surface type (Stumn et al., 1985) as well as ligand types (Zahang et al., 1985) . Mobilization of iron from its oxides can occur by complex formation with organic ligands (e.g., o xalic acid, citric acids, siderophores and other dissolved forms of carbon (C) that can occur in soils ) and inorganic ligands (e.g., phosphate, arsenate, sulphate and sulphide) depending on the soil redox status and ligands availability (Cornell and Schwertmann, 2003) . The interaction be tween dissolved organic matter (DOM) and Fe oxides increases the complexity of adsorption/desorption mechanisms, as DOM is rich in ligands of various size fractions, elemental composition and functional groups. For example , at pH 4, C from a larger hydrop hobic size fraction and aromatic C moieties adsorbed to a greater extent than the C of small er hydrophilic size fractions on the hematite surface has been observed (Gu et al., 1995) . The explanation for that observation was larger hydrophobic fractions contain fewer O containing functional
80 groups because of higher degree of humification. Conversely, the smaller size fraction has more reactive hydroxyl and carboxy lic groups and so a lower amount of the small size fraction was required to cover the Fe oxide surface. Iron can act as a terminal electron acceptor during microb e mediated processes , increas ing Fe mineral solubility and availability in soil solution (Forsmann and Kjaergaard, 2014) . Though Fe oxide dissolution and release of metals like Al into soil solution is well known , the factors behind the release of Al are not well understood for soil systems. The DOM metal complexation followed by hydrologically mediated drop in redox potential could promote loss of Fe , leaving Al more vulnerable for organo complexation and podzolization. T his could explain the dominance of Al over Fe in Bh horizons and associatio n of Spodosols with poorly drained landscapes in the sandy coastal plain of the SE USA. The present study is focused on possible mechanism s of Al release into soil solution from soils containing different amount s of crystalline and non crystalline Fe oxide s using different extracting procedures. The hypotheses considered in this study are as follows: (1) Aluminum release into soil solution inversely relates to crystalline Fe content because surfaces of these oxides mitigate organo Al complexing potential. ( 2) Iron release, as by biochemical reduction, will be accompanied by Al release due to loss of Fe oxide surfaces that can otherwise compete with Al oxides for ligand reactions. Material s and Methods Sites and Soils Three sites located in Florida (Figure 5 1) were chosen for this study, each in sandy soils providing an anticipated range in Fe concentration and crystallinity. The somewhat poorly drained soils in a depression and adjacent to a l ake. All soil materials
81 were collected from E horizons of the sandhill part using auger. Eighteen samples were collected from various depth s (Table 5 1), ten of which from The University of Florida Ordway Swisher Biological Station in Pu tnam C ounty three samples from the University of Florida Natural Area Teaching Laborato r Alachua C ounty and five samples from an agricultural field near the town of Bronson in Levy C o unty These materials were chosen with sand grain coatings with a range of chroma, using chroma as an indicator of Fe oxide presence and abundance. All samples were stored and transferred to laborato r y in Ziploc bags and kept refrigerated until analyzed . Basic Soil Characterization Moisture content of all samples was determined gravimetrically after oven drying. Sample pH was determined using silver/silver chloride reference electrode in double de ioni zed (DDI) water by equilibrating at 1:1 by mass ratio and electrical conductivity (EC) was measured following equilibrati on with DDI in 1:2 ratio by mass and using conductivity meter. All the extractions were done with field moist soils so as not to risk alteration of Fe crystalinity. Non crystalline secondary oxides of Fe and Al were extracted by acid (pH 3) ammonium oxalate (aao) in the dark (McKeague and Day, 1966a) and s odium citrate dithionite bicarbonate (cdb) (Mehra and Jackson, 1960) extraction was done for total crystalline secondary oxides of iron (Fe), respectively. Inductively c oupled p lasma spectroscopy (ICP AES) was used to determined metal content of extracted solution s .
82 Extractions To test the hypothesis that crystalline Fe in soil could inhibit release of Al to soil solution, t hree extraction procedures were used and their relative Al extracting capability compared for these soils containing different total crystalline and non crystalline secondary Fe. Two c hemical extractants , oxalic acid and HCl , were chosen to provide understand ing of well defined system s and soil derived dissolved organic matter (DOM ; pH ~ 4 ) was chosen as a natural complex ing extractant that more closely simulates potential soil reactions (Table 5 2) . Chemical extract ants Extractions were performed using 1mmol HCl or oxalic acid, both having a pH of ~2.9 and EC of ~ 400ÂµS. Oxalic acid was chosen as a surrogate for well defined low molecular weight (LMW) organic acid s that occur naturally in soils at concentrations as high as 1mmol (Fox and Comerfor d, 1990) and that can strongly complex metals. The purpose of HCl extraction was to provide a comparison of its extraction efficacy, as a non complexing inorganic acid influencing metal release exclusively through pH, with that of oxalic acid. Natural e xtractants Aquod surface horizon material was collected at A ustin C ary F orest , along transect 3 of the study reported in Chapter 3, and stored under laboratory condition in closed plastic container. Dissolved OM was separated from that Aquod A horizon mate rial . The soil solution was decanted after incubating the soil with DI water in 1:2 ratio for three days under lab condition and filtered through 0.45Âµm filter paper before addition to soil samples . The pH, EC, dissolved organic C (DOC) , Fe and Al were det ermined for this DOM solution.
83 Extracting Procedure About 5 g of each sample was placed in a 100 mL plastic centrifuge tube containing 50 ml of each extracted solution ( 1 mmol oxalic acid, HCl and DOM), shaken on a reciprocating shaker for 30 minutes, allowed to settle for 30 minutes and filtered through 0.45Âµm filter . Filtrates were analyzed for C via liquid CO 2 analyzer by flash combustion and Fe and Al via ICP. The metal extraction s by DOM were conducted following incubations of 1, 3 and 7 days to c heck Fe and Al release for a range of DOM metal interaction times . Statistical Analysis All statistical analyses were performed using JMP software (version 10, 2013, SAS Institute, Cary, NC). The statistical significance test for difference s between treat Results and Discussion Metal Extraction Dissolution of soil Fe and Al relates to pH , but a ~pH 3 HCl water ( 1 mmol ) mixture failed to extract a detectable amount Fe from any soil of this study. However, it extracted det ectable levels (10 50 mg kg 1 ) of Al from the Ordway samples, for which the Al released by HCl correlated significantly with Al aao (R 2 = 0.51; P < 0.05) . Hence there is no indication from these results that Al released by proton s relates to soil Fe content/ release . However, the pH 3 one mmol oxalic acid extracted a higher amount of Al (Al ox ) than did HCl. The Al ox significantly (R 2 = 0. 48 ; P < 0.00 5 ) related to the fraction of non crystalline Fe ex tracted by the oxalic acid (Figure 5 2 ). The metal s released by oxalic acid extraction w ere much higher for Ordway than other two sites. The DOM also
84 extracted a significantly higher amount of Al and Fe than did HCl (Table 5 3). The greater extraction efficacy of the organic components relative to HCl verif y that organo complexation of metals is a major factor in their release. DOC Adsorption and Metals Release Relations between DOC sorption and concentrations of extractable Fe and Al differ ed between samples from different sites (Figure 5 3 ) and also with interaction time (Figure 5 4 ) . The DOC adsorption for Bronson and NATL soils related significantly to Al aao but not to crystalline o r non crystalline Fe (Figure 5 3 A ). Conversely, the DOC adsorption for Ordway soils related significantly to crystalline Fe but not to Al aao (Figure 5 3 B ). The DOC adsorption on metal oxide surface s likely depend s more on the nature of the oxide s urface and specific surface area than on t he total metal content . A dsor ption of DOC would also be influenced by magnitude of electrostatic attraction arising from pH dependent surface charge. When the metal ligand bond is strong and at the same time metals are stable within the oxide structure th en both metal and ligand remain associated with the solid phase . Conversely, if metals are loosely bound then there is a greater chance for them to be released to the solution in association with DOM, a s the data indicated (Figure 5 5 ), confounding the rel ation between DOM sorption and metal concentrations. The DOM solution could have extracted some metal s from the exchangeable sites or that were loosely bound to the surface. Interaction T ime Fe R eduction L ow redox potentials were a likely consequence of longer incubation time , bringing Fe into solution and possibly enhancing release of the associated Al (Figure 5 4 ). However, release of Al was also significantly higher after longer incubation and related signifi cantly to Fe release for the whole sample s et (Figure 5 5 A ) as well as for samples
85 within i ndividual locations (Figure 5 5B, C ). The proportion of DOM extractable Al was much higher than DOM extractable Fe after the 7 day incubation. Release of the Al substituted in Fe oxides upon reduction would b e trivial relative to Fe and thus could not explain the observed relation between Fe and Al release. The positive relation between Al aao and DOM extractable Al released after 7 days suggest that longer interaction time with DOC ligands and subsequent disso lution from the noncrystalline phase via organo Al complexation are factors in Al release during DOM incubation (see 3.2.3 below) (Figure 5 6 ). Another possible mechanism is that Fe oxide surfaces retain organic ligands sufficiently to reduce the efficacy of the ligands for Al complexation until the surfaces are eliminated via chemical reduction and dissolution of the Fe oxide. Iron and Al released from Bronson NATL soils were much higher than that for Ordway soils under longer ( 7day ) incubation with Al release being ~ 6 times higher than Fe for Ordway (Figure 5 5 B ) and ~14 ti mes for Bronson NATL (Figure 5 5 C ). The1h soil DO M interaction was not sufficient to produce significant difference s between sites in relation to metal released into solution. Poss ible Adsorption Release Mechanism Shorter interaction time was enough for DOC to sorb on solid surfaces depending on the solid phase structures but was not enough to induce immediately the release of Al and Fe from surfaces. Possibly, the DOC was coordinated with surface exposed Fe atoms that were bound strongly to the oxide structure, making the DOC less available or reducing its chelating activity. Hence, t he first necessary step in releasing Fe into solution is to make Fe less stable in its soli d phase by converting Fe I I I to Fe II . T he release could be specifically related to the number of unstable Fe III atoms bound in an
86 oxide phase. A d rop in redox potential could enhance the release of Fe into DOC solution from a set of samples depending on met al stability at a specific oxide site. The change in oxidation state could be the rate limiting step of Fe release into solution. Though Al is not a redox sensitive element, it forms stable complexes with DOC under natural conditions. Release of DOC result ing from Fe oxide dissolution could elevate complexing pressure on Al, bringing it into solution in temporal correspondence with Fe. Summary Though metal extracting efficacy of LMW oxalic acid was greater than that of soil DOM, the latter was more effective than HCl in extracting soil Fe and Al. Therefore, r elease of Fe and Al in extracting solution depends on metal chelating ability of extracting solution rather than pH only. The significant adsorption of DOC in the short term but not after longer incubation times supports the idea that was significantly adsorbed by soil surfac es in the short term supports the idea that stability of metals in their solid phases decreases with longer metal ligand interaction and results in release of metals into extracting solution. Results of this study are consistent with the idea that the pres ence of Fe oxides, but not necessarily their concentration, inhibits organo Al complexation, possibly via surface retention of organic ligands until redox related dissolution. Results are important in explaining response of Al to water table rise or soil r educing condition and thus increased organo metal complexation in flatwood landscapes.
87 T able 5 1 . Description of soils and their p hysicochemical characteristics . EC = electrical conductivity; Fe cdb = citrate dithionate bi carbonate extractable Fe ; Fe aao = acid ammonium oxalate extractable Fe and Al aao = acid ammonium oxalate extractable Al Site Sample # Horizon Depth (cm) Color Moisture (%) pH EC ÂµS Fe cdb Fe aao Al aao mmol kg 1 Ordway 1 E1 40 70 10YR 5/4 4.09 5.9 12.6 10.4 8.58 18.96 2 E2 71 103 10YR 5/6 3.83 6.0 10.5 11.0 8.32 20.05 3 E3 103 125+ 10YR 5/8 3.52 5.8 9.4 9.6 7.86 17.62 4 E1 50 85 10YR 6/6 3.99 5.0 12.3 9.6 8.25 18.58 5 E2 85 124+ 10YR 5/6 3.62 5.2 12.9 8.1 5.08 7.88 6 EA 12 31 10YR 5/4 4.39 4.6 25.5 4.3 5.35 8.32 7 E 31 80+ 10YR 5/6 5.15 4.7 22.3 6.9 8.37 23.08 8 E1 10 40 10YR 4/4 4.46 5.1 19.2 16.3 9.14 27.02 9 E2 Up 40 80 10YR 5/6 3.98 5.4 15.5 16.2 7.60 21.44 10 E2 Low 80 120+ 7.5YR 5/6 3.49 5.0 12.3 16.6 6.75 17.48 Bronson 11 E 50 95 10YR 6/6 3.60 6.6 28.0 10.5 7.09 23.41 12 E 65 105 10YR 7/2 1.03 5.3 28.7 0.6 0.79 18.90 13 E 50 90 10YR 5/4 3.22 4.8 48.3 11.9 8.04 36.66 14 E1 25 100 10 YR 5/4 1.25 6.0 28.9 9.9 7.67 34.42 15 E2 150 215 10YR 6/8 4.43 5.2 28.8 14.2 8.06 41.44 NATL 16 E1 12 97 10YR 6/4 1.68 5.3 14.2 9.6 9.23 19.82 17 E2 97 145 10YR 6/3 2.59 5.6 10.1 7.5 6.24 10.69 18 E 15 27 10YR 5/4 3.04 6.5 28.9 22.0 17.46 23.62
88 T able 5 2. Chemical characterization data for extractants used Extractants pH E 465 /E 665 DOC Fe Al mg L 1 1 mmol oxalic acid 2.99 400 0.00 NA * B D L B D L 1 mmol HCl 2.9 428 0.00 0.00 B D L B D L DOM1 3.8 99 7.74 74.7 0.2 0.36 DOM2 4.2 117.6 4.03 73.1 0.04 0.24 *B D L = Below detection limit DOM1 used for 1 h total extraction time DOM2 used for 1,3, 7 days of total extraction time
89 Table 5 3. Data for Fe and Al extracted by 1 mmol HCl , 1 mmol oxalic acid (ox) , and soil derived dissolved organic matter (DOM ) Sample # Fe HCl Al HCl Fe ox Al ox Fe DOM Al DOM mg kg 1 1 * BDL 10.8 24.1 104.3 9.3 25.8 2 BDL 21.5 27.0 107.0 10.7 27.6 3 BDL 17.1 27.1 95.3 8.4 26.7 4 BDL 23.5 29.5 136.3 7.8 27.0 5 BDL 20.0 25.0 106.1 4.8 23.4 6 BDL 11.6 50.2 86.7 5.3 18.9 7 BDL 35.0 22.3 140.2 5.2 28.2 8 BDL 48.7 17.9 138.0 6.4 32.4 9 BDL 28.9 10.9 91.7 4.8 24.9 10 BDL 29.0 11.1 96.6 4.5 24.1 11 BDL 1.4 19.3 84.6 8.9 27.3 12 BDL 17.1 11.1 32.4 2.5 30.8 13 BDL 24.8 1.8 21.2 3.3 25.8 14 BDL 9.7 2.7 26.8 7.1 25.9 15 BDL 15.5 3.1 21.6 3.4 25.8 16 BDL 15.5 2.2 1.2 8.8 27.3 17 BDL 12.7 4.3 3.4 8.0 25.6 18 BDL 0.2 6.8 2.3 27.6 35.0 *BDL = Below detection limit
90 Figure 5 1. Locations of samples collected for the study in Florida, USA . Bronson, NATL, and Ordway samples were in Levy, Alachua, and Putnam Counties, respectively.
91 Figure 5 2 . Relation between Fe and Al extracted by 1mmol oxalic acid, by locations. Excluding t he circled data point R 2 = 0.73, and P < 0.0001.
92 Figure 5 3 . Relation between DOC adsorption and A ) non crystalline Al and B ) crystalline Fe concentrations during 1 h extraction by sample site. Adsorption of DOC did not relate to Fe aao for either location.
93 Figure 5 4 . DOM extractable A) Al and B ) Fe , as released over time A B
94 Figure 5 5. Relation between DOM extractable Fe and Al re leased after7 d incubation of A) all samples B) Ordway samples and C ) Bronson NATL samples.
95 Figure 5 6 . Relation between acid ammonium oxalate (Al aao ) and DOM extractable Al (Al DOM7 ) after 7 d incubation.
96 CHAPTER 6 SUMMARY AND CONCLUSIONS The processes of podzolization include the association of organic C with Al and Fe and the mutual accumulation of these elements in soil subsurface horizons. Spodosols of the SE USA are generally fo rmed under flatwood ecosystems. They are sandy in texture, associated with fluctuating water tables, and predominantly classified as Alaquods. The dominant vegetation associated with these poorly drained flatwoods includes Pine ( Pinus palustrus ) and Saw palmett o ( Serenoa repens ) species. Abundance of Al relative to Fe and poorly drained conditions are typical features of these flatwood Alaquods. Theories of Spodosol formation are well established for northern Spodosols, which are commonly well drained. However, restriction of southern Spodosols mainly to poor drainage and fluctuating water table conditions has not been thoroughly explained. Furthermore, the movement and redistribution of Al in the podzolization process is somehow triggered by wetness in spite of its redox insensitivity. The research of this dissertation focused on two overarching questions: (i) How is seasonal saturation promoting Al mobilization in the formation of southern Spodosols? (ii) How is direction and magnitude of C flux affected by dept h to water table fluctuations? These questions are pertinent to deep (below the A horizon) soil C sequestratrion as influenced by climate driven changes in hydrologic regime for soils of the SE USA coastal plain. The field study (Chapter 3) probed the rel ation between C , Al and Fe along hydrolo gic transitions from better drained U du lts to poorly drained A quod s and to evaluate mechanistic implications. Iron was found to be a possible inhibitor of podzolization since its deplet ion from drier to wetter soils along the transitions paralleled
97 a progressive strengthening of podzolic features. Iron depletion with increasing wetness is consistent with dissolution of Fe oxides via biogeochemical reduction and subsequent mobilization in Fe 2+ form. Iron shows no relat ion with Al and C in well expressed Bh horizons, whereas Al and C are significantly correlated for Bh horizons collectively. These results indicate that C and Al have undergone joint podzolic vertical redistribution, as by organo Al complexation, whereas F e was mainly depleted preemptively via redox processes. The A horizons had markedly lowe r Al p / C p and C p /C t ra tios than Bh horizons. These ratios can be used to confidently confirm whether the horizon is an A or Bh in cases where the latter is shallow enough to be mistaken for an A. A net increase in total C was observed when calcula ted on a comparable depth basis in the transition from the drier (Paleudult) to wetter (Alaquod) conditions. The potential explanatio n for results of the field study is that stable Fe oxides serve as an inhibitor of organo Al complexation and podzolization via competitive sorption of organic ligands. Fluctuating condition of near surface saturation depletes Fe oxides to a threshold valu e at which they are no longer inhibiting organo Al complexation and mobilization. The Fe remaining at this position is mobilized by organo complexation along with Al to form shallow, weakly expressed Bh horizons. However, with increased wetness Fe has been thoroughly depleted and does not participate in podzolic redistribution. The field study provided only circumstantial evidence for Fe serving as inhibitor of organo Al complexation; further exploration of that possibility (Chapter5) is summarized below. T he podzolization process in these soils involves clay as well as organo metal eluviation s .
98 The primary objective of the column study (Chapter 4) was to evaluate influence of fluctuating water table depth on direction and magnitude of C flux and on redist ribution of metals and < 50 Âµm components. This study revealed that the same soil materials when subjected to different fluctuating wat er table conditions can respond differently in releasing C to the atmosphere and as DOC . B io c hemical responses to the mo isture change s could be the primary f acto r responsible for C flux differences . The C mineralization rate was consistently higher for DWT treatments than SWT. With d ecreasing readily decomposable materials in columns there was a slight decrease in atmosphe ric CO 2 release with time. A production of significantly higher DOC in SWT soil pore wat er than DWT was observed in the column study. However, the net release of C into the atmosphere was less for SWT columns, favoring net C accumulation. The shallow sat urated condition also provided an opportunity for higher C metal interaction which was reflected in a significant C metal relation in the SWT pore water. The fluctuating SWT condition significantly increases downward movement of < 50 Âµm material. The sign ificant relation between pyrophosphate C and < 50 Âµm suggests that the downward movement of C was also influenced by fluctuating water table. Results of this study imply that the potential for downward flux of C and either its accumulation in the subsurfa ce (e.g., as a Bh horizon) or less to groundwater and streams is promoted by near surface water table fluctuation. The final study was conducted to explore plausible mechanism s of Al release into soil solution from sandy soil horizons containing different amount s of crystalline and non crystalline Fe oxides . This study placed specific emphasis on the prospect of Fe
99 oxides acting as inhibitors of Al release via organo complexation. Samples were extracted with a complexing organic acid (oxalic) , a strong inor ganic acid (HCl) of equivalent concentration, and soil DOM . A lso, metal release and C sorption were determined after incubation with DOM over time . Organic complexing agents (oxalic and DOM) were consistent ly more effective in releasing Fe and Al than HCl , verifying that organo metal complexation is a major factor in metal release relative to pH, per se. Significantly higher metal release was induced with longer (7 day) incubation than with shorter times (up to 3 days). Iron release after longer incubation may relate to the biogeochemical reduction of Fe oxides. However, Al release was several times greater than Fe release with 7 day incubation even though Al is not directly redox sensitive, suggesting that Fe dissolution may somehow foster Al release. Corre lated release of Al release by this mechanism would be much less than what was observed. Therefore any linkage between Fe and Al release would have to relate to a di fferent mechanism. Longer interaction between Al and DOM could be another factor promoting release of Al; Al aao was significantly related to Al in extracting solution. Release of Fe from its oxide surface could have resulted in increased activity of solubl e organo complexing components in the DOM and subsequent increased Al complexation pressure on Al. Results of the dissertation research suggest a modified Fe inhibition mechanism whereby Fe could initially inhibit Al release but once Fe is destabilized from its oxide surface it actually enhances Al release into extracting solution. Abundant well organized crystalline Fe could adsorb strong ly complexing ligands and reduce their solution activity , thereby inhibiting their complexation and mobilization of Al. A t the same time C -
100 chain s associated with these ligands become more vulnerable to mineralization when exposed to soil biological activity , result ing in increase d atmospheric C flux . As long as Fe remains abundant and well crystalline it can serve to inhibit complexation and mobilization of Al. However, o nce crystalline Fe is transformed to amorphous Fe by fluctuating WT its ligand sorption capacity is reduced or eliminated thus favori ng release of complexed Fe in to solution . These complexes can be los t from the system or end up in the incipient Bh. Aluminum r elease d by dilute organic acids could come from three different sites, (1) easily available exchangeable or surface Al, (2) diffi cultly available amorphous Al with some Fe III substitution and (3) very tightly bound crystalline Al. The exchangeable surface Al would be stable in an acidic soil environment unless it becomes chelated. Replenishment of the exchangeable Al fraction can oc cur from amorphous and other pools of soil Al. Exchangeable Al could vary with geographic location, possibly explaining geographic distinctions observed in this research. Unlike Al, exchangeable/surface Fe III is ephemeral as it can be easily transformed to Fe II and lost from the soil system and thus release of the exchangeable Al is independent of Fe release. However, amorphous Al phases could have substituted Fe, making the surface vulnerable to dissolution after loss of Fe by reduction. The degree of such substitution could vary with geographic locations of soils. During soil OM decomposition organic chelating agents, including LMW organic acids and long chain fulvic acids / humic acids, are released. The high affinity of Al for orgno complexation renders inorganic Al species to be negligible in soil solution. The chelating functional groups of DOM first form complexes with readily available Al and Fe in exchangeable form, after which chelating substances get adsorbed to metal oxide surfaces. The lower Fe O bond energy relative
101 to that of Al O and the susceptibility of Fe to chemical reduction results in amorph ous Fe being released first and weakening the entire Al structure. Loss of Fe and the high stability of organo Al complexes result in partial structur e collapse and release of Al into solution. As mentioned previously, the Fe/Al ratio in this amorphous structure can differ with soils geographical position and that makes a difference in the concentration of these metals released into soil solution. So me unexpected but consistent observations made during this dissertation research generated new research questions. The observations for both column and batch studies that release of Fe and Al into solution are significantly correlated, but yet release is m uch higher for Al, spawned several questions of relevance to the overall understanding of biogeochemical influences on these metals: (i) Is there an amorphous phase with Fe and Al sharing same surface structure but with different metal ratios? (ii) Is ther e significant Fe substitution for Al in the amorphous Al oxide phases? (iii) Do longer organo metal interactions favor Al over Fe, bringing more Al into solution independent of soil Fe concentrations? (iv) Is fluctuating SWT promoting complexation of both metals but organo Fe complexes are more susceptible to microbial decomposition relative to but organo Al complexes due to Al toxicity of the latter? (v) Is the low Fe concentration in solution due to lowering of proton activity with decreasing Eh reducing Fe release and organo Fe complexation? Another observation was that the E 4 /E 6 is higher in SWT pore water than DWT pore water, giving rise to the question of whether SWT conditions promote DOM more prone to Al than Fe complexation. The three related studie s of this dissertation were complementary in explaining mechanisms by which water saturation and organo metal interactions influence C and
102 metal transport and subsurface C sequestration as Bh horizons in sandy coastal plain soils. The field study provided consistent evidence of increasing Fe loss with decreasing depth of water table fluctuation and increasing podzolic C accumulation. The column study confirmed effects of fluctuating water table depth on direction and magnitude of C flux under controlled con ditions that would not be possible at field scale but that were consistent with field observations. It also verified that shallower water table fluctuation is more effective in promoting higher organo metal interaction and downward movement of C and other colloidal and non colloidal materials. Relationships between Al and Fe release were revealed in the final study via batch extractions. These relationships suggested a linkage between Fe and Al release with increased wetness similar to what could be inferre d from results of the field study. A significant relation between Fe and Al for weakly expressed (incipient, drier) Bh horizons along with higher Fe content of better drained soils that were sampled during this research provide evidence for overlap of redo ximorphic and podzolic processes along these drainage transitions (Figure 6 1). Results of this dissertation are pertinent to effects of hydrologic change on C and metal distributions in soil s of S.E. USA coastal plain landscapes.
103 Figure 6 1. Schematic e xample of a landscape ecological transition from better drained sandhill (left) to poorly drained flatwoods (right), showing a soil Fe depletion gradient, water table position and a downward strengthening Bh. A possib le explanation of this transition is that h igher crystalline Fe content of better drained sandhill soils may inhibit podzolic mobilization of Al by reducing its exposure to destabilizing components of DOM produced i n the soil. Incipient podzolization occur s at a threshold where the inhibitory effect of Fe on Al complexation is lost such that Al can be mobilized. Some Fe is still present in this incipient zone and undergoes podzolic redistribution along with Al. With relative rise in water table (moving tow ards the surface) and higher production of DOM in the flatwoods , Fe loss is nearly complete . Hence, subsurface accumulation of C via podzolization in strongly expressed Bh horizons occurs mainly in association with Al.
104 LIST OF REFERENCES Abrahamson, W.G., Hartnett, D.C., 1990. Pine flatwoods and dry prairies. In: R.L. Myers, J.J. Ewel (Eds.), Ecosystems of Florida. Univ. of Central Florida Press, Orlando, FL, pp. 143 149. Anderson, J.P.E., 1994. Soil respiration. In: A.L. Page, R.H. Miller, D.R. Keeney (Eds.), Methods of Soil Analysis, Part 2: Chemical and microbiological properties. ASA SSSA Inc., Madison, WI, pp. 831 866. Andersson, S., Nilsson, S.I., Saetre, P., 2000. Leaching of dissolved organic carbon (DOC) and dissol ved organic nitrogen (DON) in mor humus as affected by temperature and pH. Soil Biol. & Biochem. 32 (1), 1 10. Aran, D., Gury, M., Jeanroy, E., 2001. Organo metallic complexes in an Andosol: a comparative study with a Cambisol and Podzol. Geoderma 99 (1 2) , 65 79. Banik, C., Harris, W.G., Ellis, R., Hurt, W., Balboa, S., 2014. Relations of iron, aluminum, and carbon along transitions from Udults to Aquods. Geoderma 226 227, 332 339. Bardy, M., Bonhomme, C., Fritsch, E., Maquet, J., Hajjar, R., Allard, T., Derenne, S., Calas, G., 2007. Al speciation in tropical podzols of the upper Amazon Basin: A solid state 27Al MAS and MQMAS NMR study. Geochimica et Cosmochimica Acta 71, 3211 3222. Boudot, J.P., 1992. Relative efficiency of complexed aluminum, noncrystall ine A1 hydroxide, allophane and imogolite in retarding the biodegradation of citric acid. Geoderma 52, 29 39. Boudot, J.P., Bel Hadj, B.A., Chone, T., 1986. Carbon mineralization in Andosols and Aluminium rich highland soils. Soil Biol. Biochem 18, 457 461 . Boudot, J.P., Bel Hadj, B.A., Steiman, R., Siegle Murandi, F., 1989. Biodegradation of synthetic organo metallic complexes of iron andaluminum with selected metal to carbon ratios. Soil Biol. Biochem 21, 961 966. Brady, M., Derenne, S., Allard, T., Bened etti, F.M., Fritsch, E., 2011. Podzolisation and exportation of organic matter in black waters of the Rio Negro (upper Amazon basin, Brazil). Biogeochemistry 106, 71 88. Brasfield, J.F., Carlisle, V.W., Johnson, R.W., 1973. Spodosols Soils with a spodic ho rizon. In: S.W. Buol (Ed.), Soils of the Southern States and Puerto Rico. 174 of Southern Cooperative series bulletin. USDA, pp. 57 60. Buckau, G., Artinger, R., Fritz, P., Geyer, S., Kim, J.I., Wolf, M., 2000. Origin and mobility of humic colloids in the Gorleben aquifer system. Appl. Geochem. 15, 171 179.
105 Chen, Z., Cai, Y., Liu, G., Solo Gabriele, H., Snyder, G.H., Cisar, J.L., 2008. Role of soil derived dissolved substances in arsenic transport and transformation in laboratory experiments. Sci. Total. En viron 406, 180 189. Chivers, M.R., Turetsky, M.R., Waddington, J.M., Harden, J.W., McGuire, A.D., 2009. Effects of experimental water table and temperature manipulations on ecosystem CO 2 fluxes in an Alaskan rich fen. Ecosystems 12, 1329 1342. Chow, A.T., Tanji, K.K., Gao, S., R.A, D., 2006. Temprature, water content and wet dry cycle effects on DOC production and carbon mineralization in agric ultural peat soils. Soil Biol. & Biochem. 38, 477 488. Coelho, M.R., Vidal Torrado, P., Perez, X.L.O., Martins, V.M ., Vazquez, F.M., 2010. Selectivity of sodium pyrophosphates and unbuffered chlorides (CuCl 2 and LaCl 3 ) as extractants of aluminum bound to soil organic matter in soils of the sandy coastal plain in Sao Paulo state. Revista Brasileira de Ciencia Do Solo 34 , 1561 1572. Collins, M.E., 1990. Aquods. In: J.M. Kimble, R.D. Yeck (Eds.), Proceedings of the fifth international soil correlation meeting, characterization, classification, and utilization of spodosols. USDS NRCS, U.S. Govt. Printing Office, Washington, DC, pp. 105 122. Cornell, R.M., Schwertmann, U., 2003. The iron oxides. Wiley VCH GmbH &Co. KGaA, Weinheim. Cox, A.C., Gordon, D.R., Slapcinsky, J.L., Seamon, G.S., 2004. Understory restoration in longleaf pine sandhills. Natural Areas J. 24, 4 14. Daniel s, R.B., Gamble, E.E., Holzhey, C.S., 1975. Thick Bh horizons in the North Carolina coastal plain: I. Morphology and relation to texture and soil ground water. Soil Sci. Soc. of Am. Proc. 39, 1177 1181. DeConinck, F., 1980. Major mechanisms in formation of spodic horizons. Geoderma 24, 108 128. Desphande, T.L., Greenland, D.J., Quirk, J.P., 1968. Changes in soil properties associated with the removal of iron and aluminium oxides. J. Soil Sci. 19, 108 122. Ekstrom, E.B., Learman, D.R., Madden, A.S., Hansel, C.M., 2010. Contrasting effects of Al substitution on microbial reduction of Fe(III) (hydr)oxides. Geochimica et Cosmochimica Acta 74, 7086 7099. Farmer, V.C., Russell, J.D., Berrow, M.L., 1980. Imogolite and proto imogolite allophane in Spodic Horizons:Ev idence for a mobile aluminium silicatecomplex i n Podzol formation. Soil Sci. 31, 673 684.
106 Fierer, N., Schimel, J.P., Holden, P.A., 2003a. Controls on microbial CO 2 production: a comparison of surface and subsurface soil horizons. Global Change Biol 9, 1322 1332. Fierer, N., Schimel, J.P., Holden, P.A., 2003b. Influence of drying rewetting frequency on soil bacteria l community structure. Microb. Ecol. 45, 63 71. Forsmann, D.M., Kjaergaard, C., 2014. Phosphorus release from anaerobic peat soils during convective discharge Effect of soil Fe:P molar ratio and preferential flow. Geoderma 223 225, 21 32. Fox, T.R., Comerford, N.B., 1990. Low molecular weight organic acids in selected forests soils of the southeastern USA. Soil Sci. Soc. Am. J. 54, 1139 11 44. Franzmeier, D.P., Hajek, B.F., Simonson., C.H., 1965. Use of amorphous material to identi fy spodic horizons. Soil Sci. Soc. of Am. Proc. 29, 737 743. Garman, C.R., Carlisle, V.W., Zelazny, L.W., Beville, B.C., 1981. Aquiclude related spodic horizon dev elopment. Soil and Crop Sci. Soc. Florida Proc. 40, 105 110. Gorham, E., 1991. Northern peatlands: Role in the carbon cycle and probable responses to climatic warming. Ecol. Appl. 1, 182 195. Gu, B., Schmitt, J., Chen, Z., Liang, L., McCarthy, J.F., 1995. Adsorption and desorption of different organic matter fractions on iron oxide. Geochimica et Cosmochimica Acta 59, 219 229. Guillet, B., Rouiller, J., Souchier, B., 1975. Podzolization and clay migration, in: Spodosols of eastern France. Geoderma 14, 223 2 45. Hansson, K., Kleja, D.B., Kalbitz, K., Larsson, H., 2009. Amounts of carbon mineralised and leached as DOC during decomposition of Norway spruce need les and fine roots. Soil Biol. & Biochem. 42 (2), 178 185. Harris, W.G., 2001. Hydrologically linked Sp odosol formation in the Southeastern United States. In: J.L. Richardson, M.J. Vepraskas (Eds.), Wetland soils: their genesis, hydrology, landscape, and separation into hydric and nonhtdric soils. CRC Press Boca Raton, FL, pp. 331 342. Harris, W.G., Carlisl e, V.W., Chesser, S.L., 1987a. Clay Mineralogy as related to morphology of Florida soils with sandy epipedons. Soil Sci. Soc. Am. J. 51, 1673 1677. Harris, W.G., Carlisle, V.W., Rees, K.C.J.V., 1987b. Pedon zonation of hydroxy interlayered minerals in Ulti c Haplaquods. Soil Sci. Soc. Am. J. 51, 1367 1372.
107 Harris, W.G., Crownover, S.H., Comerford, N.B., 1995. Experimental formation of Aquod like features in sandy coastal plain soils. Soil Sci. Soc. Am. J. 59, 877 886. Harris, W.G., Crownover, S.H., Hinchee, J., 2004. Problem arising from fixed depth assessment of deeply weathered sandy soils. Geoderma 126 (1 2), 161 165. Harris, W.G., Ellis, R., Hurt, W., 2013. Reconciling podzolic and redoximorphic responses to seasonal saturation in sandy coastal plain soil s, Final report submitted to USDA NRCS at completion of funded project. Contract number: 68 7482 10 532. Harris, W.G., Hollien, K.A., 1999. Changes in quantity and composition of crystalline clay across E Bh boundaries of Alaquods. Soil Science 164, 602 608. Harris, W.G., Hollien, K.A., 2000. Changes across artificial E Bh boundaries formed under simulated fluctuating water tables. Soil Sci. Soc. Am. J. 64, 967 973. Harris, W.G., Rischar, C.A., 2012. Factors related to Bh horizon depth for artificial and natural E Bh sequences. Geoderma 189 190, 502 507. Hodgkinson, A., 1977. Oxalic acid in biology and medicine. Academic Press, London. Holzhey, C.S., Daniels, R.B., Gamble, E.E., 1975. Thick Bh horizons in North Carolina Coastal Plain : II.Physical and chem ical properties and rates of organic additions from surface sources. Soil Sci. Soc. Am. J. 39, 1182 1187. Johnson, D.W., Murphy, J.D., Rau, B.M., Miller, W.W., 2011. Subsurface carbon contents: some case studies in forest soils. Forest Science 57, 3 10. Jo ner, E.J., Eldhuset, T.D., Lange, H., Frostegard, A., 2005. Changes in the microbial community in a forest soil amendedwith aluminium in situ. Plant and Soil 275, 295 304. Knoche, W., LÃ³pez Quintela, M.A., 1983. Stability constants of Al(III) complexes in aqueous solution obtained from kinetic measurements. Thermochimica Acta 62, 295 306. Kononova, M.M., 1966. Soil organic matter. Pergamon Press, Oxford, NY, pp. 400 404. Lamparter, A., Bachmann, J., Goebel, M. O., Woche, S.K., 2009. Carbon mineralization in soil: Impact of wetting drying, aggregation and water repellency. Geoderma 150, 324 333. Lin, C.W., Hseu, Z.Y., Chen, Z.S., 2002. Clay mineralogy of Spodosols with high clay contents in the subalpine forests of Taiwan. Clay and Clay Minerals 50 (6), 726 7 35.
108 Lundstrom, U.S., van Breemen, N., Bain, D., 2000. The podzolization process. A review. Geoderma 24, 91 107. Maynard, J.J., Dahlgren, R.A., O ' Geen, A.T., 2011. Soil carbon cycling and sequestration in a seasonally saturated wetland receiving agricultura l runoff. Biogeosciences 8, 3391 3406. McKeague, J.A., Day, D.H., 1966a. Dithionite and oxalate extractable Fe and Al as aids in differentiating various classes of soils. Canadian J. Soil Sci 46, 13 22. McKeague, J.A., Day, D.H., 1966b. Dithionite and oxal ate extractable Feand Al as aids in differentiating various classes of soils. Canadian J. Soil Sci . 46, 13 22. McKeague, J.A., DeConinck, F., Franzmeier, D.P., 1983. Spodosols. In: L.P. Wilding, N.E. Smeck, G.F. Hall (Eds.), Pedogenesis and Soil Taxonomy I I. The Soil Orders. Elsevier, New York, pp. 217 252. Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by dithionite citrate system buffered with sodium bicarbonate, Proc. 7th Natl. Conf. On Clay and Clay Minerals. Permagon Press, New York, pp. 317 327. Nelson, D.W., Sommers, L.E., 1996. Total carbon, organic carbon, and organic matter. In: D.L. Sparks (Ed.), Methods of Soil Analysis, Part 3 Chemical Methods. Soil Sci. Soc. of Am. Inc., Madison, WI, pp. 961 1010. Olivas, C.P., Oberb auer, F.S., Tweedie, C.E., Oechel, C.W., Kuchy, A., 2010. Responses of CO 2 flux components of Alaskan Coastal Plain tundra to shifts in wat er table. J. of Geophy. Res. 115 (G00I05). Olphen, H.V., 1963. An intoduction to clay colloid chemistry. Interscience Publishers, NY. Petersen, L., 1976. Podzols and Podzolization. Royal Veterinary and Agric. Univ., Copenhagen, Denmark. Petersen, L., 1984. The podzol concept. In: P. Buurman (Ed.), Podzols. Van Nostrand Reinhold Co, New York, NY, pp. 12 20. Pohlman, A.A., McColl, J.G., 1988. Soluble organic from forest litter and thei role in metal dissolution. Soil Sci. Soc. Am. J. 52, 265 271. Qualls, R.G., Richardson, C.J., 2008. Carbon cycling and dissolved organic matter export in the northern Everglades. . In: R. C.J (Ed.), The Everglades Experiments. Springer, NY. Rav Acha, C., Rebhun, M., 1992. Binding of organic solutes to dissolved humic substances and its effect on adsorption and transport in the environment. Water Qual. Res. J. Can 38, 267 281.
109 Rousk, J., BaÂ°aÂ°t h, E., Brookes, C.P., Lauber, L.C., Lozupone, C., Caporaso, G.J., Knight, K., Fierer, N., 2010. Soil bacterial and fungal communities across a pH gradien t in an arable soil. Int. Soc. for Microb. Ecol. 1340 1351. Salm, J.O., Kimmel, K., Uri, V., Mander, U. , 2009. Global warming potential of drained and undrained peatlands in Estonia: A Synthesis. Wetlands 29 (4), 1081 1092. Schaetzl, R., Harris, W.G., 2011. Spodosols. In: P.M. Huang, Y. Li., M.E. Sumner (Eds.), Handbook of Soil Science. CRC Press, New York, pp. 33 113 to 133 127. Seo, D.C., deLaune, R.D., 2010. Effect of redox condition on bacterial and fungal biomass and carbon dioxide production in Louisiana coasta l swamp forest sediment. Sci. of the Total Environ. 408, 3623 3631. Singh, B., Gilkes, R.J., 1992. properties and distribution of iron oxides and their association with minor elements in the soils of south western Australia. J. Soil Sci. 43, 77 98. Singh, B.K., Bardgett, R.D., Smith, P., Reay, D.S., 2010. Microorganisms and climate change: terrest rial feedbacks and mitigation options. Nature Reviews Microbiology 8, 779 790. Soil Survey Staff, 2004. Soil survey investigation report No. 42. , Lincoln, NE. Stevenson, F.J., 1967. Organic acids in soil. In: A.D. MvLaren, G.H. Pete rson (Eds.), Soil Bioch em . Marcel Dekker, New York, pp. 119 146. Stone, E.L., Harris, W.G., Brown, R.B., Kuehl, R.J., 1993. Carbon storage in Florida Spodosols. Soil Sci. Soc. of Am. J. 57, 179 182. Stumn, W., Furrer, G., Wieland, E., Zinder, B., 1985. The Effects of Complex For ming Ligands on the Dissolution of Oxides and Aluminosilicates. In: J.I. Drever (Ed.), The chemistry of weathering. D.Reidel, Dordrecht, The Netherlands, pp. 55 74. Tan, H.T., 2003. Humic Matter in Soil and the Environment: Principles and Controversies. De kker, M. Inc., Basel, Switzerland. Tan, Z., Harris, W.G., Mansell, R.S., 1999. Watertable dynamics across an Aquod Udult transition in Florida flatwoods. Soil Science 164, 10 17. Thompson, A., Chadwick, O.A., Rancourt, D.G., Chorover, J., 2006. Iron oxide crystallinity increases during soil redox oscillations. Geochimica et Cosmochimica Acta 70, 1710 1727. Thompson, C.H., 1992. Genesis of Podzols on coastal dunes in southern Queensland: I. Field relationships and profile morphology. Austr. J. Soil Res 30, 5 93 613.
110 Tipping, E., Berggren, D., Mulder, J., Woof, C., 1995. Modelling the solid solution distribution of protons, aluminium, base cations and humic substances in acid soils. European J. of Soil Sci. 46, 77 94. Ugolini, F.C., Dawson, H., Zachara, J., 1 977. Direct evidence of particle migration in the soil solution of a podzol. Science 198, 603 605. Van Hees, P.A.W., LundstrÃ¶m, 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. Vicca, S., Janssens, I.A., Flessa, H., Fiedler, S., Jungkunst, H.F., 2009. Temperature dependence of greenhouse gas emissions from three hydromorphic soils at differen t groundwater lev els. Geobiol. 7, 465 476. Watts, F.C., Collins, M.E., 2008. Soils of Florida. Book and Multimedia Publishing Committee, Madison, WI. Wickman, T., 1996. Weathering assesment and nutrient availability in coniferous forest. Dissertation Thesis, Royal Institut e of technology, Stockholm. Wright, J.R., Schnitzer, M., 1963. Metallo organo interactions associated with podzolization. Soil Sci. Soc. of Am. Proc. 27, 171 176. Wright, W.R., Foss, J.E., 1968. Movement of silt sized particl es in sand columns. Soil Sci. Soc. of Am. Proc. 32, 446 448. Yamamoto, A., Hirota, M., Suzuki, S., Oe, Y., Zhang, P., Mariko, S., 2009. Effects of tidal fluctuations on CO 2 and CH 4 fluxes in the littoral zone of a brackish water lake. Limnology 10, 229 237. Yuan, T.L., 1966. Characteri zation of surface and spodic horizons of some spodosols. Soil Crop Sci. Soc. Florida Proc. 26, 163 171. Zahang, Y., Kallay, N., Matijevic, E., 1985. Interactions of metal hydroxides with chelating agents. 7. Hematite oxalic and citric acid systems. Langmui r 1, 201 206.
111 BIOGRAPHICAL SKETCH Chumki Banik was born on 1983 in Raiganj , West Bengal, India. She received a Bachelor of Science degree in chemistry with honors from Department of Chemistry in 2005 and Master of Science degree from the Department of Agricultural Chemistry and Soil Science in 2007, both from University of Calcu tta, India. She began her Ph.D. study in the Department of Soil and Water Science, University of Florida, in 2010 and received her degree in 2014. Since 2010, she has worked as research assistant in Pedology Mineralogy lab at the UF campus in Gainesville, FL. She conducted research on the mechanistic linkage between water saturation and organo metal interactions in sandy 2010 2011 and 2012 2013 from the University of Florid a. She received V. W. Carlisle Fellowship from the department in the year 2011 2012. She received travel grants to attend international annual meetings. She is a member of the American Society of Agronomy (ASA), Crop Science Society of America (CSSA), and Soil Science Society of America (SSSA). She has presented several posters and talks about her research at the ASA CSSA SSSA International Annual Meetings and in the Graduate Research Forum, Department of Soil and Water Science, University of Florida.