<%BANNER%>

Genotypic and Forest Management Effects on Size-Density Fractionation of Soil Carbon in a Forested Spodosol

HIDE
 Title Page
 Dedication
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 List of abbreviations
 Abstract
 Introduction
 Effects of forest management on...
 Soil aggregation and aggregate...
 Genotypic and forest management...
 Summary and conclusions
 Appendix A: Experimental proto...
 Appendix B: Aggregates observed...
 References
 Biographical sketch
 

PAGE 1

1 GENOTYPIC AND FOREST MANAGE MENT EFFECTS ON SIZE-DENSITY FRACTIONATION OF SOIL CARB ON IN A FORESTED SPODOSOL By DEOYANI VINAYAK SARKHOT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

2 Copyright 2006 by Deoyani Vinayak Sarkhot

PAGE 3

3 This work is dedicated to my grandfather and to Dr. George Washington Carver, who taught me the worth and fun of worki ng with plants and soil.

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor, Dr. Nick Co merford, for all his help during the last four years. He taught me the meaning and the scie ntific rigor of the terms objectives and conclusions. He was always there to talk wh en I was confused, to encourage when I was nervous (before the seminars and the qualifying oral exam, for example) and to coax when I was lazy. I wish to thank Adriana and him for their ho me-like care when I was hit by a car. It really made a difference. My co-advisor, Dr. Eric Joke la, was the first person I met, when I came in December 2002 to find the campus nearly deserted because of Christmas. He not only welcomed me warmly, but also took me to the Social S ecurity office and helped me through the paperwork. It was perhaps a little thing for him, but becau se I was so far away from home (and it was my birthday), it meant a lot. I wish to express my gratitude to him for all his encouragement and support continued throughout this time. I would lik e to thank Dr. Willie Harris for his help with the scanning electron microscope and also for the fun during the pedology course, when I was his teaching assistant. I also want to thank Dr. Wendell Cr opper and Dr. Yuncong Li, my committee members, for their time and contributions in this wor k. I would like to thank Dr. Jim Reeves from Maryland, for the work on DRIFTS. It added a lot of scientific value to my thesis and whetted my appetite for more spectroscopic work. The Forest Biology Research C ooperative deserves a special thank you, as they provided me the funding, the study site to work on managed by International Paper C o. and also four great annual meetings, one of them in the midst of a b eautiful forest in Texas, a place I will always remember. I also wish to acknowledge the Major Analytical Instrumentat ion Center, Department of Materials Science and Engineer ing, University of Florida, for the use of the SEM. I had always wanted to work on the scanning electron mi croscope and it was great to be able to see it and work on it.

PAGE 5

5 I would like to thank Dave Nolle tti from school of forest res ources and conservation for his help with the C and N analysis. I also wish to thank Mary McCloude, Sally Wu and Aja Stoppe, our lab managers during the last four years, and all my lab mates for their support. Last but not least, I would like to th ank my parents and my younger brothe r Yogesh for their love and support even though from far away.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 LIST OF ABBREVIATIONS........................................................................................................12 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 INTRODUCTION................................................................................................................. .15 Pine Plantations in the Southeastern United States................................................................16 Need for Carbon-wise Management.......................................................................................16 Challenges and Opportunities in Sandy Soils.........................................................................17 2 EFFECTS OF FOREST MANAGEMENT ON SOIL CARBON AND NITROGEN IN A NORTH FLORIDA SANDY SPODOSOL........................................................................21 Introduction................................................................................................................... ..........21 Materials and Methods.......................................................................................................... .23 Experimental Site............................................................................................................23 Laboratory Methods........................................................................................................25 Statistical Analysis..........................................................................................................2 6 Results........................................................................................................................ .............27 Dry vs. Wet Sieving........................................................................................................27 Characterization of Size Fractions...................................................................................28 Effect of Management Intensity and Soil Depth.............................................................29 Discussion..................................................................................................................... ..........29 Dry vs. Wet Sieving........................................................................................................30 Distribution of C and N among Size Fractions...............................................................30 Impact of Management Intensity.....................................................................................32 Conclusions.................................................................................................................... .........33 3 SOIL AGGREGATION AND AGGREG ATE CARBON IN A FORESTED SOUTHEASTERN COASTAL PLAIN SPODOSOL...........................................................44 Introduction................................................................................................................... ..........44 Materials and Methods.......................................................................................................... .46 Experimental Site............................................................................................................46 Laboratory Methods........................................................................................................48 Statistical Analysis..........................................................................................................5 0

PAGE 7

7 Results........................................................................................................................ .............51 Aggregate Morphology...................................................................................................51 Quantifying Organic C in Aggregates.............................................................................51 Effect of Management Intensity......................................................................................53 Discussion..................................................................................................................... ..........53 Aggregate Morphology, Stab ility and OM Content........................................................53 Effect of Management Intensity......................................................................................55 Methodological Considerations.......................................................................................55 Aggregate Structure in Coastal Plain Spodosols Additional Considerations...............56 Conclusions.................................................................................................................... .........58 4 GENOTYPIC AND FOREST MANAGE MENT EFFECTS ON SIZE-DENSITY FRACTIONATION OF SOIL CARB ON IN A FORESTED SPODOSOL..........................69 Introduction................................................................................................................... ..........69 Materials and Methods.......................................................................................................... .71 Experimental Site............................................................................................................71 Laboratory Methods........................................................................................................73 Size fractionation......................................................................................................73 Density fractionation................................................................................................73 Sonication.................................................................................................................74 Statistical Analysis..........................................................................................................7 5 Results........................................................................................................................ .............75 Management Intensity and Family Effects......................................................................75 Distribution of C and N in the Size-density Fractions....................................................76 Effect of Depth................................................................................................................ 77 Discussion..................................................................................................................... ..........77 Effects of Family.............................................................................................................7 8 Fraction Characteristics and Effect of Depth..................................................................79 Methodological Considerations.......................................................................................80 Conclusions.................................................................................................................... .........81 5 SUMMARY AND CONCLUSIONS.....................................................................................92 Methodological Contributions................................................................................................92 Aggregation and Physical Protection......................................................................................93 Influence of Management Intensity and Family.....................................................................94 Active and Passive C Pools Identified In This Study.............................................................96 Additional Research Needs.....................................................................................................9 8 Modeling C Dynamics............................................................................................................ 98 APPENDIX A EXPERIMENTAL PROTOCOLS.......................................................................................103 B AGGREGATES OBSERVED IN THE SANDY SPODOSOLS.........................................108

PAGE 8

8 LIST OF REFERENCES............................................................................................................. 114 BIOGRAPHICAL SKETCH.......................................................................................................121

PAGE 9

9 LIST OF TABLES Table page 2.1 The effect of wet and dry sieving on C content (g C in size fraction per 100g of whole soil) in soil size fractions for a sandy Spodosol in north Florida............................34 2.2 The distribution of N concentration and content among the soil size fractions for a sandy Spodosol in north Florida........................................................................................35 2.3 Discriminant analysis of the DRIFTS sp ectra for a sandy Spodosol in north Florida.......36 3.1 Amount of organic C held in soil aggregat es for a sandy Spodosol in north Florida........60 3.2 Energy output of the sonicator probe and the amount of organic matter lost from each soil size fraction as aggregate orga nic matter (2000 to 250; 250 to 150 and 150 to 53 m) for each energy level for a sandy Spodosol in north Florida............................61 3.3 Effects of forest management intensity a nd soil depth on aggregate organic matter in the 2000 to 250, 250 to 150 and 150 to 53 m fractions for a sandy Spodosol in north Florida........................................................................................................................ ........62 4.1 Characteristics of the size fractions for a sandy Spodosol in north Florida.......................82 4.2 Characteristics of the density fractions for a sandy Spodosol in north Florida.................83 4.3 Distribution of aggregate organic matte r (AOM) and particulate organic matter (POM) in the different size-density fractio ns for a sandy Spodosol in north Florida........84 A.1 Amplitude and time combinations an d energy outputs used for this study.....................106

PAGE 10

10 LIST OF FIGURES Figure page 2.1 Carbon concentrations (% of fraction) of size fractions as affected by the dry and wet sieving for a sandy Spodosol in north Florida...................................................................37 2.2 The interaction between fraction size and sieving method on the ratio of C to organic matter (C:OM) for a sandy Spodosol in north Florida.......................................................38 2.3 Diffuse Reflectance Infrared Fourier Tran sform Spectroscopy spect ra showing effect of sieving method (150 to 53 and <53 m frac tions, low intensity, 5 to 10 cm depth) for a sandy Spodosol in north Florida................................................................................39 2.4 Diffuse Reflectance Infrared Fourier Tran sform Spectroscopy spect ra showing effect of fraction size (dry sieving, high intensity, 0 to 5 cm depth) for a sandy Spodosol in north Florida.................................................................................................................. .....40 2.5 Effect of management intensity and so il depth on the C content of the soil size fractions for a sandy Spodosol in north Florida.................................................................41 2.6 Effect of management intensity and soil depth on N content (g N in a soil fraction per 100g of whole soil) in soil size fractions for a sandy Spodosol in north Florida........42 2.7 Diffuse Reflectance Infrared Fourier Tran sform Spectroscopy spectra showing effect of intensity (2000 to 250 m fraction, dry sieving, 0 to 5 cm depth) for a sandy Spodosol in north Florida...................................................................................................43 3.1. Observations of soil aggregation in a sandy surface horizon of a Coastal Plain Spodosol....................................................................................................................... ......63 3.2 Effect of sonication energy input on th e loss of aggregate organic matter (AOM, % of total OM in size fraction) after sonication of the 150 to 53 m fraction.......................65 3.3 Loss of aggregate organic matter (AOM, % of total OM in size fraction) with increasing energy for the vari ous soil size fractions..........................................................66 3.4 Diffusive Reflectance Infra-red Fourie r Transformed Spectra (DRIFTS) showing characteristics of particulate organic matter (POM) and aggregate organic matter (AOM) of the 250 to 150 m fraction fo r a sandy Spodosol in north Florida...................67 3.5 Effect of forest management intensit y on the amount of aggregate organic matter (AOM, % of total OM in si ze fraction) for the 2000 to 250 m fraction for a sandy Spodosol in north Florida...................................................................................................68 4.1 Effects of family on the C content of the 2000 to 250 m light density fraction for a sandy Spodosol in north Florida........................................................................................85

PAGE 11

11 4.2 Family x Depth interaction in aggreg ate organic matter (AOM) of the medium density, 250 to 150 m fraction for a sandy Spodosol in north Florida............................86 4.3 Effects of management intensity and fam ily on N concentrations in the 2000 to 250 m medium density fraction for a sandy Spodosol in north Florida.................................87 4.4 Family x management intensity interacti on in aggregate organi c matter (AOM) of the heavy density, 2000 to 250 m fraction for a sandy Spodosol in north Florida................88 4.5 Effects of soil depth on C and N contents among the various size fractions for a sandy Spodosol in north Florida........................................................................................89 4.6 Distribution of C among the various density fractions for a sandy Spodosol in north Florida........................................................................................................................ ........90 4.7 Distribution of N among the various density fractions for a sandy Spodosol in north Florida........................................................................................................................ ........91 5.1 Carbon profile at age 4 for a sandy Spodosol in north Florida........................................101 5.2 Carbon profile at age 6 for a sandy Spodosol in north Florida........................................102 A.1 Effect of sieving time for dry sieving on the weight distribution across size fractions...106 A.2 Funnel assembly used for density fractionation...............................................................107 B.1 Image and elemental dot map of aggregate (2000 to 250 m)........................................109 B.2 EDX spectra for the aggregate shown in B.1...................................................................109 B.3 Image and elemental dot map of aggregate (2000 to 250 m)........................................110 B.4 EDX spectra for the aggregate shown in B.3...................................................................110 B.5. Image and elemental dot map of aggregate (2000 to 250 m)........................................111 B.6. EDX spectra for the aggregate shown in B.5...................................................................111 B.7. Image and elemental dot map of aggregate (250 to 150 m)..........................................112 B.8. EDX spectra for the aggregate shown in B.7...................................................................112 B.9. Image and elemental dot map of aggregate (250 to 150 m)..........................................113 B.10. EDX spectra for the aggregate shown in B.9...................................................................113

PAGE 12

12 LIST OF ABBREVIATIONS SOC: soil organic carbon OM: organic matter PPINES: Pine Productivity Inte ractions Experimental Study DRIFTS: Diffuse Reflectance Infrared Fourier Transform Spectroscopy AOM: aggregate organic matter POM: particulate organic matter SEM: scanning electron microscope SOM: soil organic matter

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy GENOTYPIC AND FOREST MANAGE MENT EFFECTS ON SIZE-DENSITY FRACTIONATION OF SOIL CARB ON IN A FORESTED SPODOSOL By Deoyani Vinayak Sarkhot December 2006 Chair: Nicholas. B. Comerford Cochair: Eric. J. Jokela Major Department: Soil and Water Science Soil C accounts for 75 to 85% of terrestrial C, making soil C sequestration an important ecosystem service. This study wa s undertaken to char acterize the soil organic carbon (SOC) pools in a sandy Spodosol of north Florida; to study the aggregate C pool in soils that are considered to have weak aggregation; and to investigate the influence of intensive forest management on SOC pools. A loblolly pine ( Pinus taeda L.) plantation under two levels of forest management (fertilization and understory control) was evaluated. Dry and wet sieving methods were compared for thei r applicability in size fractionation. Dry sieving was found to be satisfactory for these soils, as it preserved mo re structure and the wate r-soluble SOC components such as esters and amides. The use of organic matter (OM) release associated with aggregate breakdown by sonication allowed simultaneous meas urement of aggregate strength and amount of aggregate OM. Aggregate C was found to be an important C pool in these soils, accounting for nearly half of total SOC. Diffuse Refl ectance Infrared Fourier Transform Spectroscopy (DRIFTS) spectra showed presence of recently added OM in the 2000 to 250 m fraction, and more decomposed OM in the <53 m fraction. The spectra showed clear separation between aggregate and particulate OM. Aggregate OM wa s characterized by higher esters, amides and

PAGE 14

14 polysaccharides than particulate OM, indicating it s susceptibility to decomposition in case of aggregate destruction. The 2000 to 250 m fraction, especially its light and medium density components, was found to be the C pool most responsive to short-term management-related changes. Intensive management reduced the >53 m fraction C as well as aggregate C, probably due to the reduced input of understory roots. The best family (chosen a priori based on growth) added more decomposable C as shown by the high er C content in the light density fraction, and higher N concentration in the medium density u nder intensive management. The medium family encouraged aggregation as shown by the higher aggregate OM values. These methods and the sensitivity of these SOC pools to the short-term management-related changes offer a promise for better understanding SOC dynamics in sandy Spodosols of Florida.

PAGE 15

15 CHAPTER 1 INTRODUCTION What of thee, O earth, (I) dig out, let that quickly grow over let me not hit thy vitals, nor thy heart…. O cleansing one! Atharva Veda (Whitney, 1905) The need for conserving natural resources ha d been recognized by civilizations for over 5000 years. However, the management of these resources for sustainabl e production is still a challenge. With the rapidly increasing worl d population, demand for forest products is increasing, while large tracts of fo restland are being converted to other uses or being degraded (Forest Resources Assessment [FRA], 2005). The al arming rate of deforestation, especially in the developing countries, demands greater concer n for the existing forests. So worldwide, emphasis is being given to the preservation of na tural forests and their non-commodity functions (FRA, 2005). Timber harvest has been restricted in many of the world’s natural forests and consequently the onus of providing the world’ s increasing demand for wood and fiber is now on plantations managed especially for timber and pulp production. Therefore, it is important to increase the productivity of these areas. Concentrating tim ber production on the best sites will allow the world’s wood and fiber demands to be met on fewer acres. It will also allow large areas of native forests to be conserved or preserved (FRA, 2005).

PAGE 16

16 Pine Plantations in the So utheastern United States Pine plantations are important commercial ecosystems, covering more than 12 million ha in the southern United States (Wear and Grei s 2002). Apart from their economic significance, these ecosystems are important for their ecos ystem services, such as maintenance of ground water and air quality. Above and below ground C st orage in standing biomass and roots and their contribution to soil organic carbon (SOC) is anothe r service that will like ly become an economic commodity (e.g., C credits) in the near future. In the last few decades, increased management intensity has resulted in an unprecedented increase in the yield of w oody biomass and litterfall (Dalla-Tea and Jokela, 1991, McCrady and Jokela, 1998). This suggests an even higher potential for C sequestra tion. However, the longterm impacts of these management changes on SOC pools and on the soil’s C sequestration potential are not well documented. There are four important elements of intensive management that have the potential to cha nge nutrient cycling and SOC seques tration: fertilization, chemical understory control, deployment of improved familie s and site preparation. Th e first three of those elements are a part of this dissertation. Fertilizat ion has been shown to increase N and P mineralization rates (Polglase et al., 1992b), whil e understory control has been shown to reduce the SOC (Echeverria et al., 2004). However, no wo rk has addressed genotypic effects of family or their interaction with fertilization and weed control on the distribution and characteristics of SOC. Need for Carbon-wise Management Soil organic matter is a key constituent of the complex below-ground ecosystem, affecting a multitude of physical, chemical and biological soil properties. As an important binding agent for soil aggregates, SOC plays a role in soil aeration, water holding cap acity and permeability through its effect on soil struct ure and water holding characteris tics. Soil organic matter is

PAGE 17

17 important for soil fertility. Humus is a buffer th at influences soil pH. It also has high cation exchange capacity, surpassed only by 2:1 expandi ng clay minerals that have a higher CEC per unit volume. In most soils, humus accounts for 20 to 90% of the soil CEC (Brady, 2000). SOC is also an important storehouse for N, P, S, and micronutrients. It protects nutrients like N from being leached and protects P and mi cronutrients from fixation so th at they remain available for plant uptake for a longer period. Chelating functiona l groups in SOC, improve the availability of nutrient cations while reducing the toxicity of others. In sandy so ils with low inherent fertility and little clay to perform some of these functi ons, the significance of SOC in sustaining the productivity of these ecosystems and the need for understanding the C dynamics are even greater. However, total SOC has been found to be an insensitive indicator of management related changes in sandy soils (Harding and Jokela, 1994) Measurement of changes in C is difficult against the large amount of recalc itrant or inert material already present in the soil. This is especially true for detecting th e changes in nutrient supply charac teristics of SOC in which the actively cycling fractions of SOC play a ma jor role (Brady, 2000). Physical fractionation techniques based on differences in size and density may provide grea ter insight into this question by separating the active a nd inert C pools. Measurement of thes e pools can also give insight into the mechanisms through which management activities affect total SOC. For example, an increase in light density C suggests that trees may be a dding more decomposable organic matter since this fraction is an active form of C (Swanston et al., 2005). Challenges and Opportunities in Sandy Soils Sandy soils are generally found to be more res ilient to compaction by ti llage, but they are probably more susceptible to losses of SOC as ther e is very little clay to protect it (Shan et al., 2001). Carbon is sequestered when it is protected from decomposition. Protection is generally

PAGE 18

18 provided by sorption onto clay, by incorporation into aggregates or as a decomposition product that is resistant to further microbial attack. Th e low clay content of sandy surface soils dictates against the first option, leaving the latter two protection mechan isms as possibilities. Even though the potential for aggregation through th e formation of clay-cation-OM complexes (Edwards and Bremner, 1967) is small in thes e soils, aggregation thr ough the action of fungal hyphae and roots (Tisdall and Oades, 1982) is possi ble. This suggests a lower potential for C sequestration in the sandy soils, but offers the opportunity to study physical protection mechanisms without the confounding influence of clay. Unfortunately, very little is documented about the biochemical protection of SOC in soils that support sout hern pine ecosystems. Soil aggregation studies have been conducted in soils with higher clay contents. Therefore, the aggregation in these sandy so ils deserves investigation. This study was conducted with the following main objectives: To examine soil C forms in the surface horizon of a sandy Flatwoods Spodosol, To study aggregation in this soil; and To determine the management-related changes in the different soil C pools. Three laboratory methods, viz. size fractio nation, density fracti onation and sonication, were used to study C pools and to determine the effects of forest management on these pools. The results of this study are provided in thr ee chapters. Chapter 2 focuses on two different size fractionation methods, dry and wet sieving, and the C and N characteristics of the size fractions. The hypotheses were that: 1) dr y sieving would be a suitable method of fractionation for studying extremely sandy soils with inheren tly weak structure; 2) the 2000 to 250 m fraction would be the most important C fraction due to the input of roots and litter; 3) intensive management that uses fertilization and understory competition control to increase aboveground biomass would reduce the C content in the 2000 to 250 m size fraction due to reduction in

PAGE 19

19 understory root inputs; and 4) fertilizer N adde d through forest management would be reflected in higher N content in the size fractions. Chapter 3 focuses on the aggregate morphology, C content, and stability in the surface horizons of sandy Spodosols. Soni cation (breakdown of aggregates using ultrasonic energy) was used to measure aggregate stability and the patt ern of C release from aggregates of varying stability. The hypotheses tested were that: 1) Su rface horizons in a north Florida Spodosol would have more aggregation than has been describe d through soil mapping due to the high input of above ground litter and roots and higher activity of biological ag ents of aggregation such as fungal hyphae; 2) aggregate stability would incr ease with decreasing aggregate size due to increasing surface area of mineral matter availabl e to the action of organic binding agents; 3) particulate organic matter would be the dominan t C form over aggregate organic matter because of low clay content to suppor t aggregation; and 4) intens ive management would reduce aggregation in the short-term due to the reduc tion in the input of unde rstory fine roots. Chapter 4 focuses on the effects of full sib loblolly pine families (families with both parents known, chosen a priori based on their grow th performance) and their interactions with management intensity on size-density fractions of SOC. The hypotheses test ed in this study were that: 1) the best growing family would exhibi t the highest SOC, especially under intensive management, due to high litterfall inputs and th e responsiveness of this family to intensive management; 2) the light and medium density fractio ns would be the main reservoirs of C, since the mineral matter in these soils is predominan tly quartz sand with little or no C adsorption capacity; and 3) the light and me dium density of the 2000 to 250 m fraction would be the most responsive pool for detecting family effects b ecause earlier results show ed that the 2000 to 250 m fraction SOC was the most responsive size fraction and because the light and medium

PAGE 20

20 density fractions were most likel y to show the effects of recen tly added organic matter (Romkens et al., 1999). Chapter 5 summarizes the most important findi ngs of this study. It identifies opportunities for future research and suggests a possible course of action for unde rstanding the impact of forest management on SOC forms and functions.

PAGE 21

21 CHAPTER 2 EFFECTS OF FOREST MANAGEMENT ON SOIL CARBON AND NITROGEN IN A NORTH FLORIDA SANDY SPODOSOL Introduction The soil C density of Spodosols in Florida (20 kg m-2; Stone et al., 1993) is higher than the soil C density in a majority of the life zones st udied (Post et al., 1982), yet little research has been conducted to study the C profile of thes e soils. Many Florida Spodosols support southern pine plantations (Adegbidi et al., 2002), whic h have the potential for significant above ground storage of C (Richter et al., 1995). During the la st few decades, intensive management of these plantations has resulted in an unprecedented in crease in litterfall and yield of woody biomass (Dalla-Tea and Jokela, 1991), suggesting the potential to fu rther increase above ground C storage. However, the short and long-term impa cts of forest management on soil organic carbon (SOC) pools and the C sequestration potential of these soils are poorly documented. Fertilization and chemical understory control are common silvicultura l practices used to increase yields when managing southern pine stands in the southeastern U.S. Chemical understory competition control has been show n to reduce total SOC (Shan et al., 2001; Echeverria et al., 2004) and the mineralization of C and N in both whole soil and soil density fractions (Polglase et al., 1992a, b; Echeverria et al., 2004). Fert ilization has been reported to increase mineralization rates, especially for P (Polglase et al., 1992a, b). Yet, previous studies have shown no significant effect of fertilization on total SOC (Har ding and Jokela, 1994; Shan et al., 2001), probably because the increased organic matter inputs associated with large growth responses compensated the losses by increased mineralization rates. Physical fractionation of soil into size and dens ity fractions has been an effective technique for studying the forms and cycling of soil C (C hristensen, 1992; Ellert and Gregorich, 1995). Organic matter (OM) changes in the sand fractio n have been useful as early indicators of

PAGE 22

22 management-related C changes. For example, two years of barley ( Hordeum vulgare L.) and alfalfa ( Medicago sativa L.) cultivation, compared to a bare soil, increased sand fraction C (24 to 60%), N (36 to 45%) and carbohydrates (46 to 83%; Angers and Mehuys 1990). Likewise, OM (>150 m fraction) was depleted under continuous maize cultivation, yet r ecovered rapidly when the land use was returned to pasture (Romkens et al., 1999). In the extremely sandy Spodosols of the southeastern U.S., which ofte n contain less than 5% silt + cl ay, the sand fraction C is likely to equal total C since the clay fraction of these soils contains mainly quartz and kaolinite with very low C sorption capacity (Harri s and Carlisle, 1987). It follows that further fractionation of the sand size C would be required to detect the short-term management related changes in the soil C pools. Fractionation of the SOC into m eaningful pools also help s in understanding the processes underlying any possible changes. For example, identifying the pools responsible for different functions such as shor t-term nutrient turnover (e.g., sa nd size fraction) and long-term C storage (e.g., fine silt fraction; Liu et al., 2003) may help in the development of more accurate C dynamics models. Wet sieving is the standard technique for soil size fractionation (Yoder, 1936; Marx et al., 2005). Even though it disrupts macroaggregates (Angers and Giroux, 1996), wet sieving is necessary when working with high clay soils to break the strong aggregates that are >2 mm in diameter. In contrast, the sandy nature of many Coastal Plain soils hinders the formation of large stable macroaggregates. Under thes e conditions, dry sieving may offe r a viable alternative to the more time consuming wet sieving method. This study was conducted to defi ne the distribution of soil C and N in the surface horizon of a representative sandy Coastal Plain Spodosol a nd to evaluate the shortterm impacts of forest management activities on these soil characteristics. The firs t study objective compared wet

PAGE 23

23 sieving and dry sieving as alte rnative methods for investigati ng C and N distribution among soil size fractions. The hypothesis was that, given the low clay content and weak, single grain to crumb aggregate structure, dry sieving would be a suitable and rapid technique for size fractionation. The second objective was to esta blish the C and N distri bution, as well as the chemical fingerprint of OM in the various si ze fractions. The assumption was that short-term management impacts would be more identifiable w ithin a size fraction than in the whole soil. We expected the results to show that the major ity of the C and N would be in the 2000 to 250 m fraction, assuming that this fraction received the greatest inputs from roots and aboveground litter. We also expected this soil to contrast with those having c onsiderably more clay and larger C pools in the <53 m fraction. The third objective was, th rough the use of data generated by the first two objectives, to provide an evaluation of the suitability of these methods for determining the short-term impacts of two contrasting forest management intensities. Intensive management, especially with sustained understory competition control, would be expected to reduce SOC content in the 2000 to 250 m fraction due to the reduction in r oot C inputs. We anticipated that there would be enhanced N incorporated into all size fractions because of the fertilizer N inputs. Materials and Methods Experimental Site The study site was a loblolly pine ( Pinus taeda L.) plantation in north Florida (30o24’N lat; 82 o33’W long) managed by the Forest Biology Res earch Cooperative at the University of Florida as part of the Pine Productivity Interactions Experi mental Study (PPINES). This longterm study aims at understanding the family x envi ronment interactions in full-sib families of loblolly and slash pine ( P. elliottii Engelm. var. elliottii ). The climate is warm, humid subtropical, with 1,394-mm average annual rainfall, 270C average annual maximum temperature, and 130C average annual minimum temperature (NOAA, 2002). The soil is classified as a Leon

PAGE 24

24 series (sandy, siliceous, thermic Aeric Al aquod), with <5% silt + clay, and <10 cmolc kg-1 of cation exchange capacity. Trees were planted in January 2000 in four re plicates using a randomized complete block, split plot design. Prior to planting, the entire study was double bedded and treated with the herbicides Arsenal (imazapyr – 1.02 L ha-1) and Garlon (triclopyr – 7.02 L ha-1) to remove the understory vegetation and to reduce competition. The experimental design was a 2 x 2 x 8 factorial, which included two planting densities (close spacing at 1.3 x 3 m and wide spacing at 3 x 3 m), two management regimes (high and low i nputs) and six elite loblolly pine full-sib families, a mix of these elite families, and one poor growing family. We chose treatments that maximized the difference in C inputs to the soil in order to evaluate the capacity for short-term SOC changes, based on biomass production. The high intensity treatment included the most productive family and consisted of sustained und erstory competition control using herbicides, and annual fertilization using a complete fertilizer In the high intensity treatment, Arsenal (0.28 L ha-1) and Oust (0.14 L ha-1) were also applied as necessary to provide sustained understory competition control. The low intensity treatment, planted with the poorest growing family, was chosen for comparison. The families were designated a priori based on their growth performance in long-term genetic experiements. This management regime consisted of a onetime fertilizer application and the aforementi oned time of planting understory competition control treatment. At age four, when sampling was conducted, the fertili zer added to the high intensity treatment totaled 368 kg ha-1 N and 128 kg ha-1 P plus most other essential nutrients (i.e., 121 kg ha-1 K, 45 kg ha-1 Mg, 45 kg ha-1 Ca, 35 kg ha-1 S, 0.89 kg ha-1 B, 3 kg ha-1 Zn, 2 kg ha-1 Mn, 16 kg ha-1 Fe, 4 kg ha-1 Cu, 0.01 kg ha-1 Mo), while the low intensity treatment had 45 kg ha-1 N and 50.6 kg ha-1 P applied as diammonium phosphate. Both treatments were planted at

PAGE 25

25 the 1.3 x 3 m spacing (close spacing 2,900 trees ha-1) and each treatment plot was 480 m2 in size. The entire study was treated when necessary with insecticides (Dimilin, Pounce or Mimic) for tip moth ( Rhyacionia spp.) control during the first gr owing season. Furthe r details of the study site, including the stem volume and above ground biomass are discussed by Roth et al. (2006). Soil samples were collected from the A horizon at depth increments of 0 to 5 and 5 to 10 cm from each treatment plot in three replicate blocks in September 2003. One combined soil sample for each depth of each plot (treatment w ithin a block) came from four individual soil samples. The four individual soil samples were collected from alternate rows (interbed position), while within an interbed the sample locations were chosen randomly. Laboratory Methods Since the soil moisture content is highly va riable and “air-dry” condi tions are possible in these surface soils und er field conditions, soil samples we re air-dried an d passed through 8000 m and 2000 m sieves. The <2000 m fraction was further size-fractio nated in order to contrast dry and wet sieving methods into four size fractions, a macroaggregate fraction 2000 to 250 m, two microaggregate fractions 250 to 150 m and 150 to 53 m, as well as a <53 m fraction. For dry sieving, samples were sieved in a mechan ical shaker for 5 minutes. The purpose was to accomplish size fractionation with minimum destru ction of soil structure. A previous study determined that there was no significant differe nce in weight distribut ion among the four size classes after sieving for 5 minutes (unpublished data). Therefore, this time frame was used for all dry sieving. Wet sieving followed the procedure of Cambardella and Elliott (1993) without prewetting because large aggregates were not common. For each sample, 100 g of soil were added to the coarsest sieve (250 m) in a tray so that there was standing water 2 cm above the sieve screen. The sample was allowed to stand for 5 minutes and then the sieve was moved up and

PAGE 26

26 down 50 times (for approximately 2 minutes) taki ng care that the sieve screen broke the water surface every time. The soil suspension in the water tray was then transferred to the next size sieve and the procedure was repeated for each successive sieve (150 and 53 m). Each sieve, when removed from the water tray, was allowed to dry for 24 hours. The < 53 m fraction was collected by sedimentation for 48 hours follo wed by oven drying. The size fractions were analyzed for total C and N concentrations with a Carlo-Erba CN Anal yzer (CE Instruments, model NCS-2500). The >8000 m and 8000 to 2000 m fractions were also ground and analyzed for total C and N concentrations and used for calculating C and N contents. Since these soils have low clay content and no carbonates, loss on ignition (LOI) was used to measure OM content (at 5500C for 6h). The ratio of analyzer C to LOI was used to estimate the C content and C:OM ratio in each size fr action. Thermal Gravimetric Analysis (Omnitherm 951 TGA; Dupont Co., Wilmington, DE) confirme d complete combustion of C during LOI. The chemical fingerprints of the size fractions were investigated using Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFT S). Samples were scanned before and after ashing (at 5500C for 6h) in the mid-infrared on a Va rian Digilab FTS-7000 Fourier Transform Mid-infrared Spectrometer (Walnut Creek, CA). Samples were scanned from 4000 to 400 cm-1 at 4 cm-1 resolution using a KBr beamsplitter and DTGS detector a nd a Pike Autodiff autosampler (Pike Technologies, Madison, WI) using ground, non-KBr diluted samples. Statistical Analysis The statistical significance of treatments was analyzed using PROC GLM (SAS, 1996), with forest management intensity, soil depth, soil size fraction, and si eving method as fixed effects and with block as a random effect. The di fferences were considered significant at p< 0.05. Since the initial analyses showed a significant depth x fraction inte raction, further analyses were

PAGE 27

27 carried out for the individual size fractions. A multiple comparison procedure with a TukeyCramer adjustment was used for the post-hoc mean separation. For the DRIFTS spectra, spectral subtraction of ashed samples from non-ashed samples was used to accentuate OM characterization using GRAMS/AI software Ver. 7.02 (Thermo Galactic, Salem, NH). Discriminant analysis was conducted using SAS (SAS Institute, Cary, NC) partial least squares (PLS) using a modified version of a custom made SAS program for spectral pre-treatments including gap derivatives scatter correction, a nd spectral data point averaging (Reeves and Delwiche, 2 003). This modified version allo wed discriminant analysis to be carried out using SAS PLS while the origin al program was develope d only for quantitative regression analysis. The higher the R2 of the discriminant analysis and the lower the Root Mean Square Deviation, the better was the separation between main e ffects as indicated by DRIFTS spectra. Results Dry vs. Wet Sieving Wet sieving reduced the C concentration by 8% (% of fraction, Fig. 2.1) and C content by 30% (% of whole soil; Table 2.1) in the 150 to 53 m fraction. There was also a significant reduction in the weight of this fraction due to we t sieving (16% of whole soil in dry sieving vs. 14% in wet sieving). In the <53 m fraction, we t sieving increased the C content by 26% (Table 2.1), but reduced the C concentration by 13% (F ig. 2.1) and C:OM ratio by 12% (Fig. 2.2). Wet sieving reduced the C concentra tion and C:OM ratio in the 2 50 to 150 m fraction, although its C content was not statistically affected by sieving method. Similarly, the 2000 to 250 m fraction was not affected by the sieving met hod. The total recovery of C as shown by mass

PAGE 28

28 balance for both methods was not st atistically different. Nitrogen c oncentration or content (Table 2.2) was not affected by the sieving method in any of the size fractions examined. The effect of sieving method was evident on th e chemical fingerprint of the size fractions as measured by DRIFTS (ash-subtracted spectra R2 = 0.94; Table 2.3). Wet sieving reduced C in all four size fractions across the entire spectra as i ndicated by spectral peak heights (Fig. 2.3). Differences in the 2000 to 53 m fractions were more pr onounced for esters (1730 cm-1), amides (1650 cm-1) and aromatic compounds (1530 cm-1). On the other hand, in the <53 m fraction, the difference between sieving methods was more pronounced in the aliphatic C-H (2870, 2930 cm1), and polysaccharide (1160 cm-1) peaks. Characterization of Size Fractions Carbon concentration (Fig. 2.1) was highest in the <53 m fraction (7.8 to 8.8%) and lowest in the 250 to 150 m fraction (0.6 to 0.7% ). Carbon content (Table 2.1) on the other hand, was highest in the 2000 to 250 m fraction and lowest in the <53 m fraction. More than 65% of the total C (Table 2.1) in thes e soils was found in the 2000 to 53 m fractions, of which about 39% was in the 2000 to 250 m fraction. N content (Table 2.2) followed a trend similar to C. The C:OM ratios for the 2000 to 250 m and 150 to 53 m fractions were 8 to 27% higher than the remaining two fractions (Fig. 2.2), which had ratios equal to the standard Van Bemmelen factor (0.58). The C:N ratios of the fractions were not affected by any of the factors and therefore, are not reported. The DRIFTS spectra (Fig. 2.4) confirmed ch emical fingerprint differences among size fractions (ash-subtracted spectra R2 = 0.94 to 0.98; Table 2.3). However, the differences were mainly in the peak heights corresponding to the amount of C in each fraction. The 2000 to 250 m fraction exhibited the highest aliphatic C-H peaks (2870, 2930 cm-1), while esters (1730 cm1), amides (1650 cm-1) and polysaccharides (1160 cm-1) were high in both 2000 to 250 and < 53

PAGE 29

29 m fractions. The peak for aromatic rings at 1530 cm-1 was similar in these latter two fractions, but the peak at 1580 cm-1 was absent in the 2000 to 250 m fr action. The ester and amide peaks were absent in the 250 to 150 m fraction. Effect of Management In tensity and Soil Depth In the 5 to 10 cm depth, low intensity manage ment had 12 and 5% higher C contents in the 2000 to 250 and 250 to 150 m fractions, respecti vely, than under the intensive management regime (Fig. 2.5). Low intensity management also had a 5% higher C cont ent (Fig. 2.5) and 7% higher N content (Fig. 2.6) than intensive management for the 0 to 5 cm depth in the 150 to 53 m fraction. In contrast, the high intensity ma nagement showed 2% higher C content and 5% higher N content in the < 53 m fracti on at the 5 to 10 cm depth. The > 2000 m fraction exhibited 22 to 48% higher C content at 0 to 5 cm depth compared to the 5 to 10 cm depth, but C content in this fraction wa s not significantly influenced by management intensity. Soils under the low intensity management regime were higher in OM as shown by consistently higher peak height s of the DRIFTS spectra (Fig. 2.7; ash-subtracted spectra R2 = 0.71; Table 2.3). At the 0 to 5 cm depth, differences were observed fo r all the four size fractions, while at the 5 to 10 cm depth, a differe nce was observed only in the 2000 to 53 m fractions. This result was observed acro ss the entire spec trum of OM in the size fractions. Discussion This study was undertaken with the objective of understanding the distribution of C across the size fractions in a forested Spodosol, with the assumption that the size fractions approximated distinct C pools; and to determine the impact of forest management on these C pools. Southern pine plantations in the southeastern U.S. repres ent important regional sinks for C (Richter et al., 1995) and unders tanding the soil C dynamics is an essential step towards sustainable management of these ecosystems.

PAGE 30

30 Dry vs. Wet Sieving The reduction in C content in the 150 to 53 m fraction and the increase in C content in the <53 m fraction (Table 2.1) indicat ed a significant transfer of C with wet sieving. This was due to breakdown of water-dispersible aggregates in the 150 to 53 m fraction, where the C and a significant amount of soil mass was washed into th e finest fraction. This finding was consistent with previous work (Carter 1992; Angers and Giroux 1996). The reduction in C concentration (Fig. 2.1) and C:OM ratio (Fig. 2.2) of the <53 m fraction indicated that C was lost even from this fraction as water-soluble C. This interpretation was suppor ted by the DRIFTS data, where esters, amides and aromatic compounds were lo st from the 150 to 53 m fraction and some polysaccharides and aliphatic –CH compounds were lost from the <53 m fraction (Fig. 2.3). However, the statistically equivalent C recovery of the two sieving methods showed that the loss of water-soluble C was small, with amounts being within experimental error. The N concentration and content, on the other hand, was not affected by sieving method, indicating that N in these soils was not likely in a significant water-soluble or water-dispersible (aggregate) form. Therefore, our hypothesis of equivalence between dry and wet sieving was accepted for N distribution and rejected for C di stribution, as wet sieving resulted in a redistribution of C. Dry sieving, which is an easier and quicker method th an wet sieving, was consid ered superior to wet sieving for the extremely sandy soils examined in this study, as it preserve d the water-dispersible aggregates and water-soluble C. Distribution of C and N among Size Fractions More than 65% of the C and N was found in the 2000 to 53 m fractions (Table 2.1, 2.2), which supported the hypothesis of the importance of these fractions for C content. The <53 m accounted for less than 10% of the total C. This is unlike soils having higher clay contents, where the C contents tend to be highest in th e silt + clay size fracti ons. Hassink et al. (1997)

PAGE 31

31 found that more than 50% of the soil C was in the < 20 m fraction (65% clay). Feller (1993) and Bronick and Lal (2005) reported similarly hi gh C contents in the silt + clay fraction. Though the C contents reported for extr emely sandy soils vary, there is ev idence that forested sandy soils in other parts of the world have similarly hi gh C content within the sand size fractions. For example, forested Spodosols in France (<10% silt + clay) were shown to have 50% of the total C within the sand size fraction (Jol ivet et al., 2003). Quideau et al. (1998) reported that, under hardwood forests, 45 to 55% of the total C wa s associated with the sand size fraction (58% sand). In contrast, when under maize cultivation, a Spodosol (<10% silt + clay) had only 17% of its total C in the sand size fraction (Quenea et al., 2006). The nature of OM, as evidenced by C:OM ra tios (Fig. 2.2), differed among size fractions. The DRIFTS data confirme d this result (Fig. 2.4). The high peaks of esters, amides and aliphatic C in spectra of the 2000 to 250 m fraction indicated recently added undecomposed OM and the high C:OM ratio of this fraction indicated presence of C rich or ganic matter. On the other hand, the high aromatic C peak at 1580 cm-1 in the < 53 m fraction indicated more decomposed OM in this fraction, but also showed high peaks of esters and amides, both of which are easily decomposable C forms. The C content of organic matter, as measur ed by the C:OM ratio, is known to change according to soil type and profile depth (Nel son and Sommers, 1982). However, this study indicated that it also change s with fraction size and sievi ng method used (Fig. 2.2). A 6% underestimation of C would have resulted if th e C content of the 2000 to 250 m fraction was estimated from the amount of OM using the Van Bemmelen factor (0.58). From a methodological perspective, these results indica te that one should c onsider using fractionspecific conversion ratios when loss on igni tion is used to estimate soil fraction C.

PAGE 32

32 Impact of Management Intensity The higher soil C content associated with th e low intensity management regime for the three >53 m fractions indicated that intensive forest management reduced the soil C content in as little as 4 years (Fig 2.5) The DRIFTS spectra showed C lo ss across the entire spectrum of organic matter (Fig. 2.7). A possi ble cause of this dec line would be the redu ction of root input caused by the understory c ontrol practice. Shan et al. (2001) and Echeverria et al. (2004), working on similar soils in Georgia, have also reported a decrease in SOC due to intensive management, especially due to chemical under story control. Therefor e, the hypothesis of C reduction under the high intensity management regime was accepted. However, it is uncertain whether this C reduction would be compensated by the increase in litterfall under the intensive management regime in the longer-term. The hypothe sis of higher N content in all fractions was not accepted, since N content showed only a small increase (5%) in the finest fraction (Fig. 2.6). The fate of fertilizer N applied in the inte nsive management regime is also uncertain. Immobilization of fertilizer N in standing bioma ss, litter layer, or stor ed in the subsoil is probably responsible for this. This interpretati on is supported by the work of Will et al. (2006), who reported that 68% of applied fertilizer N wa s stored in the aboveground biomass and forest floor, while only 21% was in the surface soil in l oblolly pine stands growing in Georgia. Size fractionation proved to be a more sensi tive method than total SOC measurement for investigating changes in SOC. Management i nduced change in the 2000 to 250 m fraction was 23% of the fraction C concentration (Fig. 2.5), wh ich represented greater than 12% change in total SOC, illustrating the sensitivity of this size fraction for assessing impacts of forest management. It is also notewort hy that these differences were obs erved just four years after the treatments were imposed, supporting the hypothesis th at size fractionation e nhances the detection of the short-term management induced changes in SOC.

PAGE 33

33 Given the presence of large palmetto roots ( Serenoa repens (B.) Small.) and other understory plants under the low intensity manage ment regime (Roth et al., 2006), we expected that the fractions >2000 m would decrease significantly with increased management intensity. The lack of a statistical C response to manageme nt in these fractions, th e reduced C content in the 2000 to 250 m fraction in response to chemical weed control, and the “f resher” nature of the OM as suggested by DRIFTS imply that the la rge sand fraction accepts the greatest detrital inputs. This interpretation is consistent with the work of Van Rees and Comerford (1986) and Escamilla et al. (1991), who have reported palmetto and other understory root biomass in the 2000 to 250 m size class under southern pine plantations. Conclusions Dry sieving was found to be a useful met hod for size fractionation for sandy Spodosols when compared to wet sieving, as it preserved more structure and the water-soluble components such as esters and amides. The size fractions were significantly different in all the properties studied. The 2000 to 250 m fraction was the most important fraction in these soils, as it contained nearly half of the to tal SOC and was sensitive to the management related changes. The DRIFTS spectra were useful for describing the changes in SOM chemical composition and indicated presence of recently added organic matte r in the large sand fraction. Intensive forest management reduced soil C in the >53 m fractions, and particularly in the 2000 to 250 m fraction, in just 4 years, probably due to the reduction in understory roots.

PAGE 34

34 Table 2.1. The effect of wet and dry sieving on C content (g C in size fr action per 100g of whole soil) in soil size fractions for a sa ndy Spodosol in north Florida. Fraction >8000 m 8000 to 2000 m 2000 to 250 m 250 to 150 m 150 to 53 m < 53 m Total Dry 0.31† (0.02) ‡ 0.47 (0.02) 1.1 a (0.1) 0.36 a (0.02) 0.39 b (0.02) 0.15 a (0.01) 2.8 (0.3) Wet 0.31§ (0.02) 0.47 (0.02) 1.3 a (0.1) 0.32 a (0.02) 0.30 a (0.01) 0.19 b (0.01) 2.9 (0.3) 11 17 40 13 14 5 100 % of Total 10 16 46 11 10 7 100 †Each value is a mean of 36 observations aver aged across treatment intensity and soil depth. ‡ Values in parentheses re present standard error. § The >8000 and 8000 to 2000 m fractions, though not wet sieved, are necessary for total C content calculations. Within a size fraction, the means followed by diffe rent letters are statistically different at p < 0.05 showing effect of both depth and sieving method.

PAGE 35

35 Table 2.2. The distribution of N concentration and content among the soil size fractions for a sandy Spodosol in north Florida. Property >8000 m 8000 to 2000 m 2000 to 250 m 250 to 150 m 150 to 53 m < 53 m Whole soil Concentration (% of fraction) 0.370† e (0.018) 0.386 e (0.015) ‡ 0.116 c§ (0.005) 0.030 a (0.002) 0.088 b (0.003) 0.290 d (0.004) 0.082 (0.004) Content (% of whole soil) 7 a 12 b 39 d 19 c 16 c 7 a 100 † Each value is a mean of 72 observations aver aged across sieving methods, treatment intensity and soil depth. ‡ Values in parentheses re present standard error. § The means followed by different letters are statistically different at p < 0.05

PAGE 36

36 Table 2.3 Discriminant analysis of the DRIFTS spectra for a sandy Spodosol in north Florida Effect R2 Root Mean Square Deviation Extent of separation between spectra Block† 0.25 to 0.55 0.29 to 0.38 Very little separation Sieving method‡ 0.94 0.12 Excellent separation Management intensity§ 0.71 0.27 Some separation Depth 0.37 0.40 No separation Fraction# 0.93 to 0.98 0.06 to 0.12 Excellent separation † Four blocks (replications across space) ‡ Dry and wet sieving methods § High and low management intensities Depths 0 to 5 cm and 5 to 10 cm # Fractions 2000 to 250, 250 to 150, 150 to 53 and <53 m

PAGE 37

37 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 2000 to 250 m250 to 150 m150 to 53 m<53 mC (% of fraction) Dry sieving Wet sieving e e ba dc g f Figure 2.1. Carbon concentrations (% of fraction) of size fractions as affected by the dry and wet sieving for a sandy Spodosol in north Florid a. Means followed by different letters are statistically different at p < 0.05. The error bars represent standard error. Each value is a mean of 36 observations averaged acr oss treatment intensity and soil depth.

PAGE 38

38 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 2000 to 250 m250 to 150 m150 to 53 m<53 mC:OM Ratio Dry sieving Wet sieving c a b c b c b b Figure 2.2. The interaction between fraction size and sieving method on the ratio of C to organic matter (C:OM) for a sandy Spodosol in north Florida. The means followed by different letters are statistica lly different at p < 0.05. The e rror bars represent standard errors. Each value is a mean of 36 obs ervations averaged across management intensity and soil depth.

PAGE 39

39 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 050010001500200025003000350040004500 Wave number (cm-1)Peak Height <53 m, Dry Sieving <53 m, Wet Sieving a c d e b 150 to 53 m, Dry Sieving 150 to 53 m Wet Sievin g Figure 2.3. Diffuse Reflectance Infra red Fourier Transform Spectros copy spectra showing effect of sieving method (150 to 53 and <53 m frac tions, low intensity, 5 to 10 cm depth) for a sandy Spodosol in north Florida. a. polysaccharides, b. aromatic compounds, c. amides, d. esters, e. aliphatic –CH.

PAGE 40

40 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 050010001500200025003000350040004500 Wave number (cm-1)Peak Heighta b c d e Figure 2.4. Diffuse Reflectance Infra red Fourier Transform Spectros copy spectra showing effect of fraction size (dry sieving, high intensity, 0 to 5 cm depth) for a sandy Spodosol in north Florida. a. polysaccharides, b. arom atic compounds, c. amides, d. esters, e. aliphatic –CH. 2000 to 250 m < 53 m 150 to 53 m 250 to 150 m

PAGE 41

41 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 HighLowHighLowHighLowHighLowHighLowHighLow >8000m8000 to 2000m 2000 to 250m 250 to 150m150 to 53m<53m Carbon content (g C in size fraction per 100 g of whole soil) 0 to 5 cm (D1) 5 to 10 cm (D2) a a b a a a b a a b a c b c a b a b a b a b a a a a a D1 > D2 D1 > D2 Intensity by Depth interaction Figure 2.5. Effect of management intensity and soil depth on the C content of the soil size fractions for a sandy Spodosol in north Flor ida. Within a size fraction, the means followed by different letters ar e statistically different at p < 0.05, showing the effect of both management intensity and depth. E ach value is a mean of 18 observations averaged across sieving methods.

PAGE 42

42 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0 to 5 cm 5 to 10 cm 0 to 5 cm 5 to 10 cm 0 to 5 cm 5 to 10 cm 0 to 5 cm 5 to 10 cm 0 to 5 cm 5 to 10 cm 0 to 5 cm 5 to 10 cm >80008000 to 20002000 to 250250 to 150150 to 53<53 Nitrogen content (g N in size fraction per 100 g of whole soil) High (H) Low (L) b ab a ab b a a a a a a a a a a a b aa c c b a b H > L L > H H > L Figure 2.6. Effect of management intensity and so il depth on N content (g N in a soil fraction per 100g of whole soil) in soil si ze fractions for a sandy Spodosol in north Florida. Within a size fraction, the means followe d by different letter s are statistically different at p < 0.05, showing the effect of both management intensity and depth. Each value is a mean of 18 observati ons averaged across sieving methods.

PAGE 43

43 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 050010001500200025003000350040004500 Wave number (cm-1)Peak Heights High intensity Low intensitya b c d e Figure 2.7. Diffuse Reflectance Infra red Fourier Transform Spectros copy spectra showing effect of intensity (2000 to 250 m fraction, dry sieving, 0 to 5 cm depth) for a sandy Spodosol in north Florida. a. polysacchari des, b. aromatic compounds, c. amides, d. esters, e. aliphatic –CH.

PAGE 44

44 CHAPTER 3 SOIL AGGREGATION AND AGGREGATE CARB ON IN A FORESTED SOUTHEASTERN COASTAL PLAIN SPODOSOL Introduction Carbon-wise soil management requires an unde rstanding of the processes by which soil C is sequestered. Yet, understanding of these processes for forested soils of the southeastern USA is limited. Secondary forests in the southeastern USA accumulate C at a rate greater than 70 MT y-1 (Richter et al., 1995), which implicates them as important regional C sinks. Southern pine plantations are important components of these fo rests, covering more than 12 million ha (Wear and Greis, 2002). Many of these plantations ar e underlain by sandy Spodosols (Adegbidi et al., 2002), which represent the dominant so il order in Florida, covering 27% of the state (Stone et al., 1993). Many Spodosols are exceptionally sandy with less than 5% silt + clay and less than 10 cmolc kg 1 of cation exchange capacity (Carlisle et al., 1981, 1988, 1989; Sodek et al., 1990). Forest fertilization and chemical weed c ontrol are two management inputs that have increased productivity of southern pines in th ese landscapes. However, fertilization has not promoted an increase in soil C (Harding and J okela, 1994; Shan et al., 2001); while chemical weed control, presumably by reduc ing detrital inputs of understory plants, has reduced the soil C content (Shan et al., 2001; Eche verria et al., 2004). The eff ects of these practices on the development of aggregates and the soil C contai ned within them have yet to be considered. Soil organic carbon (SOC) can be protected fr om decomposition through four mechanisms: sorption onto clay particles (chemical protecti on), incorporation into aggregates (physical protection), movement to subsoils (translocation ), and biochemical transformation into products that are resistant to microbial attack (biochemi cal protection; Six et al ., 2002; Blanco-Canqui and Lal, 2004; Jimnez and Lal, 2006). The soil structure of Florida’s Spodosols is described as weak crumb to granular or single grain (Carlis le et al., 1981, 1988, 1989; Sodek et al., 1990),

PAGE 45

45 suggesting poor soil aggregation. In these soils, the low cation content limits aggregate formation through clay-polyvalent cation-organic matter complexes (Edwards and Bremner, 1967), though aggregates can form under the influence of mi crobial products, fungal hyphae, and roots (Tisdall and Oades, 1982; Blanco-Canqui and Lal, 2004). Th e potential for chemical protection of soil C is also limited by the low clay content. Given thes e factors, the interest in aggregation in sandy Spodosols in the southeastern U.S. has been low as evidenced by the few studies addressing this topic. In Russian Spodosols, aggregation did res pond to agricultural mana gement. Water-stable aggregation and total soil C cont ent increased after just two year s under the influence of a grassclover mixture, but decreased when followed by spring wheat (Buchkina and Balashov, 2001). In Florida, soil aggregation was found to be in fluential in P dynamics. Higher water-extractable P and heavy metals, along with slower rate s of release, were found in the 500 to 250 m and 250 to 125 m aggregates compared to the smaller si ze fractions (Zhang et al., 2003). Sandy Spodosols in Florida represent a unique soil condition and it was deemed necessary to better understand aggregation under these conditions. The purpose of this research was to study aggregation and its relation to SOC in a representative forested Spodosol of northern Florida. The first objective was to observe the aggregation present in these soil s. We hypothesized that aggrega tion, albeit weak, was present in the < 2 mm fraction in these extremely sandy soils because of the high input from root turnover and aboveground litter. The second obj ective was to determine the stre ngth of aggregates in the < 2 mm fraction, and quantify the amount of aggreg ate C. The hypotheses related to this objective were that: (1) aggregate strength, as measured by an aggregat e’s resistance to dispersion, would increase with decreasing aggregate size, and (2) that particulate C would be the dominant pool of

PAGE 46

46 SOC. The first hypothesis grew from the concept th at smaller particles ha ve greater surface area available for binding; hence, the aggregate’s strength/stability would be greater. The second hypothesis recognized that SOC c ould be found as either particulate organic matter (POM) or aggregate organic matter (AOM). Given the weak structure often described for Coastal Plain Spodosols, the dominance of POM would be expected. The third objective of this research was to provide preliminary information on the shortterm influence of two contrasting management intensities on the amount and distribution of AOM and POM. The hypothesis was that intensive management (more fertilization and chemical weed control) would equate to decreases in ag gregation. It was expect ed that reduced root turnover of understory plants, re sulting from sustained chemical control, would cause reduced aggregation and short-term reductions in soil C. Materials and Methods Experimental Site A loblolly pine ( Pinus taeda L.) plantation in north Florida (30o24’N lat; 82o33’W long) was the study site and it was managed by the Forest Biology Research Cooperative at the University of Florida as part of the Pine Produc tivity Interactions Experimental Study (PPINES; Roth et al., 2006). This long-te rm study aims at understanding the family x environment interactions in full-sib lo blolly and slash pine ( P. elliottii Engelm. var. elliottii ) families. The climate is warm, humid subtropical, with 1,394 mm average annual rainfall, 27oC average annual maximum temperature, and 13oC average annual minimum temper ature (NOAA, 2002). The soil is classified as a Leon series (sandy, siliceous, thermic Aeric Al aquod), with < 5% silt + clay, and < 10 cmolc kg-1 of cation exch ange capacity. The trees were planted in January 2000 in f our replicates using a randomized complete block, split plot design. Prior to planting, the entire study was double bedded and treated with the

PAGE 47

47 herbicides Arsenal (imazapyr, 1.02 L ha-1) and Garlon (triclopyr, 7.02 L ha-1) to remove the understory vegetation and to provide a competition free environment. The experimental design was a 2 x 2 x 8 factorial, which included two pl anting densities (close spacing at 1.3 x 3 m and wide spacing at 3 x 3 m), two management regime s (high and low inputs), and six elite loblolly pine full-sib families, a mix of these elite families, and one poor growing family. The families were designated a priori based on their growth performance in long-term genetic experiments. Two treatment combinations representing the ma ximum differences in biomass production were selected in order to evaluate the capacity for short-term SOC changes, with the idea that the differences in input would be reflected in th e differences in SOC pools. The high intensity treatment included the most productive family under sustained understory competition control using herbicides and annual fert ilization with a complete fertili zer. The low intensity treatment was planted with the poorest performing fam ily and managed with a one-time fertilizer application and understory competition control at planting. The fertiliz er added to the high intensity treatment totaled 368 kg ha-1 N and 128 kg ha-1 P plus nearly all other essential nutrients (i.e., 121 kg ha-1 K, 45 kg ha-1 Mg, 45 kg ha-1 Ca, 35 kg ha-1 S, 0.89 kg ha-1 B, 3 kg ha-1 Zn, 2 kg ha-1 Mn, 16 kg ha-1 Fe, 4 kg ha-1 Cu, 0.01 kg ha-1 Mo); while the low intensity treatment included 45 kg ha-1 N and 50.6 kg ha-1 P applied as diammonium phos phate. In the high intensity treatment only, Arsenal (0.28 L ha-1) and Oust (0.14 L ha-1) provided sustained understory competition control. Both treatments were planted at the 1.3m x 3 m spacing (close spacing 2,900 trees ha-1) and each treatment plot was 480 m2 in size. The entire study was treated when necessary with insecticides (Dimilin Pounce or Mimic) for tip moth ( Rhyacionia spp.) control during the first growing season.

PAGE 48

48 Soil samples were collected from the A horizon at depth increments of 0 to 5 cm and 5 to 10 cm from each treatment plot in three replic ate blocks in September 2003. One combined soil sample from each depth of each pl ot (treatment within a block) came from four individual soil samples. The four individual soil samples were collected from alternate interbed rows, while within an interbed, sample lo cations were chosen randomly. Laboratory Methods Soil samples were air-dried and passed through a 2 mm sieve. They were then dry sieved to separate the aggregates into four size cla sses with minimum disruption of aggregation (a macroaggregate fraction 2000 to 250 m, two microaggregate fractions 250 to 150 m and 150 to 53 m as well as a <53 m). A preliminary study determined that there was no significant difference in weight distributi on among the four size classes af ter shaking the sieves for 5 minutes. Therefore, this time frame was used for all dry sieving. The first part of the inves tigation included a microscopi c examination of aggregate morphology in the dry sieved size fractions. The in tent was to determine if there was identifiable aggregation and a positive result would justify examining othe r objectives. Aggregate samples were examined and photographed using a dissec ting light microscope (Carl Zeiss 475003-9902) with a mounted digital camera (Sony MVC FD90 ). A scanning electron microscope (SEM) (JEOL JSM 6400) equipped with an energydispersive x-ray fl uorescence elemental microanalysis (EDX) system was used for obt aining images and silica dot maps within aggregates. Samples were prepared for SEM by mounting on C stubs and coating with C. Upon finding aggregates and noting that they were a significant component of the soil matrix, the next step was to examine aggregat e strength and to determine the quantity of C contained in stable aggregates. This an alysis was performed on the three >53 m fractions using sonication to input energy into a water-soil sy stem. Sonication has been used for aggregate

PAGE 49

49 disruption (North, 1976; Six et al., 2001; Swan ston et al., 2005) because, unlike chemical dispersion techniques, it avoids chemical modification of the or ganic matter. It also allows measurement of aggregate strength on an en ergy basis, which allo ws for a quantitative comparison of samples. Energy inputs to the si ze fractions ranged from 0 to 27,000 J and were achieved with a Sonic Dismembrator (Fisher Scientific, model 500) by using a range of amplitude (20 to 60%) and time (1 to 7 min) combinations. The energy output was given by the sonicator and was calculated by internal software using voltmeter readings recorded every 10 seconds (Fisher Scientific, personal communicat ion, 2006). The energy output thus calculated was replicable (coefficient of variation < 10 %). The pulse method (60 sec ON and 30 sec OFF) was used to avoid an excessi ve rise in temperature. Each size fraction was sonicated at increm ental energy levels. This was accomplished by using one sub-sample for each energy level until complete aggregate breakdown was achieved. Microscopic observation and releas e of soil organic matter (SOM) was used to ensure that all the aggregates were disrupted. This analysis was repeated on the >53 m fractions of 12 soil samples representing three replic ations of both management intensi ties and soil depths. For each sample, sonication was done at 9 to 11 energy levels ( one sub-sample for each energy level). For each sub-sample, 2 g soil was weighed into a 250 mL beaker to which 100 mL water was added. The suspension was sonicated at the desired energy leve l. Depth of immersion of the sonicator probe was kept constant at 10 mm, as this variable is known to infl uence the degree of disruption (North, 1976). The suspension was then passed th rough the same size sieve used to obtain the size fraction (e.g., 250 m sieve for the 2000 to 250 m fraction). The SOM remaining on the sieve and the SOM passing through the sieve were measured by loss on ignition.

PAGE 50

50 Organic matter passing through the sieve afte r sonication was termed aggregate organic matter (AOM), as it contained finer organic matter held inside the aggregates that was released after aggregate disr uption. It was expressed as percent of total organic matter in each sample. Organic matter remaining on the sieve after soni cation was termed particulate organic matter (POM) which, after complete aggregate dispersi on, contained SOM of the same size. Energy input was plotted against AOM to obtain the resp onse curve for each sample. The AOM lost at 0 J represented the organic matter associated with water-dispersible aggregates. The POM data are not reported because the focus of this paper is aggregation and also because POM can be derived from the AOM data (POM = 100 – AOM). Upon finding that AOM was a significant component of total organic ma tter, the chemical nature of AOM and POM was investigated thro ugh the use of Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). Sample s were scanned before and after ashing (at 550oC for 6 h) in the mid-infrared on a Dig ilab FTS-7000 Fourier Transform Mid-infrared Spectrometer (Varian Instruments; Walnut Creek, CA). Samples were scanned from 4000 to 400 cm-1 at 4 cm-1 resolution using a KBr beamsplitter and DTGS detector and a Pike Autodiff autosampler (Pike Technologies, Madison, WI) using ground, non-KBr diluted samples. Spectral subtraction of ashed samples from non-ashed samp les was performed to accentuate differences in organic matter characterization using GRAMS/AI software Ver. 7.02 (Thermo Galactic, Salem, NH). Statistical Analysis The equation, y = (a*x) (b + x)-1 was used to fit the energy vs. AOM release data for each size fraction; where, a = maximum AOM lost for the size fraction, b = energy level at 0.5*a, x = sonicator energy output (J) and y = AOM. The energy output for each individual run was used for this analysis.

PAGE 51

51 PROC MIXED (SAS, 1996) was used to contrast the effects of management intensity and depth based on a completely randomized de sign, with energy level (i.e., each unique combination of amplitude and time), size fraction, management intensity and soil depth as fixed effects and block as a random effect. The differe nces were considered significant at p < 0.05. There was a significant energy level x management intensit y x size fracti on interaction. Therefore, further statistical analyses were r un separately for each size fraction. A multiple posthoc comparison procedure with the T ukey-Cramer adjustment was used. Results Aggregate Morphology Microscopic observation identified two qualita tive categories of aggregation. Irregular shaped aggregates (Fig. 3.1.1 – 3.1.4) had mine ral matter, organic debris and fungal hyphae and/or fine roots enmeshed toge ther. Spherical aggregates exhi bited mineral matter encrusted on organic debris or plant remains (Fig. 3.1.5) with or without funga l hyphae and/or fine roots (Fig. 3.1.6). The internal structure of the aggregat es, as shown by the silicon dot maps (Fig. 3.1.1 – 3.1.4), indicated mineral matter embedded in the or ganic matter. Images suggested a role of fungal hyphae and fine roots in aggregate forma tion, either through mechanical enmeshing (Fig. 3.1.1 3.1.5) or through encrustation of mine ral matter on plant remains (Fig. 3.1.7). Quantifying Organic C in Aggregates On average, 46% of the total soil C (Table 3.1) was contained in the soil aggregates. As the energy input to the soil sample increased, aggreg ates were destroyed, in creasing the amount of AOM removed from that size class of soil materi al (Fig. 3.2). Eventually a plateau was reached, indicating that all the organic matter that coul d be removed from the aggregates was removed (Fig. 3.3). This finding was supported by micr oscopic observations (Fig. 3.1.8, 3.1.9), which indicated that only POM was present when the plateau was reached. The energy level at which

PAGE 52

52 this plateau was reached exceeded the strength of the most stable aggregates in the size fraction studied. The energy required for the complete br eakdown of aggregates (Fig. 3.3) increased in the order 250 to 150 m (8500 J) < 2000 to 250 m (13300 J) < 150 to 53 m (16422 J). The shapes of the size fraction response curv es were different (Fig. 3.3) among the three size fractions, which indicated dive rsity in the nature of aggregat es. The aggregates in the 250 to 150 m fraction were the least stab le among the three fractions, lo sing the highest proportion of its organic matter (y) at all en ergy levels (x). The 150 to 53 m fraction had the most stable aggregates (Fig. 3.3, Table 3.2). The 2000 to 250 m and 250 to150 m fractions (Fig. 3.3) exhibited a step-wise loss of organic matter, with steps at 2500 J and 6000 J (Fig. 3.3). The 150 to 53 m fraction (Fig. 3.2, 3.3) exhi bited a continuous spectrum of organic matter loss with increasing energy input. The variab ility (Table 3.2) of organic matter release in AOM decreased with decreasing fraction size, wh ile within a fraction, the variab ility decreased with increasing energy output. The highest variability was obser ved at 0 J (i.e., in the water-dispersible aggregates). All three size fractions were su sceptible to organic matter loss by wetting, with values ranging from 5 to 17% (Table 3.2). The selected equation proved satisfactory to model the response of AOM to sonication energy output for 250 to150 m (R2 = 0.93) and 150 to 53 m (R2 = 0.96) fractions (Table 3.2). For the 2000 to 250 m fraction, however, the equation did not work as well (R2 = 0.84) due to the well-defined steps in the loss of organic matter. The DRIFTS spectra separated the POM and AOM fractions (Fig. 3.4). The AOM fraction exhibited higher quantities of polysaccharides (1160 cm-1), aromatic rings (1500 cm-1), esters (1730 cm-1) and amides (1650 cm-1) than the POM fraction. Th e negative peak at 1350 cm-1 on the POM spectra was due to distor tions caused by the spectral s ubtraction and should be ignored.

PAGE 53

53 The peaks at 2880 cm-1 and 2950 cm-1 were due to aliphatic -CH groups. Peak heights indicated higher content of these materials in the AOM component within the size fractions. Effect of Management Intensity The effects of energy input and size frac tions on AOM in the size fractions were significant at p < 0.01 (Table 3.3). There was al so a significant energy level x management intensity x size fraction interaction in the AOM data. The effect of management intensity on AOM was statistically signi ficant in the 2000 to 250 m fraction, with an energy level x management intensity interaction present (F ig. 3.5). At lower energy levels, treatment differences were small. However, beyond about 6,000 J the low intensity treatment had higher AOM. The energy level x management intensity in teraction was also significant for the 250 to 150 m fraction, but the absolute difference between the management intensities was less than 4%. The 150 to 53 m fraction was not affected by the intensity of management. Discussion Forested Spodosols of southeastern U.S. are important regional C si nks (Richter et al., 1995) and understanding the C sequ estration mechanisms in these soils can help in maintaining and improving the C storage potential of these fore sts. Physical protectio n of C by incorporation into aggregates is an important mechanism of C sequestration and this study was conducted with an objective of understanding the different aspects of aggregation in these soils. Aggregate Morphology, Stability and OM Content The microscopic observations confirmed the presence of aggregates and a range of aggregate forms (Fig. 3.1). Thus, the hypothesis regarding the presence of aggregates was accepted. The images further implicated the role of fungal hyphae and fine roots in aggregation (Fig. 3.1.1-3.1.5), while the DRIFTS spectra suggested polysacchar ides as an aggregate binding agent (Fig. 3.4). Oades (1993) disc ussed the importance of biologi cal agents of aggregation in

PAGE 54

54 sandy soils, since the abiotic mechanisms of aggr egation are most important in soils with clay contents greater than 15% (Horn, 1990). The binding action of polysaccharides, secreted by fungi or bacteria, has also been previously reported (Tisdall, 1994; Caesar-Tonthat, 2002; Blanco-Canqui and Lal, 2004). Our observations provide justification to further study the significance of roots, fungi and polysaccharides as biological agents for aggregation in these sandy soils, while questioning th eir role in sequestering C. Sonication facilitated making quan titative estimates of aggregate stability and aggregate C. The hypothesis (second objective) s uggesting an inverse relationshi p between aggregate size and stability was rejected because the 250 to150 m fraction was less stable than both the 2000 to 250 m and 150 to 53 m fractions (Fig. 3.3). The higher ag gregate strength of the 2000 to150 m fraction, which was counter to our hypothe sis, appeared to be a function of a microaggregate/macroaggregate hierarchical stru cture (see discussion below; Oades and Waters, 1991). Aggregates, through physical occlusion, protec t organic matter from destructive agents such as physical breakdown by tillage, removal of finer particles by erosion, or decomposition by soil organisms of different sizes. Not all aggregates offer protection from a ll of these agents. The extent of protection depends on the size of pores within the aggregates and the strength of the aggregates, which in turn depends on the binding agents and the size of the primary particles in the aggregate. The aggregate strength should in dicate the extent of m echanical protection (e.g., from breakdown by tillage), but the extent of protect ion from the soil microbes is not certain. In the absence of sufficient clay, it is possible that one form of organic matter may be protected by another form. For example, a rind of biochemica lly inert material such as aromatic compounds may discourage the entry of microbe s inside the aggregates and pr otect the labile organic matter

PAGE 55

55 inside. Sonication, as proposed in this study, combined with chemical analyses of the C lost from aggregates of differential stability, offers the opportunity to advance these studies. The hypothesis suggesting the dominance of POM was only partially supported, since this fraction accounted for just over half of the total organic matter. On the other hand, the possibility of partial breakdown of POM during sonication does suggest that a greater proportion of total SOM may be in this pool. Though AOM is an im portant pool of sequestered C, POM may be a more important in determining the short-term turnover of essential nut rients (Haynes, 2005). Effect of Management Intensity The AOM in the 2000 to 250 m fraction was reduced by intensive management (Fig 3.5). Therefore, the hypothesis of a short-term reductio n in aggregation by intensive management was accepted. The effect of management intensity could be partly attributed to changes in fine root biomass. Intensive management, especially chem ical control of understo ry plants, has been reported to decrease the fine root biomass and length (Escamilla et al., 1991; Shan et al., 2001). Higher decomposability due to fe rtilization (Polglase et al., 1 992a) may also contribute to reductions in AOM. However, it is unclear whether these diffe rences between management intensities will be sustained over time. The hi gher levels of productivity and C inputs reported under intensive management (Dalla Tea and Joke la, 1991; Jokela and Ma rtin, 2000) could result in longer-term opportunities for higher aggregati on; especially after canopy closure, when nutrient deficiencies reduce site productivity le vels under the low intens ity management regime (Jokela et al., 2004). Methodological Considerations The response of AOM in the 2000 to 250 m size fraction to management intensity (Table 3.1, Fig 3.5) illustrated the sensitivity of the soni cation technique to soil organic matter changes in as few as four years after treatment. It also used operationally defined C fractions that could be

PAGE 56

56 related to meaningful C pools. The two pools sepa rated by this procedure within a size fraction, POM and AOM, showed a marked difference in th eir chemical composition as indicated by the DRIFTS spectra (Fig. 3.4). The AOM showed highe r content of polysacchari des, phenols, esters and amides; of which polysaccharides have alr eady been shown to function as binding agents (Tisdall, 1994). The higher aromatic C cont ent also suggested th e presence of more biochemically inert organic matter. Higher amo unts of esters and amides, on the other hand, suggested that this fraction is susceptible to decomposition if the aggregates were destroyed, since esters and amides are highly reactive C forms. The differences in chemical composition indicated that this method should be useful in separating soil C into more functional pools and lead to better conceptualization of the cycling of SOC when used in conjunction with chemical decomposition/mineralization studies. North (1976) used a similar method of aggreg ate strength measurement, which has been reproduced by others (Schmidt et al., 1999; Roscoe et al., 2000). However, North (1976) used the amount of clay lost as an indicator of aggr egate destruction, which made it unsuitable for the highly sandy soils examined in this study. Clay can become saturated with organic matter (Hassink et al., 1997) and so it is not necessarily an appropriate measure for protection of C in soils with low clay content. This indicates the su itability of organic matter release instead of clay release in the sonication techni que. The method described here can also be used to study aggregates of varying stability by sonicating the soil at different energy le vels and analyzing the remaining aggregates for properties such as age or mineralizability. Aggregate Structure in Coastal Plain Spodosols – Additional Considerations The structure in Spodosols has been descri bed as weak (Carlisle et al., 1981, 1988, 1989; Sodek et al., 1990). Consequently, aggregation in these soils has received little, if any, attention. Oades and Waters (1991) studied the patterns of aggregate breakdown in Mollisols, Alfisols and

PAGE 57

57 Oxisols. In Oxisols, the aggr egates broke down to release primary particles. In Mollisols and Alfisols, the authors reported a hierarchical stru cture; where the larger, weaker aggregates broke down to release smaller, stronger aggregates, before breaking down into primary particles. Results from this study indica ted that the surface horizon of sandy Coastal Plain Spodosols also had an aggregate hierarchy, as exhibited by the step-wise breakdown of ag gregates in the 2000 to 250 m and 250 to 150 m fractions (Fig. 3.3). In the 2000 to 250 m fraction, the least stable aggregates were disr upted at energies 2500 J. At energy levels grea ter than 2500 J, a plateau in percent OM lost (AOM) indicated that there wa s no further aggregate destruction until 3700 J. We interpret this energy level as the threshol d for aggregates of second order, which were destroyed between 3700 J to 6000 J. A second plateau was observed between 6000 J to 7500 J. This was the threshold for the third order of aggregates, which started breaking down at 7500 J and the loss went on until 13300 J when all the aggreg ates in this fraction were destroyed. In the 2000 to 250 m fraction, the well-defined steps indicated a well-developed structure. In the 250 to 150 m fraction, although the first tw o steps were observed at the same energy levels, they were less distinct and all the aggregates were destroyed at 8500 J, indi cating a poorly developed structure. However, the 150 to 53 m fraction exhibited a continuous spectrum, suggesting that this fraction was simply a continuum of aggregates of different stabilities. This behavior appears unique for these Spodosols and requires further study, especially since this C pool remained unaffected by the intensive management regime. Edwards and Bremner (1967) defined microaggr egates as the water stable aggregates bound by strong clay-polyvalent metal-organic matter complexes. The authors used 250 m as the separation point between macro and microagg regates. Tisdall and Oades (1982) also used 250 m as the separation point between micro and macroaggregates. They described

PAGE 58

58 microaggregates as those stabilized by organomineral complexes and resistant to disruption by wetting, cultivation or other disturbances. In cont rast, macroaggregates were described as those stabilized by roots and funga l hyphae, having varying stability depending upon management and other factors. Apparently, the separation point at 250 m was chosen based on the increase in aggregate strength and difference in binding agents. However, the similar behavior of the 2000 to 250 and 250 to150 m fractions, as well as the low stabil ity and presence of roots and fungal hyphae in the 250 to 150 m fraction, suggests that 150 m is a more appropriate separation point for our soils. The 150 to 53 m fraction, although stable and unaffected by management, was susceptible to loss by wetting (Table 3.2). Th is indicates that size alone is insufficient to define aggregates types in th ese sandy textured soils. A quan titative measure of aggregate strength such as the energy required to achieve complete aggregate disruption can be used to improve the definition. It should be noted that sonication was used to measure mechanical strength. Chemical modifications may break aggregates before the threshold for mechanical failure is reached. These factor s, in addition to the limited pot ential for clay or cation binding, indicate the necessity of a different approach for studyi ng aggregation in sandy Spodosols. Conclusions This study found that aggregates form in sandy Spodosols and they have a hierarchal structure in the large soil size fractions. The use of organic matter release in stead of clay release after aggregate breakdown by soni cation was useful for studying aggr egate properties. It allowed simultaneous measurement of aggregate strength and amount of aggregate organic matter. Aggregate C was an important pool in these soils and the intensiv e management regime used to enhance pine plantation productivity reduced this C pool significan tly in the short-term. Results from this study highlight the necessity of using a different approach for aggregate classification in sandy soils as well as for quantification of aggregate characteristic s like stability. Finally,

PAGE 59

59 results from this study highlight the need to as sess the long-term, management-related changes using quantifiable C pools for assessi ng the metrics of sustainability.

PAGE 60

60 Table 3.1. Amount of organic C held in soil aggr egates for a sandy Spodosol in north Florida. Size fraction ( m) C content in size fraction† (g C in size fraction 100 g-1 soil) C content in soil aggregates (g C in fraction 100 g-1 soil) 2000 to 250 1.1 0.69 (63%)‡ 250 to150 0.36 0.30 (84%) 150 to 53 0.39 0.28 (72%) < 53 0.15 NA§ Total 2.8 1.28 (46%) † The C content was measured by loss on ignition and the standard Van Bemmelen factor (0.58) was used for conversion of organic matter into C content. ‡ The values in parentheses represent the averag e proportion of organic matter lost from the size fraction after sonication at the highest energy level. The percent loss remained the same for organic matter and C. See Table 3.2 for the statistical analysis. § NA: Not Applicable, since this fractio n was not analyzed for aggregation.

PAGE 61

61 Table 3.2. Energy output of the sonicator probe and the amount of organic matter lost from each soil size fraction as aggregat e organic matter (2000 to 250; 250 to 150 and 150 to 53 m) for each energy level for a sandy Spodosol in north Florida. 2000 to 250 m 250 to 150 m 150 to 53 m Energy (J) Aggregate OM (%†) Energy (J) Aggregate OM (%) Energy (J) Aggregate OM (%) 0 (0) 5‡ a§ (1) 0 (0) 17 a (2) 0 (0) 6 a (1) 945 (6) 14 b (1) 936 (11) 39 b (1) 2813 (34) 23 b (1) 2729 (64) 22 c (2) 2468 (55) 54 c (1) 5997 (55) 45 c (2) 3793 (21) 23 c (2) 3805 (19) 60 c (1) 9445 (57) 59 d (1) 5973 (35) 45 d (2) 5969 (44) 76 d (1) 12186 (84) 65 e (1) 7446 (47) 47 de (3) 7450 (52) 78 d (1) 13513 (270) 67 ef (1) 8481 (42) 55 ef (2) 8467 (54) 82 d (1) 16422 (362) 70 f (1) 9489 (67) 58 fg (3) 9467 (65) 83 d (1) 20447 (536) 71 f (1) 10505 (55) 60 fg (2) 10516 (44) 84 d (1) 22297 (208) 72 f (1) 13367 (55) 60 fg (3) N A†† N A N A N A 19691 (312) 63 g (2) N A N A N A N A y =(95.7*x)/(7508.3+x) # R2 = 0.84 y =(97.1*x)/(1773.2+x) R2 = 0.93 y =(98.4*x)/(7038.2+x) R2 = 0.96 † The aggregate OM or AOM is expressed as a percent of the total OM in the size fraction. ‡ Each value is a mean of 12 samples. § Within a size fraction, the means followed by differe nt letters are statistically different at p < 0.05 and show the effect of sonication energy le vel. The Tukey-Cramer adjustment for mean separation was used. Values in parentheses re present standard error. # The equations are in the form y = (a*x)/ (b + x); where, a = maximum AOM, b = energy level at 0.5*a, x = sonicator energy output (J) and y = AOM. †† NA = Not applicable.

PAGE 62

62 Table 3.3. Effects of forest management intensit y and soil depth on aggregate organic matter in the 2000 to 250, 250 to 150 and 150 to 53 m fractions for a sandy Spodosol in north Florida. Main Effects / size fraction 2000 to 250 m 250 to 150 m 150 to 53 m Management intensity p = 0.03 NS NS Soil depth NS† NS NS Intensity*depth NS NS NS Energy level p < 0.01 Fraction size p < 0.01 Energy Level x Intensity x Fraction p = 0.02 †NS = Not significant (p>0.05)

PAGE 63

Figure 3.1. Observations of soil aggregation in a sandy surface horizon of a Coastal Plain Spodosol. Figures 3.1.1 to 3.1.4 are irregularly shaped aggregat es showing fungal hyphae, organic debris and mineral matter enmeshed together. Magnification: 1.1 = 87X, 1.2=55X, 1.3 = 295X, 1.4 = 217X. Figures 3.1.5 to 3.1.7 show spherical aggregates, which are encrustations of mi neral matter on organic debris combined with fungal hyphae/fine roots. Figures 3.1.8 and 3.1.9 are photographs of the 250 to 150 m fraction. Figure 3.1.8 shows the soil after dry sieving, while Figure 3.1.9 is after sonication and shows clean sand grai ns and particulate organic matter but no aggregates.

PAGE 64

64 (1) (2) (3) (4) (5) (6) ( 7 ) ( 8 ) ( 9 ) 150 m 250 m 250 m

PAGE 65

65 Figure 3.2. Effect of sonication en ergy input on the loss of aggreg ate organic matter (AOM, % of total OM in size fraction) af ter sonication of the 150 to 53 m fraction. The error bars represent the range of values (n = 12 samples), while the box represents interquartile range (upper quartile = 75th percentile, lower quartile = 25th percentile). The plus sign in the box represents the mean and the line in the box represents the median.

PAGE 66

66 0 10 20 30 40 50 60 70 80 90 0500010000150002000025000 Energy (J)AOM (% of total in size fraction) 2000 to 250 m 250 to 150 m 150 to 53 m Figure 3.3. Loss of aggregate organic matter (AOM % of total OM in size fraction) with increasing energy for the various soil size fractions. The vertical lines indicate the steps in the continuity of organic matter lost from aggregates for the 2000 to 250 and 250 to 150 m fractions. Each data point represents the mean of 12 samples.

PAGE 67

67 -1 -0.5 0 0.5 1 1.5 050010001500200025003000350040004500 Wave Number cm-1Peak Height POM AOM a b c d e Figure 3.4. Diffusive Reflectance Infra-red Fourier Transformed Spectra (DRIFTS) showing characteristics of particulate organic matter (POM) and aggregate organic matter (AOM) of the 250 to 150 m fraction for a sandy Spodosol in north Florida. The AOM spectra show peaks are: (a) polysaccharid es (b) aromatics (c) esters (d) amides (e) C-H bonds.

PAGE 68

68 0 10 20 30 40 50 60 70 0500010000150002000025000 Energy (J)AOM(% of total in size fraction) High Low Management intensity si gnificant at p < 0.01 Figure 3.5. Effect of forest management intens ity on the amount of aggregate organic matter (AOM, % of total OM in si ze fraction) for the 2000 to 250 m fraction for a sandy Spodosol in north Florida. Each data poi nt represents the mean of 6 samples.

PAGE 69

69 CHAPTER 4 GENOTYPIC AND FOREST MANAGE MENT EFFECTS ON SIZE-DENSITY FRACTIONATION OF SOIL CARB ON IN A FORESTED SPODOSOL Introduction Genetic deployment of fast growing and dis ease resistant families is a key factor for enhancing forest plantation productivity (McKeand et al., 2003). Contrasti ng tree species are also known to differentially influence aspects of soil C accumulation and cycling. For example, acidic soil condition beneath spruce ( Picea spp ) canopies has reportedly lowered microbial biomass and produced lower rates of microbial CO2 than the soils beneath beech ( Fagus spp ) or oak ( Quercus spp ) forests (Anderson and Do msch, 1993). Scrub oak ( Q. dumosa Nutt) also has been shown to contain more soil organic carbon (SOC) in all soil size fractions compared to soils under coulter pine ( Pinus coulteri B. Don ); while litter C was higher beneath the pine forests (Quideau et al., 2000). The chemical character of SOC can also be influenced by species. In the last example, O-alkyl C decrea sed from the litter to the macro organic matter (water-floatable) and fine silt size fractions fo r both species, but the decrease was more pronounced under pine, indicating greater decomposition of cellulo se and hemicellulose in pine litter. Families within a single species may also differ in their adaptative capabilities and biomass accumulation. For example, loblolly pine ( Pinus taeda L.) is an important plantation species in the United States, Brazil and Arge ntina, among other countries (Schultz, 1999). Families of loblolly pine vary in commercially and ecologically importa nt qualities, such as biomass production, light use efficiency, and nutri ent use efficiency (Pope, 1979; Crawford et al., 1991; McCrady and Jokela, 1998 ) as well as fusiform rust resistance (Schmidt, 2003) and response to tropospheric ozone (Taylor, 1994). Ho wever, it is not well-known if and how families affect the accumulation of soil C. The differential growth response of families to intensive forest management argues that they ma y be a factor influencing amount and quality of

PAGE 70

70 SOC. Since genetically improved families of loblo lly pine are an important element of intensive management, a better understanding of the influence of genetics, a nd their interaction with forest management practices on SOC is warranted. One of the major mechanisms involved in management-related changes in SOC is the change in quality and quantity of organic matte r inputs. Density fractionation has been used by many researchers to understand th ese changes (Romkens et al., 1999; Echeverria et al., 2004), as this technique separates the organic matter from the mineral matter and accentuates the differences in organic matter. The light and me dium density pools are reported to be actively recycling fractions with higher C an d N concentrations and faster tu rnover rates. In contrast, the heavy density fraction is reported to be passive, characterized by low C and N concentrations and slow turnover rates (Swanston et al., 2005). As families can be e xpected to differentially impact SOC through litter and root C inputs, density frac tionation may offer sufficient sensitivity to detect these differences. This study was undertaken with the overall objective of closing this gap in knowledge. The first objective was to study the short-term effects of family and family x management interactions on the surface soil C pools. The hypothesis related to th is objective was that the best family (family designations chosen a priori based on growth performa nce in long-term genetic experiments) would promote the greatest increase in SOC, a nd the effect would be most pronounced under intensive management. This hypot hesis was based on the observation that the best family used in this study was highly res ponsive to fertilization, and produced the most litterfall under in tensive management (E.J. Jokela, unpublished data). The second objective was to determine the pr ofile of C and N in a typical forested Spodosol and identify the size-densi ty fractions most responsive to the varying levels of forest

PAGE 71

71 management. The first hypothesis related to this objective was that the medium and light density fractions would be the main reservoirs of C a nd N in these soils. This was based on the premise that the mineral fraction in these soils is pr edominantly quartz sand, with little or no sorption capacity for SOC. The second hypothesis was that th e light and medium density fractions in the 2000 to 250 m fraction would be the responsive pools fo r detecting genotypic differences in SOC as influenced by varying management inte nsity. This hypothesis was formulated on the observation that in preliminar y studies (Chapter 2, 3); this fraction responded most to management intensity. It was expected that th e effects on organic matter would be accentuated by separating the density fractions. Materials and Methods Experimental Site A loblolly pine plantation in north Florida (30o24’N lat; 82 o33’W long) was the study site. It is managed by the Forest Biology Research Cooperative at the Univer sity of Florida, as part of the Pine Productivity Interacti ons Experimental Study (PPINES). This long-term study aims at understanding the family x environment interactions in full-sib families of loblolly and slash pine ( P. elliottii Engelm. var. elliottii ). The climate is warm, humi d subtropical, with 1,394 mm average annual rainfall, 270C average annual maximum temperature and 130C average annual minimum temperature (NOAA, 2002) The soil is classified as a Leon series (sandy, siliceous, thermic Aeric Alaquod), with <5% silt + clay and <10 cmolc kg-1 of cation exchange capacity. The trees were planted in January 2000 in f our replicates using a randomized complete block, split plot design. Prior to planting, the entire study was double bedded and treated with the herbicides Arsenal (imazapyr – 1.02 L ha-1) and Garlon (triclopyr – 7.02 L ha-1) to remove the understory vegetation and to reduce competition. The experimental design was a 2 x 2 x 8 factorial including two planting de nsities (close spacing at 3 x 1.3 m and wide spacing at 3 x 3

PAGE 72

72 m), two management regimes (high and low input) a nd six elite loblolly pine full-sib families, a mix of these elite families and one poor growing family. Out of these, six treatment combinations (1 x 2 x 3) representing the close spacing, both management regimes and three families were chosen. The full-sib families were chosen a priori based on their above-ground growth performance in long-term genetic experi ments and included the best grower, a medium grower and the poorest grower (abov e ground biomass 45.2, 42.6 and 40.6 Mg ha-1 respectively at age 5; Roth et al., 2006). The high intensity management regime consisted of sustained understory competition control and annual fertilization us ing a complete fertilizer. The low intensity management regime consisted of a one-time fertili zation and understory competition control at planting. At age 6, when the soil was sampled, the fertilizer adde d to the high intensity treatment totaled 368 kg ha-1 N and 128 kg ha-1 P plus nearly all other e ssential nutrients (i.e., 121 kg ha-1 K, 45 kg ha-1 Mg, 45 kg ha-1 Ca, 35 kg ha-1 S, 0.89 kg ha-1 B, 3 kg ha-1 Zn, 2 kg ha-1 Mn, 16 kg ha-1 Fe, 4 kg ha-1 Cu, 0.01 kg ha-1 Mo). The low intensity treatment had 45 kg ha-1 N and 50.6 kg ha-1 P applied as diammonium phosphate. In the high intensity treatment, the herbicides Velpar or Oust and Glyphosate applied at labeled rate s provided sustained understory co mpetition control. The entire study was treated when necessary with insecticides (Dimilin, P ounce or Mimic) for tip moth ( Rhyacionia spp.) control during the fi rst growing season. Each tr eatment plot was 480 m2 in size. Soil samples were collected from the A horizon of interbeds at soil depth increments of 0 to 5 and 5 to 10 cm from each treatment plot in three of the four replicate blocks in March 2005. Two composite samples for each depth of each plot (treatment within a block) came from about six individual soil samples. The individual soil samples for each composite sample were

PAGE 73

73 collected from alternate rows (interbed position) in two lines diagonally across the plot. The core method was used to measure bulk density (0 to 5 cm depth) based on two samples per treatment plot. Laboratory Methods Size fractionation Soil samples were air-dried and passed thr ough a 2000 m sieve. The <2000 m fraction was further size fractionated using dry sievi ng (see Chapter 2). Four size fractions, a macroaggregate fraction 2000 to 250 m, two microaggregate fractions (250 to150 m and 150 to 53 m), as well as a <53 m fraction were obtained. The size fractions, including the > 2 mm size fraction, were ground to a fine powder and analyzed for total C and N concentrations on a Carlo-Erba Analyzer (CE Instruments, model NCS-2500). The three 2000 to 53 m fractions were also analyzed for organic ma tter content using loss on ignition. Density fractionation The most widely used liquids for density se paration are water, sodium polytungstate and Ludox, an inert silica suspension (Christensen, 1992; Cambardella and Elliott, 1993; Meijboom et al., 1995). Sodium polytungstate is expensive, toxic and is reported to hinder mineralization (Sollins et al., 1984). Ludox was test ed initially in this study but appeared to dissolve some C from the size fractions, as shown by the consid erably darkened supernatant. Therefore, a modified density separation procedure based on Me ijboom et al. (1995) was used, with water as a separating liquid. For each size fraction, 10 g soil was added to a 50 mL beaker. Twenty-five mL water was added to this beaker and the organic matter was separated by swirling and decanting into a 600mL beaker. The process was repeated until no more organic matter could be visually separated. The light fraction floating on top of th e water in the 600 mL beaker was periodically

PAGE 74

74 transferred to another beaker to avoid co-precipitation of the li ght density fraction (settling down under the weight of medium density fraction). Organic matter, which could not be separated from mineral material, was termed the heavy fr action. The suspension in the 600 mL beaker was added to a 250 mL funnel with the suspension of light density material added on top. Organic matter, which settled in water, was termed th e medium density fraction, while organic matter, which floated on water, was termed th e light density fraction. The 2000 to 250 m fraction needed 24 hrs to separate light and medium densities, while the 250 to 150 m fraction needed 12 hrs to achieve a clear separation. The separa ted fractions were passe d through the same size sieve used for obtaining the size fraction (e.g., 250 m sieve for the 2000 to 250 m fraction) in order to separate the water-dispersible aggreg ate fraction. The water-di spersible fraction was defined as the organic matter passing through th e sieve after the size fr actions were densityfractionated with water. The samples were th en air-dried and used for further analysis. Preliminary studies indicate d that in the 150 to 53 m fraction, more than 90% of the SOC was in the medium density fraction. Therefore, this fraction was not density fractionated. Sonication The medium density and heavy density fractions of the 2000 to 250 and 250 to 150 m fractions, as well as the whole 150 to 53 m fraction, were further fr actionated into aggregate organic matter (AOM) and particulate organic ma tter (POM) using sonicati on. In this analysis, a 5 g sample was used for the heavy density fractions, a 1 g sample for the medium density fractions and a 2 g sample for the 150 to 53 m whole fractions. The energy level for sonication was chosen based on the preliminary an alysis (Chapter 3). The 2000 to 250 m and 150 to 53 m fractions were sonicated at 20000J for (60% amplitude, 5 minutes; pulse method 60 sec ON and 30 sec OFF), while the 250 to 150 m fraction was sonicated at 10000J (40% amplitude, 5

PAGE 75

75 minutes; pulse method 60 sec ON and 30 sec OFF). Loss on ignition was used to measure the AOM and POM separated by sonication. Statistical Analysis The statistical significan ce was analyzed using ‘PROC MIXED’ (SAS, 1996), with management intensity, depth, size fraction, and fam ily as fixed effects and block and replication (nested within block) as random effects. Differe nces were considered significant at p<0.05. As the initial analyses showed that the size and de nsity fractions were signi ficantly different from each other with interactions betw een depth, intensity and family, further analyses were carried out for the individual fractions. To analyze the si gnificant interactions within each size-density fraction, the contrast procedure was used for post-hoc comparis ons to ascertain differences among least square means. Results Management Intensity and Family Effects Family and forest management intensity did not affect SOC and N among the size fractions, except for the >2000 m size fraction. In this fraction, the high intensity management regime reduced the C:N ratio by 17% (60 in high intensity vs. 73 in low intensity). However, the density fractions of both the 2000 to 250 m and 250 to 150 m size fractions exhibited significant effects of manageme nt intensity and family. In the 2000 to 250 m light density fraction, the best family produced higher C content than the poor family (Fig 4.1). On the other hand, the medium family exhibited higher AOM in the 250 to 150 m me dium density fraction (Fig 4.2). There was also an intensity x family interaction for N concentr ation in the 2000 to 250 m medium density (Fig 4.3) and AOM in the heavy density fractions (Fig 4.4). The medium family had significantly higher soil N concen tration when grown under the low intensity management regime (Fig 4.3). The best family trended toward higher N concentrations under the

PAGE 76

76 high intensity management regime, with the diffe rence being statistically significant at p = 0.06. In the 2000 to 250 m heavy density fraction (Fig 4.4), the medium family showed higher AOM under the high intensity management regime (b est family 68% vs. medium family 80%). Distribution of C and N in the Size-density Fractions Among the size fractions, C concentration (% of fraction) was highe st in the >2000 m fraction (Table 4.1), while C content (% of w hole soil), was highest in the 2000 to 250 m fraction (Fig 4.5). Among the dens ity fractions, C content was highest in the medium density (47-85%; Fig 4.6) in both the 2000 to 250 and 250 to 150 m fractions. The C concentration was highest in the light density 2000 to 250 m fraction (Table 4.2). The light density material of the 250 to 150 m fraction accounted for <1% of the fraction’s weight and C content. However, it accounted for 6 to 10% of the C in the 2000 to 250 m fraction. The water-dispersible fraction accounted for 7 to 16% of the tota l C in the size fractions. The me dium density of the 250 to 150 m fraction showed the highest AOM (81 to 82%; Table 4.3), while the medium density of the 2000 to 250 m fraction showed the lowest amounts (65 to 66%). Nitrogen concentrations in the size fractions followed the same trend as C (Table 4.1), but the N content was highest in the 250 to 150 m fraction (Fig 4.5). The C: N ratio was highest in the >2000 m fraction (55-73; Table 4.1) a nd lowest in the 250 to 150 m fraction (9-16). Among the density fractions, N content was highest in the heavy density, es pecially in the 250 to 150 m fraction (84 to 92% of total N in the frac tion; Fig 4.7). The heavy density also exhibited extremely low C: N ratios (2.2-3, Table 4.2). The C: N ratio of th e light density in the 2000 to 250 m fraction was highest (56 to 58, Table 4.2) among the density fractions and was similar to the >2000 m fraction (55 to 73, Table 4.1).

PAGE 77

77 Effect of Depth The 0 to 5 cm depth had higher C and N con centrations in all size fractions except the >2000 m fraction, which had its higher C concentr ation at the 5 to 10 cm depth (Table 4.1). However, the effect was more prominent for C th an for N. There was also a significant depth x fraction interaction in C and N contents of the soil fractions (Fig 4.5). The 2000 to 250 m fraction exhibited higher content at the 0 to 5 cm depth, while all other fr actions exhibited higher content at the 5 to 10 cm depth. The 250 to 150 m fraction in part icular exhibited significantly higher N content (42 vs 28% of tota l soil N) at the 5 to 10 cm depth. There was also a depth x density fraction in teraction. In both size fractions, the medium density had higher C concentration at the 0 to 5 cm depth, while th e other three density fractions had higher C concentration at the 5 to 10 cm depth (Table 4.2). The 0 to 5 cm depth also showed higher AOM in the 2000 to 250 m heavy density fraction, while the 5 to 10 cm depth showed higher AOM in the 150 to 53 m fraction and higher POM in the 2000 to 250 m heavy density fraction (Table 4.3). Discussion This study was undertaken with the objective of understanding how forest management activities and genetic de ployment affect the C in forested Spodosols of the southeastern U.S. Utilization of genetically improved seedlings for growth and disease resistance, coupled with management practices that include understory co mpetition control and fertilizer applications, are common approaches used for increasing the pro ductivity of managed plantation forests in the southeastern U.S. Yet, the impacts of mana gement practices and ge netics on soil related processes are still poorly documented.

PAGE 78

78 Effects of Family The best performing family added more decom posable organic matter to the soil as shown by the higher C content in the light density fracti on (Fig 4.1). The light density fraction has been reported to be highly mineralizable due to its hi gh C and N concentration and lack of mineral protection (Romkens et al., 1999; Swanston et al ., 2005). The best family also showed a trend toward higher soil N concentrations under the hi gh intensity management regime (Fig 4.3). This family was characterized by much higher yield under intensive management. It had 31% higher stem volume and 11% more aboveground biomass as compared to the poor performing family at age 5 (Roth et al., 2006). It also exhibited higher N concentrations in the foliage (Jokela, E.J., unpublished data). These factors suggest that a comb ination of the best family and the intensive management regime may be favorable for faster C turnover rates. Although this is a positive factor for availability of nutrients such N, P a nd S in poor fertility soils, the consequences for long-term C sequestration are unc ertain. These data document early trends through age 6 and it will be necessary to determine whether these effects are sustained over time. The medium performing family had higher leve ls of aggregation as shown by the higher AOM values in the 250 to 150 m medium density fraction (Fig 4.2). A po ssible explanation for this effect is the difference in root biomass and root architecture of these families, since fine roots are important aggregating agents (Tisdall and Oades, 1982). This family also exhibited higher N under the low intensity management re gime (Fig 4.3), which was reported to have greater aggregation (Chapter 3). This suggests a role of aggregat es in N storage. In the 2000 to 250 m heavy density fraction, this family show ed higher aggregation under the high intensity management regime (Fig 4.4), suggesting that the positive impact of this family on aggregation was independent of management intensity. It is possible that the genotypic effects were better expressed under intensive management as it ha s been reported to reduce the environmental

PAGE 79

79 variation (Lopez-Upton et al., 1999 ). This result also offers a possibility for improving the C sequestration potential of these ecosystems wit hout sacrificing yield th rough choices of family deployment. Fraction Characteristics and Effect of Depth The high C:N ratios of the medium and light density in the 2000 to 250 m fraction (Table 4.2), which were similar to those of the > 2000 m fraction (Table 4.1), indicated that OM in these fractions mainly came from recently added lit ter and root biomass. Similar high C:N ratios, especially for the light density fraction have been reported by many authors (Swanston et al., 2004; Gregorich et al., 2006; Liao et al., 2006) This interpretation was supported by the higher C content in the 2000 to 250 m size fraction (Table 4.1), as well as in the medium density of this fraction at the 0 to 5 cm depth(Ta ble 4.2), which can be expected to receive higher root and litter inputs. This was in accordance with the conclusion in Chapter 2 that the 2000 to 250 m fraction received fresh organic matter. The high C:N rati o of the 250 to 150 m medium density fraction (Table 4.2) indicated that th is fraction also received fr esh organic matter inputs. The high N content of the 250 to 150 m fracti on (42% of total N at 5 to 10 cm depth; Table 4.1) indicated that it may be responsible for th e storage of N. It al so exhibited the highest proportion of aggregate C as shown by the high AOM values (Table 4.2), which supports the importance of aggregates for the N storage in th ese soils. Within this size fraction, the heavy density showed the highest N content (84 to 92% of the N; Fig 4.7) and lowe st C: N ratios (2.2 to 3, Table 4.3). The heavy density 2000 to 250 m fr action also had low C:N ratios. This finding highlights the importance of heavy density soil ma terial in the N storage of these soils. Zhong and Makeschin (2006) also reported that the heavy density fraction contained more labile N than the light density fraction in temperate forest so ils. Stable sorption of nitrogenous compounds like amides on the mineral surfaces (Sollins et al., 2006) can be one of the mechanisms responsible

PAGE 80

80 for the high N content and low C/N ratios of the heavy density fractions. Presence of microbial N protected by the aggregates may represent an other mechanism. Foster (1988) reported that aggregates offered prime sites for microbes due to the protection from sudden changes in moisture and from predation by the protozoa and other larg e predators. Therefore, it is likely that the N stored in the aggregates is of microbia l origin. Although these ra tios are unusually low, Tscherko et al. (2003) re ported similar values for microbial biomass (lowest value reported 1.4). A comparison of the N content in AOM and mine ral fraction or the measurement of microbial C/N ratios by fumigation extracti on would be necessary to provide further insight into this phenomenon. Methodological Considerations The modified density fractionation proce dure offered many advantages. There was minimal chance of chemical alte ration of the organic matter since only water was used during the procedure. This also made the procedure inexpe nsive and easy to use in any laboratory. It allowed fraction-wise estimation of the water-dispersible aggregat es, for which there is no other method available at present. In agricultural soils, only the waterstable aggregat es are usually measured, since these soils are continually expo sed to splash and wind erosion. However, the forested soils are protected from the erosive forc es by a thick litter layer. Therefore, the waterdispersible aggregates are likely to be important for C dynamics as well as the short-term cycling of N, P and S in these soils. The medium and li ght density fractions sepa rated by this procedure showed effects of family and management intens ity on the SOC pools as early as six years after the treatments were imposed. This indicated th e applicability of this method for detecting management related changes. Other researchers, using different procedur es, have also reported the sensitivity of light and medium density fractions to the short and long-term management related changes (Romkens et al., 1999; Echeverria et al., 2004).

PAGE 81

81 Conclusions The 2000 to 250 m fraction contained the most C, while the 250 to 150 m fraction was most important for N storage. Of the density frac tions studied, the medium density contained the most C, while the heavy density contained the most N. The be st growing family added more decomposable C as indicated by high er C in the light density 2000 to 250 m fraction, while the medium family encouraged better aggregation, as show n by higher AOM. These findings indicate that understandi ng the short-term as well as longterm genotypic effects on soil C would be necessary for carbon-wise management of th e forested Spodosols in north Florida. Waterdispersible aggregates were an important compone nt of these soils, as they accounted for 7 to 16% of the C in the two largest size fractions. Th e modified density fractionation procedure used in this study offered an inexpensive and e ffective way for separating soil C pools and for detecting management related changes in them.

PAGE 82

82 Table 4.1. Characteristics of the size fracti ons for a sandy Spodosol in north Florida Variable Depth >2000 m 2000 to 250m 250 to 150m 150 to 53m <53m 0 to 5 cm 18.8† h (1.0) 2.21 e (0.13) ‡ 0.43 b (0.02) 1.48 d (0.06) 9.99 g (0.36) C concentration (% of fraction) 5 to 10 cm 23.6 i (1.1) 0.89 c (0.08) 0.17 a (0.02) 0.66 c (0.04) 5.83 f (0.26) 0 to 5 cm 0.35 de (0.02) 0.071 c (0.004) 0.029 ab (0.002) 0.063 c (0.003) 0.365 e (0.016) N concentration (% of fraction) 5 to 10 cm 0.33 de (0.01) 0.030 ab (0.002) 0.024 a (0.002) 0.037 b (0.002) 0.228 d (0.009) 0 to 5 cm 55 d (1) 31 c (1) 16 b (1) 24 c (1) 28 c (2) C:N Ratio 5 to 10 cm 73 e (2) 31 c (1) 9 a (1) 18 b (2) 27 c (2) †Each value is a mean of 36 observations averaged across treatment intensity and families. ‡ Values in parentheses repr esent the standard error. § Within a variable, the means followed by different letters are statistica lly different at p < 0.05 showing the effects of both depths and size fraction.

PAGE 83

83 Table 4.2. Characteristics of the density frac tions for a sandy Spodosol in north Florida Variable Fraction Depth Heavy Medium Light 0 to 5 cm 0.12† a§ (0.01) ‡ 29.7 b (0.7) 35.5 c (0.9) 2000 to 250 m 5 to 10 cm 0.10 a (0.01) 30.7 bc (0.7) 34.5 bc (0.9) 0 to 5 cm 0.08 a (0.02) 22.4 b (0.5) 20.6 (4.5) C concentration (% of fraction) 250 to 150 m 5 to 10 cm 0.07 a (0.01) 19.9 b (0.9) 17.6 NA 0 to 5 cm 0.04 a (0.001) 0.75 c (0.01) 0.66 bc (0.04) 2000 to 250 m 5 to 10 cm 0.04 a (0.001) 0.64 b (0.02) 0.61 b (0.04) 0 to 5 cm 0.041 b (0.001) 0.55 c (0.02) 1.05 (0.12) N concentration (% of fraction) 250 to 150 m 5 to 10 cm 0.036 a (0.001) 0.48 c (0.03) 0.82 NA 0 to 5 cm 3.0 a (0.3) 40.3 c (0.8) 55.8 d (2.9) 2000 to 250 m 5 to 10 cm 2.6 a (0.4) 50.8 b (1.7) 58.5 d (2.5) 0 to 5 cm 2.2 a (0.6) 42.2 b (1.6) 18.7 (2.6) C:N Ratio 250 to 150 m 5 to 10 cm 2.3 a (0.4) 45.4 b (2.0) 21.5 NA †Each value is a mean of 36 observations averaged across treatment intensity and families. ‡ Values in parentheses re present standard error. § Within a variable, the means followed by different letters are statistica lly different at p < 0.05 showing effect of both depth and density, within a size fraction. The light density in 250 to 150 m was less than <0.01% of the fract ion weight and only 5 samples had enough weight for analysis. Therefore, no statistical analysis wa s carried out for this fraction. # The weight and C content of the water-disper sible fraction was measur ed by difference and not analyzed directly. Therefore, the character istics of this fraction were not reported.

PAGE 84

84 Table 4.3. Distribution of aggregat e organic matter (AOM) and pa rticulate organic matter (POM) in the different size-density fractions for a sandy Spodosol in north Florida 2000 to 250m 250 to 150 m 150 to 53 m Variable Depth Medium Heavy Medium Heavy Whole 0 to 5 cm 65† a (2) ‡ 77 b (1) 82 a (1) 73 a (2) 79.5 a (0.5) AOM (% of total OM in size fraction) 5 to 10 cm 66 a (1) 70 a (1) 81 a (1) 63 a (3) 81.3 b (0.8) 0 to 5 cm 35 a (2) 23 a (1) 18 a (1) 27 a (2) 20.5 a (0.5) POM (% of total OM in size fraction) 5 to 10 cm 34 a (1) 30 b (1) 19 a (1) 37 a (3) 18.7 a (0.8) †Each value is a mean of 34-36 observations aver aged across treatment intensity and families. ‡ Values in parentheses re present standard error. § Within a size-density fraction, the means followed by different letters are statistically different at p < 0.05 showing the effect of depth.

PAGE 85

85 0 0.05 0.1 0.15 0.2 0.25 Best growerMedium growerPoor grower FamilyCarbon (g C/ 100 g fraction) a ab b Figure 4.1. Effects of family on the C content of the 2000 to 250 m light density fraction for a sandy Spodosol in north Florida. The m eans followed by different letters are statistically different at p < 0.05. The error bars represent the standard errors. Each value represents the mean of 24 observations averaged across management intensity and soil depth.

PAGE 86

86 70 72 74 76 78 80 82 84 86 88 90 0 to 5 cm5 to 10 cm Soil DepthAOM (% of total OM in size fraction) Best grower Medium grower Poor grower Figure 4.2. Family x Depth interaction in aggr egate organic matter (AOM) of the medium density, 250 to 150 m fraction for a sandy S podosol in north Florid a. The error bars represent the standard errors. Each valu e represents the mean of 12 observations averaged across management intensities.

PAGE 87

87 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Best growerMedium growerPoor grower FamilyNitrogen concentration (% of fraction) High Low Figure 4.3. Effects of management intensity and family on N concentrations in the 2000 to 250 m medium density fraction for a sandy Spodos ol in north Florida. The error bars represent the standard errors. Each valu e represents the mean of 12 observations averaged across management intensity and soil depth.

PAGE 88

88 0 10 20 30 40 50 60 70 80 90 HighLow Management intensityAOM (% of total OM in size fraction) Best grower Medium grower Poor grower Figure 4.4. Family x management intensity interac tion in aggregate organi c matter (AOM) of the heavy density, 2000 to 250 m fraction for a sandy Spodosol in north Florida. The error bars represent the standard errors Each value represents the mean of 12 observations averaged across soil depths.

PAGE 89

89 0% 20% 40% 60% 80% 100% 0 to 5 cm5 to 10 cm0 to 5 cm5 to 10 cm Carbon contentNitrogen content Content (% of whole soil) 2000 to 250 m 250 to 150 m 150 to 53 m <53 m Figure 4.5. Effects of soil depth on C and N contents among the va rious size fractions for a sandy Spodosol in north Florida.

PAGE 90

90 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 to 5 cm5 to 10 cm 0 to 5 cm5 to 10 cm 2000 to 150m250 to 150m OM content (% of total in size fraction) heavy light medium water-dispersible Figure 4.6. Distribution of C among the various density fractions for a sandy Spodosol in north Florida.

PAGE 91

91 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 to 5 cm5 to 10 cm0 to 5 cm5 to 10 cm 2000 to 150m250 to 150m N content (% of N in fraction) Heavy Medium Light Figure 4.7. Distribution of N among the various density fractions for a sandy Spodosol in north Florida.

PAGE 92

92 CHAPTER 5 SUMMARY AND CONCLUSIONS Soil organic carbon (SOC) and its chemical natu re in sandy Spodosols under loblolly pine plantations in the southeastern U.S. are poorly described. The objectives of this dissertation were to investigate aggregation in a repres entative Spodosol in north Florida under pine management, to characterize SOC pools in the A horizon, and to determine the impact of management activities and genetics on the SOC pr ofile. This research requ ired an adaptation of the size-density fractionation methods to suit the sandy nature of the surface soils, since the weak structure required less invasive methods for size and density frac tionation. The low clay content also ruled out using clay rele ase after sonication as an indi cator of aggregate destruction. Described below is a critical evaluation of the main findings of this dissertation with focus on how this research contributed to under standing soil C dynamics in forest soils. Methodological Contributions Dry sieving was shown to function as well as, or better than, the more widely used wet sieving technique. Dry sieving pr eserved more of the weak structure and the water-soluble components within the soil, such as esters and amid es (Chapter 2). As dry sieving is considerably faster than wet sieving, it offe rs economy of time and effort for sandy textured soils. When sonication was used to disrupt soil aggr egates, the loss of orga nic matter instead of clay was used as a measure of aggregate stability. The use of organic matter was a superior procedure because (1) there is li ttle clay to measure in these soils and (2) organic matter in aggregates and aggregate stability can be measur ed together (Chapter 3). This procedure also offers opportunities for understa nding the nature of aggregates in these soils as it allows investigation of aggregate hier archy when using incremental energy input and organic matter

PAGE 93

93 release. This procedure should also be useful in analyzing the chemical nature of organic matter in aggregates of differing strength. The density fractionation procedure used in th is study is an inexpensive and chemically mild technique of organic matter fractionation, since only water was used during the procedure (Chapter 4). The three density fr actions; heavy, medium and light, we re different from each other in their appearance as well as C and N concentra tions. They also differed in their response to changes in forest management. It is clear that this procedure separates three distinct C pools of different chemical characteristics. Use of DRIFTS spectra, especially the spectra l subtraction technique used in this study (Chapter 2), proved useful in identifying di fferences in organic matter pools and in understanding the changes in chemical com position resulting from sieving method, soil aggregate size, and forest management. This tech nique enabled identifica tion of major functional compounds that differed among C pool s. In particular, it was usef ul in separating aggregate C and particulate C, showing highe r concentration of esters, amides and polysaccharides in aggregate C. The fractionation method also proved useful in detecting SOC changes that occurred due to forest management practices in as little as 4 to 6 y. Although this study focused on C, these fractionation methods could also be used to study the distribution and dynamics of other soil nutrients and pollutants. Aggregation and Physical Protection The structure of the surface soil of Florid a’s Spodosols has been described as weak. Therefore, aggregation has receiv ed little, if any, a ttention. This was the first study to focus on the soil aggregate C in a Florida Spodosol. Result s suggested that nearly half of the C was held by the aggregates (Chapter 3). This study also show ed that, in spite of th eir mechanical stability,

PAGE 94

94 the aggregates were susceptible to manageme nt related influences. Therefore, C-wise management will require an understanding of the natu re and stability of these aggregates. Further studies on differences in aggregate stability a nd response to alternative management practices could offer further insights into the soils’ po tential to protect SOC through aggregates as a physical protection mechanism. Even though this st udy measured the stability of aggregates and the amount of C held inside the aggregates, quan tification of the minera lization potential and C dating of these pools would be necessary before the C sequestration potential of the various C pools could be established. Influence of Management Intensity and Family The distribution of C across different fractions and the responsiveness of these fractions to management intensity and families are summarized in Figs 5.1 and 5.2. Intensive management did not reduce SOC throughout all soil C pools, but only in specific size-density fractions; especially the 2000 to 250 m frac tion (Chapter 2, 3, 4). It is r easonable to assume that this change resulted from reduced C input when the understory root tu rnover was reduced by sustained weed control practices. The response of this fraction to management and family was unexpected in some ways. First, the manageme nt effect was most pronounced in the stable aggregates that had strength of 6000 J or higher (C hapter 3). This indicates that measuring the mechanical strength of aggregates would not be sufficient to determine the mechanisms through which intensive management practices impact a ggregates. An investigation of the chemical changes in these aggregates due to management practices would be necessary. For example, a comparison of the DRIFTS spectra of the aggr egate organic matter would indicate whether intensive management reduces the amount of po lysaccharides, which are reported to work as binding agents (Tisdall, 1994).

PAGE 95

95 In the 2000 to 250 m fraction, the heavy densit y, which is usually considered a passive fraction (Romkens et al., 1999), also responded to management (C hapter 4). This was probably due to predominance of quartz sand in these soils which has little or no adsorption capacity to offer chemical protection. When modeling C dynamics in sandy soils, one will need to reconsider the usual definitions of active and passive fractions. Another surprising result was the effect of ma nagement intensity on the chemical character of SOM as evidenced by the DRIFTS spectra (Chapter 2). Manageme nt intensity was not expected to affect aromatic C in such a short time frame. Change in the chemical composition of the inputs, either litter or roots, might be responsible for this effect and a spectroscopic analysis of the OM inputs under differing management intens ities should help to address this question. The best growing family added more N at the 0 to 5 cm depth and ha d greater C content in the 2000 to 250 m light density fraction than th e other two families (Chapter 4). The medium growing family, on the other hand, exhibited hi gher AOM in the 250 to 150 m medium density fraction and the 2000 to 250 m heavy density fractions. This illustrates how families may differentially influence soil properties. The best fa mily is reported to have higher foliar N (Jokela E.J., unpublished data) and above ground biomass pr oduction (Roth et al., 2006) in response to N fertilization. This opens an avenue for objectivebased deployment of fam ilies for a variety of management objectives and ecosystem services (e.g., long-term C sequestration versus improved nutrient turnover). However, family did not significantly affect the total soil C content of the size fractions. As this study assessed treatment and family differences through age 6, a follow-up investigation repeated at age 12 to 15 y may be more definitive in examining changes in soil C due to management activities.

PAGE 96

96 Active and Passive C Pools Identified In This Study Identifying the most active C pools would help in understanding the mechanisms behind the total SOC changes and aid in efficient resour ce allocation for future research efforts. The characteristics of the size-density fractions used in this study indicate their functionality, which will help in relating the operationally defi ned C fractions to meaningful C pools. The 2000 to 250m fraction was found to be the C pool most responsive to forest management activities. The C content of the whol e fraction, as well as th e aggregate C in this fraction, responded to management intensity (Chapt er 2, 3). In contrast, the light and medium density fraction in this size fraction responded to family differences (Chapter 4). The chemical nature of SOC, as shown by the peaks for este rs, amides and aliphatic C compounds, indicated the presence of recently added, undecomposed organic matter in this fraction (Chapter 2). Since this fraction accounted for nearly half of the total SOC, the changes in this fraction should be useful for documenting expected changes in total SOC. The 250 to 150 m fraction was also sensitiv e to the effects of management intensity (Chapter 2) and family (Chapter 4). However, the DRIFTS spectra of this fraction did not offer any explanation for the differences. Even though th e C and N concentrations were the lowest in this fraction, it accounted for more than 50% of the soil weight and contained 42% of total soil N (Chapter 4). This highlights the importance of this fraction for SOC and N dynamics. Since this fraction contained the highest pr oportion of aggregate C (84%; Ch apter 2), further study of the aggregates should help in understanding the natu re of C in this fract ion and the role of aggregation in soil C cycling. For example, measurement of the amount of glomalin in aggregates of different stability or under di fferent management intensities may help in identifying the exact effect of forest management activities on aggregation.

PAGE 97

97 The <53m fraction was also responsive to fo rest management activities, even though the magnitude of change was small (2% change in C c ontent, Chapter 2) and exhibited high peaks of esters and amides. These peaks, combined with the presence of a signifi cant amount of aromatics further indicated the presence of decomposed OM combined with fresh undecomposed OM (Chapter 2). However, this fracti on was not studied in detail, as it accounted for less than 10% of the total SOC. The 150 to 53m fraction on the other hand, was probably the most stable C pool, as shown by its lack of response to management th rough age 6 and its high aggregate stability at age 4 (Chapter 3). This fraction also exhibite d a unique behavior in terms of response to sonication energy output because it did not exhibi t aggregate hierarchy. However, this fraction had a small (5%) decrease in SOC content under the intensive management regime at age 4 (Chapter 2), indicating the pres ence of more decomposable C forms. Similar effects were not detected at age 6. A future comparison of the ch emical composition of th is fraction at the two sampling times should help explain these differences. Of the three density fractions studied, the me dium density fractions were most responsive to both forest management intensity and family in both the 2000 to 250m and 250 to 150 m fractions (Chapter 4). The medium density fraction contained the most C. The light density in the 2000 to 250 m fraction accounted for 7 to 16% of the fraction C and also responded to family differences. The high C: N ratios of these two fr actions suggested the pr esence of relatively undecomposed organic matter and the importance of these two density fractions in C cycling. In contrast, the heavy density fraction was found to be important for N dynamics, as it contained the highest amount of N. The extremel y low C: N ratios (2 to 3) implied that the N in this fraction

PAGE 98

98 was probably microbial N. A study of the funga l and bacterial populations in these density fractions could provide further insights into the C and N dynamics of these soils. Additional Research Needs This study could have been improved. The sa mples were prepared by air drying for two weeks and sieving, which is a standard method of sample preparation. However, the soils might have been passed through the 2 mm sieve when stil l field moist (with light pressure). While these soils can be near air-dry during dry down periods, they were moist when sampled. It is not know if sieving in a field-moist c ondition would have given different results. I susp ect the results would have been similar; yet it would be wo rth running a comparative evaluation. Use of the bulk density cores for the measurement of the >2 mm fraction would also have been preferable, especially for the volumetric C measurements. Another possible improvement would have b een to collect the water-dispersible C by sedimentation during the density fractionation procedure (Chapter 4) instead of measuring it by difference. This fraction is likely to be important for the short-te rm recycling of C and nutrients and it may be a worthwhile component in future studies. Modeling C Dynamics Even though modeling carbon dynamics was not a part of this study, th e results suggested that use of the carbon pools id entified in this study could aid carbon-modeling efforts. The general approach is to use c onceptual carbon pools in modeling (e.g., Century model, Parton et al., 1987; see Smith et al., 1997 for a comparison of various carbon models). However, in most cases the actual measurement of soil organic carbon is limited to total soil carbon. The physical fractionation techniques, such as size and density fractionation, can improve the accuracy of these models by using operationally defined and hence directly measurable carbon fractions to represent the conceptual carbon pools. For example, the Century model (Parton et al., 1987) used

PAGE 99

99 three soil carbon pools, ac tive, slow and passive, based on tur nover rates. The ch aracteristics of different size-density fractions in this study suggested that th e carbon content in the light and medium density of the 2000 to 250 and 250 to 150 m fractions could be used to represent the active pool, whereas the car bon content of the 150 to 53 and <53 m fractions could be used to represent the slow or passive pools. Similarly, Hassink and Whitmore (1997) used two pools of carbon, protected and non-protecte d, for modeling the buildup a nd decline of organic carbon with and without addition of or ganic matter. Their model was ba sed on the concept that clay plays an important role in physical protection of organic matter, but the pr otection capacity of the soil is limited. The carbon content of the <53m fraction in this study c ould represent the clayprotected carbon pool, while the >53m fractions can represent the unprotected carbon pool. However, more study of the tur nover rates and chemical characte ristics of these size-density fractions would be necess ary for this purpose. The following model, though far from adequate, re presents an initial ou tline. It uses the continuity equation as the basis and instead of focusing on the ri ght side of the equation, which incorporates various processes and flows, it fo cuses on the left side of the equation, which represents the effect of these processes on changes in carbon pool s. Thus, the changes in total SOC can be separated into changes in labile C (C associated with the active fractions) and protected C (C associated with passive fractio ns). The equation can be rewritten as follows: (dC/dt)x = dClabile + dCprotected (dC/dt)x = dClabile + dCprotected (physic al) + dCprotected (che mical) + dCprotected (biochemical) where,

PAGE 100

100 Clabile = light fraction + partic ulate organic matter in medium and heavy density in both 2000 to 250 m and 250 to 150 m fractions. Cprotected (physical) = aggregate organic matter as measured by loss on sonication in medium and heavy density of the 2000 to 250 m and 250 to 150 m fractions + C in 150 to 53 m fraction Cprotected (chemical) = negligible Cprotected (biochemical) = f(polyphenol content, lignin/N ratio) However, it must be reiterated that even though AOM and the heavy density organic matter are considered passive C pools, the AOM in the medium density of 250 to 150 m fraction and the AOM in the heavy density of 2000 to 250 m fraction were influenced by management intensity and family in as few as 6 years. Also, the particul ate organic matter, which is considered to be a labile poo l, did not respond to any of these treatment effects. Therefore, a study of the mineralization potentials and C dati ng is required to confirm the C protection potential of these size-density fractions. In conclusion, the major contributions of this study were 1) devel opment of C fractionation techniques suitable for the sandy soils, 2) esta blishing the importance of aggregates in these soils, which was a neglected subject until now, 3) establishing the profile of C in terms of content distribution and, 4) iden tification of the most responsive C pools to the managementrelated changes as well as the pools that ha ve potential for long-term C sequestration.

PAGE 101

101 Figure 5.1. Carbon profile at age 4 for a sandy Spodosol in north Florida. >8000 m (11.2%) 8000 to 2000 m (16.9%) 2000 to 250 m (39.6%) 250 to 150 m 12.9% 150 to 53 m 14% <53 m 5.4% POM (14.6%) AOM (24.9%) AOM (10.9%) POM (2.1%) AOM (10.1%) POM (3.9%) Responsive to management intensity

PAGE 102

102 Figure 5.2. Carbon profile at age 6 fo r a sandy Spodosol in north Florida Light 3. 4 % Medium Heavy 1.3% Waterdispersible 3.7% Light 0.1% Medium 10.7% Heavy 3.2% Waterdispersible 1.8% 16.1% <53 m Responsive to management intensity Responsive to genotype 2000 to 250 m POM POM 250 to 150 m POM POM 150 to 53 m 12.6% AOM POM AOM AOM AOM AOM

PAGE 103

103 APPENDIX A EXPERIMENTAL PROTOCOLS Wet sieving – Used in Chapter 1 1. Wet sieving was used for size fractionation in th is study. Soil samples obtained from the field were first air-dried and passed th rough a 2-mm sieve. They were then sieved to separate the aggregates into four size classes, using thr ee 8” diameter brass sieves of sizes 250, 150 and 53 m). 2. Place top sieve (250 m) in a water tray and pour water into the tray until it stands about 2 cm above the sieve mesh. Any container with su fficient height and diameter/width to fit the sieve can be used as a water tray. 3. Add 100 g of soil to the sieve and swirl the sieve to allow even distribution of the soil. Let it stand for 5 minutes, so that the soil sample is thoroughly soaked. 4. Move the sieve up and down 50 times taking car e that every time the water surface is broken. This should take about 2 minutes (Cambardella and Elliot, 1993). 5. Swirl as necessary to avoid collection of soil in the middle of the sieve. 6. Take the sieve out of the water, wash sieve’s si de with a spray bottle to ensure that all the finer soil is collected in the tray, then allow the sieve dry for 24 hours. 7. Transfer the soil water mixture in the tray to the next size (150 m) sieve and repeat the procedure. Dry sieving – Used in Chapters 1, 2, 3 1. Soil samples obtained from the field were firs t air-dried and passed th rough a 2-mm sieve. They were then sieved to separate the aggreg ates into four size classes, using three 8” diameter brass sieves of sizes 250, 150 and 53 m.). 2. For this, 100 g of air-dried soil was placed in th e large sieve of a stack of sieves with sieve sizes mentioned above. The sieves were plac ed on a horizontal mechanical shaker for 5 minutes at 75 rpm. This duration was used to achieve size fractionation with minimum disruption of aggregates, and also because, for th e soil used in this st udy, there is very little difference in the size distribution af ter 5 min and 20 min of shaking. Sonication: Used in Chapters 2 and 3 Sonication was used to breakdown soil aggregates, to measure the strength of aggregates, and to determine the amount of C held inside the aggregates. Three size fractions, 2000 to 250, 250 to150 and 150 to 53 m, were used in this study. The same technique can be used for whole soil if desired. 1. Setting the sonicator: Sonic Dismembrat or (Fisher Scientific, model 500) a. Adjust the amplitude using the circul ar knob on left of the control box. The amplitude ranges from 0 to 100%, but for the model mentioned above, a different probe is necessary for amplitude greater than 60%. Choice of amplitude depends on the energy output required. See Table 1 for the values used in this study.

PAGE 104

104 b. Adjust the ON and OFF timing (59.9 sec maximum) and total duration. Save the settings for future use. In this st udy, the pulse method (59.9 sec ON and 30 sec OFF) was used to avoid an excessive rise in temperature. However, if temperature is not a concern, using the con tinuous method will save time. c. Make sure the printer and out put settings are enabled. For this, use the arrow keys on the control box to go to the printer/out put options and click enter on desired option. 2. Sonication: a. Weigh 10 g soil in a 250 mL beaker. Make sure that you use the same shape and size of beaker every time because the he ight of immersion of probe affects the extent of sonication (North, 1976). b. Add 100 mL distilled water to the beaker slowly to avoid formation of bubbles. c. Put the beaker in the sound box. Immerse the probe in the soil-water suspension taking care that it is in the center and not touching the beaker walls. Use a clamp or another inverted beaker to maintain a constant depth of immersion. Close the sound box securely and push START. d. After the sonicator shows ‘experiment co mpleted’ message, take out the beaker. Wash the particles on the probe into the beaker using a spray bottle. Dry the probe gently. 3. Post-sonication sieving: a. This is done to separate the material re leased by the aggregates after sonication from the POM and OM still in aggregates (if not all aggregates have been destroyed) b. Pour the contents of the beaker onto the sieve taking care that no particles are sticking to the beaker walls. Use the same size sieve used for sieving before sonication (e.g. 250m sieve for the 2000 to 250m fraction). c. Spray some water lightly onto the sieve to spread the soil evenly on it and let it dry for 24 hours. d. Measure C remaining after sonication us ing loss on ignition or any other chosen method of C measurement. Density separation: Used in Chapter 3 1. In this study, two size fractions, 2000-250 m and 250-150 m were density fractionated, but the same procedure can be used for whole soil. 2. Weigh 10 g of sample in a small (50 mL) beaker. 3. Add approximately 10 mL of water and swirl the beaker to mix it thoroughly. 4. Add approximately an additional 15 mL of water in order to get a clear delineation between heavier mineral soil and th e floating soil organic matter. 5. Decant the supernatant into a 600 mL beaker. Repeat the process until no more organic matter can be visually separated. (Twenty-five milliliter water was added to the beaker and the organic matter was separa ted by swirling and decantation.) 6. The organic matter that cannot be separated from the mineral matter is defined as the heavy fraction (HF).

PAGE 105

105 7. Transfer the contents of the 600mL beaker to the funnel assembly (See the Photograph) and let it settle till the water is clear. It ta kes 12 to 24 hrs, depending on the fraction size. Finer fractions need more time. 8. After 24 hours, the fraction floating on water is called the light dens ity fraction (LF) and the fraction settled at the bottom of the funne l is called the medium density fraction (MF). 9. Collect the medium-density fraction in a beaker by opening the st opcock and allowing the suspension to pass through. 10. Collect the light fraction in another beaker. 11. The three fractions thus separated can be drie d with or without sieving. Sieving is done for measurement of water-dispersible aggreg ates if desired. Simply pass the fraction through the same size sieve used for sieving. Soil passing through the sieve is called the water-dispersible fraction. Alternatively, the water-dispersible fraction can also be measured by difference (Total soil used for density fraction – [HF + MF + LF])

PAGE 106

106 Table A.1. Amplitude and time combinations and energy outputs used for this study Amplitude Time Energy (J) Peak Power (W) 20 1 954 16 20 2 1911 16 20 3 2871 16 20 4 3830 16 20 5 4785 16 30 1 1523 26 30 2 3004 26 30 3 4588 26 30 4 5981 26 30 5 7602 26 40 1 2093 36 40 2 4129 36 40 3 6183 36 40 4 8863 36 40 5 10432 36 50 5 13828 48 Note: Energy output can vary depending upon the peak power. The values given here are those observed in this study. 0 10 20 30 40 50 60 2000-250250-150150-53<53 5 min 10 min 15 min 20 min Figure A.1. Effect of sieving time for dry sieving on the weight distribution across size fractions

PAGE 107

107 Figure A.2. Funnel assembly used for density fractionation.

PAGE 108

108 APPENDIX B AGGREGATES OBSERVED IN THE SANDY SPODOSOLS This appendix shows the morphology and chemi cal nature of the aggregates in a sandy Spodosol of north Florida under forest management Aggregates were observed using a Scanning electron microscope (SEM) (JEOL JSM 6400) e quipped with an energy-dispersive x-ray fluorescence elemental microanalysis (EDX) system. The elemental dot maps were used to determine the internal structure, that is, the spa tial distribution of the different elements in the aggregates. The edx graphs were used to determ ine the relative concentration of the different elements in the aggregates.

PAGE 109

109 Figure B.1. Image and elemental dot map of aggregate (2000 to 250 m). The upper left panel is the SEM photo, the upper right panel is th e dot map for aluminum, the lower right panel is the dot map for silica and the lowe r right panel is the dot map for calcium. Figure B.2. EDX spectra for the aggregate shown in B.1.

PAGE 110

110 Figure B.3. Image and elemental dot map of aggregate (2000 to 250 m). The upper left panel is the SEM photo, the upper right panel is th e dot map for aluminum, the lower right panel is the dot map for silica and the lowe r right panel is the dot map for calcium. Figure B.4. EDX spectra for the aggregate shown in B.3.

PAGE 111

111 Figure B.5. Image and elemental dot map of aggregate (2000 to 250 m). The upper left panel is the SEM photo, the upper right panel is th e dot map for aluminum, the lower right panel is the dot map for silica and the lowe r right panel is the dot map for calcium. Figure B.6. EDX spectra for the aggregate shown in B.5.

PAGE 112

112 Figure B.7. Image and elemental dot map of aggregate (250 to 150 m). The upper left panel is the SEM photo, the upper right panel is th e dot map for aluminum, the lower right panel is the dot map for silica and the lowe r right panel is the dot map for calcium. Figure B.8. EDX spectra for the aggregate shown in B.7.

PAGE 113

113 Figure B.9. Image and elemental dot map of aggregate (250 to 150 m). The upper left panel is the SEM photo, the upper right panel is th e dot map for aluminum, the lower right panel is the dot map for silica and the lowe r right panel is the dot map for calcium. Figure B.10. EDX spectra for the aggregate shown in B.9.

PAGE 114

114 LIST OF REFERENCES Adegbidi, H.G., E.J. Jokela, N.B. Comerford, and N.F. Barros. 2002. Biomass development for intensively managed loblolly pine plantati ons growing on Spodosols in the southeastern USA. For. Ecol. Manage. 167:91-102. Anderson, T. H., and K.H. Domsch. 1993. The meta bolic quotient for CO2 (qCO2) as a specific activity parameter to assess the effects of environmental conditions, such as pH, on the microbial biomass of forest soil s. Soil Biol. Biochem. 25:393-395. Angers, D.A., and G.R. Mehuys. 1990. Barley and alfalfa cropping effects on carbohydrate content of a clay soil and its size fr actions. Soil Biol. Biochem. 22: 282-288. Angers, D.M., and M. Giroux. 1996. Recently depos ited organic matter in soil water-stable aggregates. Soil Sci. Soc. Am. J. 60:1547-1551. Blanco-Canqui, H., and R. Lal. 2004. Mechanisms of Carbon Sequestration in Soil Aggregates. Crit. Rev. Plant Sci. 23:481–504. Brady, N.C. 2000. The Nature and Properties of Soils. 10th edition. Macmillan Publishing Company, New York, NY. Bronick, C.J., and R. Lal. 2005. Manuring and ro tation effects on soil organic C concentration for different aggregate size fractions on two so ils in northeastern Ohio, USA. Soil Tillage Res. 81:239–252. Buchkina, N.P., and E.V. Balashov. 2001. The Influence of a grass–clover mixture on soil organic matter and aggregation of a podzolic loamy sand soil. p 214-219. In R.M. Rees et al. (ed.), Sustainable Management of Soil Organic Matter, CABI Publishing, Wallingford, UK. Caesar-Tonthat, T.C. 2002. Soil binding propert ies of mucilage produced by a basidiomycete fungus in a model system. Mycol. Res. 106: 930-937. Cambardella, C.A., and E.T. Elliott. 1993. Methods of physical separation and characterization of soil organic matter fractions. Geoderma 56:449-457. Carlisle, V.W., C.T. Hallmark, F. Sodek III, R.E. Caldwell, L.C. Hammond, and V.E. Berkheiser. 1981. Characterization data for sel ected Florida soils. Soil Science Research Report Number 81-1. Compiled by Soil Charac terization Laboratory, Soil Science Department, University of Florida in coopera tion with U.S. Department of Agriculture – Soil Conservation Service. Carlisle, V.W., F. Sodek III, M.E. Collins, L.C. Hammond, and W.G. Harris. 1988. Characterization data for selected Florida soils. Soil Science Research Report Number 881. Compiled by Soil Characterization Laboratory, Soil Science Department, University of Florida in cooperation with U.S. Department of Agriculture – Soil Conservation Service.

PAGE 115

115 Carlisle, V.W., F. Sodek III, M.E. Collins, L.C. Hammond, and W.G. Harris. 1989. Characterization data for selected Florida soils. Soil Science Research Report Number 891. Compiled by Soil Characterization Laboratory, Soil Science Department, University of Florida in cooperation with U.S. Department of Agriculture – Soil Conservation Service. Carter, M.R. 1992. Influence of reduced tillage systems on organic matter, microbial biomass, macro-aggregate distribution and structural stab ility of the surface soil in a humid climate. Soil Tillage Res. 23:361-372. Christensen, B.T. 1992. Physical fr actionation of soil and organic ma tter in primary particle size and density separates. Adv. Soil Sci. 20: 1-89. Crawford, D.T., B.G. Lockaby, and G.L. Somers 1991. Family nutrition interactions in fieldplanted loblolly-pine. Can. J. For. Res. 21: 1523-1532. Dalla-Tea, F., and E.J. Jokela. 1991. Needlefall, canopy light interception, and productivity of young intensively managed slash and loblolly pine stands. For. Sci. 37:1298–1313. Echeverria, M.E., D. Markewitz, L.A. Morris, and R.L. Hendrick. 2004. Soil organic matter fractions under managed pine plantations of th e southeastern USA. Soil Sci. Soc. Am. J. 68:950-958. Edwards, A.P., and J.M. Bremner. 1967. Microa ggregates in soils. J. Soil Sci. 18: 64-73. Ellert, B.H., and E.G. Gregorich. 1995. Manageme nt-induced changes in the actively cycling fractions of soil organic matter. p. 119–138. In W.W. McFee and J.M. Kelly (ed.) Carbon Forms and Functions in Forest Soils. Soil Science Society of America, Madison, WI. Escamilla, J.A., N.B. Comerford, and D.G. Neary. 1991. Soil core break method to estimate pine root distribution. Soil Sc i. Soc. Am. J. 55: 1722-1726. Feller, C. 1993. Organic inputs, so il organic matter and functional soil organic compartments in low capacity clay soils in tropical zones. p. 77-88. In K. Mulongoy and R. Merckx (ed.) Soil Organic Matter Dynamics and Sustainabi lity of Tropical Agri culture. IITA/K.A. Leuven. John Wiley & Sons, Chichester, NY. Foster, R.C. 1988. Microenvironments of soil-mi croorganisms. Biol. Fertil. Soils. 6: 189-203. FRA, Global Forest Resources Assessment. 2005. www.fao.org/forestry/site/fra2005/en Last assessed in November 2006. Gregorich, E.G., M.H. Beare, U.F. Mckim, a nd J.O. Skjemstad. 2006. Chemical and biological characteristics of physically uncomplexed organic matter. Soil Sci. Soc. Am. J. 70: 975985. Harding, R.B., and E.J. Jokela. 1994. Long-term e ffects of forest ferti lization on site organic matter. Soil Sci. Soc. Am. J. 58:216-221.

PAGE 116

116 Harris, W.G., and V.W. Carlisle. 1987. Clay mine ralogical relationships in Florida Haplaquods. Soil Sci. Soc. Am. J. 51: 481-484. Hassink, J., A.P. Whitmore, and J. Kubat. 1997. Si ze and density fractionation of soil organic matter and the physical capacity of soils to protect organic matter. Eur. J. Agron. 7:189199. Haynes, R.J. 2005. Labile organic matter fractions as central components of the quality of agricultural soils: An overview. Adv. Agron. 85: 221-268. Horn, R. 1990. Aggregate characterization as compar ed to soil bulk properties. Soil Tillage Res. 17: 265-289. Jimenez, J.J., and R. Lal. 2006. Mechanisms of C sequestration in soils of Latin America. Crit. Rev. Plant Sci. 25: 337-365. Jokela, E.J., and T.A. Martin. 2000. Effects of ontogeny and soil nutrient supply on production, allocation, and leaf area efficiency in loblolly and slash pine st ands. Can. J. For. Res. 30: 1511-1524. Jokela, E.J., P.M. Dougherty, and T.A. Mar tin. 2004. Production dynamics of intensively managed loblolly pine stands in the southern United States: a synthesis of seven long-term experiments. For. Ecol. Manage. 192: 117-130. Jolivet, C., D. Arrouays, J. Leveque, F. A ndreux, and C. Chenu. 2003. Organic C dynamics in soil particle-size separates of sandy Spodosols when forest is cleared for maize cropping. Eur. J. Soil Sci. 54: 257–268. Liao, J.D., T.W. Boutton, and J.D. Jastrow. 2006. Storage and dynamics of carbon and nitrogen in soil physical fractions following woody plant invasion of grassland. Soil Biol. Biochem. 38: 3184-3196. Liu, Q.M., S.J. Wang, H.C. Piao, and Z.Y. Ouya ng. 2003. The changes in soil organic matter in a forest-cultivation sequence traced by stable C isotopes. Aust. J. Soil Res. 41: 1317-1327. Lopez-Upton, J., T.L. White, and D.A. Huber. 1999. Effects of site and intensive culture on family differences in early growth and rust incidence of loblolly and slash pine. Silvae Genetica 48: 284-293. Marx, M.C., E. Kandeler, M. Wood, N. Werm bter, and S.C. Jarvis. 2005. Exploring the enzymatic landscape: distribution and kinetics of hydrolytic enzymes in soil particle-size fractions. Soil Biol. Biochem. 37:35-48. McCrady, R.L., and E.J. Jokela. 1998. Canopy dyna mics, light interception, and radiation use efficiency of selected loblolly pine families. For. Sci. 44: 64-72. McKeand, S., T. Mullin, T. Byram, and T. Wh ite. 2003. Deployment of genetically improved loblolly and slash pines in the south. J. for. 101: 32-37.

PAGE 117

117 Meijboom, F.W., J. Hassink, and M.V. Noor dwijk. 1995. Density fractionation of soil macroorganic matter using silica suspen sions. Soil Biol. Biochem. 27:1109-1111. Nelson, D.W., and L.E. Sommers. 1982. Total C, organic C, and organic matter. p. 539-579. In A.L. Page et al. (ed.) Methods of Soil An alysis, Part 2, Chemical and Microbiological Properties. American Society of Agronomy, Madison, WI. NOAA. 2002. State, regional, a nd national monthly precipita tion: area-weighted monthly normals, 1971-2000 (and previous periods), 4-2. Historical Climatography Series. North, P.F. 1976. Towards an absolute measuremen t of soil structural st ability using ultrasound. J. Soil Sci. 27:451-459. Oades, J.M. 1993. The role of biology in form ation, stabilization and degradation of soil structure. Geoderma 56: 377–400. Oades, J.M., and A.G. Waters. 1991. Aggregate hi erarchy in soils. Aust. J. Soil Res. 29: 815– 828. Parton, W.J., D.S. Schimel, C.V. Cole, and D.S. O jima. 1987. Analysis of factors controlling soil organic matter levels in Gr eat Plains grasslands. Soil Sci. Soc. Am. J. 51: 1173-1179. Polglase, P.J., E.J. Jokela, and N.B. Comerf ord. 1992a. Nitrogen and phosphorus release from decomposing needles of southern pine plantations. Soil Sci. Soc. Am. J. 56:914–920. Polglase, P.J., N.B. Comerford, and E.J. Jokela, 1992b. Mineralization of nitrogen and phosphorus from soil organic matter in southern pine plantations. Soil Sci. Soc. Am. J. 56: 921–927. Pope, P.E. 1979. Effect of family on biomass a nd nutrient content in 11year-old loblolly-pine plantations. Can. J. For. Res. 9: 224-230. Post, W.M., W.R. Emanuel, P.J. Zinke, a nd A.G. Stangenberger. 1982. Soil carbon pools and world life zones. Nature (London) 298 (5870): 156-159. Quenea, K., C. Largeau, S. Derenne, R. Spacci ni, G. Bardoux, and A. Mariotti. 2006. Molecular and isotopic study of lipids in particle size fractions of a sandy cultivated soil (Cestas cultivation sequence, southwest France): S ources, degradation, and comparison with Cestas forest soil. Org. Geochem. 37: 20–44. Quideau, S.A., M.A. Anderson, R.C. Graham, O. A. Chadwick, and S.E. Trumbore. 2000. Soil organic matter processes: characterization by 13C NMR and 14C measurements. For. Ecol. Manage. 138:19-27. Quideau, S.A., R.C. Graham, O.A. Chadwic k, H.B. Wood. 1998. Organi c C sequestration under chaparral and pine after four decades of soil development. Geoderma 83: 227-242.

PAGE 118

118 Reeves, III, J. B., and S. R. Delwiche. 2003. SAS partial least squares regr ession for analysis of spectroscopic data. J. Near Infrared Spectrosc. 11:415-431. Richter, D.D., D. Markewitz, J.K. Dunsomb, P.R. Heine, C.G. Wells, A. Stuanes, H.L. Allen, B. Urrego, K. Harrison, G. Bonani. 1995. Carbon cycling in a loblolly pine forest: implications for the mission carbon sink and for the concept of soil. p. 223-251. In : W.W. McFee and J.M. Kelly (ed.), Carbon Forms and Functions in Forest Soils. Soil Science Society of America, Madison, WI. Romkens, P.F.A.M., J. van der Plicht, and J. Hassink. 1999. Soil organic matter dynamics after the conversion of arable land to pa sture. Biol. Fertil. Soils. 28:277-284 Roscoe, R., P. Buurman, and E. J. Velthorst 2000. Disruption of soil aggregates by varied amounts of ultrasonic energy in fractionation of organic matter of a clay Latosol: carbon, nitrogen and 13C distribution in particle-size fr actions. Eur. J. Soil Sci. 51: 445-454. Roth, B. E., E. J. Jokela, T. A. Martin, D. A. Huber, and T. L. White. 2006. Family x environment interactions for sele cted loblolly and slash pine pl antations in the southeastern United-States. For. Ecol. Manage. In press. SAS, 1996. SAS/STAT Software, Changes a nd Enhancement through release 6.11. p 1104. SAS Institute Inc., Cary, NC. Schmidt, M.W.I., C. Rumpel, and I. Kgel-Kna bner. 1999. Evaluation of an ultrasonic dispersion procedure to isolate primary organomineral comp lexes from soils. Eur. J. Soil Sci. 50: 8794. Schmidt, R.A. 2003. Fusiform rust of southern pines: A major success for forest disease management. Phytopathology 93: 1048-1051. Schultz, R.P. 1999. Loblolly the pine for th e twenty-first century. New For. 17: 71-88. Shan, J.P., L.A. Morris, and R.L. Hendrick. 2001 The effects of management on soil and plant carbon sequestration in slash pine pl antations. J. Appl. Ecol. 38:932-941. Six, J., G. Guggenberger, K. Paustian, L. Haum aier, E.T. Elliott, and W. Zech. 2001. Sources and composition of soil organic matter fractions between and within soil aggregates. Eur. J. Soil Sci. 52:607-618. Six, J., R.T. Conant, E.A. Paul, and K. Paustia n. 2002. Stabilization mechanisms of soil organic matter: Implications for C-satura tion of soils. Plant Soil 241:155-176. Smith, P., J.U. Smith, D.S. Powlson, W.B. McGill, J.R.M. Arah, O.G. Chertov, K. Coleman, U. Franko, S. Frolking, D.S. Jenkinson, L.S. Jens en, R.H. Kelly, H. Klein-Gunnewiek, A.S. Komarov, C. Li, J.A.E. Molina, T. Mueller, W.J. Parton, J.H.M. Thornley, and A.P. Whitmore. 1997. A comparison of the performa nce of nine soil organic matter models using datasets from seven long-te rm experiments. Geoderma 81:153-225.

PAGE 119

119 Sodek, III, F., V.W. Carlisle, M.E. Co llins, L.C. Hammond, and W.G. Harris. 1990. Characterization data for selected Florida soils. Soil Science Research Report Number 901. Compiled by Soil Characterization Laboratory, Soil Science Department, University of Florida in cooperation with U.S. Department of Agriculture – Soil C onservation Service. Sollins, P., C. Swanston, M. Kleber, T. Filley, M. Kramer, S. Crow, B.A. Caldwell, K. Lajtha, and R. Bowden. 2006. Organic C and N stabiliz ation in a forest soil: Evidence from sequential density fractionation. Soil Biol. Biochem. 38: 3313–3324. Sollins, P., G. Spycher, and C.A. Glassman. 198 4. Net nitrogen mineralisation from light and heavy fraction forest soil organic matter. Soil Biol. Biochem. 16: 31-37. Stone, E.L., W.G. Harris, R.B. Brown, and R.J. Kuehl. 1993. Carbon storage in Florida Spodosols. Soil Sci. Soc. Am. J. 57: 179-182. Swanston, C.W., M.S. Torn, P.J. Hanson, J.R. S outhon, C.T. Garten, E.M. Hanlon, and L. Ganio. 2005. Initial characterization of processes of soil carbon stabi lization using forest standlevel radiocarbon enrichment. Geoderma 128: 52–62. Swanston, C., P.S. Homann, B.A. Caldwell, D.D. Myrold, L. Ganio, and P. Sollins. 2004. Longterm effects of elevated nitrogen on forest soil organic matter stability. Biogeochemistry 70: 227-250. Taylor, G.E. 1994. Role of family in the response of loblolly-pine to tropo spheric ozone effects at the whole-tree, stand, and regional level. J. Environ. Qual. 23: 63-82. Tisdall, J.M. 1994. Possible role of soil microorga nisms in aggregation in soils. Plant Soil 159: 115–121. Tisdall, J.M., and J.M. Oades. 1982. Organic matter and water-stable aggregates in soils. J. Soil Sci. 33:141-163. Tscherko, D., J. Rustemeier, A. Richter, W. Wa nek, and E. Kandeler. 2003. Functional diversity of the soil microflora in primary succession across two glacier forelands in the Central Alps. Eur. J. Soil Sci. 54: 685-696. Vanrees, K.C.J., and N.B. Comerford. 1986. Vertical root distribution and strontium uptake of a slash pine stand on a Florida Spodosol Soil Sci. Soc. Am. J. 50: 1042-1046. Wear, D. N., and J.G. Gries. (ed.) 2002. Southe rn forest resource assessment. Gen. Tech. Rep. SRS–53. 635 p. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. Will R.E., D. Markewitz, R.L. Hendrick, D.F. Meason, T.R. Crocker, and B.E. Borders. 2006. Nitrogen and phosphorus dynamics for 13-year-old loblolly pine stands receiving complete competition control and annual N fertil izer. For. Ecol. Manage. 227: 155-168.

PAGE 120

120 Whitney, W.D. 1905. Atharva Veda Samhita, Harv ard Oriental Series. Authorised Indian Reprint, Motilal Banarasidas. 1984. Delhi, India. Yoder, R.E. 1936. A direct method of aggregate an alysis of soils and a study of the physical nature of erosion losses. J. Am. Soc. Agron. 28:337-351. Zhang, M.K., Z.L. He, D.V. Calvert, P.J. St offella, X.E. Yang, and Y.C. Li. 2003. Phosphorus and heavy metal attachment and release in sandy soil aggregate fractions. Soil Sci. Soc. Am. J. 67: 1158-1167. Zhong, Z.K., and F. Makeschin. 2006. Comparison of soil nitrogen availab ility indices under two temperate forest types. Pedosphere 16: 273-283.

PAGE 121

121 BIOGRAPHICAL SKETCH Deoyani Vinayak Sarkhot has an undergraduate degree in agriculture and a master's degree in soil science, both earned in India. A love for nature and for the plants led her to agriculture, and her interest in soil and water conservation, which d eepened with every passing year, naturally led to soil science. She has worked as a teacher, teaching ba sic soil science in a farmers’ school, as an extension worker in a salin e land reclamation project and as a journalist in an agri-business publication. Her main ambition is doing some work that will make a difference for the farmers and for her beloved nature. She is immensely grateful for all the opportunities she has had so far and hopes that this good fortune will continue to be with her as she begins the new phase of her life.


Permanent Link: http://ufdc.ufl.edu/UFE0017363/00001

Material Information

Title: Genotypic and Forest Management Effects on Size-Density Fractionation of Soil Carbon in a Forested Spodosol
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0017363:00001

Permanent Link: http://ufdc.ufl.edu/UFE0017363/00001

Material Information

Title: Genotypic and Forest Management Effects on Size-Density Fractionation of Soil Carbon in a Forested Spodosol
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0017363:00001


This item has the following downloads:


Table of Contents
    Title Page
        Page 1
        Page 2
    Dedication
        Page 3
    Acknowledgement
        Page 4
        Page 5
    Table of Contents
        Page 6
        Page 7
        Page 8
    List of Tables
        Page 9
    List of Figures
        Page 10
        Page 11
    List of abbreviations
        Page 12
    Abstract
        Page 13
        Page 14
    Introduction
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
    Effects of forest management on soil carbon and nitrogen in a north Florida sandy spodosol
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
    Soil aggregation and aggregate carbon in a forested southeastern coastal plain spodosol
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
    Genotypic and forest management effects on size-density fractionation of soil carbon in a forested spodosol
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
    Summary and conclusions
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
    Appendix A: Experimental protocols
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
    Appendix B: Aggregates observed in the sandy spodosols
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
    References
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
    Biographical sketch
        Page 121
Full Text





GENOTYPIC AND FOREST MANAGEMENT EFFECTS ON SIZE-DENSITY
FRACTIONATION OF SOIL CARBON IN A FORESTED SPODOSOL

















By

DEOYANI VINAYAK SARKHOT


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

2006

































Copyright 2006

by

Deoyani Vinayak Sarkhot
































This work is dedicated to my grandfather and to Dr. George Washington Carver, who taught me
the worth and fun of working with plants and soil.









ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Nick Comerford, for all his help during the last four

years. He taught me the meaning and the scientific rigor of the terms "objectives" and

"conclusions." He was always there to talk when I was confused, to encourage when I was

nervous (before the seminars and the qualifying oral exam, for example) and to coax when I was

lazy. I wish to thank Adriana and him for their home-like care when I was hit by a car. It really

made a difference. My co-advisor, Dr. Eric Jokela, was the first person I met, when I came in

December 2002 to find the campus nearly deserted because of Christmas. He not only welcomed

me warmly, but also took me to the Social Security office and helped me through the paperwork.

It was perhaps a little thing for him, but because I was so far away from home (and it was my

birthday), it meant a lot. I wish to express my gratitude to him for all his encouragement and

support continued throughout this time. I would like to thank Dr. Willie Harris for his help with

the scanning electron microscope and also for the fun during the pedology course, when I was

his teaching assistant. I also want to thank Dr. Wendell Cropper and Dr. Yuncong Li, my

committee members, for their time and contributions in this work. I would like to thank Dr. Jim

Reeves from Maryland, for the work on DRIFTS. It added a lot of scientific value to my thesis

and whetted my appetite for more spectroscopic work.

The Forest Biology Research Cooperative deserves a special "thank you," as they provided

me the funding, the study site to work on managed by International Paper Co. and also four great

annual meetings, one of them in the midst of a beautiful forest in Texas, a place I will always

remember. I also wish to acknowledge the Major Analytical Instrumentation Center, Department

of Materials Science and Engineering, University of Florida, for the use of the SEM. I had

always wanted to work on the scanning electron microscope and it was great to be able to see it

and work on it.









I would like to thank Dave Nolletti from school of forest resources and conservation for his

help with the C and N analysis. I also wish to thank Mary McCloude, Sally Wu and Aja Stoppe,

our lab managers during the last four years, and all my lab mates for their support. Last but not

least, I would like to thank my parents and my younger brother Yogesh for their love and support

even though from far away.









TABLE OF CONTENTS



A C K N O W L E D G M E N T S ..............................................................................................................4

L IS T O F T A B L E S ................................................................................................. ..................... 9

LIST OF FIGURES ............................................. .. ......... ............ ............... 10

L IST O F A B B R E V IA T IO N S ......................................................................... ...... ............... 12

A B S T R A C T ............... ................................................................ .......................................... 13

CHAPTER

1 INTRODUCTION ............................................... ............................. 15

Pine Plantations in the Southeastern United States ........................................... ................ 16
N eed for C arbon-w ise M anagem ent........................................ ....................... ............... 16
Challenges and O opportunities in Sandy Soils.................................................... ................ 17

2 EFFECTS OF FOREST MANAGEMENT ON SOIL CARBON AND NITROGEN IN
A NORTH FLORIDA SANDY SPODOSOL....................................................21

Introduction ........................................................ .................. 21
M materials and M methods .............. .............................................................................. 23
Experimental Site ............... ................. .. .......... ............................. 23
L ab oratory M methods .................................................. .............................................. 2 5
S tatistic al A n aly sis ..........................................................................................................2 6
R results ......................... ........................ ................. .................... 27
D ry v s. W et S iev in g ........................................................................................................2 7
Characterization of Size Fractions ...................................................................................28
Effect of M management Intensity and Soil Depth..........................................................29
Discussion ....................................................... .................. 29
D ry v s. W et S iev in g ................... ... ..............................................................................3 0
Distribution of C and N among Size Fractions ............................................................30
Im pact of M anagem ent Intensity ..................................................................................32
Conclusions ...................................................... .................. 33

3 SOIL AGGREGATION AND AGGREGATE CARBON IN A FORESTED
SOUTHEASTERN COASTAL PLAIN SPODOSOL......................................... ...............44

Introduction ........................................................ .................. 44
M materials and M methods .................................................................................................... 46
E x p erim en tal S ite ............................................................................................................4 6
L ab o rato ry M eth o d s ........................................................................................................4 8
Statistical A n aly sis ................................................................................................ 50


6









R e su lts.. . ............. ........................................................... .........................................5 1
A ggregate M orphology .......................................................................... ................ 51
Quantifying Organic C in Aggregates....................................................................51
E effect of M anagem ent Intensity ....................................... ....................... ................ 53
D iscu ssio n .................... ........................... .. ...... ...................................................... ......... 5 3
Aggregate Morphology, Stability and OM Content..................................................... 53
E effect of M anagem ent Intensity ....................................... ....................... ................ 55
M ethodological C onsiderations.................................................................. ................ 55
Aggregate Structure in Coastal Plain Spodosols Additional Considerations ............56
C o n clu sio n s............................................................................................................ ........ .. 5 8

4 GENOTYPIC AND FOREST MANAGEMENT EFFECTS ON SIZE-DENSITY
FRACTIONATION OF SOIL CARBON IN A FORESTED SPODOSOL ..........................69

In tro du ctio n ............................................................................................................. ........ .. 6 9
M materials and M methods .............. .............................................................................. 7 1
Experimental Site ............................... .. .......... ............................. 71
L ab oratory M ethods .................................................. .............................................. 73
Size fractionation......................... .... ........... .......... ............... 73
D ensity fractionation ... ................................................................... .. ............. 73
S o n ic atio n ............................................................................................................. .. 7 4
S statistical A n aly sis .......................................................................................................... 7 5
R esu lts................ ..... .......... ... ............................................................................... . 7 5
Management Intensity and Family Effects....................................................75
Distribution of C and N in the Size-density Fractions ...............................................76
E effect o f D ep th ................................................................................................................ 7 7
D iscu ssio n .............................................................................................................. ........ .. 7 7
Effects of Family ....................... .............. ...... ........... ............... 78
Fraction Characteristics and Effect of Depth .............................................................79
M ethodological C onsiderations........................................ ....................... ................ 80
C o n clu sio n s............................................................................................................. ........ .. 8 1

5 SUMMARY AND CONCLUSIONS.............................................................................92

M ethodological C contributions ....................................................................... ................ 92
A ggregation and Physical Protection....................................... ....................... ................ 93
Influence of Management Intensity and Family................................................................94
Active and Passive C Pools Identified In This Study ........................................................96
A additional R research N eeds................................................... ............................................ 98
M modeling C D ynam ics ..................................................... ............................................... 98

APPENDIX

A EXPERIMENTAL PROTOCOLS .......................................................................................103

B AGGREGATES OBSERVED IN THE SANDY SPODOSOLS................ ................... 108









L IST O F R E F E R E N C E S ....................................................... ................................................ 114

B IO G R A PH IC A L SK E T C H .................................................... ............................................. 12 1




















































8









LIST OF TABLES


Table page

2.1 The effect of wet and dry sieving on C content (g C in size fraction per 100g of
whole soil) in soil size fractions for a sandy Spodosol in north Florida.........................34

2.2 The distribution of N concentration and content among the soil size fractions for a
sandy Spodosol in north Florida ........................................ ........................ ................ 35

2.3 Discriminant analysis of the DRIFTS spectra for a sandy Spodosol in north Florida.......36

3.1 Amount of organic C held in soil aggregates for a sandy Spodosol in north Florida........60

3.2 Energy output of the sonicator probe and the amount of organic matter lost from
each soil size fraction as aggregate organic matter (2000 to 250; 250 to 150 and 150
to 53 tm) for each energy level for a sandy Spodosol in north Florida .........................61

3.3 Effects of forest management intensity and soil depth on aggregate organic matter in
the 2000 to 250, 250 to 150 and 150 to 53 [tm fractions for a sandy Spodosol in north
F lo rid a ........................................................................................................... . ....... .. 6 2

4.1 Characteristics of the size fractions for a sandy Spodosol in north Florida....................82

4.2 Characteristics of the density fractions for a sandy Spodosol in north Florida .................83

4.3 Distribution of aggregate organic matter (AOM) and particulate organic matter
(POM) in the different size-density fractions for a sandy Spodosol in north Florida........ 84

A. 1 Amplitude and time combinations and energy outputs used for this study .................. 106









LIST OF FIGURES


Figure page

2.1 Carbon concentrations (% of fraction) of size fractions as affected by the dry and wet
sieving for a sandy Spodosol in north Florida .............................................. ................ 37

2.2 The interaction between fraction size and sieving method on the ratio of C to organic
matter (C:OM ) for a sandy Spodosol in north Florida.................................. ................ 38

2.3 Diffuse Reflectance Infrared Fourier Transform Spectroscopy spectra showing effect
of sieving method (150 to 53 and <53 [m fractions, low intensity, 5 to 10 cm depth)
for a sandy Spodosol in north Florida........................................................... ................ 39

2.4 Diffuse Reflectance Infrared Fourier Transform Spectroscopy spectra showing effect
of fraction size (dry sieving, high intensity, 0 to 5 cm depth) for a sandy Spodosol in
n north F lorida.................................................................................................... ....... .. 4 0

2.5 Effect of management intensity and soil depth on the C content of the soil size
fractions for a sandy Spodosol in north Florida............................................ ................ 41

2.6 Effect of management intensity and soil depth on N content (g N in a soil fraction
per 100g of whole soil) in soil size fractions for a sandy Spodosol in north Florida ........42

2.7 Diffuse Reflectance Infrared Fourier Transform Spectroscopy spectra showing effect
of intensity (2000 to 250 [tm fraction, dry sieving, 0 to 5 cm depth) for a sandy
Spodosol in north F lorida.. ......................................................................... ................ 43

3.1. Observations of soil aggregation in a sandy surface horizon of a Coastal Plain
S p o d o so l ......................................................................................................... ........ .. 6 3

3.2 Effect of sonication energy input on the loss of aggregate organic matter (AOM, %
of total OM in size fraction) after sonication of the 150 to 53 tm fraction................ 65

3.3 Loss of aggregate organic matter (AOM, % of total OM in size fraction) with
increasing energy for the various soil size fractions..................................... ................ 66

3.4 Diffusive Reflectance Infra-red Fourier Transformed Spectra (DRIFTS) showing
characteristics of particulate organic matter (POM) and aggregate organic matter
(AOM) of the 250 to 150 [m fraction for a sandy Spodosol in north Florida................67

3.5 Effect of forest management intensity on the amount of aggregate organic matter
(AOM, % of total OM in size fraction) for the 2000 to 250 [im fraction for a sandy
Spodosol in north F lorida... ........................................................................ ................ 68

4.1 Effects of family on the C content of the 2000 to 250 [m light density fraction for a
sandy Spodosol in north Florida. ................. ........................................................... 85









4.2 Family x Depth interaction in aggregate organic matter (AOM) of the medium
density, 250 to 150 [m fraction for a sandy Spodosol in north Florida .........................86

4.3 Effects of management intensity and family on N concentrations in the 2000 to 250
rm medium density fraction for a sandy Spodosol in north Florida ...............................87

4.4 Family x management intensity interaction in aggregate organic matter (AOM) of the
heavy density, 2000 to 250 [m fraction for a sandy Spodosol in north Florida ................88

4.5 Effects of soil depth on C and N contents among the various size fractions for a
sandy Spodosol in north Florida. ................. ........................................................... 89

4.6 Distribution of C among the various density fractions for a sandy Spodosol in north
F lo rid a ............................................................................................................. ....... .. 9 0

4.7 Distribution of N among the various density fractions for a sandy Spodosol in north
F lo rid a ............................................................................................................ ........ .. 9 1

5.1 Carbon profile at age 4 for a sandy Spodosol in north Florida.................................. 101

5.2 Carbon profile at age 6 for a sandy Spodosol in north Florida...................................102

A. 1 Effect of sieving time for dry sieving on the weight distribution across size fractions... 106

A.2 Funnel assembly used for density fractionation...... .......... ....................................... 107

B.1 Image and elemental dot map of aggregate (2000 to 250 tm). .................................109

B.2 EDX spectra for the aggregate shown in B.1...................................................... 109

B.3 Image and elemental dot map of aggregate (2000 to 250 tm). .................................110

B.4 EDX spectra for the aggregate shown in B.3....... ... ......................................... 110

B.5. Image and elemental dot map of aggregate (2000 to 250 tm).. .....................111

B .6. ED X spectra for the aggregate show n in B .5................................................................. 111

B.7. Image and elemental dot map of aggregate (250 to 150 tm). ............... ...................112

B.8. EDX spectra for the aggregate shown in B.7............... ........................ 112

B.9. Image and elemental dot map of aggregate (250 to 150 m). ............... ...................113

B. 10. EDX spectra for the aggregate shown in B.9............... ........................ 113









LIST OF ABBREVIATIONS

* SOC: soil organic carbon
* OM: organic matter
* PPINES: Pine Productivity Interactions Experimental Study
* DRIFTS: Diffuse Reflectance Infrared Fourier Transform Spectroscopy
* AOM: aggregate organic matter
* POM: particulate organic matter
* SEM: scanning electron microscope
* SOM: soil organic matter









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

GENOTYPIC AND FOREST MANAGEMENT EFFECTS ON SIZE-DENSITY
FRACTIONATION OF SOIL CARBON IN A FORESTED SPODOSOL


By

Deoyani Vinayak Sarkhot

December 2006

Chair: Nicholas. B. Comerford
Cochair: Eric. J. Jokela
Major Department: Soil and Water Science

Soil C accounts for 75 to 85% of terrestrial C, making soil C sequestration an important

ecosystem service. This study was undertaken to characterize the soil organic carbon (SOC)

pools in a sandy Spodosol of north Florida; to study the aggregate C pool in soils that are

considered to have weak aggregation; and to investigate the influence of intensive forest

management on SOC pools. A loblolly pine (Pinus taeda L.) plantation under two levels of

forest management (fertilization and understory control) was evaluated. Dry and wet sieving

methods were compared for their applicability in size fractionation. Dry sieving was found to be

satisfactory for these soils, as it preserved more structure and the water-soluble SOC components

such as esters and amides. The use of organic matter (OM) release associated with aggregate

breakdown by sonication allowed simultaneous measurement of aggregate strength and amount

of aggregate OM. Aggregate C was found to be an important C pool in these soils, accounting

for nearly half of total SOC. Diffuse Reflectance Infrared Fourier Transform Spectroscopy

(DRIFTS) spectra showed presence of recently added OM in the 2000 to 250 [tm fraction, and

more decomposed OM in the <53 pm fraction. The spectra showed clear separation between

aggregate and particulate OM. Aggregate OM was characterized by higher esters, amides and









polysaccharides than particulate OM, indicating its susceptibility to decomposition in case of

aggregate destruction. The 2000 to 250 [tm fraction, especially its light and medium density

components, was found to be the C pool most responsive to short-term management-related

changes. Intensive management reduced the >53 [tm fraction C as well as aggregate C, probably

due to the reduced input of understory roots. The best family (chosen a priori based on growth)

added more decomposable C as shown by the higher C content in the light density fraction, and

higher N concentration in the medium density under intensive management. The medium family

encouraged aggregation as shown by the higher aggregate OM values. These methods and the

sensitivity of these SOC pools to the short-term management-related changes offer a promise for

better understanding SOC dynamics in sandy Spodosols of Florida.









CHAPTER 1
INTRODUCTION

What of thee, 0 earth, (I) dig out,

let that quickly grow over

let me not hit thy vitals,

nor thy heart....

0 cleansing one!

Atharva Veda (Whitney, 1905)

The need for conserving natural resources had been recognized by civilizations for over

5000 years. However, the management of these resources for sustainable production is still a

challenge. With the rapidly increasing world population, demand for forest products is

increasing, while large tracts of forestland are being converted to other uses or being degraded

(Forest Resources Assessment [FRA], 2005). The alarming rate of deforestation, especially in

the developing countries, demands greater concern for the existing forests. So worldwide,

emphasis is being given to the preservation of natural forests and their non-commodity functions

(FRA, 2005).

Timber harvest has been restricted in many of the world's natural forests and consequently

the onus of providing the world's increasing demand for wood and fiber is now on plantations

managed especially for timber and pulp production. Therefore, it is important to increase the

productivity of these areas. Concentrating timber production on the best sites will allow the

world's wood and fiber demands to be met on fewer acres. It will also allow large areas of native

forests to be conserved or preserved (FRA, 2005).









Pine Plantations in the Southeastern United States

Pine plantations are important commercial ecosystems, covering more than 12 million ha

in the southern United States (Wear and Greis 2002). Apart from their economic significance,

these ecosystems are important for their ecosystem services, such as maintenance of ground

water and air quality. Above and below ground C storage in standing biomass and roots and their

contribution to soil organic carbon (SOC) is another service that will likely become an economic

commodity (e.g., C credits) in the near future.

In the last few decades, increased management intensity has resulted in an unprecedented

increase in the yield of woody biomass and litterfall (Dalla-Tea and Jokela, 1991, McCrady and

Jokela, 1998). This suggests an even higher potential for C sequestration. However, the long-

term impacts of these management changes on SOC pools and on the soil's C sequestration

potential are not well documented. There are four important elements of intensive management

that have the potential to change nutrient cycling and SOC sequestration: fertilization, chemical

understory control, deployment of improved families and site preparation. The first three of those

elements are a part of this dissertation. Fertilization has been shown to increase N and P

mineralization rates (Polglase et al., 1992b), while understory control has been shown to reduce

the SOC (Echeverria et al., 2004). However, no work has addressed genotypic effects of family

or their interaction with fertilization and weed control on the distribution and characteristics of

SOC.

Need for Carbon-wise Management

Soil organic matter is a key constituent of the complex below-ground ecosystem, affecting

a multitude of physical, chemical and biological soil properties. As an important binding agent

for soil aggregates, SOC plays a role in soil aeration, water holding capacity and permeability

through its effect on soil structure and water holding characteristics. Soil organic matter is









important for soil fertility. Humus is a buffer that influences soil pH. It also has high cation

exchange capacity, surpassed only by 2:1 expanding clay minerals that have a higher CEC per

unit volume. In most soils, humus accounts for 20 to 90% of the soil CEC (Brady, 2000). SOC is

also an important storehouse for N, P, S, and micronutrients. It protects nutrients like N from

being leached and protects P and micronutrients from fixation so that they remain available for

plant uptake for a longer period. Chelating functional groups in SOC, improve the availability of

nutrient cations while reducing the toxicity of others. In sandy soils with low inherent fertility

and little clay to perform some of these functions, the significance of SOC in sustaining the

productivity of these ecosystems and the need for understanding the C dynamics are even

greater.

However, total SOC has been found to be an insensitive indicator of management related

changes in sandy soils (Harding and Jokela, 1994). Measurement of changes in C is difficult

against the large amount of recalcitrant or inert material already present in the soil. This is

especially true for detecting the changes in nutrient supply characteristics of SOC in which the

actively cycling fractions of SOC play a major role (Brady, 2000). Physical fractionation

techniques based on differences in size and density may provide greater insight into this question

by separating the active and inert C pools. Measurement of these pools can also give insight into

the mechanisms through which management activities affect total SOC. For example, an increase

in light density C suggests that trees may be adding more decomposable organic matter since this

fraction is an active form of C (Swanston et al., 2005).

Challenges and Opportunities in Sandy Soils

Sandy soils are generally found to be more resilient to compaction by tillage, but they are

probably more susceptible to losses of SOC as there is very little clay to protect it (Shan et al.,

2001). Carbon is sequestered when it is protected from decomposition. Protection is generally









provided by sorption onto clay, by incorporation into aggregates, or as a decomposition product

that is resistant to further microbial attack. The low clay content of sandy surface soils dictates

against the first option, leaving the latter two protection mechanisms as possibilities. Even

though the potential for aggregation through the formation of clay-cation-OM complexes

(Edwards and Bremner, 1967) is small in these soils, aggregation through the action of fungal

hyphae and roots (Tisdall and Oades, 1982) is possible. This suggests a lower potential for C

sequestration in the sandy soils, but offers the opportunity to study physical protection

mechanisms without the confounding influence of clay. Unfortunately, very little is documented

about the biochemical protection of SOC in soils that support southern pine ecosystems. Soil

aggregation studies have been conducted in soils with higher clay contents. Therefore, the

aggregation in these sandy soils deserves investigation.

This study was conducted with the following main objectives:

* To examine soil C forms in the surface horizon of a sandy Flatwoods Spodosol,
* To study aggregation in this soil; and
* To determine the management-related changes in the different soil C pools.

Three laboratory methods, viz. size fractionation, density fractionation and sonication,

were used to study C pools and to determine the effects of forest management on these pools.

The results of this study are provided in three chapters. Chapter 2 focuses on two different size

fractionation methods, dry and wet sieving, and the C and N characteristics of the size fractions.

The hypotheses were that: 1) dry sieving would be a suitable method of fractionation for

studying extremely sandy soils with inherently weak structure; 2) the 2000 to 250 [tm fraction

would be the most important C fraction due to the input of roots and litter; 3) intensive

management that uses fertilization and understory competition control to increase aboveground

biomass would reduce the C content in the 2000 to 250 tm size fraction due to reduction in









understory root inputs; and 4) fertilizer N added through forest management would be reflected

in higher N content in the size fractions.

Chapter 3 focuses on the aggregate morphology, C content, and stability in the surface

horizons of sandy Spodosols. Sonication (breakdown of aggregates using ultrasonic energy) was

used to measure aggregate stability and the pattern of C release from aggregates of varying

stability. The hypotheses tested were that: 1) Surface horizons in a north Florida Spodosol would

have more aggregation than has been described through soil mapping due to the high input of

above ground litter and roots and higher activity of biological agents of aggregation such as

fungal hyphae; 2) aggregate stability would increase with decreasing aggregate size due to

increasing surface area of mineral matter available to the action of organic binding agents; 3)

particulate organic matter would be the dominant C form over aggregate organic matter because

of low clay content to support aggregation; and 4) intensive management would reduce

aggregation in the short-term due to the reduction in the input of understory fine roots.

Chapter 4 focuses on the effects of full sib loblolly pine families (families with both

parents known, chosen a priori based on their growth performance) and their interactions with

management intensity on size-density fractions of SOC. The hypotheses tested in this study were

that: 1) the best growing family would exhibit the highest SOC, especially under intensive

management, due to high litterfall inputs and the responsiveness of this family to intensive

management; 2) the light and medium density fractions would be the main reservoirs of C, since

the mineral matter in these soils is predominantly quartz sand with little or no C adsorption

capacity; and 3) the light and medium density of the 2000 to 250 tm fraction would be the most

responsive pool for detecting family effects because earlier results showed that the 2000 to 250

tm fraction SOC was the most responsive size fraction and because the light and medium









density fractions were most likely to show the effects of recently added organic matter (Romkens

et al., 1999).

Chapter 5 summarizes the most important findings of this study. It identifies opportunities

for future research and suggests a possible course of action for understanding the impact of forest

management on SOC forms and functions.









CHAPTER 2
EFFECTS OF FOREST MANAGEMENT ON SOIL CARBON AND NITROGEN IN A
NORTH FLORIDA SANDY SPODOSOL

Introduction

The soil C density of Spodosols in Florida (20 kg m-2; Stone et al., 1993) is higher than the

soil C density in a majority of the life zones studied (Post et al., 1982), yet little research has

been conducted to study the C profile of these soils. Many Florida Spodosols support southern

pine plantations (Adegbidi et al., 2002), which have the potential for significant above ground

storage of C (Richter et al., 1995). During the last few decades, intensive management of these

plantations has resulted in an unprecedented increase in litterfall and yield of woody biomass

(Dalla-Tea and Jokela, 1991), suggesting the potential to further increase above ground C

storage. However, the short and long-term impacts of forest management on soil organic carbon

(SOC) pools and the C sequestration potential of these soils are poorly documented.

Fertilization and chemical understory control are common silvicultural practices used to

increase yields when managing southern pine stands in the southeastern U.S. Chemical

understory competition control has been shown to reduce total SOC (Shan et al., 2001;

Echeverria et al., 2004) and the mineralization of C and N in both whole soil and soil density

fractions (Polglase et al., 1992a, b; Echeverria et al., 2004). Fertilization has been reported to

increase mineralization rates, especially for P (Polglase et al., 1992a, b). Yet, previous studies

have shown no significant effect of fertilization on total SOC (Harding and Jokela, 1994; Shan et

al., 2001), probably because the increased organic matter inputs associated with large growth

responses compensated the losses by increased mineralization rates.

Physical fractionation of soil into size and density fractions has been an effective technique

for studying the forms and cycling of soil C (Christensen, 1992; Ellert and Gregorich, 1995).

Organic matter (OM) changes in the sand fraction have been useful as early indicators of









management-related C changes. For example, two years of barley (Hordeum vulgare L.) and

alfalfa (Medicago sativa L.) cultivation, compared to a bare soil, increased sand fraction C (24 to

60%), N (36 to 45%) and carbohydrates (46 to 83%; Angers and Mehuys, 1990). Likewise, OM

(>150 [m fraction) was depleted under continuous maize cultivation, yet recovered rapidly when

the land use was returned to pasture (Romkens et al., 1999). In the extremely sandy Spodosols of

the southeastern U.S., which often contain less than 5% silt + clay, the sand fraction C is likely

to equal total C since the clay fraction of these soils contains mainly quartz and kaolinite with

very low C sorption capacity (Harris and Carlisle, 1987). It follows that further fractionation of

the sand size C would be required to detect the short-term management related changes in the

soil C pools. Fractionation of the SOC into meaningful pools also helps in understanding the

processes underlying any possible changes. For example, identifying the pools responsible for

different functions such as short-term nutrient turnover (e.g., sand size fraction) and long-term C

storage (e.g., fine silt fraction; Liu et al., 2003) may help in the development of more accurate C

dynamics models.

Wet sieving is the standard technique for soil size fractionation (Yoder, 1936; Marx et al.,

2005). Even though it disrupts macroaggregates (Angers and Giroux, 1996), wet sieving is

necessary when working with high clay soils to break the strong aggregates that are >2 mm in

diameter. In contrast, the sandy nature of many Coastal Plain soils hinders the formation of large

stable macroaggregates. Under these conditions, dry sieving may offer a viable alternative to the

more time consuming wet sieving method.

This study was conducted to define the distribution of soil C and N in the surface horizon

of a representative sandy Coastal Plain Spodosol and to evaluate the short-term impacts of forest

management activities on these soil characteristics. The first study objective compared wet









sieving and dry sieving as alternative methods for investigating C and N distribution among soil

size fractions. The hypothesis was that, given the low clay content and weak, single grain to

crumb aggregate structure, dry sieving would be a suitable and rapid technique for size

fractionation. The second objective was to establish the C and N distribution, as well as the

chemical fingerprint of OM in the various size fractions. The assumption was that short-term

management impacts would be more identifiable within a size fraction than in the whole soil. We

expected the results to show that the majority of the C and N would be in the 2000 to 250 [tm

fraction, assuming that this fraction received the greatest inputs from roots and aboveground

litter. We also expected this soil to contrast with those having considerably more clay and larger

C pools in the <53 [m fraction. The third objective was, through the use of data generated by the

first two objectives, to provide an evaluation of the suitability of these methods for determining

the short-term impacts of two contrasting forest management intensities. Intensive management,

especially with sustained understory competition control, would be expected to reduce SOC

content in the 2000 to 250 [tm fraction due to the reduction in root C inputs. We anticipated that

there would be enhanced N incorporated into all size fractions because of the fertilizer N inputs.

Materials and Methods

Experimental Site

The study site was a loblolly pine (Pinus taeda L.) plantation in north Florida (3024'N lat;

82 33'W long) managed by the Forest Biology Research Cooperative at the University of

Florida as part of the Pine Productivity Interactions Experimental Study (PPINES). This long-

term study aims at understanding the family x environment interactions in full-sib families of

loblolly and slash pine (P. elliottii Engelm. var. elliottii). The climate is warm, humid

subtropical, with 1,394-mm average annual rainfall, 270C average annual maximum temperature,

and 130C average annual minimum temperature (NOAA, 2002). The soil is classified as a Leon









series (sandy, siliceous, thermic Aeric Alaquod), with <5% silt + clay, and <10 cmolo kg-1 of

cation exchange capacity.

Trees were planted in January 2000 in four replicates using a randomized complete block,

split plot design. Prior to planting, the entire study was double bedded and treated with the

herbicides Arsenal (imazapyr 1.02 L ha-1) and Garlon (triclopyr 7.02 L ha-1) to remove the

understory vegetation and to reduce competition. The experimental design was a 2 x 2 x 8

factorial, which included two planting densities (close spacing at 1.3 x 3 m and wide spacing at 3

x 3 m), two management regimes (high and low inputs) and six elite loblolly pine full-sib

families, a mix of these elite families, and one poor growing family. We chose treatments that

maximized the difference in C inputs to the soil in order to evaluate the capacity for short-term

SOC changes, based on biomass production. The high intensity treatment included the most

productive family and consisted of sustained understory competition control using herbicides,

and annual fertilization using a complete fertilizer. In the high intensity treatment, Arsenal (0.28

L ha-1) and Oust (0.14 L ha-1) were also applied as necessary to provide sustained understory

competition control. The low intensity treatment, planted with the poorest growing family, was

chosen for comparison. The families were designated a priori based on their growth

performance in long-term genetic experiments. This management regime consisted of a one-

time fertilizer application and the aforementioned time of planting understory competition

control treatment. At age four, when sampling was conducted, the fertilizer added to the high

intensity treatment totaled 368 kg ha-1 N and 128 kg ha-1 P plus most other essential nutrients

(i.e., 121 kg ha-1 K, 45 kg ha-1 Mg, 45 kg ha-1 Ca, 35 kg ha-1 S, 0.89 kg ha-1 B, 3 kg ha-1 Zn, 2 kg

ha-1 Mn, 16 kg ha-1 Fe, 4 kg ha-1 Cu, 0.01 kg ha-1 Mo), while the low intensity treatment had 45

kg ha-1 N and 50.6 kg ha-1 P applied as diammonium phosphate. Both treatments were planted at









the 1.3 x 3 m spacing (close spacing 2,900 trees ha-1) and each treatment plot was 480 m2 in

size. The entire study was treated when necessary with insecticides (Dimilin, Pounce or Mimic)

for tip moth (Rhyacionia spp.) control during the first growing season. Further details of the

study site, including the stem volume and above ground biomass are discussed by Roth et al.

(2006).

Soil samples were collected from the A horizon at depth increments of 0 to 5 and 5 to 10

cm from each treatment plot in three replicate blocks in September 2003. One combined soil

sample for each depth of each plot (treatment within a block) came from four individual soil

samples. The four individual soil samples were collected from alternate rows (interbed position),

while within an interbed the sample locations were chosen randomly.

Laboratory Methods

Since the soil moisture content is highly variable and "air-dry" conditions are possible in

these surface soils under field conditions, soil samples were air-dried and passed through 8000

[m and 2000 [tm sieves. The <2000 [m fraction was further size-fractionated in order to contrast

dry and wet sieving methods into four size fractions, a macroaggregate fraction 2000 to 250 am,

two microaggregate fractions 250 to 150 [tm and 150 to 53 [tm, as well as a <53 gm fraction. For

dry sieving, samples were sieved in a mechanical shaker for 5 minutes. The purpose was to

accomplish size fractionation with minimum destruction of soil structure. A previous study

determined that there was no significant difference in weight distribution among the four size

classes after sieving for 5 minutes (unpublished data). Therefore, this time frame was used for all

dry sieving. Wet sieving followed the procedure of Cambardella and Elliott (1993) without pre-

wetting because large aggregates were not common. For each sample, 100 g of soil were added

to the coarsest sieve (250 am) in a tray so that there was standing water 2 cm above the sieve

screen. The sample was allowed to stand for 5 minutes and then the sieve was moved up and









down 50 times (for approximately 2 minutes) taking care that the sieve screen broke the water

surface every time. The soil suspension in the water tray was then transferred to the next size

sieve and the procedure was repeated for each successive sieve (150 and 53 tpm). Each sieve,

when removed from the water tray, was allowed to dry for 24 hours. The < 53 [im fraction was

collected by sedimentation for 48 hours followed by oven drying. The size fractions were

analyzed for total C and N concentrations with a Carlo-Erba CN Analyzer (CE Instruments,

model NCS-2500). The >8000 [tm and 8000 to 2000 [tm fractions were also ground and analyzed

for total C and N concentrations and used for calculating C and N contents.

Since these soils have low clay content and no carbonates, loss on ignition (LOI) was used

to measure OM content (at 5500C for 6h). The ratio of analyzer C to LOI was used to estimate

the C content and C:OM ratio in each size fraction. Thermal Gravimetric Analysis (Omnitherm

951 TGA; Dupont Co., Wilmington, DE) confirmed complete combustion of C during LOI.

The chemical fingerprints of the size fractions were investigated using Diffuse Reflectance

Infrared Fourier Transform Spectroscopy (DRIFTS). Samples were scanned before and after

ashing (at 5500C for 6h) in the mid-infrared on a Varian Digilab FTS-7000 Fourier Transform

Mid-infrared Spectrometer (Walnut Creek, CA). Samples were scanned from 4000 to 400 cm-1

at 4 cm-1 resolution using a KBr beamsplitter and DTGS detector and a Pike Autodiff

autosampler (Pike Technologies, Madison, WI) using ground, non-KBr diluted samples.

Statistical Analysis

The statistical significance of treatments was analyzed using PROC GLM (SAS, 1996),

with forest management intensity, soil depth, soil size fraction, and sieving method as fixed

effects and with block as a random effect. The differences were considered significant at p< 0.05.

Since the initial analyses showed a significant depth x fraction interaction, further analyses were









carried out for the individual size fractions. A multiple comparison procedure with a Tukey-

Cramer adjustment was used for the post-hoc mean separation.

For the DRIFTS spectra, spectral subtraction of ashed samples from non-ashed samples

was used to accentuate OM characterization using GRAMS/AI software Ver. 7.02 (Thermo

Galactic, Salem, NH). Discriminant analysis was conducted using SAS (SAS Institute, Cary,

NC) partial least squares (PLS) using a modified version of a custom made SAS program for

spectral pre-treatments including gap derivatives, scatter correction, and spectral data point

averaging (Reeves and Delwiche, 2003). This modified version allowed discriminant analysis to

be carried out using SAS PLS while the original program was developed only for quantitative

regression analysis. The higher the R2 of the discriminant analysis and the lower the Root Mean

Square Deviation, the better was the separation between main effects as indicated by DRIFTS

spectra.



Results

Dry vs. Wet Sieving

Wet sieving reduced the C concentration by 8% (% of fraction, Fig. 2.1) and C content by

30% (% of whole soil; Table 2.1) in the 150 to 53 .im fraction. There was also a significant

reduction in the weight of this fraction due to wet sieving (16% of whole soil in dry sieving vs.

14% in wet sieving). In the <53 [im fraction, wet sieving increased the C content by 26% (Table

2.1), but reduced the C concentration by 13% (Fig. 2.1) and C:OM ratio by 12% (Fig. 2.2). Wet

sieving reduced the C concentration and C:OM ratio in the 250 to 150 .im fraction, although its

C content was not statistically affected by sieving method. Similarly, the 2000 to 250 tm

fraction was not affected by the sieving method. The total recovery of C as shown by mass









balance for both methods was not statistically different. Nitrogen concentration or content (Table

2.2) was not affected by the sieving method in any of the size fractions examined.

The effect of sieving method was evident on the chemical fingerprint of the size fractions

as measured by DRIFTS (ash-subtracted spectra R2 = 0.94; Table 2.3). Wet sieving reduced C in

all four size fractions across the entire spectra as indicated by spectral peak heights (Fig. 2.3).

Differences in the 2000 to 53 [tm fractions were more pronounced for esters (1730 cm-1), amides

(1650 cm-1) and aromatic compounds (1530 cm-1). On the other hand, in the <53 rm fraction, the

difference between sieving methods was more pronounced in the aliphatic C-H (2870, 2930 cm-

1), and polysaccharide (1160 cm-1) peaks.

Characterization of Size Fractions

Carbon concentration (Fig. 2.1) was highest in the <53 rm fraction (7.8 to 8.8%) and

lowest in the 250 to 150 [m fraction (0.6 to 0.7%). Carbon content (Table 2.1) on the other hand,

was highest in the 2000 to 250 [m fraction and lowest in the <53 [m fraction. More than 65% of

the total C (Table 2.1) in these soils was found in the 2000 to 53 [im fractions, of which about

39% was in the 2000 to 250 [m fraction. N content (Table 2.2) followed a trend similar to C.

The C:OM ratios for the 2000 to 250 [m and 150 to 53 [m fractions were 8 to 27% higher than

the remaining two fractions (Fig. 2.2), which had ratios equal to the standard Van Bemmelen

factor (0.58). The C:N ratios of the fractions were not affected by any of the factors and

therefore, are not reported.

The DRIFTS spectra (Fig. 2.4) confirmed chemical fingerprint differences among size

fractions (ash-subtracted spectra R2 = 0.94 to 0.98; Table 2.3). However, the differences were

mainly in the peak heights corresponding to the amount of C in each fraction. The 2000 to 250

[im fraction exhibited the highest aliphatic C-H peaks (2870, 2930 cm-1), while esters (1730 cm-

1), amides (1650 cm-1) and polysaccharides (1160 cm1) were high in both 2000 to 250 and < 53









lm fractions. The peak for aromatic rings at 1530 cm-1 was similar in these latter two fractions,

but the peak at 1580 cm1 was absent in the 2000 to 250 [m fraction. The ester and amide peaks

were absent in the 250 to 150 [im fraction.

Effect of Management Intensity and Soil Depth

In the 5 to 10 cm depth, low intensity management had 12 and 5% higher C contents in the

2000 to 250 and 250 to 150 [m fractions, respectively, than under the intensive management

regime (Fig. 2.5). Low intensity management also had a 5% higher C content (Fig. 2.5) and 7%

higher N content (Fig. 2.6) than intensive management for the 0 to 5 cm depth in the 150 to 53

[im fraction. In contrast, the high intensity management showed 2% higher C content and 5%

higher N content in the < 53 [m fraction at the 5 to 10 cm depth. The > 2000 [tm fraction

exhibited 22 to 48% higher C content at 0 to 5 cm depth compared to the 5 to 10 cm depth, but C

content in this fraction was not significantly influenced by management intensity.

Soils under the low intensity management regime were higher in OM as shown by

consistently higher peak heights of the DRIFTS spectra (Fig. 2.7; ash-subtracted spectra R2

0.71; Table 2.3). At the 0 to 5 cm depth, differences were observed for all the four size fractions,

while at the 5 to 10 cm depth, a difference was observed only in the 2000 to 53 [tm fractions.

This result was observed across the entire spectrum of OM in the size fractions.

Discussion

This study was undertaken with the objective of understanding the distribution of C across

the size fractions in a forested Spodosol, with the assumption that the size fractions

approximated distinct C pools; and to determine the impact of forest management on these C

pools. Southern pine plantations in the southeastern U.S. represent important regional sinks for C

(Richter et al., 1995) and understanding the soil C dynamics is an essential step towards

sustainable management of these ecosystems.









Dry vs. Wet Sieving

The reduction in C content in the 150 to 53 [im fraction and the increase in C content in the

<53 rm fraction (Table 2.1) indicated a significant transfer of C with wet sieving. This was due

to breakdown of water-dispersible aggregates in the 150 to 53 [im fraction, where the C and a

significant amount of soil mass was washed into the finest fraction. This finding was consistent

with previous work (Carter 1992; Angers and Giroux 1996). The reduction in C concentration

(Fig. 2.1) and C:OM ratio (Fig. 2.2) of the <53 rm fraction indicated that C was lost even from

this fraction as water-soluble C. This interpretation was supported by the DRIFTS data, where

esters, amides and aromatic compounds were lost from the 150 to 53 .im fraction and some

polysaccharides and aliphatic -CH compounds were lost from the <53 [im fraction (Fig. 2.3).

However, the statistically equivalent C recovery of the two sieving methods showed that the loss

of water-soluble C was small, with amounts being within experimental error. The N

concentration and content, on the other hand, was not affected by sieving method, indicating that

N in these soils was not likely in a significant water-soluble or water-dispersible (aggregate)

form. Therefore, our hypothesis of equivalence between dry and wet sieving was accepted for N

distribution and rejected for C distribution, as wet sieving resulted in a redistribution of C. Dry

sieving, which is an easier and quicker method than wet sieving, was considered superior to wet

sieving for the extremely sandy soils examined in this study, as it preserved the water-dispersible

aggregates and water-soluble C.

Distribution of C and N among Size Fractions

More than 65% of the C and N was found in the 2000 to 53 [im fractions (Table 2.1, 2.2),

which supported the hypothesis of the importance of these fractions for C content. The <53 tm

accounted for less than 10% of the total C. This is unlike soils having higher clay contents,

where the C contents tend to be highest in the silt + clay size fractions. Hassink et al. (1997)









found that more than 50% of the soil C was in the < 20 [im fraction (65% clay). Feller (1993)

and Bronick and Lal (2005) reported similarly high C contents in the silt + clay fraction. Though

the C contents reported for extremely sandy soils vary, there is evidence that forested sandy soils

in other parts of the world have similarly high C content within the sand size fractions. For

example, forested Spodosols in France (<10% silt + clay) were shown to have 50% of the total C

within the sand size fraction (Jolivet et al., 2003). Quideau et al. (1998) reported that, under

hardwood forests, 45 to 55% of the total C was associated with the sand size fraction (58%

sand). In contrast, when under maize cultivation, a Spodosol (<10% silt + clay) had only 17% of

its total C in the sand size fraction (Quenea et al., 2006).

The nature of OM, as evidenced by C:OM ratios (Fig. 2.2), differed among size fractions.

The DRIFTS data confirmed this result (Fig. 2.4). The high peaks of esters, amides and aliphatic

C in spectra of the 2000 to 250 [tm fraction indicated recently added undecomposed OM and the

high C:OM ratio of this fraction indicated presence of C rich organic matter. On the other hand,

the high aromatic C peak at 1580 cm-1 in the < 53 [im fraction indicated more decomposed OM

in this fraction, but also showed high peaks of esters and amides, both of which are easily

decomposable C forms.

The C content of organic matter, as measured by the C:OM ratio, is known to change

according to soil type and profile depth (Nelson and Sommers, 1982). However, this study

indicated that it also changes with fraction size and sieving method used (Fig. 2.2). A 6%

underestimation of C would have resulted if the C content of the 2000 to 250 [m fraction was

estimated from the amount of OM using the Van Bemmelen factor (0.58). From a

methodological perspective, these results indicate that one should consider using fraction-

specific conversion ratios when loss on ignition is used to estimate soil fraction C.









Impact of Management Intensity

The higher soil C content associated with the low intensity management regime for the

three >53 rm fractions indicated that intensive forest management reduced the soil C content in

as little as 4 years (Fig 2.5). The DRIFTS spectra showed C loss across the entire spectrum of

organic matter (Fig. 2.7). A possible cause of this decline would be the reduction of root input

caused by the understory control practice. Shan et al. (2001) and Echeverria et al. (2004),

working on similar soils in Georgia, have also reported a decrease in SOC due to intensive

management, especially due to chemical understory control. Therefore, the hypothesis of C

reduction under the high intensity management regime was accepted. However, it is uncertain

whether this C reduction would be compensated by the increase in litterfall under the intensive

management regime in the longer-term. The hypothesis of higher N content in all fractions was

not accepted, since N content showed only a small increase (5%) in the finest fraction (Fig. 2.6).

The fate of fertilizer N applied in the intensive management regime is also uncertain.

Immobilization of fertilizer N in standing biomass, litter layer, or stored in the subsoil is

probably responsible for this. This interpretation is supported by the work of Will et al. (2006),

who reported that 68% of applied fertilizer N was stored in the aboveground biomass and forest

floor, while only 21% was in the surface soil in loblolly pine stands growing in Georgia.

Size fractionation proved to be a more sensitive method than total SOC measurement for

investigating changes in SOC. Management induced change in the 2000 to 250 [im fraction was

23% of the fraction C concentration (Fig. 2.5), which represented greater than 12% change in

total SOC, illustrating the sensitivity of this size fraction for assessing impacts of forest

management. It is also noteworthy that these differences were observed just four years after the

treatments were imposed, supporting the hypothesis that size fractionation enhances the detection

of the short-term management induced changes in SOC.









Given the presence of large palmetto roots (Serenoa repens (B.) Small.) and other

understory plants under the low intensity management regime (Roth et al., 2006), we expected

that the fractions >2000 [tm would decrease significantly with increased management intensity.

The lack of a statistical C response to management in these fractions, the reduced C content in

the 2000 to 250 .im fraction in response to chemical weed control, and the "fresher" nature of the

OM as suggested by DRIFTS imply that the large sand fraction accepts the greatest detrital

inputs. This interpretation is consistent with the work of Van Rees and Comerford (1986) and

Escamilla et al. (1991), who have reported palmetto and other understory root biomass in the

2000 to 250 [im size class under southern pine plantations.

Conclusions

Dry sieving was found to be a useful method for size fractionation for sandy Spodosols

when compared to wet sieving, as it preserved more structure and the water-soluble components

such as esters and amides. The size fractions were significantly different in all the properties

studied. The 2000 to 250 [tm fraction was the most important fraction in these soils, as it

contained nearly half of the total SOC and was sensitive to the management related changes. The

DRIFTS spectra were useful for describing the changes in SOM chemical composition and

indicated presence of recently added organic matter in the large sand fraction. Intensive forest

management reduced soil C in the >53 [tm fractions, and particularly in the 2000 to 250 [tm

fraction, in just 4 years, probably due to the reduction in understory roots.









Table 2.1. The effect of wet and dry sieving on C content (g C in size fraction per 100g of whole
soil) in soil size fractions for a sandy Spodosol in north Florida.
Fraction >8000 [tm 8000 to 2000 to 250 to 150 to < 53 [tm Total
2000 pm 250 pm 150 pm 53 pm
Dry 0.31t 0.47 1.1 a 0.36 a 0.39 b 0.15 a 2.8
(0.02) 1 (0.02) (0.1) (0.02) (0.02) (0.01) (0.3)
Wet 0.31 0.47 1.3 a 0.32 a 0.30 a 0.19 b 2.9
_(0.02) (0.02) (0.1) (0.02) (0.01) (0.01) (0.3)
% of Total 11 17 40 13 14 5 100
10 16 46 11 10 7 100

tEach value is a mean of 36 observations averaged across treatment intensity and soil depth.
1 Values in parentheses represent standard error.
The >8000 and 8000 to 2000 [tm fractions, though not wet sieved, are necessary for total C
content calculations.
T Within a size fraction, the means followed by different letters are statistically different at p <
0.05 showing effect of both depth and sieving method.









Table 2.2. The distribution of N concentration and content among the soil size fractions for a
sandy Spodosol in north Florida.
8000 to 2000 to 250 to 150 to .
Property >8000 tm 2000 tm 250 tm 150 m 53 m < 53 tm Whole soil
___ 2000m 250a m 150 pm 53 pm
Concentration 0.370t e 0.386 e 0.116 c 0.030 a 0.088 b 0.290 d 0.082
(% of fraction) (0.018) (0.015) 1 (0.005) (0.002) (0.003) (0.004) (0.004)
Content
(% of whole 7 a 12 b 39 d 19 c 16 c 7 a 100
soil)

t Each value is a mean of 72 observations averaged across sieving methods, treatment intensity
and soil depth.
1 Values in parentheses represent standard error.
The means followed by different letters are statistically different at p < 0.05









Table 2.3 Discriminant analysis of the DRIFTS spectra for a sandy Spodosol in north Florida
Effect R Root Mean Square Extent of separation
Deviation between spectra
Block 0.25 to 0.55 0.29 to 0.38 Very little separation
Sieving methods 0.94 0.12 Excellent separation
Management intensity 0.71 0.27 Some separation
Depth 0.37 0.40 No separation
Fraction# 0.93 to 0.98 0.06 to 0.12 Excellent separation


t Four blocks replicationss across space)
Dry and wet sieving methods
High and low management intensities
Depths 0 to 5 cm and 5 to 10 cm
# Fractions 2000 to 250, 250 to 150, 150 to 53 and <53 atm










10.00

9.00 L Dry sieving g
--|-- f
Wet sieving
8.00 -

7.00 -

6.00 -
e e
e
5.00 -

4.00 -

3.00 d c

2.00 -
b a
1.00 -

0.00
2000 to 250 pm 250 to 150 pm 150 to 53 pm <53 pm


Figure 2.1. Carbon concentrations (% of fraction) of size fractions as affected by the dry and wet
sieving for a sandy Spodosol in north Florida. Means followed by different letters are
statistically different at p < 0.05. The error bars represent standard error. Each value
is a mean of 36 observations averaged across treatment intensity and soil depth.











O Dry sieving
E Wet sieving
b


2000 to 250 ftm


250 to 150 ftm


150 to 53 tpm


Figure 2.2. The interaction between fraction size and sieving method on the ratio of C to organic
matter (C:OM) for a sandy Spodosol in north Florida. The means followed by
different letters are statistically different at p < 0.05. The error bars represent standard
errors. Each value is a mean of 36 observations averaged across management
intensity and soil depth.


C


b
3E b


0.80

0.70

0.60

* 0.50

0.40
C
-
j 0.30

0.20

0.10


c
b


<53 tpm


0.00 I













<53itm, Dry Sieving

<53[ m, Wet Sieving


0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1


2500 3000 3500 4000 4500


Wave number (cm-1)


Figure 2.3. Diffuse Reflectance Infrared Fourier Transform Spectroscopy spectra showing effect
of sieving method (150 to 53 and <53 tm fractions, low intensity, 5 to 10 cm depth)
for a sandy Spodosol in north Florida. a. polysaccharides, b. aromatic compounds, c.
amides, d. esters, e. aliphatic -CH.


150 to 53tJm, Dry Sieving
'- 150 to 53 tm, Wet Sieving


500 1000 1500 2000













2000 to 250 fm
0.8 < 53 m

0.6 a
150 to 53 fm

S0.4 -
... 250 to 150 pm
0.2 .

0 -- - I i ------- ....--------- .....-----II

0 500 1000 1500 2000 2500 3000 3500 4000 4500
-0.2

-0.4

Wave number (cm 1)


Figure 2.4. Diffuse Reflectance Infrared Fourier Transform Spectroscopy spectra showing effect
of fraction size (dry sieving, high intensity, 0 to 5 cm depth) for a sandy Spodosol in
north Florida. a. polysaccharides, b. aromatic compounds, c. amides, d. esters, e.
aliphatic -CH.











1.80 -
.1 b 0 0 to 5 cm (D1) D1 > D2
1.60 D1 > D2 5 to 10 cm (D2)
1.40 a
a a
2 1.20 I Intensity by Depth interaction:

1.00 -

I 0.80
b
S0.60 b b b

S0.40 aa a

0 0.20
0.00 ]

S High Low High Low High Low High Low High Low High Low
>80001m 8000 to 2000 to 250 to 150gm 150 to 53gm <53gm
20001m 250gm




Figure 2.5. Effect of management intensity and soil depth on the C content of the soil size
fractions for a sandy Spodosol in north Florida. Within a size fraction, the means
followed by different letters are statistically different at p < 0.05, showing the effect
of both management intensity and depth. Each value is a mean of 18 observations
averaged across sieving methods.












0.045
Sa L>H H>L
S 0.040 I
0 H>L a
[] High (H)
S 0.035 a a
I *Low (L)
0.030 -

0.025 -
a b
S0.020 -







O to5 5to10 O to 5 5to 10 O to 5 5 to10 Oto5 5to10 Oto5 5to10 Oto5 5to10
aa








Scm cm cm cm cm cm cm cm cm cm cm cm

>8000P 8000 to 2000p 2000 to 250p 250 to 150p 150 to 53p <53p



Figure 2.6. Effect of management intensity and soil depth on N content (g N in a soil fraction per
100g of whole soil) in soil size fractions for a sandy Spodosol in north Florida.
Within a size fraction, the means followed by different letters are statistically
different at p < 0.05, showing the effect of both management intensity and depth.
Each value is a mean of 18 observations averaged across sieving methods.












-High intensity Low intensity


1
C e
0.9 -

0.8 b d-
0.7 -

0.6 a


0.4 -
0.3

0.2 -

0.1

0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Wave number (cm-1)

Figure 2.7. Diffuse Reflectance Infrared Fourier Transform Spectroscopy spectra showing effect
of intensity (2000 to 250 tm fraction, dry sieving, 0 to 5 cm depth) for a sandy
Spodosol in north Florida. a. polysaccharides, b. aromatic compounds, c. amides, d.
esters, e. aliphatic -CH.









CHAPTER 3
SOIL AGGREGATION AND AGGREGATE CARBON IN A FORESTED SOUTHEASTERN
COASTAL PLAIN SPODOSOL

Introduction

Carbon-wise soil management requires an understanding of the processes by which soil C

is sequestered. Yet, understanding of these processes for forested soils of the southeastern USA

is limited. Secondary forests in the southeastern USA accumulate C at a rate greater than 70 MT

y-1 (Richter et al., 1995), which implicates them as important regional C sinks. Southern pine

plantations are important components of these forests, covering more than 12 million ha (Wear

and Greis, 2002). Many of these plantations are underlain by sandy Spodosols (Adegbidi et al.,

2002), which represent the dominant soil order in Florida, covering 27% of the state (Stone et al.,

1993). Many Spodosols are exceptionally sandy with less than 5% silt + clay and less than 10

cmolc kg 1 of cation exchange capacity (Carlisle et al., 1981, 1988, 1989; Sodek et al., 1990).

Forest fertilization and chemical weed control are two management inputs that have

increased productivity of southern pines in these landscapes. However, fertilization has not

promoted an increase in soil C (Harding and Jokela, 1994; Shan et al., 2001); while chemical

weed control, presumably by reducing detrital inputs of understory plants, has reduced the soil C

content (Shan et al., 2001; Echeverria et al., 2004). The effects of these practices on the

development of aggregates and the soil C contained within them have yet to be considered.

Soil organic carbon (SOC) can be protected from decomposition through four mechanisms:

sorption onto clay particles (chemical protection), incorporation into aggregates (physical

protection), movement to subsoils (translocation), and biochemical transformation into products

that are resistant to microbial attack (biochemical protection; Six et al., 2002; Blanco-Canqui and

Lal, 2004; Jimenez and Lal, 2006). The soil structure of Florida's Spodosols is described as weak

crumb to granular or single grain (Carlisle et al., 1981, 1988, 1989; Sodek et al., 1990),









suggesting poor soil aggregation. In these soils, the low cation content limits aggregate formation

through clay-polyvalent cation-organic matter complexes (Edwards and Bremner, 1967), though

aggregates can form under the influence of microbial products, fungal hyphae, and roots (Tisdall

and Oades, 1982; Blanco-Canqui and Lal, 2004). The potential for chemical protection of soil C

is also limited by the low clay content. Given these factors, the interest in aggregation in sandy

Spodosols in the southeastern U.S. has been low as evidenced by the few studies addressing this

topic.

In Russian Spodosols, aggregation did respond to agricultural management. Water-stable

aggregation and total soil C content increased after just two years under the influence of a grass-

clover mixture, but decreased when followed by spring wheat (Buchkina and Balashov, 2001).

In Florida, soil aggregation was found to be influential in P dynamics. Higher water-extractable

P and heavy metals, along with slower rates of release, were found in the 500 to 250 [tm and 250

to 125 tm aggregates compared to the smaller size fractions (Zhang et al., 2003). Sandy

Spodosols in Florida represent a unique soil condition and it was deemed necessary to better

understand aggregation under these conditions.

The purpose of this research was to study aggregation and its relation to SOC in a

representative forested Spodosol of northern Florida. The first objective was to observe the

aggregation present in these soils. We hypothesized that aggregation, albeit weak, was present in

the < 2 mm fraction in these extremely sandy soils because of the high input from root turnover

and aboveground litter. The second objective was to determine the strength of aggregates in the <

2 mm fraction, and quantify the amount of aggregate C. The hypotheses related to this objective

were that: (1) aggregate strength, as measured by an aggregate's resistance to dispersion, would

increase with decreasing aggregate size, and (2) that particulate C would be the dominant pool of









SOC. The first hypothesis grew from the concept that smaller particles have greater surface area

available for binding; hence, the aggregate's strength/stability would be greater. The second

hypothesis recognized that SOC could be found as either particulate organic matter (POM) or

aggregate organic matter (AOM). Given the weak structure often described for Coastal Plain

Spodosols, the dominance of POM would be expected.

The third objective of this research was to provide preliminary information on the short-

term influence of two contrasting management intensities on the amount and distribution of

AOM and POM. The hypothesis was that intensive management (more fertilization and chemical

weed control) would equate to decreases in aggregation. It was expected that reduced root

turnover of understory plants, resulting from sustained chemical control, would cause reduced

aggregation and short-term reductions in soil C.

Materials and Methods

Experimental Site

A loblolly pine (Pinus taeda L.) plantation in north Florida (3024'N lat; 8233'W long)

was the study site and it was managed by the Forest Biology Research Cooperative at the

University of Florida as part of the Pine Productivity Interactions Experimental Study (PPINES;

Roth et al., 2006). This long-term study aims at understanding the family x environment

interactions in full-sib loblolly and slash pine (P. elliottii Engelm. var. elliottii) families. The

climate is warm, humid subtropical, with 1,394 mm average annual rainfall, 27C average annual

maximum temperature, and 13C average annual minimum temperature (NOAA, 2002). The soil

is classified as a Leon series (sandy, siliceous, thermic Aeric Alaquod), with < 5% silt + clay,

and < 10 cmolo kg-1 of cation exchange capacity.

The trees were planted in January 2000 in four replicates using a randomized complete

block, split plot design. Prior to planting, the entire study was double bedded and treated with the









herbicides Arsenal (imazapyr, 1.02 L ha-1) and Garlon (triclopyr, 7.02 L ha-1) to remove the

understory vegetation and to provide a competition free environment. The experimental design

was a 2 x 2 x 8 factorial, which included two planting densities (close spacing at 1.3 x 3 m and

wide spacing at 3 x 3 m), two management regimes (high and low inputs), and six elite loblolly

pine full-sib families, a mix of these elite families, and one poor growing family. The families

were designated a priori based on their growth performance in long-term genetic experiments.

Two treatment combinations representing the maximum differences in biomass production were

selected in order to evaluate the capacity for short-term SOC changes, with the idea that the

differences in input would be reflected in the differences in SOC pools. The high intensity

treatment included the most productive family under sustained understory competition control

using herbicides and annual fertilization with a complete fertilizer. The low intensity treatment

was planted with the poorest performing family and managed with a one-time fertilizer

application and understory competition control at planting. The fertilizer added to the high

intensity treatment totaled 368 kg ha-1 N and 128 kg ha-1 P plus nearly all other essential

nutrients (i.e., 121 kg ha-1 K, 45 kg ha-1 Mg, 45 kg ha-1 Ca, 35 kg ha-1 S, 0.89 kg ha-1 B, 3 kg ha-1

Zn, 2 kg ha-1 Mn, 16 kg ha-1 Fe, 4 kg ha-1 Cu, 0.01 kg ha-1 Mo); while the low intensity treatment

included 45 kg ha-1 N and 50.6 kg ha-1 P applied as diammonium phosphate. In the high intensity

treatment only, Arsenal (0.28 L ha-]) and Oust (0.14 L ha-]) provided sustained understory

competition control. Both treatments were planted at the 1.3m x 3 m spacing (close spacing -

2,900 trees ha-1) and each treatment plot was 480 m2 in size. The entire study was treated when

necessary with insecticides (Dimilin, Pounce or Mimic) for tip moth (Rhyacionia spp.) control

during the first growing season.









Soil samples were collected from the A horizon at depth increments of 0 to 5 cm and 5 to

10 cm from each treatment plot in three replicate blocks in September 2003. One combined soil

sample from each depth of each plot (treatment within a block) came from four individual soil

samples. The four individual soil samples were collected from alternate interbed rows, while

within an interbed, sample locations were chosen randomly.

Laboratory Methods

Soil samples were air-dried and passed through a 2 mm sieve. They were then dry sieved

to separate the aggregates into four size classes with minimum disruption of aggregation (a

macroaggregate fraction 2000 to 250 atm, two microaggregate fractions 250 to 150 atm and 150

to 53 am as well as a <53 am). A preliminary study determined that there was no significant

difference in weight distribution among the four size classes after shaking the sieves for 5

minutes. Therefore, this time frame was used for all dry sieving.

The first part of the investigation included a microscopic examination of aggregate

morphology in the dry sieved size fractions. The intent was to determine if there was identifiable

aggregation and a positive result would justify examining other objectives. Aggregate samples

were examined and photographed using a dissecting light microscope (Carl Zeiss 475003-9902)

with a mounted digital camera (Sony MVC FD90). A scanning electron microscope (SEM)

(JEOL JSM 6400) equipped with an energy-dispersive x-ray fluorescence elemental

microanalysis (EDX) system was used for obtaining images and silica dot maps within

aggregates. Samples were prepared for SEM by mounting on C stubs and coating with C.

Upon finding aggregates and noting that they were a significant component of the soil

matrix, the next step was to examine aggregate strength and to determine the quantity of C

contained in stable aggregates. This analysis was performed on the three >53 am fractions using

sonication to input energy into a water-soil system. Sonication has been used for aggregate









disruption (North, 1976; Six et al., 2001; Swanston et al., 2005) because, unlike chemical

dispersion techniques, it avoids chemical modification of the organic matter. It also allows

measurement of aggregate strength on an energy basis, which allows for a quantitative

comparison of samples. Energy inputs to the size fractions ranged from 0 to 27,000 J and were

achieved with a Sonic Dismembrator (Fisher Scientific, model 500) by using a range of

amplitude (20 to 60%) and time (1 to 7 min) combinations. The energy output was given by the

sonicator and was calculated by internal software using voltmeter readings recorded every 10

seconds (Fisher Scientific, personal communication, 2006). The energy output thus calculated

was replicable (coefficient of variation < 10%). The pulse method (60 sec ON and 30 sec OFF)

was used to avoid an excessive rise in temperature.

Each size fraction was sonicated at incremental energy levels. This was accomplished by

using one sub-sample for each energy level until complete aggregate breakdown was achieved.

Microscopic observation and release of soil organic matter (SOM) was used to ensure that all the

aggregates were disrupted. This analysis was repeated on the >53 [tm fractions of 12 soil samples

representing three replications of both management intensities and soil depths. For each sample,

sonication was done at 9 to 11 energy levels (one sub-sample for each energy level). For each

sub-sample, 2 g soil was weighed into a 250 mL beaker to which 100 mL water was added. The

suspension was sonicated at the desired energy level. Depth of immersion of the sonicator probe

was kept constant at 10 mm, as this variable is known to influence the degree of disruption

(North, 1976). The suspension was then passed through the same size sieve used to obtain the

size fraction (e.g., 250 [tm sieve for the 2000 to 250 [tm fraction). The SOM remaining on the

sieve and the SOM passing through the sieve were measured by loss on ignition.









Organic matter passing through the sieve after sonication was termed aggregate organic

matter (AOM), as it contained finer organic matter held inside the aggregates that was released

after aggregate disruption. It was expressed as percent of total organic matter in each sample.

Organic matter remaining on the sieve after sonication was termed particulate organic matter

(POM) which, after complete aggregate dispersion, contained SOM of the same size. Energy

input was plotted against AOM to obtain the response curve for each sample. The AOM lost at 0

J represented the organic matter associated with water-dispersible aggregates. The POM data are

not reported because the focus of this paper is aggregation and also because POM can be derived

from the AOM data (POM = 100 AOM).

Upon finding that AOM was a significant component of total organic matter, the chemical

nature of AOM and POM was investigated through the use of Diffuse Reflectance Infrared

Fourier Transform Spectroscopy (DRIFTS). Samples were scanned before and after ashing (at

550C for 6 h) in the mid-infrared on a Digilab FTS-7000 Fourier Transform Mid-infrared

Spectrometer (Varian Instruments; Walnut Creek, CA). Samples were scanned from 4000 to 400

cm-1 at 4 cm-1 resolution using a KBr beamsplitter and DTGS detector and a Pike Autodiff

autosampler (Pike Technologies, Madison, WI) using ground, non-KBr diluted samples. Spectral

subtraction of ashed samples from non-ashed samples was performed to accentuate differences in

organic matter characterization using GRAMS/AI software Ver. 7.02 (Thermo Galactic, Salem,

NH).

Statistical Analysis

The equation, y = (a*x) (b + x)-1 was used to fit the energy vs. AOM release data for each

size fraction; where, a = maximum AOM lost for the size fraction, b = energy level at 0.5*a, x =

sonicator energy output (J) and y = AOM. The energy output for each individual run was used

for this analysis.









PROC MIXED (SAS, 1996) was used to contrast the effects of management intensity and

depth based on a completely randomized design, with energy level (i.e., each unique

combination of amplitude and time), size fraction, management intensity and soil depth as fixed

effects and block as a random effect. The differences were considered significant at p < 0.05.

There was a significant energy level x management intensity x size fraction interaction.

Therefore, further statistical analyses were run separately for each size fraction. A multiple post-

hoc comparison procedure with the Tukey-Cramer adjustment was used.

Results

Aggregate Morphology

Microscopic observation identified two qualitative categories of aggregation. Irregular

shaped aggregates (Fig. 3.1.1 3.1.4) had mineral matter, organic debris and fungal hyphae

and/or fine roots enmeshed together. Spherical aggregates exhibited mineral matter encrusted on

organic debris or plant remains (Fig. 3.1.5) with or without fungal hyphae and/or fine roots (Fig.

3.1.6). The internal structure of the aggregates, as shown by the silicon dot maps (Fig. 3.1.1 -

3.1.4), indicated mineral matter embedded in the organic matter. Images suggested a role of

fungal hyphae and fine roots in aggregate formation, either through mechanical enmeshing (Fig.

3.1.1 3.1.5) or through encrustation of mineral matter on plant remains (Fig. 3.1.7).

Quantifying Organic C in Aggregates

On average, 46% of the total soil C (Table 3.1) was contained in the soil aggregates. As the

energy input to the soil sample increased, aggregates were destroyed, increasing the amount of

AOM removed from that size class of soil material (Fig. 3.2). Eventually a plateau was reached,

indicating that all the organic matter that could be removed from the aggregates was removed

(Fig. 3.3). This finding was supported by microscopic observations (Fig. 3.1.8, 3.1.9), which

indicated that only POM was present when the plateau was reached. The energy level at which









this plateau was reached exceeded the strength of the most stable aggregates in the size fraction

studied. The energy required for the complete breakdown of aggregates (Fig. 3.3) increased in

the order 250 to 150 tm (8500 J) < 2000 to 250 tm (13300 J) < 150 to 53 tm (16422 J).

The shapes of the size fraction response curves were different (Fig. 3.3) among the three

size fractions, which indicated diversity in the nature of aggregates. The aggregates in the 250 to

150 pm fraction were the least stable among the three fractions, losing the highest proportion of

its organic matter (y) at all energy levels (x). The 150 to 53-pm fraction had the most stable

aggregates (Fig. 3.3, Table 3.2). The 2000 to 250 pm and 250 tol50 pm fractions (Fig. 3.3)

exhibited a step-wise loss of organic matter, with steps at 2500 J and 6000 J (Fig. 3.3). The 150

to 53 pm fraction (Fig. 3.2, 3.3) exhibited a continuous spectrum of organic matter loss with

increasing energy input. The variability (Table 3.2) of organic matter release in AOM decreased

with decreasing fraction size, while within a fraction, the variability decreased with increasing

energy output. The highest variability was observed at 0 J (i.e., in the water-dispersible

aggregates). All three size fractions were susceptible to organic matter loss by wetting, with

values ranging from 5 to 17% (Table 3.2).

The selected equation proved satisfactory to model the response of AOM to sonication

energy output for 250 tol50 pm (R2 = 0.93) and 150 to 53 pm (R2 = 0.96) fractions (Table 3.2).

For the 2000 to 250 tm fraction, however, the equation did not work as well (R2 = 0.84) due to

the well-defined steps in the loss of organic matter.

The DRIFTS spectra separated the POM and AOM fractions (Fig. 3.4). The AOM fraction

exhibited higher quantities of polysaccharides (1160 cm-1), aromatic rings (1500 cm-1), esters

(1730 cm-1) and amides (1650 cm-1) than the POM fraction. The negative peak at 1350 cm-1 on

the POM spectra was due to distortions caused by the spectral subtraction and should be ignored.









The peaks at 2880 cm-1 and 2950 cm1 were due to aliphatic -CH groups. Peak heights indicated

higher content of these materials in the AOM component within the size fractions.

Effect of Management Intensity

The effects of energy input and size fractions on AOM in the size fractions were

significant at p < 0.01 (Table 3.3). There was also a significant energy level x management

intensity x size fraction interaction in the AOM data. The effect of management intensity on

AOM was statistically significant in the 2000 to 250 [tm fraction, with an energy level x

management intensity interaction present (Fig. 3.5). At lower energy levels, treatment

differences were small. However, beyond about 6,000 J the low intensity treatment had higher

AOM. The energy level x management intensity interaction was also significant for the 250 to

150 [m fraction, but the absolute difference between the management intensities was less than

4%. The 150 to 53 tm fraction was not affected by the intensity of management.

Discussion

Forested Spodosols of southeastern U.S. are important regional C sinks (Richter et al.,

1995) and understanding the C sequestration mechanisms in these soils can help in maintaining

and improving the C storage potential of these forests. Physical protection of C by incorporation

into aggregates is an important mechanism of C sequestration and this study was conducted with

an objective of understanding the different aspects of aggregation in these soils.

Aggregate Morphology, Stability and OM Content

The microscopic observations confirmed the presence of aggregates and a range of

aggregate forms (Fig. 3.1). Thus, the hypothesis regarding the presence of aggregates was

accepted. The images further implicated the role of fungal hyphae and fine roots in aggregation

(Fig. 3.1.1-3.1.5), while the DRIFTS spectra suggested polysaccharides as an aggregate binding

agent (Fig. 3.4). Oades (1993) discussed the importance of biological agents of aggregation in









sandy soils, since the abiotic mechanisms of aggregation are most important in soils with clay

contents greater than 15% (Horn, 1990). The binding action of polysaccharides, secreted by

fungi or bacteria, has also been previously reported (Tisdall, 1994; Caesar-Tonthat, 2002;

Blanco-Canqui and Lal, 2004). Our observations provide justification to further study the

significance of roots, fungi and polysaccharides as biological agents for aggregation in these

sandy soils, while questioning their role in sequestering C.

Sonication facilitated making quantitative estimates of aggregate stability and aggregate C.

The hypothesis (second objective) suggesting an inverse relationship between aggregate size and

stability was rejected because the 250 tol50 [tm fraction was less stable than both the 2000 to

250 [tm and 150 to 53 [tm fractions (Fig. 3.3). The higher aggregate strength of the 2000 tol50

[tm fraction, which was counter to our hypothesis, appeared to be a function of a

microaggregate/macroaggregate hierarchical structure (see discussion below; Oades and Waters,

1991).

Aggregates, through physical occlusion, protect organic matter from destructive agents

such as physical breakdown by tillage, removal of finer particles by erosion, or decomposition by

soil organisms of different sizes. Not all aggregates offer protection from all of these agents. The

extent of protection depends on the size of pores within the aggregates and the strength of the

aggregates, which in turn depends on the binding agents and the size of the primary particles in

the aggregate. The aggregate strength should indicate the extent of mechanical protection (e.g.,

from breakdown by tillage), but the extent of protection from the soil microbes is not certain. In

the absence of sufficient clay, it is possible that one form of organic matter may be protected by

another form. For example, a rind of biochemically inert material such as aromatic compounds

may discourage the entry of microbes inside the aggregates and protect the labile organic matter









inside. Sonication, as proposed in this study, combined with chemical analyses of the C lost from

aggregates of differential stability, offers the opportunity to advance these studies.

The hypothesis suggesting the dominance of POM was only partially supported, since this

fraction accounted for just over half of the total organic matter. On the other hand, the possibility

of partial breakdown of POM during sonication does suggest that a greater proportion of total

SOM may be in this pool. Though AOM is an important pool of sequestered C, POM may be a

more important in determining the short-term turnover of essential nutrients (Haynes, 2005).

Effect of Management Intensity

The AOM in the 2000 to 250 Im fraction was reduced by intensive management (Fig 3.5).

Therefore, the hypothesis of a short-term reduction in aggregation by intensive management was

accepted. The effect of management intensity could be partly attributed to changes in fine root

biomass. Intensive management, especially chemical control of understory plants, has been

reported to decrease the fine root biomass and length (Escamilla et al., 1991; Shan et al., 2001).

Higher decomposability due to fertilization (Polglase et al., 1992a) may also contribute to

reductions in AOM. However, it is unclear whether these differences between management

intensities will be sustained over time. The higher levels of productivity and C inputs reported

under intensive management (Dalla Tea and Jokela, 1991; Jokela and Martin, 2000) could result

in longer-term opportunities for higher aggregation; especially after canopy closure, when

nutrient deficiencies reduce site productivity levels under the low intensity management regime

(Jokela et al., 2004).

Methodological Considerations

The response of AOM in the 2000 to 250 Itm size fraction to management intensity (Table

3.1, Fig 3.5) illustrated the sensitivity of the sonication technique to soil organic matter changes

in as few as four years after treatment. It also used operationally defined C fractions that could be









related to meaningful C pools. The two pools separated by this procedure within a size fraction,

POM and AOM, showed a marked difference in their chemical composition as indicated by the

DRIFTS spectra (Fig. 3.4). The AOM showed higher content of polysaccharides, phenols, esters

and amides; of which polysaccharides have already been shown to function as binding agents

(Tisdall, 1994). The higher aromatic C content also suggested the presence of more

biochemically inert organic matter. Higher amounts of esters and amides, on the other hand,

suggested that this fraction is susceptible to decomposition if the aggregates were destroyed,

since esters and amides are highly reactive C forms. The differences in chemical composition

indicated that this method should be useful in separating soil C into more functional pools and

lead to better conceptualization of the cycling of SOC when used in conjunction with chemical

decomposition/mineralization studies.

North (1976) used a similar method of aggregate strength measurement, which has been

reproduced by others (Schmidt et al., 1999; Roscoe et al., 2000). However, North (1976) used

the amount of clay lost as an indicator of aggregate destruction, which made it unsuitable for the

highly sandy soils examined in this study. Clay can become saturated with organic matter

(Hassink et al., 1997) and so it is not necessarily an appropriate measure for protection of C in

soils with low clay content. This indicates the suitability of organic matter release instead of clay

release in the sonication technique. The method described here can also be used to study

aggregates of varying stability by sonicating the soil at different energy levels and analyzing the

remaining aggregates for properties such as age or mineralizability.

Aggregate Structure in Coastal Plain Spodosols Additional Considerations

The structure in Spodosols has been described as weak (Carlisle et al., 1981, 1988, 1989;

Sodek et al., 1990). Consequently, aggregation in these soils has received little, if any, attention.

Oades and Waters (1991) studied the patterns of aggregate breakdown in Mollisols, Alfisols and









Oxisols. In Oxisols, the aggregates broke down to release primary particles. In Mollisols and

Alfisols, the authors reported a hierarchical structure; where the larger, weaker aggregates broke

down to release smaller, stronger aggregates, before breaking down into primary particles.

Results from this study indicated that the surface horizon of sandy Coastal Plain Spodosols also

had an aggregate hierarchy, as exhibited by the step-wise breakdown of aggregates in the 2000 to

250 [tm and 250 to 150 [tm fractions (Fig. 3.3). In the 2000 to 250 [tm fraction, the least stable

aggregates were disrupted at energies < 2500 J. At energy levels greater than 2500 J, a plateau in

percent OM lost (AOM) indicated that there was no further aggregate destruction until 3700 J.

We interpret this energy level as the threshold for aggregates of second order, which were

destroyed between 3700 J to 6000 J. A second plateau was observed between 6000 J to 7500 J.

This was the threshold for the third order of aggregates, which started breaking down at 7500 J

and the loss went on until 13300 J when all the aggregates in this fraction were destroyed. In the

2000 to 250 atm fraction, the well-defined steps indicated a well-developed structure. In the 250

to 150 atm fraction, although the first two steps were observed at the same energy levels, they

were less distinct and all the aggregates were destroyed at 8500 J, indicating a poorly developed

structure. However, the 150 to 53 atm fraction exhibited a continuous spectrum, suggesting that

this fraction was simply a continuum of aggregates of different stabilities. This behavior appears

unique for these Spodosols and requires further study, especially since this C pool remained

unaffected by the intensive management regime.

Edwards and Bremner (1967) defined microaggregates as the water stable aggregates

bound by strong clay-polyvalent metal-organic matter complexes. The authors used 250 atm as

the separation point between macro and microaggregates. Tisdall and Oades (1982) also used

250 atm as the separation point between micro and macroaggregates. They described









microaggregates as those stabilized by organo-mineral complexes and resistant to disruption by

wetting, cultivation or other disturbances. In contrast, macroaggregates were described as those

stabilized by roots and fungal hyphae, having varying stability depending upon management and

other factors. Apparently, the separation point at 250 [tm was chosen based on the increase in

aggregate strength and difference in binding agents. However, the similar behavior of the 2000 to

250 and 250 tol50 [tm fractions, as well as the low stability and presence of roots and fungal

hyphae in the 250 to 150 tm fraction, suggests that 150 tm is a more appropriate separation

point for our soils. The 150 to 53 tm fraction, although stable and unaffected by management,

was susceptible to loss by wetting (Table 3.2). This indicates that size alone is insufficient to

define aggregates types in these sandy textured soils. A quantitative measure of aggregate

strength such as the energy required to achieve complete aggregate disruption can be used to

improve the definition. It should be noted that sonication was used to measure mechanical

strength. Chemical modifications may break aggregates before the threshold for mechanical

failure is reached. These factors, in addition to the limited potential for clay or cation binding,

indicate the necessity of a different approach for studying aggregation in sandy Spodosols.

Conclusions

This study found that aggregates form in sandy Spodosols and they have a hierarchal

structure in the large soil size fractions. The use of organic matter release instead of clay release

after aggregate breakdown by sonication was useful for studying aggregate properties. It allowed

simultaneous measurement of aggregate strength and amount of aggregate organic matter.

Aggregate C was an important pool in these soils and the intensive management regime used to

enhance pine plantation productivity reduced this C pool significantly in the short-term. Results

from this study highlight the necessity of using a different approach for aggregate classification

in sandy soils as well as for quantification of aggregate characteristics like stability. Finally,









results from this study highlight the need to assess the long-term, management-related changes

using quantifiable C pools for assessing the metrics of sustainability.









Table 3.1. Amount of organic C held in soil aggregates for a sandy Spodosol in north Florida.
Size fraction C content in size fraction C content in soil aggregates
(pm) (g C in size fraction 100 g-1 soil) (g C in fraction 100 g-1 soil)
2000 to 250 1.1 0.69 (63%)l
250 to150 0.36 0.30 (84%)
150 to 53 0.39 0.28 (72%)
< 53 0.15 NA
Total 2.8 1.28 (46%)

t The C content was measured by loss on ignition and the standard Van Bemmelen factor (0.58)
was used for conversion of organic matter into C content.
1 The values in parentheses represent the average proportion of organic matter lost from the size
fraction after sonication at the highest energy level. The percent loss remained the same for
organic matter and C. See Table 3.2 for the statistical analysis.
NA: Not Applicable, since this fraction was not analyzed for aggregation.









Table 3.2. Energy output of the sonicator probe and the amount of organic matter lost from each
soil size fraction as aggregate organic matter (2000 to 250; 250 to 150 and 150 to 53
lm) for each energy level for a sandy Spodosol in north Florida.
2000 to 250 m 250 to 150 _m 150 to 53 mr__
Energy Aggregate Energy Aggregate Energy Aggregate
(J) OM (%t) (J) OM (%) (J) OM (%)
0 51 a 0 17 a 0 6 a
(0) (1) (0) (2) (0) (1)
945 14 b 936 39b 2813 23 b
(6) (1) (11) (1) (34) (1)
2729 22 c 2468 54 c 5997 45 c
(64) (2) (55) (1) (55) (2)
3793 23c 3805 60c 9445 59 d
(21) (2) (19) (1) (57) (1)
5973 45 d 5969 76 d 12186 65 e
(35) (2) (44) (1) (84) (1)
7446 47 de 7450 78 d 13513 67 ef
(47) (3) (52) (1) (270) (1)
8481 55 ef 8467 82 d 16422 70 f
(42) (2) (54) (1) (362) (1)
9489 58 fg 9467 83 d 20447 71 f
(67) (3) (65) (1) (536) (1)
10505 60 fg 10516 84 d 22297 72 f
(55) (2) (44) (1) (208) (1)
13367 60 fg NAff NA NA NA
(55) (3)
19691 63g NA NA NA NA
(312) (2)
y =(95.7*x)/(7508.3+x) # y =(97. 1*x)/(1773.2+x) y =(98.4*x)/(7038.2+x)
R2 = 0.84 R2 = 0.93 R2 = 0.96

t The aggregate OM or AOM is expressed as a percent of the total OM in the size fraction.
1 Each value is a mean of 12 samples.
Within a size fraction, the means followed by different letters are statistically different at p <
0.05 and show the effect of sonication energy level. The Tukey-Cramer adjustment for mean
separation was used.
Values in parentheses represent standard error.
# The equations are in the form y = (a*x)/ (b + x); where, a = maximum AOM, b = energy level
at 0.5*a, x = sonicator energy output (J) and y = AOM.
ft NA = Not applicable.









Table 3.3. Effects of forest management intensity and soil depth on aggregate organic matter in
the 2000 to 250, 250 to 150 and 150 to 53 [m fractions for a sandy Spodosol in north
Florida.
Main Effects / size fraction 2000 to 250 250 to 150 150 to 53
pm rm rm
Management intensity p = 0.03 NS NS
Soil depth NSt NS NS
Intensity*depth NS NS NS
Energy level p < 0.01
Fraction size p < 0.01
Energy Level x Intensity x Fraction p = 0.02


tNS = Not significant (p>0.05)









Figure 3.1. Observations of soil aggregation in a sandy surface horizon of a Coastal Plain
Spodosol. Figures 3.1.1 to 3.1.4 are irregularly shaped aggregates showing fungal
hyphae, organic debris and mineral matter enmeshed together. Magnification: 1.1 =
87X, 1.2=55X, 1.3 = 295X, 1.4 = 217X. Figures 3.1.5 to 3.1.7 show spherical
aggregates, which are encrustations of mineral matter on organic debris combined
with fungal hyphae/fine roots. Figures 3.1.8 and 3.1.9 are photographs of the 250 to
150 rm fraction. Figure 3.1.8 shows the soil after dry sieving, while Figure 3.1.9 is
after sonication and shows clean sand grains and particulate organic matter but no
aggregates.



















(1) (2) (3) (4)


(5) (6)


(7) (8) (9)




























0 2813 5997 9445 12186 13513 16422 20447 22297
Energy (J)

Figure 3.2. Effect of sonication energy input on the loss of aggregate organic matter (AOM, % of
total OM in size fraction) after sonication of the 150 to 53 [tm fraction. The error bars
represent the range of values (n = 12 samples), while the box represents interquartile
range (upper quartile = 75th percentile, lower quartile = 25th percentile). The plus sign
in the box represents the mean and the line in the box represents the median.


E =-;


Mt












80 -


S60 -

.S 50

40 2000 to 250 m

S3250 to 150Im
S-150 to 53f1m
20 -

10 -

0
0 5000 10000 15000 20000 25000
Energy (J)



Figure 3.3. Loss of aggregate organic matter (AOM, % of total OM in size fraction) with
increasing energy for the various soil size fractions. The vertical lines indicate the
steps in the continuity of organic matter lost from aggregates for the 2000 to 250 and
250 to 150 [tm fractions. Each data point represents the mean of 12 samples.












d


POM -AOM


500 1000


2000 2500 3000 3500 4000 4500


Wave Number cm


Figure 3.4. Diffusive Reflectance Infra-red Fourier Transformed Spectra (DRIFTS) showing
characteristics of particulate organic matter (POM) and aggregate organic matter
(AOM) of the 250 to 150 [m fraction for a sandy Spodosol in north Florida. The
AOM spectra show peaks are: (a) polysaccharides (b) aromatics (c) esters (d) amides
(e) C-H bonds.


1


0.5

o



-0.5


-1













60 -

50 Management intensity significant at p < 0.01

S40

30 Hig

20 -E- Low
I10

0

0 5000 10000 15000 20000 25
Energy (J)


Figure 3.5. Effect of forest management intensity on the amount of aggregate organic matter
(AOM, % of total OM in size fraction) for the 2000 to 250 rm fraction for a sandy
Spodosol in north Florida. Each data point represents the mean of 6 samples.


000


h
Vl









CHAPTER 4
GENOTYPIC AND FOREST MANAGEMENT EFFECTS ON SIZE-DENSITY
FRACTIONATION OF SOIL CARBON IN A FORESTED SPODOSOL

Introduction

Genetic deployment of fast growing and disease resistant families is a key factor for

enhancing forest plantation productivity (McKeand et al., 2003). Contrasting tree species are also

known to differentially influence aspects of soil C accumulation and cycling. For example, acidic

soil condition beneath spruce (Picea spp.) canopies has reportedly lowered microbial biomass

and produced lower rates of microbial CO2 than the soils beneath beech (Fagus spp.) or oak

(Quercus spp.) forests (Anderson and Domsch, 1993). Scrub oak (Q. dumosa Nutt) also has been

shown to contain more soil organic carbon (SOC) in all soil size fractions compared to soils

under coulter pine (Pinus coulteri B. Don.); while litter C was higher beneath the pine forests

(Quideau et al., 2000). The chemical character of SOC can also be influenced by species. In the

last example, O-alkyl C decreased from the litter to the macro organic matter (water-floatable)

and fine silt size fractions for both species, but the decrease was more pronounced under pine,

indicating greater decomposition of cellulose and hemicellulose in pine litter.

Families within a single species may also differ in their adaptative capabilities and

biomass accumulation. For example, loblolly pine (Pinus taeda L.) is an important plantation

species in the United States, Brazil and Argentina, among other countries (Schultz, 1999).

Families of loblolly pine vary in commercially and ecologically important qualities, such as

biomass production, light use efficiency, and nutrient use efficiency (Pope, 1979; Crawford et

al., 1991; McCrady and Jokela, 1998) as well as fusiform rust resistance (Schmidt, 2003) and

response to tropospheric ozone (Taylor, 1994). However, it is not well-known if and how

families affect the accumulation of soil C. The differential growth response of families to

intensive forest management argues that they may be a factor influencing amount and quality of









SOC. Since genetically improved families of loblolly pine are an important element of intensive

management, a better understanding of the influence of genetics, and their interaction with forest

management practices on SOC is warranted.

One of the major mechanisms involved in management-related changes in SOC is the

change in quality and quantity of organic matter inputs. Density fractionation has been used by

many researchers to understand these changes (Romkens et al., 1999; Echeverria et al., 2004), as

this technique separates the organic matter from the mineral matter and accentuates the

differences in organic matter. The light and medium density pools are reported to be actively

recycling fractions with higher C and N concentrations and faster turnover rates. In contrast, the

heavy density fraction is reported to be passive, characterized by low C and N concentrations and

slow turnover rates (Swanston et al., 2005). As families can be expected to differentially impact

SOC through litter and root C inputs, density fractionation may offer sufficient sensitivity to

detect these differences.

This study was undertaken with the overall objective of closing this gap in knowledge.

The first objective was to study the short-term effects of family and family x management

interactions on the surface soil C pools. The hypothesis related to this objective was that the best

family (family designations chosen a priori based on growth performance in long-term genetic

experiments) would promote the greatest increase in SOC, and the effect would be most

pronounced under intensive management. This hypothesis was based on the observation that the

best family used in this study was highly responsive to fertilization, and produced the most

litterfall under intensive management (E.J. Jokela, unpublished data).

The second objective was to determine the profile of C and N in a typical forested

Spodosol and identify the size-density fractions most responsive to the varying levels of forest









management. The first hypothesis related to this objective was that the medium and light density

fractions would be the main reservoirs of C and N in these soils. This was based on the premise

that the mineral fraction in these soils is predominantly quartz sand, with little or no sorption

capacity for SOC. The second hypothesis was that the light and medium density fractions in the

2000 to 250gm fraction would be the responsive pools for detecting genotypic differences in

SOC as influenced by varying management intensity. This hypothesis was formulated on the

observation that in preliminary studies (Chapter 2, 3); this fraction responded most to

management intensity. It was expected that the effects on organic matter would be accentuated

by separating the density fractions.

Materials and Methods

Experimental Site

A loblolly pine plantation in north Florida (3024'N lat; 82 33'W long) was the study site.

It is managed by the Forest Biology Research Cooperative at the University of Florida, as part of

the Pine Productivity Interactions Experimental Study (PPINES). This long-term study aims at

understanding the family x environment interactions in full-sib families of loblolly and slash pine

(P. elliottii Engelm. var. elliottii). The climate is warm, humid subtropical, with 1,394 mm

average annual rainfall, 270C average annual maximum temperature and 13C average annual

minimum temperature (NOAA, 2002). The soil is classified as a Leon series (sandy, siliceous,

thermic Aeric Alaquod), with <5% silt + clay and <10 cmolo kg-1 of cation exchange capacity.

The trees were planted in January 2000 in four replicates using a randomized complete

block, split plot design. Prior to planting, the entire study was double bedded and treated with the

herbicides Arsenal (imazapyr 1.02 L ha-1) and Garlon (triclopyr 7.02 L ha-1) to remove the

understory vegetation and to reduce competition. The experimental design was a 2 x 2 x 8

factorial including two planting densities (close spacing at 3 x 1.3 m and wide spacing at 3 x 3









m), two management regimes (high and low input) and six elite loblolly pine full-sib families, a

mix of these elite families and one poor growing family. Out of these, six treatment

combinations (1 x 2 x 3) representing the close spacing, both management regimes and three

families were chosen. The full-sib families were chosen a priori based on their above-ground

growth performance in long-term genetic experiments and included the best grower, a medium

grower and the poorest grower (above ground biomass 45.2, 42.6 and 40.6 Mg ha-1 respectively

at age 5; Roth et al., 2006).

The high intensity management regime consisted of sustained understory competition

control and annual fertilization using a complete fertilizer. The low intensity management regime

consisted of a one-time fertilization and understory competition control at planting. At age 6,

when the soil was sampled, the fertilizer added to the high intensity treatment totaled 368 kg ha-1

N and 128 kg ha-1 P plus nearly all other essential nutrients (i.e., 121 kg ha-1 K, 45 kg ha-1 Mg, 45

kg ha-1 Ca, 35 kg ha-1 S, 0.89 kg ha-1 B, 3 kg ha-1 Zn, 2 kg ha-1 Mn, 16 kg ha-1 Fe, 4 kg ha-1 Cu,

0.01 kg ha-1 Mo). The low intensity treatment had 45 kg ha-1 N and 50.6 kg ha-1 P applied as

diammonium phosphate. In the high intensity treatment, the herbicides Velpar or Oust and

Glyphosate applied at labeled rates provided sustained understory competition control. The entire

study was treated when necessary with insecticides (Dimilin, Pounce or Mimic) for tip moth

(Rhyacionia spp.) control during the first growing season. Each treatment plot was 480 m2 in

size.

Soil samples were collected from the A horizon of interbeds at soil depth increments of 0

to 5 and 5 to 10 cm from each treatment plot in three of the four replicate blocks in March 2005.

Two composite samples for each depth of each plot (treatment within a block) came from about

six individual soil samples. The individual soil samples for each composite sample were









collected from alternate rows (interbed position), in two lines diagonally across the plot. The

core method was used to measure bulk density (0 to 5 cm depth) based on two samples per

treatment plot.

Laboratory Methods

Size fractionation

Soil samples were air-dried and passed through a 2000 tm sieve. The <2000 [tm fraction

was further size fractionated using dry sieving (see Chapter 2). Four size fractions, a

macroaggregate fraction 2000 to 250 am, two microaggregate fractions (250 tol50 am and 150

to 53 am), as well as a <53 am fraction were obtained. The size fractions, including the > 2 mm

size fraction, were ground to a fine powder and analyzed for total C and N concentrations on a

Carlo-Erba Analyzer (CE Instruments, model NCS-2500). The three 2000 to 53 am fractions

were also analyzed for organic matter content using loss on ignition.

Density fractionation

The most widely used liquids for density separation are water, sodium polytungstate and

Ludox, an inert silica suspension (Christensen, 1992; Cambardella and Elliott, 1993; Meijboom

et al., 1995). Sodium polytungstate is expensive, toxic and is reported to hinder mineralization

(Sollins et al., 1984). Ludox was tested initially in this study but appeared to dissolve some C

from the size fractions, as shown by the considerably darkened supernatant. Therefore, a

modified density separation procedure based on Meijboom et al. (1995) was used, with water as

a separating liquid.

For each size fraction, 10 g soil was added to a 50 mL beaker. Twenty-five mL water was

added to this beaker and the organic matter was separated by swirling and decanting into a

600mL beaker. The process was repeated until no more organic matter could be visually

separated. The light fraction floating on top of the water in the 600 mL beaker was periodically









transferred to another beaker to avoid co-precipitation of the light density fraction (settling down

under the weight of medium density fraction). Organic matter, which could not be separated

from mineral material, was termed the heavy fraction. The suspension in the 600 mL beaker was

added to a 250 mL funnel with the suspension of light density material added on top. Organic

matter, which settled in water, was termed the medium density fraction, while organic matter,

which floated on water, was termed the light density fraction. The 2000 to 250 tm fraction

needed 24 hrs to separate light and medium densities, while the 250 to 150 tm fraction needed

12 hrs to achieve a clear separation. The separated fractions were passed through the same size

sieve used for obtaining the size fraction (e.g., 250 tm sieve for the 2000 to 250 tm fraction) in

order to separate the water-dispersible aggregate fraction. The water-dispersible fraction was

defined as the organic matter passing through the sieve after the size fractions were density-

fractionated with water. The samples were then air-dried and used for further analysis.

Preliminary studies indicated that in the 150 to 53 tm fraction, more than 90% of the SOC was

in the medium density fraction. Therefore, this fraction was not density fractionated.

Sonication

The medium density and heavy density fractions of the 2000 to 250 and 250 to 150 [tm

fractions, as well as the whole 150 to 53 tm fraction, were further fractionated into aggregate

organic matter (AOM) and particulate organic matter (POM) using sonication. In this analysis, a

5 g sample was used for the heavy density fractions, a 1 g sample for the medium density

fractions and a 2 g sample for the 150 to 53 tm whole fractions. The energy level for sonication

was chosen based on the preliminary analysis (Chapter 3). The 2000 to 250 tm and 150 to 53 tm

fractions were sonicated at 20000J for (60% amplitude, 5 minutes; pulse method 60 sec ON and

30 sec OFF), while the 250 to 150 tm fraction was sonicated at 10000J (40% amplitude, 5









minutes; pulse method 60 sec ON and 30 sec OFF). Loss on ignition was used to measure the

AOM and POM separated by sonication.

Statistical Analysis

The statistical significance was analyzed using 'PROC MIXED' (SAS, 1996), with

management intensity, depth, size fraction, and family as fixed effects and block and replication

(nested within block) as random effects. Differences were considered significant at p<0.05. As

the initial analyses showed that the size and density fractions were significantly different from

each other with interactions between depth, intensity and family, further analyses were carried

out for the individual fractions. To analyze the significant interactions within each size-density

fraction, the contrast procedure was used for post-hoc comparisons to ascertain differences

among least square means.

Results

Management Intensity and Family Effects

Family and forest management intensity did not affect SOC and N among the size

fractions, except for the >2000 [im size fraction. In this fraction, the high intensity management

regime reduced the C:N ratio by 17% (60 in high intensity vs. 73 in low intensity). However, the

density fractions of both the 2000 to 250 [m and 250 to 150 [m size fractions exhibited

significant effects of management intensity and family. In the 2000 to 250 [m light density

fraction, the best family produced higher C content than the poor family (Fig 4.1). On the other

hand, the medium family exhibited higher AOM in the 250 to 150 [m medium density fraction

(Fig 4.2). There was also an intensity x family interaction for N concentration in the 2000 to 250

[m medium density (Fig 4.3) and AOM in the heavy density fractions (Fig 4.4). The medium

family had significantly higher soil N concentration when grown under the low intensity

management regime (Fig 4.3). The best family trended toward higher N concentrations under the









high intensity management regime, with the difference being statistically significant at p = 0.06.

In the 2000 to 250 [im heavy density fraction (Fig 4.4), the medium family showed higher AOM

under the high intensity management regime (best family 68% vs. medium family 80%).

Distribution of C and N in the Size-density Fractions

Among the size fractions, C concentration (% of fraction) was highest in the >2000 .im

fraction (Table 4.1), while C content (% of whole soil), was highest in the 2000 to 250 .im

fraction (Fig 4.5). Among the density fractions, C content was highest in the medium density

(47-85%; Fig 4.6) in both the 2000 to 250 and 250 to 150 [im fractions. The C concentration was

highest in the light density 2000 to 250 [im fraction (Table 4.2). The light density material of the

250 to 150 tm fraction accounted for <1% of the fraction's weight and C content. However, it

accounted for 6 to 10% of the C in the 2000 to 250 .im fraction. The water-dispersible fraction

accounted for 7 to 16% of the total C in the size fractions. The medium density of the 250 to 150

.im fraction showed the highest AOM (81 to 82%; Table 4.3), while the medium density of the

2000 to 250 tm fraction showed the lowest amounts (65 to 66%).

Nitrogen concentrations in the size fractions followed the same trend as C (Table 4.1),

but the N content was highest in the 250 to 150 [im fraction (Fig 4.5). The C: N ratio was highest

in the >2000 tm fraction (55-73; Table 4.1) and lowest in the 250 to 150 tm fraction (9-16).

Among the density fractions, N content was highest in the heavy density, especially in the 250 to

150 .im fraction (84 to 92% of total N in the fraction; Fig 4.7). The heavy density also exhibited

extremely low C: N ratios (2.2-3, Table 4.2). The C: N ratio of the light density in the 2000 to

250 tm fraction was highest (56 to 58, Table 4.2) among the density fractions and was similar to

the >2000 tm fraction (55 to 73, Table 4.1).









Effect of Depth

The 0 to 5 cm depth had higher C and N concentrations in all size fractions except the

>2000 [im fraction, which had its higher C concentration at the 5 to 10 cm depth (Table 4.1).

However, the effect was more prominent for C than for N. There was also a significant depth x

fraction interaction in C and N contents of the soil fractions (Fig 4.5). The 2000 to 250 [im

fraction exhibited higher content at the 0 to 5 cm depth, while all other fractions exhibited higher

content at the 5 to 10 cm depth. The 250 to 150 rm fraction in particular exhibited significantly

higher N content (42 vs 28% of total soil N) at the 5 to 10 cm depth.

There was also a depth x density fraction interaction. In both size fractions, the medium

density had higher C concentration at the 0 to 5 cm depth, while the other three density fractions

had higher C concentration at the 5 to 10 cm depth (Table 4.2). The 0 to 5 cm depth also showed

higher AOM in the 2000 to 250 [m heavy density fraction, while the 5 to 10 cm depth showed

higher AOM in the 150 to 53 [m fraction and higher POM in the 2000 to 250 [m heavy density

fraction (Table 4.3).

Discussion

This study was undertaken with the objective of understanding how forest management

activities and genetic deployment affect the C in forested Spodosols of the southeastern U.S.

Utilization of genetically improved seedlings for growth and disease resistance, coupled with

management practices that include understory competition control and fertilizer applications, are

common approaches used for increasing the productivity of managed plantation forests in the

southeastern U.S. Yet, the impacts of management practices and genetics on soil related

processes are still poorly documented.









Effects of Family

The best performing family added more decomposable organic matter to the soil as shown

by the higher C content in the light density fraction (Fig 4.1). The light density fraction has been

reported to be highly mineralizable due to its high C and N concentration and lack of mineral

protection (Romkens et al., 1999; Swanston et al., 2005). The best family also showed a trend

toward higher soil N concentrations under the high intensity management regime (Fig 4.3). This

family was characterized by much higher yield under intensive management. It had 31% higher

stem volume and 11% more aboveground biomass as compared to the poor performing family at

age 5 (Roth et al., 2006). It also exhibited higher N concentrations in the foliage (Jokela, E.J.,

unpublished data). These factors suggest that a combination of the best family and the intensive

management regime may be favorable for faster C turnover rates. Although this is a positive

factor for availability of nutrients such N, P and S in poor fertility soils, the consequences for

long-term C sequestration are uncertain. These data document early trends through age 6 and it

will be necessary to determine whether these effects are sustained over time.

The medium performing family had higher levels of aggregation as shown by the higher

AOM values in the 250 to 150 [m medium density fraction (Fig 4.2). A possible explanation for

this effect is the difference in root biomass and root architecture of these families, since fine

roots are important aggregating agents (Tisdall and Oades, 1982). This family also exhibited

higher N under the low intensity management regime (Fig 4.3), which was reported to have

greater aggregation (Chapter 3). This suggests a role of aggregates in N storage. In the 2000 to

250 rm heavy density fraction, this family showed higher aggregation under the high intensity

management regime (Fig 4.4), suggesting that the positive impact of this family on aggregation

was independent of management intensity. It is possible that the genotypic effects were better

expressed under intensive management as it has been reported to reduce the environmental









variation (Lopez-Upton et al., 1999). This result also offers a possibility for improving the C

sequestration potential of these ecosystems without sacrificing yield through choices of family

deployment.

Fraction Characteristics and Effect of Depth

The high C:N ratios of the medium and light density in the 2000 to 250 [tm fraction (Table

4.2), which were similar to those of the > 2000 [im fraction (Table 4.1), indicated that OM in

these fractions mainly came from recently added litter and root biomass. Similar high C:N ratios,

especially for the light density fraction have been reported by many authors (Swanston et al.,

2004; Gregorich et al., 2006; Liao et al., 2006) This interpretation was supported by the higher C

content in the 2000 to 250 [im size fraction (Table 4.1), as well as in the medium density of this

fraction at the 0 to 5 cm depth(Table 4.2), which can be expected to receive higher root and litter

inputs. This was in accordance with the conclusion in Chapter 2 that the 2000 to 250 [tm fraction

received fresh organic matter. The high C:N ratio of the 250 to 150 [im medium density fraction

(Table 4.2) indicated that this fraction also received fresh organic matter inputs.

The high N content of the 250 to 150 [tm fraction (42% of total N at 5 to 10 cm depth;

Table 4.1) indicated that it may be responsible for the storage of N. It also exhibited the highest

proportion of aggregate C as shown by the high AOM values (Table 4.2), which supports the

importance of aggregates for the N storage in these soils. Within this size fraction, the heavy

density showed the highest N content (84 to 92% of the N; Fig 4.7) and lowest C: N ratios (2.2 to

3, Table 4.3). The heavy density 2000 to 250 [tm fraction also had low C:N ratios. This finding

highlights the importance of heavy density soil material in the N storage of these soils. Zhong

and Makeschin (2006) also reported that the heavy density fraction contained more labile N than

the light density fraction in temperate forest soils. Stable sorption of nitrogenous compounds like

amides on the mineral surfaces (Sollins et al., 2006) can be one of the mechanisms responsible









for the high N content and low C/N ratios of the heavy density fractions. Presence of microbial N

protected by the aggregates may represent another mechanism. Foster (1988) reported that

aggregates offered prime sites for microbes due to the protection from sudden changes in

moisture and from predation by the protozoa and other large predators. Therefore, it is likely that

the N stored in the aggregates is of microbial origin. Although these ratios are unusually low,

Tscherko et al. (2003) reported similar values for microbial biomass (lowest value reported 1.4).

A comparison of the N content in AOM and mineral fraction or the measurement of microbial

C/N ratios by fumigation extraction would be necessary to provide further insight into this

phenomenon.

Methodological Considerations

The modified density fractionation procedure offered many advantages. There was

minimal chance of chemical alteration of the organic matter since only water was used during the

procedure. This also made the procedure inexpensive and easy to use in any laboratory. It

allowed fraction-wise estimation of the water-dispersible aggregates, for which there is no other

method available at present. In agricultural soils, only the water-stable aggregates are usually

measured, since these soils are continually exposed to splash and wind erosion. However, the

forested soils are protected from the erosive forces by a thick litter layer. Therefore, the water-

dispersible aggregates are likely to be important for C dynamics as well as the short-term cycling

of N, P and S in these soils. The medium and light density fractions separated by this procedure

showed effects of family and management intensity on the SOC pools as early as six years after

the treatments were imposed. This indicated the applicability of this method for detecting

management related changes. Other researchers, using different procedures, have also reported

the sensitivity of light and medium density fractions to the short and long-term management

related changes (Romkens et al., 1999; Echeverria et al., 2004).









Conclusions

The 2000 to 250 [im fraction contained the most C, while the 250 to 150 [im fraction was

most important for N storage. Of the density fractions studied, the medium density contained the

most C, while the heavy density contained the most N. The best growing family added more

decomposable C as indicated by higher C in the light density 2000 to 250 atm fraction, while the

medium family encouraged better aggregation, as shown by higher AOM. These findings

indicate that understanding the short-term as well as long-term genotypic effects on soil C would

be necessary for carbon-wise management of the forested Spodosols in north Florida. Water-

dispersible aggregates were an important component of these soils, as they accounted for 7 to

16% of the C in the two largest size fractions. The modified density fractionation procedure used

in this study offered an inexpensive and effective way for separating soil C pools and for

detecting management related changes in them.









Table 4.1. Characteristics of the size fractions for a sandy Spodosol in north Florida
Variable Depth >2000 im 2000 to 250im 250 to 150pm 150 to 53km <53rm
18.8t h 2.21 e 0.43 b 1.48 d 9.99 g
(1.0) (0.13) 1 (0.02) (0.06) (0.36)
concentration
(% of fraction) 5 to 10 cm 23.6 i 0.89 c 0.17 a 0.66 c 5.83 f
(1.1) (0.08) (0.02) (0.04) (0.26)
0.35 de 0.071 c 0.029 ab 0.063 c 0.365 e
(0.02) (0.004) (0.002) (0.003) (0.016)
concentration
Sfr0.33 de 0.030 ab 0.024 a 0.037 b 0.228 d
(% of fraction) 5 to 10 cm
(0.01) (0.002) (0.002) (0.002) (0.009)
55d 31c 16b 24c 28c
0 to 5 cm
C:N Ratio (1) (1) (1) (1) (2)
5 to 10cm 73 e 31 c 9 a 18 b 27 c
(2) (1) (1) (2) (2)

tEach value is a mean of 36 observations averaged across treatment intensity and families.
1 Values in parentheses represent the standard error.
Within a variable, the means followed by different letters are statistically different at p < 0.05
showing the effects of both depths and size fraction.









Table 4.2. Characteristics of the density fractions for a sandy Spodosol in north Florida
Variable Fraction Depth Heavy Medium Light

0.12t a 29.7 b 35.5 c
2000 to 250 (0.01) (0.7) (0.9)
Sm 0.10 a 30.7 be 34.5 be
concentration (0.01) (0.7) (0.9)
concentration
(% of fraction) 0 to 5 cm 0.08 a 22.4 b 20.6
250 to (0.02) (0.5) (4.5)
150 nm 0.07 a 19.9 b 17.6
(0.01) (0.9) NA
0.04 a 0.75 c 0.66 bc
0 to 5 cm
2000 to 250 (0.001) (0.01) (0.04)
Nm 0.04 a 0.64 b 0.61 b
N 5 to 10 cm
concenor1i(0.001) (0.02) (0.04)
concentration -
(% of fraction) 0 to 5 cm 0.041 b 0.55 c 1.05
250 to (0.001) (0.02) (0.12)
150 5m 0.036 a 0.48 c 0.82
(0.001) (0.03) NA
3.0 a 40.3 c 55.8 d
2000 to 250 (0.3) (0.8) (2.9)
Im 2.6 a 50.8 b 58.5 d
C:N Ratio (0.4) (1.7) (2.5)
C :N R atio --------------------
tocm 2.2 a 42.2 b 18.7
0 to 3 cm
250 to (0.6) (1.6) (2.6)
150 n 5m 2.3 a 45.4 b 21.5
(0.4) (2.0) NA

tEach value is a mean of 36 observations averaged across treatment intensity and families.
1 Values in parentheses represent standard error.
Within a variable, the means followed by different letters are statistically different at p < 0.05
showing effect of both depth and density, within a size fraction.
T The light density in 250 to 150 [m was less than <0.01% of the fraction weight and only 5
samples had enough weight for analysis. Therefore, no statistical analysis was carried out for this
fraction.
# The weight and C content of the water-dispersible fraction was measured by difference and not
analyzed directly. Therefore, the characteristics of this fraction were not reported.









Table 4.3.


Distribution of aggregate organic matter (AOM) and particulate organic matter (POM)
n i th diff r nt i d n it fr ti n f r nd S d l in n d


Variable Depth 2000 to 250itm 250 to 150 _m 150 to 53 mrn
Medium Heavy Medium Heavy Whole
AM 65t a 77 b 82 a 73 a 79.5 a
AOM 0 to 5 cm
(% of total OM (2) 1 (1) (1) (2) (0.5)
66 a 70 a 81 a 63 a 81.3 b
in size fraction) 5 to 10 cm (1) () () (3) (0.8)
35 a 23 a 18 a 27 a 20.5 a
POM 0 to 5 cm
(% of total OM (2) (1) (1) (2) (0.5)
34 a 30 b 19 a 37 a 18.7 a
in size fraction) 5 to 10 cm (1) (1) (1) (3) (0.8)

tEach value is a mean of 34-36 observations averaged across treatment intensity and families.
1 Values in parentheses represent standard error.
Within a size-density fraction, the means followed by different letters are statistically different
at p < 0.05 showing the effect of depth.











0.25
a

0.2


0.15 ab b






0
0.05



Best grower Medium grower Poor grower
Family


Figure 4.1. Effects of family on the C content of the 2000 to 250 [m light density fraction for a
sandy Spodosol in north Florida. The means followed by different letters are
statistically different at p < 0.05. The error bars represent the standard errors. Each
value represents the mean of 24 observations averaged across management intensity
and soil depth.


























0 to 5 cm


-4- Best grower
S --- Medium grower
-A-- Poor grower






5 to 10 cm


Soil Depth


Figure 4.2. Family x Depth interaction in aggregate organic matter (AOM) of the medium
density, 250 to 150 [m fraction for a sandy Spodosol in north Florida. The error bars
represent the standard errors. Each value represents the mean of 12 observations
averaged across management intensities.































Best grower


Medium grower
Family


Poor grower


Figure 4.3. Effects of management intensity and family on N concentrations in the 2000 to 250
[m medium density fraction for a sandy Spodosol in north Florida. The error bars
represent the standard errors. Each value represents the mean of 12 observations
averaged across management intensity and soil depth.


0.8

S0.7

0.6

0.5

- 0.4

: 0.3
0
S0.2
0
- 0.1
z


-- High

-4-Low












-*- Best grower --- Medium grower -A- Poor grower


* 80
*- 70
0 ,60

|40

30
S20
0 10
<;


High Low
Management intensity


Figure 4.4. Family x management intensity interaction in aggregate organic matter (AOM) of the
heavy density, 2000 to 250 [m fraction for a sandy Spodosol in north Florida. The
error bars represent the standard errors. Each value represents the mean of 12
observations averaged across soil depths.












100% 0


80%


60%


40%


( 20%


0%
0 to 5 cm 5 to 10 cm 0 to 5 cm 5 to 10 cm

Carbon content Nitrogen content



Figure 4.5. Effects of soil depth on C and N contents among the various size fractions for a sandy
Spodosol in north Florida.


0 250 to 150ftm Q 150 to 53ftm


S <53um


0 2000 to 250gm










E heavy U light E medium 0 water-dispersible


0 to 5 cm


5 to 10 cm


0 to 5 cm


2000 to 150tpm


5 to 10 cm


250 to 150pm


Figure 4.6. Distribution of C among the various density fractions for a sandy Spodosol in north
Florida.


100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%












0 Heavy 0 Medium U Light


11111


I'I I


100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%


0 to 5 cm 5 to 10 cm

250 to 150gm


Figure 4.7. Distribution of N among the various density fractions for a sandy Spodosol in north
Florida.


11111




11111


I'll"^^


0 to 5 cm 5 to 10 cm

2000 to 150gm









CHAPTER 5
SUMMARY AND CONCLUSIONS

Soil organic carbon (SOC) and its chemical nature in sandy Spodosols under loblolly pine

plantations in the southeastern U.S. are poorly described. The objectives of this dissertation

were to investigate aggregation in a representative Spodosol in north Florida under pine

management, to characterize SOC pools in the A horizon, and to determine the impact of

management activities and genetics on the SOC profile. This research required an adaptation of

the size-density fractionation methods to suit the sandy nature of the surface soils, since the weak

structure required less invasive methods for size and density fractionation. The low clay content

also ruled out using clay release after sonication as an indicator of aggregate destruction.

Described below is a critical evaluation of the main findings of this dissertation with focus on

how this research contributed to understanding soil C dynamics in forest soils.

Methodological Contributions

Dry sieving was shown to function as well as, or better than, the more widely used wet

sieving technique. Dry sieving preserved more of the weak structure and the water-soluble

components within the soil, such as esters and amides (Chapter 2). As dry sieving is considerably

faster than wet sieving, it offers economy of time and effort for sandy textured soils.

When sonication was used to disrupt soil aggregates, the loss of organic matter instead of

clay was used as a measure of aggregate stability. The use of organic matter was a superior

procedure because (1) there is little clay to measure in these soils and (2) organic matter in

aggregates and aggregate stability can be measured together (Chapter 3). This procedure also

offers opportunities for understanding the nature of aggregates in these soils as it allows

investigation of aggregate hierarchy when using incremental energy input and organic matter









release. This procedure should also be useful in analyzing the chemical nature of organic matter

in aggregates of differing strength.

The density fractionation procedure used in this study is an inexpensive and chemically

mild technique of organic matter fractionation, since only water was used during the procedure

(Chapter 4). The three density fractions; heavy, medium and light, were different from each other

in their appearance as well as C and N concentrations. They also differed in their response to

changes in forest management. It is clear that this procedure separates three distinct C pools of

different chemical characteristics.

Use of DRIFTS spectra, especially the spectral subtraction technique used in this study

(Chapter 2), proved useful in identifying differences in organic matter pools and in

understanding the changes in chemical composition resulting from sieving method, soil

aggregate size, and forest management. This technique enabled identification of major functional

compounds that differed among C pools. In particular, it was useful in separating aggregate C

and particulate C, showing higher concentration of esters, amides and polysaccharides in

aggregate C.

The fractionation method also proved useful in detecting SOC changes that occurred due to

forest management practices in as little as 4 to 6 y. Although this study focused on C, these

fractionation methods could also be used to study the distribution and dynamics of other soil

nutrients and pollutants.

Aggregation and Physical Protection

The structure of the surface soil of Florida's Spodosols has been described as weak.

Therefore, aggregation has received little, if any, attention. This was the first study to focus on

the soil aggregate C in a Florida Spodosol. Results suggested that nearly half of the C was held

by the aggregates (Chapter 3). This study also showed that, in spite of their mechanical stability,









the aggregates were susceptible to management related influences. Therefore, C-wise

management will require an understanding of the nature and stability of these aggregates. Further

studies on differences in aggregate stability and response to alternative management practices

could offer further insights into the soils' potential to protect SOC through aggregates as a

physical protection mechanism. Even though this study measured the stability of aggregates and

the amount of C held inside the aggregates, quantification of the mineralization potential and C

dating of these pools would be necessary before the C sequestration potential of the various C

pools could be established.

Influence of Management Intensity and Family

The distribution of C across different fractions and the responsiveness of these fractions to

management intensity and families are summarized in Figs 5.1 and 5.2. Intensive management

did not reduce SOC throughout all soil C pools, but only in specific size-density fractions;

especially the 2000 to 250 .im fraction (Chapter 2, 3, 4). It is reasonable to assume that this

change resulted from reduced C input when the understory root turnover was reduced by

sustained weed control practices. The response of this fraction to management and family was

unexpected in some ways. First, the management effect was most pronounced in the stable

aggregates that had strength of 6000 J or higher (Chapter 3). This indicates that measuring the

mechanical strength of aggregates would not be sufficient to determine the mechanisms through

which intensive management practices impact aggregates. An investigation of the chemical

changes in these aggregates due to management practices would be necessary. For example, a

comparison of the DRIFTS spectra of the aggregate organic matter would indicate whether

intensive management reduces the amount of polysaccharides, which are reported to work as

binding agents (Tisdall, 1994).









In the 2000 to 250 [m fraction, the heavy density, which is usually considered a passive

fraction (Romkens et al., 1999), also responded to management (Chapter 4). This was probably

due to predominance of quartz sand in these soils, which has little or no adsorption capacity to

offer chemical protection. When modeling C dynamics in sandy soils, one will need to

reconsider the usual definitions of active and passive fractions.

Another surprising result was the effect of management intensity on the chemical character

of SOM as evidenced by the DRIFTS spectra (Chapter 2). Management intensity was not

expected to affect aromatic C in such a short time frame. Change in the chemical composition of

the inputs, either litter or roots, might be responsible for this effect and a spectroscopic analysis

of the OM inputs under differing management intensities should help to address this question.

The best growing family added more N at the 0 to 5 cm depth and had greater C content in

the 2000 to 250 [m light density fraction than the other two families (Chapter 4). The medium

growing family, on the other hand, exhibited higher AOM in the 250 to 150 [m medium density

fraction and the 2000 to 250 [im heavy density fractions. This illustrates how families may

differentially influence soil properties. The best family is reported to have higher foliar N (Jokela

E.J., unpublished data) and above ground biomass production (Roth et al., 2006) in response to N

fertilization. This opens an avenue for objective-based deployment of families for a variety of

management objectives and ecosystem services (e.g., long-term C sequestration versus improved

nutrient turnover). However, family did not significantly affect the total soil C content of the size

fractions. As this study assessed treatment and family differences through age 6, a follow-up

investigation repeated at age 12 to 15 y may be more definitive in examining changes in soil C

due to management activities.









Active and Passive C Pools Identified In This Study

Identifying the most active C pools would help in understanding the mechanisms behind

the total SOC changes and aid in efficient resource allocation for future research efforts. The

characteristics of the size-density fractions used in this study indicate their functionality, which

will help in relating the operationally defined C fractions to meaningful C pools.

The 2000 to 250im fraction was found to be the C pool most responsive to forest

management activities. The C content of the whole fraction, as well as the aggregate C in this

fraction, responded to management intensity (Chapter 2, 3). In contrast, the light and medium

density fraction in this size fraction responded to family differences (Chapter 4). The chemical

nature of SOC, as shown by the peaks for esters, amides and aliphatic C compounds, indicated

the presence of recently added, undecomposed organic matter in this fraction (Chapter 2). Since

this fraction accounted for nearly half of the total SOC, the changes in this fraction should be

useful for documenting expected changes in total SOC.

The 250 to 150 [im fraction was also sensitive to the effects of management intensity

(Chapter 2) and family (Chapter 4). However, the DRIFTS spectra of this fraction did not offer

any explanation for the differences. Even though the C and N concentrations were the lowest in

this fraction, it accounted for more than 50% of the soil weight and contained 42% of total soil N

(Chapter 4). This highlights the importance of this fraction for SOC and N dynamics. Since this

fraction contained the highest proportion of aggregate C (84%; Chapter 2), further study of the

aggregates should help in understanding the nature of C in this fraction and the role of

aggregation in soil C cycling. For example, measurement of the amount of glomalin in

aggregates of different stability or under different management intensities may help in

identifying the exact effect of forest management activities on aggregation.









The <53 m fraction was also responsive to forest management activities, even though the

magnitude of change was small (2% change in C content, Chapter 2) and exhibited high peaks of

esters and amides. These peaks, combined with the presence of a significant amount of aromatics

further indicated the presence of decomposed OM combined with fresh undecomposed OM

(Chapter 2). However, this fraction was not studied in detail, as it accounted for less than 10% of

the total SOC.

The 150 to 53km fraction on the other hand, was probably the most stable C pool, as

shown by its lack of response to management through age 6 and its high aggregate stability at

age 4 (Chapter 3). This fraction also exhibited a unique behavior in terms of response to

sonication energy output because it did not exhibit aggregate hierarchy. However, this fraction

had a small (5%) decrease in SOC content under the intensive management regime at age 4

(Chapter 2), indicating the presence of more decomposable C forms. Similar effects were not

detected at age 6. A future comparison of the chemical composition of this fraction at the two

sampling times should help explain these differences.

Of the three density fractions studied, the medium density fractions were most responsive

to both forest management intensity and family in both the 2000 to 250[im and 250 to 150 [m

fractions (Chapter 4). The medium density fraction contained the most C. The light density in the

2000 to 250 rm fraction accounted for 7 to 16% of the fraction C and also responded to family

differences. The high C: N ratios of these two fractions suggested the presence of relatively

undecomposed organic matter and the importance of these two density fractions in C cycling. In

contrast, the heavy density fraction was found to be important for N dynamics, as it contained the

highest amount of N. The extremely low C: N ratios (2 to 3) implied that the N in this fraction









was probably microbial N. A study of the fungal and bacterial populations in these density

fractions could provide further insights into the C and N dynamics of these soils.

Additional Research Needs

This study could have been improved. The samples were prepared by air drying for two

weeks and sieving, which is a standard method of sample preparation. However, the soils might

have been passed through the 2 mm sieve when still field moist (with light pressure). While these

soils can be near air-dry during dry down periods, they were moist when sampled. It is not know

if sieving in a field-moist condition would have given different results. I suspect the results

would have been similar; yet it would be worth running a comparative evaluation. Use of the

bulk density cores for the measurement of the >2mm fraction would also have been preferable,

especially for the volumetric C measurements.

Another possible improvement would have been to collect the water-dispersible C by

sedimentation during the density fractionation procedure (Chapter 4) instead of measuring it by

difference. This fraction is likely to be important for the short-term recycling of C and nutrients

and it may be a worthwhile component in future studies.

Modeling C Dynamics

Even though modeling carbon dynamics was not a part of this study, the results suggested

that use of the carbon pools identified in this study could aid carbon-modeling efforts. The

general approach is to use conceptual carbon pools in modeling (e.g., Century model, Parton et

al., 1987; see Smith et al., 1997 for a comparison of various carbon models). However, in most

cases the actual measurement of soil organic carbon is limited to total soil carbon. The physical

fractionation techniques, such as size and density fractionation, can improve the accuracy of

these models by using operationally defined and hence directly measurable carbon fractions to

represent the conceptual carbon pools. For example, the Century model (Parton et al., 1987) used









three soil carbon pools, active, slow and passive, based on turnover rates. The characteristics of

different size-density fractions in this study suggested that the carbon content in the light and

medium density of the 2000 to 250 and 250 to 150 rm fractions could be used to represent the

active pool, whereas the carbon content of the 150 to 53 and <53 [im fractions could be used to

represent the slow or passive pools. Similarly, Hassink and Whitmore (1997) used two pools of

carbon, protected and non-protected, for modeling the buildup and decline of organic carbon

with and without addition of organic matter. Their model was based on the concept that clay

plays an important role in physical protection of organic matter, but the protection capacity of the

soil is limited. The carbon content of the <53 .m fraction in this study could represent the clay-

protected carbon pool, while the >53 .m fractions can represent the unprotected carbon pool.

However, more study of the turnover rates and chemical characteristics of these size-density

fractions would be necessary for this purpose.

The following model, though far from adequate, represents an initial outline. It uses the

continuity equation as the basis and instead of focusing on the right side of the equation, which

incorporates various processes and flows, it focuses on the left side of the equation, which

represents the effect of these processes on changes in carbon pools. Thus, the changes in total

SOC can be separated into changes in labile C (C associated with the active fractions) and

protected C (C associated with passive fractions). The equation can be rewritten as follows:

(dC/dt)x = dClabile + dCprotected

(dC/dt)x = dClabile + dCprotected (physical) + dCprotected (chemical) + dCprotected

(biochemical)

where,









Clabile = light fraction + particulate organic matter in medium and heavy density in both

2000 to 250 tm and 250 to 150 [im fractions.

Cprotected (physical) = aggregate organic matter as measured by loss on sonication in

medium and heavy density of the 2000 to 250 [im and 250 to 150 [im fractions + C in 150 to 53

lm fraction

Cprotected (chemical) = negligible

Cprotected (biochemical) = f(polyphenol content, lignin/N ratio)

However, it must be reiterated that even though AOM and the heavy density organic

matter are considered passive C pools, the AOM in the medium density of 250 to 150 itm

fraction and the AOM in the heavy density of 2000 to 250 [im fraction were influenced by

management intensity and family in as few as 6 years. Also, the particulate organic matter, which

is considered to be a labile pool, did not respond to any of these treatment effects. Therefore, a

study of the mineralization potentials and C dating is required to confirm the C protection

potential of these size-density fractions.

In conclusion, the major contributions of this study were 1) development of C fractionation

techniques suitable for the sandy soils, 2) establishing the importance of aggregates in these

soils, which was a neglected subject until now, 3) establishing the profile of C in terms of

content distribution and, 4) identification of the most responsive C pools to the management-

related changes as well as the pools that have potential for long-term C sequestration.