Characterization of Soil Organic Matter Under Varying Conservation Management Practices a Dissertation Presented to the Faculty

Home , Claypan

Characterization of

Soil Organic Matter

Under Varying

Conservation Management Practices

A Dissertation Presented to The Faculty of the Graduate School At the University of Missouri

In Partial Fulfillment Of the Requirements for the Degree Doctor of Philosophy

By KRISTEN SLOAN VEUM

Drs. Keith W. Goyne and Peter P. Motavalli, Dissertation Supervisors

May 2012

The undersigned, appointed by the Dean of the Graduate School, have examined the dissertation entitled Characterization of Soil Organic Matter

Under Varying Conservation Management Practices

Presented by Kristen Sloan Veum

A candidate for the degree of Doctor of Philosophy And hereby certify that, in their opinion, it is worthy of acceptance.

Keith W. Goyne (Chair)

Peter P. Motavalli (Co-chair)

Robert J. Kremer

Kenneth A. Sudduth

DEDICATION

This work is dedicated to

Kermit Norman Veum

and

Julia Sloan Holan

ACKNOWLEDGEMENTS

First and foremost, this work would not have been possible without the guidance and oversight of my advisor, Dr. Keith Goyne, and my other faculty mentors, Dr. Robert

Kremer, Dr. Randy Miles, Dr. Peter Motavalli and Dr. Ken Sudduth. Dr. Goyne gave me the opportunity to work in soil science, a subject area outside my previous experience and expertise, and he put up with me for five years. I am eternally grateful for the opportunity he gave me to become a soil scientist and earn my PhD. My family also deserves credit for assisting me in countless ways over the past five years. My husband, Scott, was a source of infinite encouragement and insight. My mother helped enormously after the birth of my daughter, Julia, and my father advised me based on his years of experience as a faculty member at MU.

Many of my current and former classmates helped me in the field and in the laboratory, and I would especially like to thank Meredith Albers, Satchel Gaddie, Cammy

Willett and Bei Chu for their moral support, and Jon Hendrell and Caroline Tanner for their diligent assistance in the lab. I am also grateful to Russell Dresbach and Sara

Rosenkoetter at the University of Missouri Soil Characterization Lab and to Wei Wycoff at the University of Missouri NMR Core Facility for analytical assistance. In addition, I would like to thank the University of Missouri Center for Agroforestry and the Greenley

Memorial Research Center for developing and maintaining the experimental site. This work was funded through the MU Center for Agroforestry under cooperative agreements

58-6227-1-004, 58-6227-2-008 and 58-6227-5-029 with the USDA-ARS. Any opinions, findings, conclusions or recommendations expressed in this dissertation are those of the author and do not necessarily reflect the view of the U.S. Department of Agriculture.

TABLE OF CONTENTS

Acknowledgements...... ii

Table of Contents...... iii

List of Tables ...... viii

List of Figures...... xvi

CHAPTER 1 ...... 1

Introduction...... 1

CHAPTER 2 ...... 4

Runoff and dissolved organic carbon loss from a paired-watershed study of three adjacent agricultural watersheds ...... 4

Abstract ...... 4

Introduction...... 5

Methods and Materials...... 8

Site description and management...... 8

Sample collection and analysis ...... 12

Statistical analysis ...... 15

Results and Discussion...... 16

Precipitation, runoff and DOC loss...... 16

iii

Buffer effects on runoff and DOC loss ...... 20

Crop effects on runoff and DOC loss...... 27

Seasonal effects on runoff and DOC loss...... 28

Conclusions...... 29

CHAPTER 3 ...... 31

Assessment of soil organic carbon and total nitrogen under conservation management practices in the Central Claypan Region, Missouri, USA...... 31

Abstract ...... 31

Introduction...... 32

Materials and Methods...... 38

Study site...... 38

Soil sampling and laboratory analyses...... 40

Statistical analysis ...... 44

Results...... 46

Bulk density, total SOC and TN...... 46

Particulate, adsorbed and occluded organic carbon (PAO-C) and nitrogen (PAO-N)

...... 47

SOC and TN ratios ...... 51

Discussion ...... 51

Conclusions...... 56

CHAPTER 4 ...... 57

Water stable aggregates and organic matter fractions under conservation management practices on claypan soils...... 57

Abstract ...... 57

Introduction...... 58

Materials and Methods...... 63

Study site...... 63

Soil sampling and laboratory analyses...... 64

Statistical analysis ...... 69

Results...... 70

Treatment effects on soil properties ...... 70

Comparison of whole soil to the PAO fraction...... 73

Relationships among soil properties ...... 75

Comparison of WSAFM, WSAAD and the PAO fraction ...... 77

Discussion ...... 78

Conclusions...... 82

CHAPTER 5 ...... 84

Microbial enzyme activity and labile soil organic matter under vegetative filter strips and no-till agriculture ...... 84

Abstract ...... 84

Introduction...... 85

Materials and Methods...... 90

Study site...... 90

Soil sampling and laboratory analyses...... 93

Statistical analysis ...... 100

Results...... 101

Treatment effects...... 101

Relationships among POXC, enzyme activities and other soil properties...... 108

Comparison of the PAO fraction to whole soil...... 112

Discussion ...... 113

Conclusions...... 116

CHAPTER 6 ...... 118

Spectroscopic comparison of soil organic matter and humic acid under varying conservation management practices ...... 118

Abstract ...... 118

Introduction...... 119

Materials and Methods...... 123

Study site...... 123

Soil sampling and laboratory analyses...... 125

Soil organic matter extraction ...... 126

Humic acid extraction ...... 128

Solid state 13C nuclear magnetic resonance analysis...... 130

Diffuse reflectance infrared Fourier transform analysis ...... 132

Statistical analysis ...... 133

Results...... 136

13C nuclear magnetic resonance (NMR) spectra of SOM ...... 136

13C nuclear magnetic resonance spectra of humic acid...... 139

Diffuse reflectance infrared Fourier transform spectra of SOM ...... 143

Diffuse reflectance Fourier transform infrared spectra of humic acid ...... 147

Discussion ...... 147

Conclusions...... 151

CHAPTER 7 ...... 153

Conclusions...... 153

APPENDIX...... 155

Two-Factor ANOVA with Subsamples ...... 155

Bibliography ...... 161

Vita...... 181

vii

LIST OF TABLES

Table 2.1 Characteristics of the Control, Agroforestry (AGF) and Grass watersheds

(adapted from Udawatta et al., 2006)...... 10

Table 2.2 Summary of precipitation and agricultural activities from 1991 to 2006...... 13

Table 2.3 Summary of measured runoff and estimated dissolved organic carbon (DOC)

loss...... 18

Table 2.4 Analysis of buffer type, crop and seasonal effects on runoff and dissolved

organic carbon (DOC) loss...... 24

Table 3.1 Mean soil properties in the 0–5 cm and 5–13 cm layers of the study soils by

conservation management practice. The standard error of the mean is stated in

parentheses...... 42

Table 3.2 Means of soil total organic carbon (SOC), total nitrogen (TN) and sand-free,

particulate, adsorbed and occluded soil organic carbon (PAO-C) and total

nitrogen (PAO-N) expressed on a concentration, soil volume and soil mass

(represented by the subscripts c, v and m, respectively) basis by conservation

management practice. Values followed by a different lowercase letter within

columns and for a given layer are significantly different (using Tukey’s HSD) at α

= 0.05. The standard error of the mean is stated in parentheses...... 48

Table 4.1 Mean soil properties in the surface (0–70 kg m-2) layer of the study soils by

conservation management practice. Values followed by a different lowercase

letter within rows are significantly different (using Tukey’s HSD) at α = 0.05

(n=12 per treatment). Values without lowercase letters were determined using

viii

composite samples (n=4 per treatment).The standard error of the mean is stated in

parentheses. Data partially obtained from Veum et al. (2011)...... 68

Table 4.2 Mean C:N ratios in whole soil, the particulate, adsorbed and occluded (PAO)

fraction, water-extractable organic matter from whole soil (WEOM), and water-

extractable organic matter from the particulate, adsorbed and occluded fraction

(WEOMP). Values followed by a different capital letter within columns and for a

given soil sampling depth and phase are significantly different. Values followed

by a different lowercase letter within rows are significantly different (using

Tukey’s HSD) at α = 0.05 and the standard error of the mean is stated in

parentheses...... 72

Table 4.3 Means of field-moist water stable aggregates (WSAFM) and air-dried water

stable aggregates (WSAAD) by conservation management practice expressed on a

gravimetric, soil volume, and soil mass basis, as well as WSAFM and WSAAD

normalized to soil organic carbon (SOC), water-extractable organic carbon

(WEOC), and particulate, adsorbed and occluded organic carbon (PAO-C).

Values followed by a different lowercase letter within columns and for a given

soil layer are significantly different (using Tukey’s HSD) at α = 0.05. The

standard error of the mean is stated in parentheses...... 73

Table 5.1 Mean soil properties in the surface (0–70 kg m-2) layer of the study soils by

conservation management practice. Values followed by a different lowercase

letter within rows are significantly different (using Tukey’s HSD) at α = 0.05

(n=12 per treatment). Values without lowercase letters were determined using

composite samples (n=4 per treatment).The standard error of the mean is stated in

parentheses. Data partially obtained from Veum et al. (2011)...... 95

Table 5.2 Mean soil properties in the subsurface (70–130 kg m-2) layer of the study soils

by conservation management practice. Values followed by a different lowercase

letter within rows are significantly different (using Tukey’s HSD) at α = 0.05

(n=12 per treatment). Values without lowercase letters were determined using

composite samples (n=4 per treatment).The standard error of the mean is stated in

parentheses. Data partially obtained from Veum et al. (2011)...... 96

Table 5.3 Mean surface layer values for whole soil and the particulate, adsorbed and

occluded fraction (subscript P) of KMnO4-oxidizable organic carbon (POXC),

ultraviolet absorbance at 254 nm (A254), specific ultraviolet absorbance at 254

nm (SUVA, L cm-1 m-2), the ratio of ultraviolet absorbance at 280 to 365nm

(E2:E3), and the ratio of absorbance at 253to 203 nm (ET:BZ), POXC (g POXC

-1 -1 kg soil), POXCP (g POXC kg PAO), SUVA and SUVAP are calculated on a

sand-free basis for comparison between soil fractions. Values followed by a

different capital letter within columns for a given variable and for a given soil

sampling depth are significantly different. Values followed by a different

lowercase letter within rows are significantly different (using Tukey’s HSD) at α

= 0.05. Standard errors are stated in parentheses...... 103

Table 5.4 Mean subsurface layer values for whole soil and the particulate, adsorbed and

occluded fraction (subscript P) of KMnO4-oxidizable organic carbon (POXC),

ultraviolet absorbance at 254 nm (A254), specific ultraviolet absorbance at 254

nm (SUVA, L cm-1 m-2), the ratio of ultraviolet absorbance at 280 to 365nm

(E2:E3), and the ratio of absorbance at 253 to 203 nm (ET:BZ), POXC (g POXC

-1 -1 kg soil), POXCP (g POXC kg PAO), SUVA and SUVAP are calculated on a

sand-free basis for comparison between soil fractions. Values followed by a

different capital letter within columns for a given variable and for a given soil

sampling depth are significantly different. Values followed by a different

lowercase letter within rows are significantly different (using Tukey’s HSD) at α

= 0.05. Standard errors are stated in parentheses...... 107

Table 5.5 Means of dehydrogenase activity (DH) and phenol oxidase activity (PO) by

conservation management practice expressed on a gravimetric, and soil mass

basis, as well as DH and PO normalized to total soil organic carbon (SOC),

particulate, adsorbed and occluded organic carbon (PAO-C), KMnO4-oxidizable

organic carbon (POXC), and water-extractable organic carbon (WEOC). Values

followed by a different lowercase letter within columns and for a given soil layer

are significantly different (using Tukey’s HSD) at α = 0.05. The standard error of

the mean is stated in parentheses...... 109

Table 5.6 Correlation coefficients across all surface samples (n=36) for values from

unfractionated soil of absorbance at 254 nm (A254), specific ultraviolet

absorbance (SUVA, L cm-1 m-2), ratio of absorbance at 280 to 365 nm (E2:E3),

ratio of absorbance at 253 to 203 nm (ET:BZ), water-extractable organic carbon

(WEOC, mg m-2), carbon to nitrogen ratio of water-extractable organic matter

-2 (W-C:N), KMnO4-oxidiazble organic carbon (POXC, kg m ), particulate,

adsorbed and occluded organic carbon (PAO-C, kg m-2), solid carbon to nitrogen

ratio (C:N), soil organic carbon (SOC, kg m-2), water stable aggregates (WSA kg

m-2), phenol oxidase activity (PO, mol m-2 hr-1), and dehydrogenase activity (DH,

mg m-2 hr-1). A portion of the data from Veum et al. (2011) and Chapter 4...... 110

Table 5.7 Correlation coefficients across all surface samples (n=36) for values from the

particulate, adsorbed and occluded fraction (subscript P) of absorbance at 254 nm

(A254), specific ultraviolet absorbance (SUVA, L cm-1 m-2), ratio of absorbance

at 280 to 365 nm (E2:E3), ratio of absorbance at 253 to 203 nm (ET:BZ), water-

extractable organic carbon (WEOC, mg m-2), carbon to nitrogen ratio of water-

extractable organic matter (W-C:N), KMnO4-oxidiazble organic carbon (POXC,

kg m-2), particulate, adsorbed and occluded organic carbon (PAO-C, kg m-2), solid

carbon to nitrogen ratio (C:N), soil organic carbon (SOC, kg m-2), water stable

aggregates (WSA kg m-2), phenol oxidase activity (PO, mol m-2 hr-1), and

dehydrogenase activity (DH, mg m-2 hr-1). A portion of the data from Veum et al.

(2011) and Chapter 4...... 111

Table 6.1 Mean soil properties in the surface (0–70 kg m-2) layer of the study soils by

conservation management practice. Values followed by a different lowercase

letter within rows are significantly different (using Tukey’s HSD) at α = 0.05

(n=12 per treatment). Values without lowercase letters were determined using

composite samples (n=4 per treatment).The standard error of the mean is stated in

parentheses. Data from Veum et al. (2011) and Chapters 4 and 5...... 127

Table 6.2 13C nuclear magnetic resonance (NMR) chemical shift (ppm) assignments used

in this study from Stevenson (1994) and Wilson (1987)...... 131

xii

Table 6.3 Diffuse reflectance infrared Fourier transform (DRIFT) spectral peak

assignments and associated wavenumbers used in this study (Baes and Bloom,

1989; Inbar et al., 1989; Stevenson, 1994)...... 134

Table 6.4 Diffuse reflectance Fourier transform infrared spectral peak ratios evaluated in

this study and the associated functional groups (Wander and Traina, 1996; Ding et

al., 2002)...... 135

Table 6.5 Mean proportions (%) of C structural components soil organic matter as

determined by cross-polarization, magic angle spinning 13C nuclear magnetic

resonance (CPMAS 13C NMR) spectroscopy from grass vegetative filter strip

(VFS), agroforestry VFS and no-till soil. Values followed by a different

lowercase letter within rows are significantly different (using Tukey’s HSD) at α

= 0.05 (n=4 per treatment). The standard error is stated in parentheses...... 138

Table 6.6 Pearson correlation coefficients (r) for carbon functional groups observed in

cross polarization, magic angle spinning 13C nuclear magnetic resonance

(CPMAS 13C NMR) across all samples (n=12) of soil organic matter, calculated

per unit mass of soil, with soil organic carbon (SOC), KMnO4-oxidizable carbon

(POXC), particulate, adsorbed and occluded carbon (PAO-C), water extractable

organic carbon from unfractionated soil (WEOC) and the PAO fraction

(WEOCP), the carbon to nitrogen ratio of unfractionated soil (C:N), the PAO

fraction (C:NP), water extracts of unfractionated soil (W-C:N) and the PAO

fraction (W-C:NP). Functional groups not represented in this table did not present

significant correlations...... 140

xiii

Table 6.7 Mean proportions (%) of humic acid (HA) C structural components as

determined by cross-polarization, magic angle spinning 13C nuclear magnetic

resonance (CPMAS 13C NMR) spectroscopy from grass vegetative filter strip

(VFS), agroforestry VFS and no-till soil. Values followed by a different

lowercase letter within rows are significantly different (using Tukey’s HSD) at α

= 0.05 (n=4 per treatment). The standard error is stated in parentheses...... 142

Table 6.8 Pearson correlation coefficients (r) across all samples (n=12) for carbon

functional groups, identified using 13C nuclear magnetic resonance (NMR),

carbon functional groups of humic acids with soil organic carbon (SOC), KMnO4-

oxidizable carbon (POXC), particulate, adsorbed and occluded carbon (PAO-C),

water extractable organic carbon from unfractionated soil (WEOC) and the PAO

fraction (WEOCP), the carbon to nitrogen ratio of unfractionated soil (C:N), the

PAO fraction (C:NP), water extracts of unfractionated soil (W-C:N) and the PAO

fraction (W-C:NP). Functional groups not represented in this table did not present

significant correlations...... 144

Table 6.9 Diffuse reflectance infrared Fourier transform (DRIFT) spectral peak ratios of

soil organic matter from grass vegetative filter strip (VFS), agroforestry VFS and

no-till soil. Means within rows with different lowercase letters are significantly

different (p < 0.05). The standard error is stated in parentheses...... 146

Table 6.10 Diffuse reflectance infrared Fourier transform (DRIFT) spectral peak ratios of

humic acids from grass vegetative filter strip (VFS), agroforestry VFS and no-till

soil. Means within rows with different lowercase letters are significantly different

(p < 0.05). The standard error is stated in parentheses...... 149

xiv

Table A.1.1 Comparison of p-values using a standard two-factor ANOVA with

subsamples as replicates (pseudoreplication) and a modified two-factor ANOVA

for soil organic carbon (SOC), and total nitrogen (TN) calculated on a

concentration, soil volume and soil mass basis in the surface layer (0 – 5 cm) for

no-till (NT), agroforestry (AGF) and grass (GR) soils...... 160

LIST OF FIGURES

Figure 2.1 Location of the Greenley Memorial Research Center in Missouri, USA and the

study watersheds. Location of flumes and autosamplers (triangles), elevation

contour lines (meters above sea level; black lines), grassed waterways (grey areas,

perpendicular to contours) and buffer strips (grey lines following contours) are

noted. Adapted from Udawatta et al. (2004)...... 9

Figure 2.2 Precipitation (PPT), runoff and dissolved organic carbon (DOC) loss during

the year 1993...... 17

Figure 2.3 Total measured runoff for all watersheds by year. Dashed line indicates the

year (1997) when buffers were installed...... 19

Figure 2.4 Relationship between log-transformed runoff and dissolved organic carbon

(DOC) loss for all samples from all years...... 21

Figure 2.5 Paired-watershed relationships between control and treatment watersheds for

runoff by buffer type (a-b), crop (c-d) and season (e-f)...... 22

Figure 2.6 Paired-watershed relationships between control and treatment watersheds for

dissolved organic carbon (DOC) loss by buffer type (a-b), crop (c-d) and season

(e-f)...... 23

Figure 2.7 Paired-watershed relationships between control and treatment watersheds for

(a) runoff and (b) dissolved organic (DOC) loss for all years...... 25

Figure 3.1 Greenley Memorial Research Center study site in Knox County, Missouri

(star). Black dots indicate sampling locations. Thick black lines are grass VFS,

grey lines are agroforestry VFS and dotted lines represent soil series boundaries.

...... 39

xvi

Figure 3.2 Scatter plot and regression relationship for regression of soil organic carbon

(SOC) concentration on bulk density (BD) using all subsamples (n=36) in the 0–5

cm surface layer. Note the slope parameter (SOCc) was statistically significant

(p< 0.001)...... 49

Figure 4.1 Greenley Memorial Research Center study site in Knox County, Missouri

(star). Black dots indicate sampling locations. Thick black lines are grass VFS,

grey lines are agroforestry VFS and dotted lines represent soil series boundaries

(adapted from Veum et al., 2011)...... 65

Figure 4.2 Scatter plots and regression relationships for air-dried water stable aggregates

-2 -2 (WSAAD, kg m ) and (a) total soil organic carbon (SOC, kg m ), (b) particulate,

adsorbed and occluded carbon (PAO-C, kg m-2), (c) water-extractable organic

carbon (WEOC, g m-2) and (d) water extractable organic carbon from the PAO

-2 -2 fraction (WEOCP, g m ) in the surface layer (0-70 kg soil m ) using all

subsamples (n=36). Sand-free values are presented for PAO-C and WEOCP.

Grass vegetated filter strips (VFS) are represented by black circles, agroforestry

VFS by open circles and no-till by grey triangles. The slope parameters are

statistically significant (p < 0.001)...... 76

Figure 5.1 Greenley Memorial Research Center study site in Knox County, Missouri

(star). Black dots indicate sampling locations. Thick black lines denote grass

vegetative filter strips (VFS), grey lines denote agroforestry VFS and dotted lines

represent soil series boundaries (adapted from Veum et al., 2011)...... 92

Figure 5.2 Scatterplots and regression relationships for surface soil samples by

conservation management practice for water-extractable organic carbon (WEOC)

xvii

with (a) absorbance at 254 nm (A254), (c) specific ultra-violet absorbance at 254

nm (SUVA, L cm-1 m-2) from unfractionated soil and from the particulate,

adsorbed and occluded fraction (b and d, respectively, subscript P). Grass

vegetated filter strips (VFS) are represented by black circles, agroforestry VFS by

open circles and no-till by grey triangles. Slope parameters are statistically

significant (p< 0.05)...... 105

Figure 5.3 Scatterplots and regression relationships for surface soil samples by

conservation management practice for water-extractable organic carbon (WEOC)

with (a) the ratio of ultra-violet absorbance at 280 nm to 365 nm (E2:E3) and (c)

the ratio of ultraviolet absorbance at 253 to 203 nm (ET:BZ) from unfractionated

soil and from the particulate, adsorbed and occluded fraction (b and d,

respectively, subscript P). Grass vegetated filter strips (VFS) are represented by

black circles, agroforestry VFS by open circles and no-till by grey triangles. Slope

parameters are statistically significant (p< 0.05)...... 106

Figure 6.1 Greenley Memorial Research Center study site in Knox County, Missouri

(star). Black dots indicate sampling locations. Thick black lines denote grass

vegetative filter strips (VFS), grey lines denote agroforestry VFS and dotted lines

represent soil series boundaries (adapted from Veum et al. 2011)...... 124

Figure 6.2 Examples of solid-state, cross-polarization, magic angle spinning 13C nuclear

magnetic resonance (CPMAS 13C NMR) spectra of soil organic matter from grass

vegetative filter strip (VFS), agroforestry VFS and no-till soil...... 137

xviii

Figure 6.3 Examples of solid-state, cross-polarization, magic angle spinning 13C nuclear

magnetic resonance (CPMAS 13C NMR) spectra of humic acids extracted from

grass vegetative filter strip (VFS), agroforestry VFS and no-till soil...... 141

Figure 6.4 Examples of diffuse reflectance infrared Fourier transform infrared (DRIFT)

spectra of soil organic matter extracted from grass vegetative filter strip (VFS),

agroforestry VFS and no-till soil...... 145

Figure 6.5 Examples of diffuse reflectance infrared Fourier transform (DRIFT) spectra of

humic acid from grass vegetative filter strip (VFS), agroforestry VFS and no-till

soil...... 148

xix

CHAPTER 1

INTRODUCTION

In agroecosystems, soil organic matter (SOM) plays several important roles in the biogeochemistry of soil and contributes to vital soil functions such as sustained biological activity, nutrient cycling, crop productivity and environmental quality (Doran and Parkin,

1994; Karlen et al., 1997). Agricultural practices have been affecting the landscape for more than 5,000 years and have been implicated in large-scale changes in the SOM pool

(Paul et al., 1997). Cultivation generally results in 20 – 60% bulk SOM loss in the surface layers of soil (Buyanovsky and Wagner, 1998; Guo and Gifford, 2002; West and Post,

2002); however, the rate of loss varies with management practices that alter the chemical and physical processes of transport through soils (Chantigny, 2003). Conservation management practices demonstrate a wide range of environmental benefits (Lal et al.,

1994; Holland, 2004), and have been shown to improve many aspects of SOM quantity and quality over conventional tillage (Angers et al., 1993a; Collins et al., 2000). No-till farming, which allows the soil to retain more SOM than conventional tillage practices, has become increasingly popular; greater than 35% of U.S. cropland was under no-till operations by 2009 (Horowitz et al., 2010). Vegetative filter strips (VFS) are bands of vegetation, often grasses, planted along the contour to reduce runoff and losses of sediments and nutrients (Dillaha et al., 1986). Agroforestry combines trees and/or shrubs with crops and/or livestock, and is most popular in tropical and subtropical regions. In the

U.S., agroforestry began emerging as a management practice in the late 1960s for a

variety of potential economic and environmental benefits (Rietveld and Irwin, 1996).

More studies are needed, however, to elucidate the role of management practices such as agroforestry systems on SOM cycling (Morgan et al., 2010).

Soil organic matter quantity and quality is affected by the balance between organic inputs to the soil and the rate of SOM losses through decomposition or runoff/erosional losses. Effective SOM quality indicators are biological, chemical or physical measurements that are relevant to soil functions of interest (Karlen et al., 1997;

Schoenholtz et al., 2000). Soil organic matter is often quantified by measuring soil organic carbon (SOC); however, SOC is often insensitive to short-term changes (Lal,

2004) or subtle differences among soil conservation management practices (Veum et al.,

2011). Effective SOM quality indicators capable of detecting changes under conservation management practices are essential to furthering the understanding of these practices and their relative benefits.

The primary objective of this study was to use chemical, biological and physical indicators to characterize changes in SOM quantity and quality as a result of conservation management practices currently practiced in Missouri and other Midwestern states. This research study utilized surface water runoff and soil samples collected from three small subcatchments that were no-till planted to a corn-soybean rotation with contour VFS installed in two of the subcatchments. The parent materials for the soils at the study sites are loess and loamy sediments derived from pre-Illinoian glacial till (Unklesbay and

Vineyard, 1992). Additionally, the soils are characterized as having a claypan, an abrupt argillic sub-soil horizon imparting vertic properties to the soil (Blanco-Canqui et al.,

2002). The conservation management practices (i.e., treatments) represented by these

sites include (1) no-till, (2) upland, contour, gras s VFS and (3) upland, contour, agroforestry VFS. The effects of landscape position (i.e., summit, shoulder, backslope and footslope) were also evaluated.

Chapter 2 compares the effects of installing grass and agroforestry VFS in the no-till experimental catchment on runoff and losses of dissolved organic carbon over a ten year period using a paired watershed approach. Chapter 3 develops a modified ANOVA to evaluate stocks of soil organic carbon (SOC) and aggregate-associated, particulate, adsorbed and occluded (PAO) carbon among treatments and examines the effects of calculating carbon pools on a gravimetric, soil volume or soil mass basis. Chapter 4 compares treatment effects on aggregate stability and water-extractable carbon from unfractionated soil and the PAO fraction, and explores relationships among these and other soil properties. Chapter 5 examines biological indicators of SOM quality by measuring microbial enzyme activities, the labile KMnO4-oxidizable carbon content, and comparing the chemistry of water-extractable SOM using ultraviolet spectrophotometry.

Chapter 6 utilizes nuclear magnetic resonance (NMR) spectroscopy and diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy to compare the composition and relative extent of degradation of SOM and humic acid from grass and agroforestry

VFS and no-till soil. Together, these chapters elucidate the effects of grass and agroforestry VFS in a no-till agroecosystem on portions of the terrestrial carbon cycle and SOM quality using indicators that are more sensitive and more informative than SOC measurements alone.

CHAPTER 2

RUNOFF AND DISSOLVED ORGANIC CARBON LOSS FROM A PAIRED-WATERSHED STUDY OF THREE ADJACENT AGRICULTURAL WATERSHEDS

A version of this chapter has been published: Veum, K.S., Goyne, K.W., Motavalli, P.P., Udawatta, R.P. 2009. Runoff and dissolved organic carbon loss from a paired-watershed study of three adjacent agricultural watersheds. Agric., Ecosyst. Environ. 130 (3-4), 115- 122.

Abstract

Organic matter plays several important roles in the biogeochemistry of terrestrial and aquatic ecosystems including the mobilization and transport of nutrients and pollutants.

Cropping, tillage practices and vegetative buffer strip installation affect losses of dissolved organic carbon (DOC). While many studies show reductions in pollutant export from agroecosystems where vegetative buffers have been implemented, buffer strips may be a source of DOC and contribute to surface water pollution. Using a paired-watershed approach, the objectives of this study were to determine the effect of grass and agroforestry buffers on runoff and DOC loss, compare runoff and DOC losses between the growing and fallow seasons, and investigate crop effects on runoff and DOC losses.

The study design consisted of three small agricultural watersheds in a no-till, maize- soybean rotation located in the claypan region of northeast Missouri, USA; one watershed was planted with grass buffer strips, one with agroforestry buffer strips, and one unaltered watershed served as the control. Runoff and DOC loss were measured

during a six-year calibration period (1991 – 1997) prior to buffer installation and for a nine-year treatment period (1997 – 2006). The grass buffer strips significantly decreased runoff by 8.4% (p = 0.015) during the treatment period while the agroforestry buffer system exhibited no significant change in runoff (p = 0.207). Loss of DOC was not significantly affected by grass or agroforestry buffer installation (p = 0.535 and p =

0.246, respectively). Additionally, no significant difference in runoff or DOC loss was found between crops (maize and soybean) or between seasons (growing and fallow).

Overall, this study indicates that grass buffer systems are effective at reducing runoff and that DOC contamination of surface waters is not exacerbated by either type of vegetative buffer strip.

Introduction

Soil organic matter is a determinant of soil quality and an important factor in agricultural productivity (Doran et al., 1998). Dissolved organic matter, measured as dissolved organic carbon (DOC), is a dynamic, mobile and highly reactive fraction of soil organic matter. Loss of DOC from agricultural landscapes occurs at the expense of both soil and water quality. Many studies have linked agricultural land-use to degraded water quality including sediment, nutrient and organic loading (Jones et al., 2004; Jones and

Knowlton, 2005; Monaghan et al., 2007; Valentin et al., 2008). Additionally, DOC may enhance the sorption and mobility of pesticides (Williams, et al., 2000; Flores-Cespedes et al., 2002; Li et al., 2005) and heavy metals (Li and Shuman, 1997) and lead to drinking water quality problems (Pomes et al, 1999; Aitkenhead-Peterson et al., 2003)

Concentrations of DOC vary widely in soils and range from 1 –70 mg L-1 (Zsolnay,

1996). Headwater catchments exhibit strong links between terrestrial and aquatic ecosystems and studies in North America and Europe find that DOC losses typically range from 10 – 100 kg ha−1 yr−1 with an average of 4.7 kg ha−1 yr−1. The highest DOC losses (484 kg ha−1 yr−1) are found in boreal forest and wetland ecosystems; temperate forests lose between 3.4 and 417 kg ha−1 yr−1 and grasslands export only 1.6 – 5.0 kg ha−1 yr−1 (Hope et al., 2004). In the Midwestern USA, Dalzell et al. (2007) estimated agricultural DOC losses in the range of 14.1 – 19.5 kg ha-1 yr-1, and these values are comparable to DOC export estimates (3 – 23 kg ha-1 yr-1) made by Royer and David

(2005).

Soil organic matter and DOC are lost under conversion to crop land, but the rate of loss varies with management practices that alter the chemical and physical processes of transport through and over soils (Unger, 1968; Chantigny, 2003). While the effects of conservation tillage (CT) vary according to a wide range of environmental variables, CT practices are increasingly popular throughout the USA (Lal et al., 1994) and globally

(Holland, 2004). In general, no-till may significantly reduce soil erosion (Ghidey and

Alberts, 1998), lower runoff (Johnson et al., 1984) and retain organic carbon (Unger,

1968; Bertol et al., 2007). Vegetative buffer strips, another conservation management practice, are routinely employed to reduce losses from agroecosystems. In the Missouri claypan region, grass and agroforestry buffer strips have been shown to reduce runoff and pollutant transport (Blanco-Canqui et al., 2004; Udawatta et al., 2006). Buffer strips, although shown to have high sediment retention capability, are potential sources of DOC

(Tate et al., 2004) and may contribute to DOC loss and exacerbate surface water pollution.

Crops and crop residues are also a potential source of nutrients and carbon, and may contribute to DOC loss (Lal et al., 1994; Bertol et al., 2007). The amount and composition of plant DOC exudates varies among crop species (Zsolnay, 1996;

Chantigny, 2003). Maize and soybeans, the principal field crops in Missouri, USA, are commonly grown in rotation and losses can be problematic under both crops (Lal et al.,

1994). Information on runoff and DOC loss from specific cropping systems is inconsistent and inconclusive, and these contradictory findings may be attributable to tillage practices or other management dissimilarities (Alberts et al., 1985; Ghidey and

Alberts, 1994; Moncrief et al., 2005). Overall, multiple interactive factors influence runoff and DOC loss; therefore, elucidating the role of buffer strips in agroecosystems is essential.

Paired catchment studies with calibration and treatment periods have been widely used to evaluate the effects of land-use changes on terrestrial inputs to aquatic systems

(Watson et al., 2001). Using the paired-watershed approach, residuals of the predicted and post-treatment relationship are assumed to be a measure of the treatment effect.

Therefore, both sufficient calibration data and sufficient post-treatment data are necessary to establish strong relationships between the treatment watershed(s) and the control watershed in order to detect treatment effects. While many paired-watershed studies, particularly in forested catchments, have been conducted over the past several decades, agricultural watersheds are underrepresented and results are often contradictory or fragmented (Hope et al., 1994; Valentin et al., 2008). Many studies do not include a

control watershed or a sufficient calibration period, thereby failing to account for watershed or climatic variability. Inferences drawn from such studies are questionable; thus, there is a need for paired-watershed studies of agroecosystems utilizing well- calibrated watersheds and an unaltered control or reference watershed to fully evaluate the effects of management practices or land-use changes.

Understanding the effects of buffer strips, crop and season on runoff and DOC loss in claypan soils will help clarify the role of agroecosystems with restrictive sub-surface horizons in the hydrologic and carbon cycles. Although claypan soils are not extensive globally, soils with restrictive sub-horizons cover approximately 2.9 million km2, potentially having a large impact on biogeochemical cycles (USDA-NRCS, 1998). Using a paired-watershed approach, the objectives of this study were to (1) determine the effect of grass and agroforestry buffers on runoff and DOC loss, (2) compare losses between the growing and fallow seasons, and (3) compare losses between maize and soybean crops.

Methods and Materials

Site description and management

The Greenley Memorial Research Center study area is located in Knox County,

Missouri, USA (40˚ 01’N, 92˚11’W, Fig. 2.1). It consists of three small, contiguous, agricultural watersheds under no-till management planted to a rotation of maize (Zea mays L.) and soybean [Glycine max (L.) Merr.]. Crop residues are retained at the surface after each harvest in the fall. The three watersheds have north facing aspects with gentle slopes from 0 to 3% and the catchments are designated as East (1.65 ha), Center (4.44 ha) and West (3.16 ha). Watershed characteristics are summarized in Table 2.1 after

Figure 2.1 Location of the Greenley Memorial Research Center in Missouri, USA and the study watersheds. Location of flumes and autosamplers (triangles), elevation contour lines (meters above sea level; black lines), grassed waterways (grey areas, perpendicular to contours) and buffer strips

(grey lines following contours) are noted. Adapted from Udawatta et al. (2004).

Table 2.1 Characteristics of the Control, Agroforestry (AGF) and Grass watersheds (adapted from

Udawatta et al., 2006).

Watershed

Property East (Control) Center (AGF) West (Grass)

Area (ha) 1.65 4.44 3.16

Length (m) 234 425 383

% Slope 2.1 1.3 0.9

Grass waterway (m) 102 151 104

Soil organic carbon (g kg-1)a 26.0 22.3 19.3

Soil clay content (%)a 23.4 23.4 26.2

Soil pHa 6.7 6.8 7.3 a Ap horizon.

Udawatta et al. (2006). Each watershed was instrumented with an H-flume, a flow measuring device and an autosampler in 1991. The watershed system drains to the north and discharges into the Little Fabius River.

In 1997, five grass buffer strips were installed in the West (Grass) watershed and six agroforestry buffer strips were installed in the Center (agroforestry; AGF) watershed.

Each buffer strip is approximately 4.5 m wide and planted on the contour. Overall, the buffer strips account for approximately 8 and 10% of the land area in each treatment watershed. The grass buffers consist of redtop (Agrostis gigantea Roth), brome (Bromus spp.), and birdsfoot trefoil (Lotus maizeiculatus L.). In addition to the grasses, pin oak

(Quercus palustris Muenchh.), swamp white oak (Q. bicolor Willd.) and bur oak (Q. macrocarpa Michx.) were planted in the agroforestry buffers in an alternating sequence.

The East (Control) watershed was maintained as a control throughout the study period.

Grassed waterways of variable width (7 – 10 m) extend approximately 100 m up from the flume in each watershed (Fig. 2.1). Installation of grassed waterways along the thalweg is a common management practice in the Midwest, USA to avoid gully erosion. They were included in all watersheds throughout the study period to avoid confounding the effects of the contour buffer strips. The grassed waterways are perpendicular to the contour and can be distinguished from the contour buffer strips in Figure 2.1.

The study site is located in USDA Major Land Resource Area 113, the Central

Claypan Area (USDA-NRCS, 2006), and it is characterized by an abrupt, argillic sub-soil horizon imparting vertic properties to the soil. Claypan soils, well-known for their shrink- swell and cracking behavior (Baer and Anderson, 1997), experience complex runoff and infiltration phenomena (Arnold et al., 2005) and create poorly drained soils with perched

water tables dominated by lateral, subsurface flow (Jamison and Peters, 1967; Blanco-

Canqui et al., 2002). The claypan horizon in the experimental watershed occurs from 20 –

62 cm below the surface (Udawatta et al., 2006). Soils in the study area are classified by the USDA-NRCS as Putnam silt loam (fine, smectitic, mesic Vertic Albaqualfs),

Kilwinning silt loam (fine, smectitic, mesic Vertic Epiaqualfs) and Armstrong loam (fine, smectitic, mesic, Aquertic Hapludalfs) or as Vertic Luvisols and Gleyic Luvisols under the Food and Agricultural Organization (FAO) soil classification system. The 30-year, average annual precipitation for the Knox County (Novelty, MO) weather station is 95.9 cm. For the duration of the study, annual precipitation averaged 96.2 cm and ranged from a low of 77.4 cm in 2005 up to 139 cm in 1993. Agricultural activities and precipitation data for the study area are summarized in Table 2.2.

Sample collection and analysis

Runoff samples and discharge data were collected with automatic, flow-weighted composite samplers from 1993 to 2006 from each watershed. Stevens Type F water level recorders (Stevens Water Resources, Beaverton, OR) were installed in 1991 and used year-round until Isco bubbler-flow measuring devices (Teledyne Isco, Lincoln, NE) were installed in 1995. To avoid damage from freezing, the Isco samplers were removed in late fall and reinstalled in early spring of each year. Typically, the sample collection period extended from March through November. Runoff was calculated as a flow-weighted, total event volume. The autosampler, activated by changes in stage, collected 135 ml subsamples for each 25 m3 of runoff. The subsamples were composited after each event.

Many small runoff events did not have sufficient flow to activate the autosampler, and on

Table 2.2 Summary of precipitation and agricultural activities from 1991 to 2006.

Total Planting Tillage Planting Harvest Year Crop PPTa (cm) Date Method Method Date

1991 101.5 maize late May no-till straight rowb late Oct.

1992 97.7 soybean early June cultivator straight row early Nov.

1993 130.4 maize early June no-till straight row early Nov.

1994 83.3 soybean mid-May no-till straight row mid-Oct.

1995 106.1 soybean late June no-till straight row early Nov.

1996 88.8 maize late April no-till contour early Nov.

1997 92.1 soybean midearly-May June no-till contour mid-Oct.

1998 139.0 maize mid-May no-till contour mid-Oct.

1999 77.8 soybean early June no-till contour mid-Oct.

2000 78.7 maize mid-April cultivator contour mid-Sept.

2001 104.3 soybean mid-May no-till contour late Oct.

2002 85.9 maize early June no-till contour late Oct.

2003 102.8 soybean mid-June no-till contour early Oct.

2004 93.5 maize early May no-till contour mid- Nov.

2005 77.4 soybean early May no-till contour mid-Oct.

2006 79.3 maize mid-April no-till contour late Sept. a PPT (precipitation) does not include all snowfall. b Crops planted perpendicular to slope.

occasion, consecutive events were integrated into one event. Composited samples were frozen until analysis. Precipitation data from the Novelty, MO weather station was provided by the University of Missouri Extension Commercial Agriculture Automated

Weather Station Network. From 1991 through 1994, a National Weather Service

Standard 8-inch manual observation rain gauge (NOAA-NWS, North Platte, NE, USA) was used to record precipitation. In September 1994, an automated Texas Electronics 525

Tipping Bucket weather station (Campbell Sci., Logan, UT, USA) was installed. The automated gauge does not measure snow or ice events and will measure frozen precipitation only when it begins to melt and drip through the funnel. Underestimation also occurs during extreme events when rainfall rates exceed 5 cm hr-1. Therefore, precipitation totals are likely underestimated by approximately 10% annually (pers. comm. Dr. P. Guinan).

Dissolved organic carbon analysis was performed using a wet oxidation – persulfate method on a Tekmar-Dohrmann Phoenix 8000 TOC Analyzer with a nondispersive infrared (NDIR) detector. Frozen, composited runoff samples were thawed and filtered through pre-rinsed, 0.45 µm nominal pore-diameter Whatman membrane filters to remove particulates prior to analysis. Filtrates were acidified with concentrated reagent grade phosphoric acid to pH < 2 and refrigerated at 4˚C for no more than 14 days prior to analysis. Samples were analyzed in triplicate. Standards ranging from 0.0 to 20 mg L-1 were prepared with potassium acid phthalate in ultrapure water. Calibration standards, external standards, field blanks and field duplicates were analyzed with each batch for quality control purposes. Due to small sample volumes and prior utilization of samples for other analyses, only 36% of the runoff events had corresponding DOC data.

Composite DOC concentrations were used to generate event mean concentration data for each runoff event.

Statistical analysis

All statistical analyses were performed with SASTM Statistical Software Version 9.1

(SAS Inst., 2002). The significance criterion was established at p < 0.05. For the paired- watershed analysis, calibration and post-treatment regression relationships were established between each treatment watershed and the control watershed using Proc REG.

If the relationships were strong and amenable to regression analysis, the data were subsequently compared using Proc GLM. All regression relationships for runoff and

DOC loss were significant at p < 0.0005 or better and had strong r2 values ranging from

0.79 to 0.99. When a significant difference was observed between the calibration and treatment period for runoff and DOC loss, the residuals between the predicted relationship and the post-treatment relationship served as the measure of treatment effect.

Estimates of total (9-mo) DOC losses were calculated using a log-log relationship between measured runoff and DOC loss. These estimated DOC values were not used to evaluate treatment effects in the paired-watershed analyses. For crop comparisons, data were aggregated by growing crop, either maize or soybean, extending from planting in the spring through the fallow winter period until the subsequent spring crop was planted.

Seasons were similarly delineated by planting and harvest dates, with the growing season beginning at planting in the spring and ending at harvest in the late summer or fall. The fallow season was defined as the period between harvest and the subsequent planting date.

Results and Discussion

Precipitation, runoff and DOC loss

Precipitation, runoff and DOC loss are linked through watershed hydrology and covary over time. A time-series plot of measured precipitation, runoff and DOC loss for

1993, the highest precipitation year during the study period, illustrates the relationships among these variables at our study site (Fig. 2.2). Runoff and DOC loss values visually correspond with precipitation peaks better in early spring than in late summer or fall. This seasonal phenomenon illustrates the strong temporal influence of soil and climatic factors, such as antecedent soil moisture, on runoff from claypan soils. As summer progresses and precipitation declines, claypan soils shrink to form large, characteristic cracks allowing preferential infiltration of water and dramatic reductions in runoff (Baer and Anderson, 1997; Kazemi et al., 2008). Runoff typically exhibits a log-normal distribution; therefore, large events, although infrequent, contribute an inordinate proportion of the total runoff volume. During the entire study period, 19% of the runoff events exceeded a magnitude of 250 m3 ha-1and these events represented 68% of the total measured runoff; only 5% of the runoff events exceeded 500 m3 ha-1, yet these events represented 35% of the total measured runoff. The study site was plagued by drought for most of the treatment period. Therefore, the treatment period yielded 44% less total measured runoff (74 events) than the calibration period (124 events), even though it was two years longer. On average, the treatment period experienced 56% less total measured runoff per year. This resulted in a small and skewed distribution of runoff data for the treatment period (Table 2.3). The maximum 9-mo cumulative runoff occurred in 1993, at

5425 m3 ha-1 and a minimum of 36 m3 ha-1 occurred in 1991 (Table 2.3; Fig. 2.3).

Figure 2.2 Precipitation (PPT), runoff and dissolved organic carbon (DOC) loss during the year 1993.

Table 2.3 Summary of measured runoff and estimated dissolved organic carbon (DOC) loss.

n Mean Median Maximum Minimum Runoff All Years 198 1795 1403 5425 36

(m3 ha-1 yr-1) Calibration 124 2637 2502 5425 420

Treatment 74 1141 521 5225 36

DOC All Years 72 16.6 12.9 50.0 0.4

(kg ha-1 yr-1) Calibration 44 24.3 23.2 50.0 4.1

Treatment 28 10.6 5.0 47.7 0.4

Figure 2.3 Total measured runoff for all watersheds by year. Dashed line indicates the year (1997) when buffers were installed.

Due to the log-normal distribution of runoff, a standard log-transformation was applied and an empirical relationship between runoff and DOC loss was established using

178 samples representing 72 runoff events (Fig. 2.4). This empirical runoff-DOC relationship was used to estimate 9-mo DOC losses for the study period, ranging from 0.4 kg ha-1during the drought year of 2006, up to 50 kg ha-1 in 1993, with an average of 16.6 kg ha-1 across all years (Table 2.3). This estimate is within the range of DOC losses reported by Dalzell et al. (2007) and Royer and David (2005) for Midwestern agricultural watersheds.

Buffer effects on runoff and DOC loss

The effects of each buffer strip type on runoff and DOC loss were evaluated using the paired-watershed method (Figs. 2.5a-b and 2.6a-b). A significant, 8.4% reduction (p =

0.015; Table 2.4) in runoff was observed during the treatment period in the Grass watershed. DOC loss was not found to be significantly different between treatment and calibration periods for either the Grass or AGF watersheds (Table 2.4). Trends based on runoff event size and DOC loss were also investigated for treatment effects. Runoff events were divided into size classes using cut-offs of 100 or 250 m3 ha-1, but reasonable regression relationships could not be established (r2 < 0.3) without regressing against the entire range of runoff values.

Overall, the Grass watershed exhibited higher runoff than the AGF watershed, while the reverse was true for DOC loss (Figs. 2.7a-b). In general, DOC loss correlates positively with runoff volume, and it should follow that the Grass watershed would

Figure 2.4 Relationship between log-transformed runoff and dissolved organic carbon (DOC) loss for all samples from all years.

Figure 2.5 Paired-watershed relationships between control and treatment watersheds for runoff by buffer type (a-b), crop (c-d) and season (e-f).

Figure 2.6 Paired-watershed relationships between control and treatment watersheds for dissolved organic carbon (DOC) loss by buffer type (a-b), crop (c-d) and season (e-f).

Table 2.4 Analysis of buffer type, crop and seasonal effects on runoff and dissolved organic carbon

(DOC) loss.

Runoff DOC Loss

Effect Watershed n p n p

Buffer West (Grass) 198 0.015 54 0.535

Center (AGF) 198 0.207 50 0.246

Crop West (Grass) 198 0.121 54 0.554

Center (AGF) 198 0.542 50 0.881

Season West (Grass) 198 0.270 50 0.378

Center (AGF) 198 0.869 54 0.912

Figure 2.7 Paired-watershed relationships between control and treatment watersheds for (a) runoff and (b) dissolved organic (DOC) loss for all years.

experience higher areal DOC loss in response to higher areal runoff. These results indicate that other soil chemical and physical properties may be influencing runoff and

DOC loss. Higher DOC loss from the AGF watershed may be due to the apparent higher soil organic carbon levels within the buffer strips relative to the grass buffers or the cropped soil (Table 2.1). Studies show that as soil organic carbon increases, a concomitant increase in DOC is generally observed (Kalbitz et al., 2000). With respect to runoff and soil hydrologic properties, a previous study conducted on soils from this site found that saturated hydraulic conductivity was greatest in the AGF buffers relative to the grass and cropped soil (Seobi et al., 2005). More recent work (Anderson et al., 2008) also found initial infiltration was enhanced within AGF buffer strips relative to the cropped areas. Steady-state infiltration rates and saturated hydraulic conductivity, however, were not significantly different among treatments.

Although lower soil water content may cause higher initial infiltration rates in the

AGF buffer strips, the presence of the claypan horizon appears to dominate the overall hydrology of the system by dramatically reducing hydraulic conductivity through the profile. The claypan horizon coupled with the low areal percentage of buffer strip coverage (8-10%) and immaturity of the AGF trees, may explain the lack of detectable differences in runoff or DOC loss at the field scale in this watershed.

In addition, antecedent soil moisture conditions, storm intensity and rainwater pH can induce chemical and physical non-equilibrium within soil and affect DOC concentrations in runoff (Jardine et al., 1990). Claypan soils similar to our study site have been shown to experience increased lateral subsurface flow known as interflow (Jamison and Peters,

1967; Blanco-Canqui et al., 2002). During periods of high antecedent soil moisture

content, subsurface flow is increased due to swelling of 2:1 layer silicate clays present in the argillic sub-horizon. In contrast, desiccation of the claypan leads to the formation of large cracks that permit preferential flow of water and solutes vertically into the soil profile (Baer and Anderson, 1997; Kazemi et al., 2008). These factors all contribute to the high variability in DOC concentrations observed for a given runoff volume, and may have reduced the power to detect treatment effects in this study.

The lack of significant buffer effects on runoff or DOC loss may also be attributed to data limitations. Our study used total runoff volumes and event mean DOC concentrations; however, it has been shown that runoff concentrations are related to rainfall intensity, duration and frequency. Large storms have more consistent DOC concentrations while small events exhibit dramatically higher DOC concentrations during the rising limb of the hydrograph and peak DOC concentrations are often decoupled from peak flows (Jardine et al., 1990). Royer et al. (2007) found a wide range in grab sample concentrations of DOC in runoff under maize crop, ranging from 2.4 – 205.4 mg L-1, illustrating the highly variable DOC content among and within storm events. Overall, these results indicate that grass buffer strips are an effective conservation practice for reducing runoff and neither buffer type contributes to DOC contamination of surface waters.

Crop effects on runoff and DOC loss

The effects of crop type on runoff and DOC loss were evaluated using the paired- watershed method, and no significant differences were found between maize and soybean

(Figs. 2.5c-d and 2.6c-d; Table 2.4). The lack of significant differences may be explained

by the influence of other management practices. The agroecosystem was in a no-till, maize-soybean rotation throughout the study. Crop rotations mask crop effects, and prior crop effects in maize-soybean rotations have been noted since the 1950s (Van Doren et al., 1950). Maize and soybean produce large amounts of residue providing a high degree of soil coverage during the fallow season and the early growing stages. Typical maize yields produce more than double the residue of soybean (Larson et al., 1978; Alberts et al., 1985). While soybean residues have a higher initial decay rate, maize and soybean residues have similar carbon content (38%) and a large proportion (> 70%) of both may remain on the soil surface after over-wintering (Ghidey et al., 1985). In laboratory studies, Kladivko (1994) found that the effects of maize and soybean residues on soil physical properties are subtle and are not sensitive enough to differentiate crop effects. In addition, Stearman et al. (1989) observed no differences in organic carbon content within plots planted to maize or soybean for seven years, although differences in organic carbon composition were noted.

Seasonal effects on runoff and DOC loss

No detectable differences between the growing and fallow seasons in runoff and DOC loss were observed (Figs. 2.5e-f and 2.6e-f; Table 2.4). Seasonal fluctuations in the organic carbon content of soils have been noted (Hope et al., 1994) and dramatic changes in soil physical, chemical and biological processes occur during the fallow season versus the growing season (Wang and Bettany, 1993; Boyer et al., 1996; Matzner and Borken,

2008). We defined our seasons as growing and fallow with the growing season extending from planting, usually in March or April, through harvest in August or September. The

fallow season extended from harvest through winter until the next crop rotation was planted, therefore, season lengths varied from year to year due to weather and crop conditions. We characterized DOC loss and runoff for the 9-mo period from March through November for which data were measured.

The lack of detectable seasonal differences in this study may be attributable to several factors. No-till systems, with high surface concentrations of crop residues and organic carbon, are known to buffer seasonal changes in soil processes (Lal et al., 1994), and the presence of grassed waterways also may have buffered the treatment effects. Studies show a dramatic reduction in runoff and erosion when grassed waterways are present

(Fiener and Auerswald, 2003). Overall, seasonal effects may have been diminished by other environmental and management factors. As an additional exploratory measure, direct comparisons were made between crops (maize and soybean), seasons (growing and fallow) and buffer treatments. Areal-weighted event means were used for runoff and

DOC loss from the entire study period, without employing the paired-watershed approach, and no significant differences were observed (data not presented).

Conclusions

A long-term, experimental record covering 16 years of runoff and DOC loss data was used to examine the effects of buffer strip installation (grass and agroforestry), crop rotation and season in an agroecosystem within the claypan region of northeastern

Missouri, USA. A paired-watershed approach was employed, comparing seven years of calibration period data and nine years of treatment data. A significant reduction in runoff

(8.4%) occurred as a result of the grass buffer installation, while no other effects were

detectable. Overall, our study indicates that grass buffers provide a significant reduction in runoff and that neither grass nor agroforestry buffers contribute to DOC contamination of surface waters. Thus, the benefits of vegetative buffer strips for mitigating runoff and pollutant export are not diminished by increased loss of DOC to surface water resources.

CHAPTER 3

ASSESSMENT OF SOIL ORGANIC CARBON AND TOTAL NITROGEN UNDER CONSERVATION MANAGEMENT PRACTICES IN THE CENTRAL CLAYPAN REGION, MISSOURI, USA.

A version of this chapter has been published: Veum, K.S., Goyne, K.W., Holan, S.H., Motavalli, P.P. 2011. Assessment of soil organic carbon and total nitrogen under conservation management practices in the Central Claypan Region, Missouri, USA. Geoderma 167/168, 188-196.

Abstract

Conservation management practices including upland vegetative filter strips (VFS) and no-till cultivation have the potential to enhance soil carbon sequestration and other ecosystem services in agroecosystems. A modified two-factor analysis of variance

(ANOVA) with subsamples was used to compare soil organic carbon (SOC) and total nitrogen (TN) on a concentration, soil volume and soil mass basis in claypan soils planted to different conservation management practices and as a function of landscape position.

The three conservation management practices (no-till cultivation, grass VFS and agroforestry VFS) and four landscape positions (summit, shoulder, backslope and footslope) investigated were compared ten years after VFS establishment in a no-till system planted to maize (Zea mays.L.) – soybean (Glycine max (L.) Merr.) rotation. Two soil depth increments (0–5 cm and 5–13 cm) were modeled separately to test for treatment effects. In the surface layer, mean SOC concentration was significantly greater in the VFS soils compared to no-till. On a soil volume or mass basis, no significant

differences in SOC stocks were found among treatment means. Concentration and mass based TN values were significantly greater in the grass VFS relative to no-till in the surface layer. A rapid slaking stability test, developed to separate particulate, adsorbed and occluded organic carbon (PAO-C) and nitrogen (PAO-N), showed that VFS soils had significantly greater mean PAO-C and PAO-N concentrations, soil volume and soil mass based stocks than no-till. In addition, comparison of SOC:TN and PAO-C:PAO-N ratios suggest reduced decomposition and mineralization of SOC in the PAO fraction. No significant treatment effects were detected in total or PAO soil fractions in the subsurface layer or among landscape position in either depth increment. Study results emphasize the need to compare soil carbon and nitrogen stocks on a soil volume and/or soil mass basis using bulk density measurements. Additionally, the rapid PAO separation technique was found to be a good indicator of early changes in SOC and TN in the systems studied.

Overall, this research indicates that grass VFS may sequester TN more rapidly than agroforestry VFS and that a greater proportion of SOC and TN may be stabilized in VFS soils compared to no-till.

Introduction

Soil organic matter, typically measured as soil organic carbon (SOC), has many different critical biogeochemical functions in the global carbon cycle. For example, SOC is an important factor in soil fertility and agricultural productivity (Doran et al., 1998).

Additionally, SOC affects the degradation, sorption and mobility of agricultural pollutants (Li et al., 2005; Chu et al., 2010) and heavy metals (Li and Shuman, 1997) and it is a potential source or sink of atmospheric carbon dioxide (Wander and Nissen, 2004).

Many environmental and anthropogenic factors interact with geomorphic processes to affect the retention, distribution and transformation of SOC and total nitrogen (TN) in soil. Agricultural practices (e.g., conventional tillage) have been implicated in SOC and

TN losses (Mann, 1986; Helfrich et al., 2006; Yimer et al., 2007; Fu et al., 2010). In contrast, conservation management practices, such as no-till cultivation and vegetative filter strips (VFS), have the potential to mitigate the effects of conventional cultivation.

Conservation management practices also provide several ecosystem services including carbon sequestration in soil and biomass (Tufekcioglu et al., 2003; Chen et al., 2004;

Puget and Lal, 2005), reduced runoff and sediment loss (Beke, 1989; Veum et al., 2009), pollutant reduction or sorption (Barling and Moore, 1994; Chu et al., 2010) and increased wildlife diversity (Semlitsch and Jensen, 2001). From 1982 to 1997, agricultural soil erosion in the USA declined by over a billion tons per year, and 25% of this decrease has been attributed to conservation management efforts (Wiebe and Gollehon, 2006).

In addition to affecting total SOC and TN concentration, agricultural cultivation may impact the distribution of SOC and TN vertically in the soil profile. In particular, reduced tillage systems allow for accumulation of organic matter at the surface while conventional tillage incorporates organic matter below the surface and homogenizes soil in the plow layer (Wander et al., 1998; Lal et al., 2007; Angers and Eriksen-Hamel, 2008;

Dou et al., 2008; Tatzber et al., 2008), especially in comparison to soils under natural vegetation (Paz-González et al., 2000). In agricultural systems, spatial variation in SOC is usually attributed to differences in vegetation, management and terrain (Bricklemyer et al., 2005), which subsequently influence other factors such as infiltration, water content, porosity, aggregation and temperature.

Many studies have compared the effects of conventional tillage to conservation management practices on SOC (Angers and Eriksen-Hamel, 2008; Blanco-Canqui et al.,

2009b; Christopher et al., 2009); however, few studies have been devoted to comparison of carbon stocks under ‘best management practices’ such as no-till or agroforestry in cropping systems (Morgan et al., 2010). Riparian filter strips, the focus of nearly all published studies of filter strips, separate aquatic and terrestrial habitats with the primary goal improving water quality (Barling and Moore, 1994; Parsons et al., 1994; Tate et al.,

2004; Liu et al., 2008), while upland filter strips are located in the terrestrial landscape and are designed to improve water quality and protect soil resources at the same time.

Upland filter strips are also usually maintained (e.g. mowing, fertilization and herbicide application) in contrast to riparian filter strips, where the vegetation is often allowed to undergo natural succession. Elucidating the effects of upland filter strips on carbon stocks is necessary to fill a current gap in our knowledge base regarding conservation management practices.

Evaluation of nutrient stocks and the ability to detect treatment effects is often hampered by large spatial variability coupled with limited experimental design (Ellert et al., 2001b; VandenBygaart et al., 2007). Even though some studies find differences after one decade, (Lal et al., 1998; Lal, 2004), up to 200 years may be required before differences in total SOC become detectable after land-use changes are implemented

(Knops and Tilman, 2000). Detecting change in nutrient storage is generally accomplished by taking measurements over time, which may not account for exogenous factors, or by comparing treatments after a period of time to a conventionally managed control (Ellert et al., 2001b). Sampling depth is also a consideration; at a minimum, soil

should be sampled to a depth that captures the effects of the management practice or horizonation, whereas a sampling depth of 30 – 50 cm is generally recommended for evaluation of total stocks (Ellert et al., 2001b; VandenBygaart et al., 2007;

VandenBygaart et al., 2011).

An additional factor making it difficult to compare research outcomes of studies examining the effects of agricultural management practices is that different methods are used to calculate SOC and TN stocks. For example, SOC data is commonly reported as concentration measurements (kg SOC per kg soil), by equivalent soil volume (kg SOC per m2 to a fixed soil depth) or by equivalent soil mass (kg SOC per m2 to a fixed soil mass). Both the equivalent soil volume and equivalent soil mass methods account for differences in soil bulk density (Gregorich et al., 1994; Ellert and Bettany, 1995; Ellert et al., 2001b; VandenBygaart and Angers, 2006). Although the equivalent soil mass based method is generally accepted as the most reliable, it may not be appropriate in areas with perennial woody vegetation such as an agroforestry system (Ellert et al., 2001b). Long- term or large-scale studies designed to capture these changes are expensive and not always feasible.

Soil bulk density, structure and aggregate stability are soil properties that may be indicative of resistance to erosion and improved stability and have been related to cultivation practices and soil organic matter content (Schlesinger, 1986; Cambardella and

Elliott, 1993; Watts and Dexter, 1997; Blanco-Canqui et al., 2009a; Ruehlmann and

Korschens, 2009). Lack of tillage combined with planting and harvesting traffic may lead to compaction and increased soil bulk density under no-till management and a strong negative correlation between SOC and bulk density is well documented in the literature

(Blanco-Canqui et al., 2009b; Ruehlmann and Korschens, 2009). A variety of laboratory aggregate stability methods are commonly used to empirically relate the structural cohesion of soil to various aspects of soil function. These methods differ primarily in the technique of soil disruption including sieving (wet and dry), slaking, shaking, sonication and use of chemical dispersants (Kemper and Rosenau, 1986; Cambardella and Elliott,

1994; Six et al., 1999; Pikul et al., 2009). The distribution of SOC and TN among soil fractions is positively correlated with aggregate stability and has implications for nutrient losses through erosion and leaching (Barthes and Roose, 2002). Therefore, subsequent analysis of stable aggregates or other fractions for SOC or TN content provides a measure of the degree of structural isolation and protection in soil (Eynard et al., 2005; Swanston et al., 2005; Virto et al., 2009) and serves as an early indicator of SOC change that is not always evident with total carbon measurements (Gregorich et al., 1994).

Another gauge of soil organic matter quality is the SOC:TN ratio. Soil nitrogen is inextricably linked to the carbon cycle through biotic processes associated with vegetative productivity and organic matter decomposition (Post et al., 1985). Differences in SOC:TN ratio may be indicative of the composition of the organic material being added to the soil or of the decomposition process. As organic matter is decomposed, SOC is mineralized to carbon dioxide while much of the nitrogen is recycled in the soil, resulting in a lower SOC:TN ratio relative to the initial organic substrate (Gilmour, 1961;

Parsons and Tinsley, 1975; Jenkinson, 1988). Therefore, a larger SOC:TN ratio indicates a retardation of the decay process while a lower SOC:TN ratio is considered indicative of more advanced stages of organic matter decomposition (e.g. mineralization).

Soil formation in a landscape is the result of pedogeomorphic processes including physical weathering, chemical weathering and losses/gains associated with solute and soil transport; thus, the occurrence and distribution of soil types and their properties are influenced by landscape position (Wilson and Gallant, 2000). Over time, soil is transported down-slope, accumulating in depressions (Minasny and McBratney, 2001) and landscape position has been shown to be a factor in the spatial distribution of SOC and TN in soils with accelerated erosion and leaching (Gessler et al., 2000; Chamran et al., 2002; Yoo et al., 2006; Hancock et al., 2010). Landscape position is also a factor affecting the rate of carbon sequestration in reclaimed prairie soils (Guzman and Al-

Kaisi, 2010) and depth to argillic horizon and associated hydraulic properties in claypan soils (Jiang et al., 2007).

A better understanding of the complex interactions among agronomic practices, conservation practices and the environment is necessary to describe and predict the impacts of land use change on soil SOC and TN. Therefore, the objectives of this study were to evaluate the effects of three conservation management practices (i.e., grass VFS, agroforestry VFS and no-till) and four landscape positions (i.e., summit, shoulder, backslope and footslope) on the concentration, soil volume based stocks and soil mass based stocks of SOC, TN, particulate, adsorbed and occluded organic carbon (PAO-C) and nitrogen (PAO-N) in an agroecosystem with claypan soils planted to a maize (Zea mays. L.) – soybean (Glycine max (L.) Merr.) rotation.

Materials and Methods

Study site

The Greenley Memorial Research Center experimental site, originally designed as a paired watershed study to investigate soil solution and runoff, is located in Knox County,

Missouri, USA (40˚ 01’N, 92˚11’W) in Major Land Resource Area 113, the Central

Claypan Area (USDA-NRCS, 2006). Claypan soils are characterized by an abrupt, argillic sub-soil horizon imparting vertic properties to the soil such as shrink-swell and cracking behavior (Baer and Anderson, 1997). These soils experience complex runoff and infiltration phenomena; they are poorly drained soils with perched water tables dominated by lateral, subsurface flow and restricted upward movement of water when surface evaporation is high (Ross et al., 1944; Blanco-Canqui et al., 2002; Arnold et al.,

2005). Even though claypan soils are not extensive globally, over 16,000 km2 are classified as Epiaqualfs (Stagnic Luvisols) in the Central Claypan Area. Worldwide, soils with restrictive sub-horizons cover approximately 2.9 million km2, potentially having a large impact on global biogeochemical cycles (USDA-NRCS, 1998; USDA-NRCS,

2006; Kitchen et al., 2010).

The study site is a small catchment (9.25 ha) facing north with gentle slopes (0 to

3%) that has been cultivated for over 60 years and consistently managed since 1991 in a no-till, maize and soybean rotation. In 1997, two types of upland contour VFS were planted; grass VFS were installed in the west subcatchment (3.16 ha), agroforestry VFS were installed in the center subcatchment (4.44 ha) and the east subcatchment (1.65 ha) was left under no-till management to serve as the control (Fig.3.1). The filter strip design exceeds recommendations for upland filter strips in cropped land (USDA-NRCS, 2010);

Figure 3.1 Greenley Memorial Research Center study site in Knox County, Missouri (star). Black dots indicate sampling locations. Thick black lines are grass VFS, grey lines are agroforestry VFS and dotted lines represent soil series boundaries.

each filter strip is approximately 4.5 m wide and planted on the contour with a VFS to crop width ratio ranging from 1:8 to 1:5. Overall, the filter strips account for approximately 8 to 10% of the land area and have a combined VFS width of 22.5 to 27 m in each subcatchment. The grass VFS consist of redtop (Agrostis gigantea Roth), brome

(Bromus spp.), and birdsfoot trefoil (Lotus corniculatus L.). The agroforestry VFS consist of pin oak (Quercus palustris Muenchh.), swamp white oak (Q. bicolor Willd.) and bur oak (Q. macrocarpa Michx.) planted approximately 3 m apart within the grass species.

The filter strips are mowed periodically and treated with herbicides before planting and after harvest to inhibit vegetative succession.

Soils in the study site were formed in loess underlain by weathered glacial till

(Watson, 1979) and are characterized by an abrupt, argillic, claypan horizon occurring from 20 – 62 cm below the surface (Udawatta et al., 2006). Soil series mapped in the study area follow the landscape and are classified by the USDA-NRCS as Putnam silt loam (fine, smectitic, mesic Vertic Albaqualfs) in the summit and shoulder positions and

Kilwinning silt loam (fine, smectitic, mesic Vertic Epiaqualfs) in the backslope and footslope positions (Fig. 3.1). The 30-year, average annual precipitation for the Knox

County (Novelty, MO) weather station is 95.9 cm. A summary description of the site and comprehensive details can be found in Veum et al. (2009) and the references therein.

Soil sampling and laboratory analyses

An exploratory semivariogram analysis of cropped soil in the study area (36 samples collected under no-till with distances varying from 3 to 30 m apart) suggested a geostatistical range of 35m for SOCc, 40m for TNc and 25m for bulk density. Subsequent

samples were collected along the contour at an interval close to the estimated geostatistical range for SOCc to avoid spatial correlation. Surface (0 – 5 cm) and subsurface (5 – 13 cm) soil samples were collected from four landscape positions

(summit, shoulder, backslope and footslope) and three conservation management practices (grass VFS, agroforestry VFS and no-till). The shallow sampling depth was deliberately selected in order to isolate the dynamic surface layer most likely to exhibit treatment effects. Three subsamples from each combination of landscape position and conservation management practice were collected, totaling 36 sample locations in the study site (Fig 1). Soil cores were collected using 7.62 cm diameter by 5.08 cm height aluminum rings for the surface (0 – 5 cm) and 7.62 cm diameter by 7.62 cm height aluminum rings for the subsurface (5 – 13 cm) following the core method of Grossman and Reinsch (2002). Surface residues and litter were removed prior to soil sample collection. Gravimetric water content of the soil in the rings was determined by placing samples in a forced-air oven set at 105 ºC. Bulk density was measured using the soil core sample and adjusting for volumetric water content. For general characterization of soil properties, one kg of bulk soil from each depth at each location was also collected using a hand trowel.

Cation exchange capacity (CEC), base saturation and particle size were determined on air-dried, ground soil samples by the University of Missouri Soil Characterization

Laboratory following USDA-NRCS procedures (Burt, 2004). All other laboratory analyses used in this study were performed in triplicate on each subsample. Samples were reanalyzed as appropriate for quality control purposes. General soil characteristics by conservation management practice are summarized in Table 3.1.

Table 3.1 Mean soil properties in the 0–5 cm and 5–13 cm layers of the study soils by conservation management practice. The standard error of the mean is stated in parentheses.

Surface (0–5 cm) Soil Property Grass Agroforestry No-Till Bulk Density (g cm-3) 1.10 (0.027) 1.02 (0.023) 1.29 (0.020) -1 a CECNaOAc (cmolc kg ) 23.3 (1.20) 23.6 (0.51) 21.0 (0.40)

Base SaturationNaOAc (%) 92.8 (2.33) 87.8 (1.58) 90.0 (2.27)

Soil pHw 7.2 (0.07) 6.7(0.15) 6.9 (0.12) Textural Class Silt Loam Silt Loam Silt Loam Clay Content (g kg-1) 242.5 (17.59) 239.2 (9.42) 218.7 (6.22) Subsurface (5–13 cm) Soil Property Grass Agroforestry No-Till Bulk Density (g cm-3) 1.30 (0.011) 1.25 (0.024) 1.39 (0.024) -1 a CECNaOAc (cmolc kg ) 20.8 (1.03) 19.9 (0.60) 19.62 (0.63)

Base SaturationNaOAc (%) 96.2 (0.58) 96.2 (2.9) 92.2 (2.46)

Soil pHw 7.5 (0.07) 7.4 (0.12) 7.4 (0.13) Textural Class Silt Loam Silt Loam Silt Loam Clay Content (g kg-1) 246.8 (17.49) 226.2 (9.53) 227.8 (0.47) aCation exchange capacity

Soil was sieved (< 2 mm), air-dried, ground and analyzed on a LECO® combustion analyzer for SOC and TN concentration measurements (SOCc and TNc). Soil volume

(SOCv and TNv) and soil mass (SOCm and TNm) based stocks were calculated as the product of gravimetric concentration, bulk density and the appropriate thickness of the soil layer (Gregorich et al., 1994; Ellert and Bettany, 1995; Ellert et al., 2001a;

VandenBygaart and Angers, 2006). In this study, all three calculations were used to evaluate treatment effects on SOC and TN in the surface (0 – 5 cm; 0 – 70 kg m-2) and subsurface (5 – 13 cm; 70 – 130 kg m-2) layers.

A rapid physical slaking test was used to isolate a physically stabilized fraction of

SOC and TN and evaluate treatment effects. This fraction includes particulate, adsorbed and occluded organic carbon (PAO-C) and nitrogen (PAO-N). The separation of PAO-C and PAO-N was performed using a slaking stability test adapted from Kemper and Koch

(1966b) and Yoder (1936). Specifically, 100 g of air-dried sieved soil (< 2 mm) was slaked in deionized water over a 23 cm diameter, 53 µm sieve and wet-sieved for two minutes at 30 seconds per cycle. The soil residue retained on the sieve was backwashed, dried at 50°C until the mass stabilized, followed by grinding with a mortar and pestle.

The proportion of residue to whole soil was calculated by mass difference on a sand-free basis with the sand content determined for each individual sample location. The SOC and

TN content of the PAO fraction was determined on a LECO® combustion analyzer and the proportion of PAO-C and PAO-N in the soil was calculated by multiplying the proportion of occluded SOC or TN in the PAO residue by the proportion of PAO residue to whole soil on a sand-free basis. Stocks in this fraction were calculated and compared

on a concentration (PAO-Cc and PAO-Nc), soil volume (PAO-Cv and PAO-Nv) and soil mass (PAO-Cm and PAO-Nm) basis.

Statistical analysis

Due to the original experimental design imposed on the study site, testing the main effects of conservation management practice and landscape position was accomplished using a two-factor ANOVA from a completely randomized design (Kuehl, 2000).

However, similar to many environmental studies, a standard statistical analysis was not possible due to the inclusion of subsamples and the lack of replication among the different treatment factors. Specifically, each landscape by conservation management practice treatment combination has no “true” replication. Instead, each experimental unit has one replicate with three subsamples.

Consequently, using a standard two-factor ANOVA with subsamples results in an error term with zero degrees of freedom and, thus, a proper hypothesis test for the landscape position by conservation management practice interaction term could not be formulated. Specifically, the interaction term served as the replication, or error term, in formulating tests for main effects.

In environmental studies, subsamples are often incorrectly treated as true replicates. This ignores the fact that the subsamples are correlated (i.e., they come from the same experimental unit) and may lead to incorrect statistical and scientific inference.

Nevertheless, as is the case with many environmental studies involving designed experiments, it is still possible to conduct a proper statistical analysis capable of drawing meaningful scientific inference. In this direction, a modified two-factor ANOVA, appropriate for this study, was employed.

The statistical analysis is based on a two-factor ANOVA with sub-samples.

Specifically, the following model is considered

where denotes the overall mean, denotes the main effect due to conservation management practice (i =1,2,3), denotes the main effect due to landscape position (j

=1,2,3,4), is the random error term and is the random effect due to subsamples (k

=1,2,3) with ~ iid N(0, ) and ~ iid N(0, ) and mutually independent. This model differs from the traditional two-factor ANOVA in that there is an additional random effect ( ) due to subsampling.

Additionally, due to the lack of replication, the two-way interaction between conservation management practice and landscape position was used as the random error term (i.e., the replication). A full justification for this model specification is provided in the Appendix of this chapter. Table A1.1 in the Appendix provides a comparison of the two-factor ANOVA with subsamples employed in this study to a two-factor ANOVA where subsamples are treated as true replicates. This comparison illustrates how treating the subsamples as true replicates could lead to incorrect scientific inference about the soil properties investigated at the selected study site.

All two-factor ANOVA analyses were conducted using SAS Proc Mixed (SAS® software, Version 9.1.3, SAS Institute Inc., Cary, NC, USA.). In addition, Tukey's HSD procedure was implemented to test for pair-wise differences in treatment means. In doing so, the Tukey's HSD statistic uses the correct estimate for the variance by utilizing the interaction term as the random error term in the ANOVA while accounting for the

subsamples (Littell et al., 2006). All analyses were conducted at the α=0.05 significance level.

To examine the relationship between bulk density and SOC, a simple linear regression was considered across all subsamples. As such, this regression does not explicitly account for possible correlation due to subsampling. In order to validate the regression assumptions a residual analysis was conducted and no departures were found.

Finally, using the Shapiro-Wilk test for normality, there was insufficient evidence to reject the null hypothesis that the residuals are normally distributed (p = 0.508). To assess the paired differences between the mean total SOC:TN ratio and mean PAO-

C:PAO-N ratio, subsamples were aggregated by group (mean of three subsamples within each conservation management practice by landscape position combination, n=12). The normality of the distribution of the paired differences was validated (Shapiro-Wilk test for normality, p = 0.323) and a paired t-test was conducted, accounting for the fact that the ratios are related (i.e., matched pairs) rather than independent variables (Mason et al.,

1989).

Results

Bulk density, total SOC and TN

In the surface layer, the mean bulk density of no-till soil was significantly greater than the grass (p = 0.004) and agroforestry (p < 0.001) VFS soils. In the subsurface layer, the mean bulk density of no-till soil was significantly greater than the agroforestry (p =

0.017) VFS soil only. Overall, the grass and agroforestry VFS soils had significantly greater mean total SOCc in the surface layer than no till (p = 0.032 and 0.034),

corresponding to an increase of 23% and 22%, respectively. When mean surface stocks of SOCv or SOCm were calculated, no detectable differences were found among conservation management practices. The grass VFS soil had significantly greater surface mean TNc and TNm than no till (p = 0.031 and p = 0.049) and no significant differences were detected in the subsurface layer or among landscape positions (either depth) for mean SOCc, TNc, SOCv, TNv, SOCm or TNm (Table 3.2).

To more closely examine the general relationship between bulk density and SOCc in the surface layer, the following simple linear regression relationship (R2 = 0.50) was determined using data from all subsamples (n=36) in the study (Fig. 3.2):

-3 -1 where BD is bulk density (g cm ) and SOCc is soil organic carbon (g OC kg soil ) and the slope parameter was significant (p < 0.001). In this study, the range of SOCc was relatively narrow in comparison to the work of Ruhelmann and Korschens (2009), allowing for the use of a linear regression model instead of a non-linear model, as was used in their meta-analysis of a wide range of soils.

Particulate, adsorbed and occluded organic carbon (PAO-C) and nitrogen (PAO-N)

A simple, rapid slaking test was used to isolate a soil fraction more sensitive to land use effects than whole soil. The residues from this method include particulate, adsorbed and occluded fractions of organic carbon (PAO-C) and nitrogen (PAO-N). The proportion of soil residue, calculated by mass difference, ranged from 60 to 80%, similar to the results reported in Kemper and Rosenau (1986). The remaining residue was

Table 3.2 Means of soil total organic carbon (SOC), total nitrogen (TN)

and sand-free, particulate, adsorbed and occluded soil organic carbon

(PAO-C) and total nitrogen (PAO-N) expressed on a concentration, soil

volume and soil mass (represented by the subscripts c, v and m,

respectively) basis by conservation management practice. Values

followed by a different lowercase letter within columns and for a given

layer are significantly different (using Tukey’s HSD) at α = 0.05. The

standard error of the mean is stated in parentheses.

Figure 3.2 Scatter plot and regression relationship for regression of soil organic carbon (SOC) concentration on bulk density (BD) using all subsamples (n=36) in the 0–5 cm surface layer. Note the slope parameter (SOCc) was statistically significant (p< 0.001).

analyzed for total organic carbon and total nitrogen on a concentration (PAO-Cc and

PAO-TNc), soil volume (PAO-Cv and PAO-Nv) and soil mass (PAO-Cm and PAO-Nm) basis and corrected for sand content. The mean sand content was not significantly different among the VFS and no-till soils, ranging from 75 g kg-1 to 85 g kg-1 in the surface layer and from 71 g kg-1 to 81 g kg-1 in the subsurface layer.

Similar to total SOC, the mean surface PAO-Cc was significantly greater in the grass

(p = 0.006) and agroforestry (p = 0.003) VFS soils than in no-till. In contrast to total

SOC, the grass VFS and agroforestry VFS had significantly greater mean PAO-Cv stocks than no-till (p = 0.017 and 0.031) and significantly greater mean PAO-Cm stocks than no- till (p = 0.005 and 0.002). These results represent a 57% increase in the grass VFS and a

69% increase in the agroforestry VFS over no-till in organic carbon in the PAO fraction on a mass basis. With respect to nitrogen, the grass and agroforestry VFS exhibited significantly greater mean PAO-Nc (p = 0.005 and 0.004), PAO-Nv (p = 0.012 and 0.049) and PAO-Nm (p = 0.008 and 0.005) stocks than no-till in the surface layer. These results correspond to a 54% increase in PAO-Nm in the grass VFS and a 60% increase in PAO-

Nm in the agroforestry VFS over no-till. No significant differences were detected in the subsurface layer or among landscape positions (either depth) for mean PAO-Cc, PAO-Nc,

PAO-Cv, PAO-Nv, PAO-Cm or PAO-Nm (Table 3.2).

SOC and TN ratios

No statistically significant differences in the total SOC:TN or PAO-C:PAO-N ratios were found among conservation management practices or landscape positions in surface or subsurface layers. An overall comparison of the total SOC:TN ratio with the PAO-

C:PAO-N ratio was conducted using a paired t-test. Subsamples were aggregated by group (mean of three subsamples within each conservation management practice by landscape position combination, n=12), and the difference between the ratios was significant (p = 0.002). This suggests that the mean PAO-C:PAO-N ratio (12.3 to 1) was greater than the mean total SOC:TN ratio (11.5 to 1) since the mean difference between

PAO-C:PAO-N and total SOC:TN ratios is greater than zero. Additionally, this is an indication that the organic matter sequestered in the PAO fraction of the soil is less decomposed and more protected than the whole soil organic matter.

Discussion

Vegetated filter strips are known to effectively increase SOCc in VFS soils compared to bare soils (Staddon et al., 2001), and research on grass and agroforestry riparian buffers by Tufekcioglu (2003) found increased fine root biomass and fine root carbon in the buffers relative to cropped soil. In general, studies find that no-till cultivation is also a highly effective conservation management practice, often acting as a carbon sink from the atmosphere as opposed to conventional tillage (Izaurralde et al., 2007; Lal et al.,

2007; Dou et al., 2008). Few studies, however, compare SOC among conservation practices and studies of agroecosystems with restrictive subsurface soil horizons are rare.

Although significant differences in mean total SOCc were found between the VFS and

no-till soils in this study, no significant differences were evident when mean stocks were calculated on a soil volume or soil mass basis. No-till soils are subject to compaction and often have greater bulk density (Schlesinger, 1986; Ruehlmann and Korschens, 2009) and this study illustrates the need to account for bulk density differences when evaluating nutrient stocks in soil.

The results of this study suggest that VFS soils either do not sequester more total

SOC stocks than no-till or that the VFS soils in this study may not have reached equilibrium levels of SOC. Hansen (1993) found soil under poplar plantations on previously cultivated land required 12 – 18 years to make significant gains in near surface

SOC and that most gains occurred at greater depths. Therefore, the ten year duration of this study may not be sufficient to detect differences when VFS are compared to another conservation management practice such as no-till. In fact, small changes in SOC generally require an unrealistic sampling intensity (hundreds to thousands) in order to achieve the statistical power to detect differences (Garten and Wullschleger, 1999; Yang et al., 2008; Kravchenko and Robertson, 2011). Moreover, this degree of sampling intensity may potentially lead to an artificial reduction in variance that comes from ignoring positive spatial correlation (VandenBygaart and Angers, 2006). Several studies indicate that more than one decade is needed to observe changes in the SOC pool without intensive sampling, even though the treatments may be acting as a carbon-sink upon establishment (Oelbermann et al., 2004; Conant et al., 2011).

In addition to the short study duration, the modest changes in SOC may be partially due to a system where crop growth and hydrology are dominated by a restrictive claypan layer. Prairie and forest soils in the claypan region are generally low in productivity,

highly leached and acid (Ross et al., 1944). Cultivated soils in this region have experienced extensive erosion and degraded soil quality over the past several decades, leading to reduced crop productivity (Kitchen et al., 1999; Kitchen et al., 2010).

Therefore, the establishment and growth of the grass and agroforestry VFS in this study may have been adversely affected by the presence of the claypan layer, protracting the time needed for changes in total SOC and TN to be detectable at this site.

The PAO fraction is therefore of particular interest in a study where small changes in

SOC are coupled with the limitations of a small sampling area. In contrast to SOC stocks, the PAO fraction exhibited significantly greater mean concentration, soil volume and soil mass based stocks in VFS soils compared to no-till after only ten years. The PAO fraction indicates that changes in the distribution and quality of SOC and TN have occurred and demonstrates that the PAO separation technique is an effective early indicator of management effects in the soils studied. The physicochemical protection afforded by aggregation inhibits the biochemical attack of otherwise decomposable organic matter, sequestering carbon in the soil (Jastrow et al., 2007). A strong feedback loop exists between organic matter and aggregate stability and the extent of protection in soil aggregates is dependent on soil properties, vegetation (in particular roots), soil microorganisms, management practices and environmental conditions (Sollins et al.,

1996; Amezketa, 1999; Six et al., 2000b; Jastrow et al., 2007).

Many soil separation methods, differing primarily in the technique of soil disruption, are currently in use as a measure of the degree of structural isolation and protection of carbon in the soil. Each of these techniques separates a fraction of soil carbon representing different biological, chemical and/or physical aspects of soil and method

selection is based on time constraints and study objectives. These include sieving (wet and dry), slaking, shaking, sonication and use of chemical dispersants (Kemper and

Rosenau, 1986; Cambardella and Elliott, 1994; Six et al., 1999; Pikul et al., 2009), all of which are more time consuming methods than the method used in this study. In contrast, the PAO fractionation method was rapid, simple and an effective early indicator of SOC protection in VFS that was not evident with total SOC measurements. Additionally, the mean difference between PAO-C:PAO-N and total SOC:TN ratios in the surface layer

(n=36) was significantly different from zero. Together, these results suggest increased protection and reduced mineralization of SOC in VFS soil compared to soil under no-till management.

In the subsurface layer, no significant differences were found among conservation management practices or landscape positions for mean SOCc, SOCv or SOCm. This is consistent with studies of no-till cultivation showing that treatment effects are limited to the surface layer where organic matter is concentrated (Angers and Eriksen-Hamel, 2008;

Blanco-Canqui and Lal, 2008; Pikul et al., 2009). Whereas strong landscape effects are found in the steeply sloping (0-66%, average 22%) study site of Gessler et al.(2000) and

Chamran et al. (2002), the lack of significant effects from landscape position in this study may be due to the gentle (0 – 3%) slope of the study site, the small catchment area, and the similarity of the soil series found at each landscape position. Further, a complication in many environmental studies is that landscape position is fixed in space and thus randomization is not possible (Essington, 2004). Therefore, significant effects due to this factor become difficult to interpret; however, this study only found significant treatment effects due to conservation management practices and, since this factor was randomized

within the study design, the results are interpretable. Moreover, this study was concerned with soil stocks only, and the sequestration potential of other carbon and nitrogen pools

(i.e. above- and below-ground biomass) was not evaluated. Under perennial vegetation, upwards of 60% of total net primary productivity occurs below ground as roots (Krug and

Hollinger, 2003; Jones and Donnelly, 2004) and it is estimated that 10-40% of biomass in forested sites is contained in root biomass (MacDicken, 1997; Young et al., 1998;

Oelbermann et al., 2004). Therefore, a large pool of organic carbon is contained in the perennial biomass of the grass and agroforestry VFS in this study. In addition, estimates of crop yield, surface residues, gaseous efflux, redistribution and leaching are important considerations for future studies at this site seeking a more comprehensive analysis of nutrient cycling at the ecosystem level.

The lack of significant differences in mean SOC and TN stocks between VFS and no- till confirms that no-till is a highly effective conservation management practice in agroecosystems. As the study continues, differences in total SOC and TN stocks between

VFS and no-till soils may emerge if carbon and nitrogen are accumulating at different rates. However, the PAO fraction has demonstrated sensitivity to management effects in the short-term. The dissimilarity in PAO stocks between VFS and no-till soils indicate that differences in carbon and nitrogen distribution have occurred after ten years.

Additional investigations into soil microbiological quality and SOC characteristics of the soils studied may also provide greater insight into the role of conservation management practices over relatively short time periods (e.g. 10 years).

Conclusions

As demonstrated in this study, ignoring the underlying experimental design by treating subsamples as true replicates (pseudoreplication) can lead to incorrect scientific inference. Critical to a proper statistical analysis was the use of a modified two-factor

ANOVA accounting for the lack of replication in the study design. Additionally, this study illustrates how the choice of SOC quantification method (concentration, soil volume or soil mass) influences the scientific inference and conclusions, emphasizing the need for bulk density measurements for valid assessment of soil nutrient stocks.

Overall, this study finds that VFS soils have not yet sequestered significantly greater stocks of total SOC than no-till soils and indicates that grass VFS may sequester TN more rapidly than agroforestry VFS. Future research is necessary to determine if the ten year duration of the study was sufficient for the soils to reach equilibrium levels of SOC or TN. In contrast to total SOC and TN, a rapid, inexpensive slaking-stability test separated a physically stable soil fraction (PAO) exhibiting significantly greater mean concentration, soil volume and soil mass based stocks in VFS soils compared to no-till.

Thus, a higher proportion of SOC and TN are contained in the physically stable fraction of soils planted to VFS. This stable fraction also exhibited a greater organic carbon to nitrogen ratio than the whole soil, suggesting increased SOC protection in this fraction.

Together, these results indicate differences in the distribution and quality of SOC and TN stocks between VFS and no-till soils. Overall, this study contributes to a richer understanding of agronomic effects on soil carbon and provides information for improved management of agricultural resources for long-term sustainability and profitability.

CHAPTER 4

WATER STABLE AGGREGATES AND ORGANIC MATTER FRACTIONS UNDER CONSERVATION MANAGEMENT PRACTICES ON CLAYPAN SOILS

This chapter has been submitted to the Soil Science Society of America Journal: Veum, K.S., Goyne, K.W., Kremer, R.J., Motavalli, P.P. (Submitted March, 2012). Water stable aggregates and organic matter fractions under agricultural conservation management practices. Soil Sci. Soc. Am. J.

Abstract

Conservation management practices may improve aggregate stability and enhance carbon sequestration relative to conventional tillage in agroecosystems, yet little is known about the differences among conservation management practices. In addition, the relationship between aggregate stability and organic carbon fractions is not well understood. The objectives of this study were to (1) evaluate the effects of three conservation management practices (grass vegetative filter strips (VFS), agroforestry

VFS and no-till) and four landscape positions (summit, shoulder, backslope and footslope) in an agroecosystem with claypan soils on field-moist and air-dry aggregate stability (WSAFM and WSAAD, respectively), cold water-extractable organic carbon from whole soil (WEOC) and the particulate, adsorbed and occluded (PAO) fraction

(WEOCP), (2) evaluate relationships among these variables and associated soil properties, and (3) compare properties of the PAO fraction to whole soil. Soils under VFS had significantly greater WSAAD, WEOC, WEOCP, and C:N ratios (solid and water extracts) relative to no-till. Measures of WEOC, WEOCP and organic carbon from the

PAO fraction (PAO-C) were highly correlated with WSAAD, suggesting that these organic carbon fractions may play an important role in aggregate stability. Overall, the

PAO fraction demonstrated greater C:N ratios in the solid and water extracts relative to whole soil. Results indicate that organic matter is preferentially sequestered and less degraded in the PAO fraction relative to whole soil and that the sequestration process is enhanced under the perennial vegetation of VFS relative to no-till. In addition, WEOC and PAO-C served as effective early indicators of soil carbon change.

Introduction

Over the past several decades, conservation management practices have become increasingly popular due to a wide range of environmental benefits including increased soil carbon (Lal et al., 1994), reduced runoff (Veum et al., 2009), reduced pollutant transport (Chu et al., 2010), increased aggregate stability (Angers et al., 1993b) and increased microbial diversity (Kennedy and Smith, 1995). Most research, however, compares conservation management practices to conventional tillage, and studies comparing soil properties among conservation management practices such as upland VFS and no-till are rare (Morgan et al., 2010). Soil organic carbon (SOC) is important component of the global C cycle and a common soil quality indicator that often requires long-term studies for detection of treatment effects (Lal et al., 1994; Ellert et al., 2001a), but long-term studies are expensive and not always feasible. Therefore, more sensitive soil quality or soil carbon indicators that are able to illuminate changes more rapidly than total SOC measurements are extremely useful.

Aggregate stability is a common indicator of soil quality that integrates several aspects of agricultural sustainability (Kemper and Rosenau, 1984; Barthes and Roose,

2002; Cantón et al., 2009). The relationship between SOM and aggregate stability has been discussed extensively in the literature (e.g., Tisdall and Oades, 1982; Amezketa,

1999; Six et al., 2004) and bonding between mineral particles and SOM is considered a primary factor for aggregation (Kemper and Koch, 1966a; Tisdall and Oades, 1982).

Although total SOM content is often positively correlated with aggregate stability

(Idowu, 2003; García-Orenes et al., 2005), the composition of SOM is also important

(Tisdall and Oades, 1982). Water-extractable organic matter (WEOM) represents a potentially mobile and labile fraction of SOM (Chantigny, 2003) that may play a role in aggregate stability (Haynes and Swift, 1990; Gregorich et al., 2003). The quantity and quality of WEOM extracted from soil are affected by the extraction conditions including temperature (Chantigny, 2003; Haynes, 2005). Although Haynes and Swift (1990) promoted usage of hot and cold WEOM extracts as indicators of aggregate stability, most studies have focused on the role of “hot” extractant solutions (ranging from 70 to 100ºC), and few studies have examined the role of “cold” or room-temperature WEOM (20 to

23ºC). Studies relating WEOM to aggregate stability generally rely on correlation analysis and have produced contradictory results. Positive correlations between WEOM constituents (e.g., organic carbon, carbohydrates or polysaccharides) and aggregate stability have been found with hot extracts (Angers and Mehuys, 1989; Ball et al., 1996;

Haynes, 2000; Shepherd et al., 2001), as well as cold water extracts (Miller and Dick,

1995; Ball et al., 1996). In contrast, many studies find no correlation between hot or cold

WEOM extracts and aggregate stability (Carter et al., 1994; Chan and Heenan, 1999;

Bouajila and Gallali, 2008). Overall, the exact nature of the relationship between WEOM and aggregate stability is still unclear.

Another gage of SOM differences or change is the organic carbon to nitrogen (C:N) ratio. Differences in C:N ratios may be attributed to contrasting organic inputs to the soil or differences in the decomposition process. As decomposition progresses, SOM is mineralized to CO2 while much of the nitrogen is retained, resulting in a lower C:N ratio relative to the initial SOM (Jenkinson, 1988). Changes in C:N ratios have been used to indicate the relative degree of decomposition among soils under various management practices and among aggregate size fractions (Elliott, 1986). Generally, larger soil aggregates have greater C:N ratios, representing more labile and less processed SOM relative to microaggregates (Elliott, 1986; Cambardella and Elliott, 1993) or the clay fraction (Oades and Waters, 1991). Similarly, a relative reduction in the C:N ratio in cold

WEOM extracts also reflects the decomposition and transformation of organic matter

(Landgraf et al., 2006).

Other important soil properties shown to correlate with aggregate stability include clay content, soil bulk density (ρb) and moisture content (Kemper and Koch, 1966a). Bulk density is often negatively correlated with aggregate stability (Clapp et al., 1986) and

SOM (Ruehlmann and Korschens, 2009; Veum et al., 2011), due in part to the collapse of macroaggregates under compaction, slaking of weakly aggregated soil at the surface, and losses of inter-aggregate porosity (Angers et al., 1987; Haynes and Beare, 1996). In addition, the primary binding agent, SOM, is less dense than the mineral fraction (Kay,

1998). The relationship between aggregate stability and water content is complex.

Overall, aggregates tend to be more stable at greater water contents than when air-dried

(Kemper and Rosenau, 1986). This has been attributed to a reduction of slaking effects when capillaries are filled with antecedent water, reducing the wetting rate (Panabokke and Quirk, 1957; Kemper and Rosenau, 1984; Caron et al., 1996). Some debate still surrounds the testing of aggregate stability on air-dried or field-moist soil and the most appropriate method of rewetting samples (tension, capillarity or rapid rewetting via submergence). Air-drying soil increases internal bonding and cohesion and offers the advantage of equivalent initial moisture conditions, although it may not closely approximate real-world conditions (Kemper and Rosenau, 1984). Rewetting of air-dried soils generally has the advantage of increased sensitivity to changes in soil quality and tends to emphasize differences in macroaggregation (Beare and Bruce, 1993). In situ initial water content has the advantage of representing more natural conditions found in the field, but interpretation of the results may be confounded by the effects of variable antecedent moisture content and may only represent soil behavior at the time of sampling

(Cousen and Farres, 1984; Bullock et al., 1988; Amezketa, 1999; Zsolnay, 2003).

Comparing soil properties on a gravimetric versus an areal basis may lead to different results and inference (Ellert and Bettany, 1995; Ellert et al., 2001a; VandenBygaart and

Angers, 2006; Veum et al., 2011). Expressing results on an areal basis by equivalent soil mass is generally accepted as the most appropriate; however, many studies lack the

2 requisite ρb data to convert gravimetric values to areal values. Areal values (e.g., per m or acre) are critical for the comparison of ecosystem-level dynamics and facilitate the evaluation of treatment effects, such as land use changes and management practices, on soil properties. For example, areal aggregate stability values provide insight into potential soil losses per acre under different management regimes. Alternatively, expressing soil

properties and evaluating treatment effects on an organic carbon basis can provide insight into the relationships among SOM fractions and soil properties and elucidate the role of

SOM fractions in soil ecology and soil function (Six et al., 2002; Shi et al., 2006).

In addition, soil formation is the result of pedogeomorphic processes; thus, the occurrence and distribution of soil types and their properties are influenced by landscape position (Wilson and Gallant, 2000). In agroecosystems, landscape position has been shown to influence SOC, aggregate stability and ρb (Guzman and Al-Kaisi, 2011), among other soil properties. Previous work at the site sampled for this study found no significant differences in SOC stocks among conservation management practices or landscape positions; however, a simple wet sieving technique, designed to isolate the particulate, adsorbed and occluded (PAO) fraction, found that VFS soils contained significantly greater stocks of organic carbon in this fraction (PAO-C) relative to no-till. In addition, the C:N ratio in the PAO fraction was significantly greater than the C:N ratio of whole soil, suggesting that the PAO fraction represents a protected and less processed pool of organic matter (Veum et al., 2011).

The objectives of this study were (1) to evaluate the effects of three conservation management practices (grass VFS, agroforestry VFS and no-till) and four landscape positions (summit, shoulder, backslope and footslope) in an agroecosystem with claypan soils on WEOM from whole soil and from the PAO fraction (WEOMP), field-moist aggregate stability (WSAFM) and air-dry aggregate stability (WSAAD) ten years after the

VFS were established, (2) to evaluate relationships among these variables and associated soil properties, and (3) to compare the properties of the PAO fraction to whole soil. We hypothesized that due to the increased plant residue accumulation and exudates under

perennial vegetation, VFS soils would exhibit greater WEOM, WEOMP, WSAAD and

WSAFM than no-till soil and that the stable aggregate fractions would exhibit strong, positive correlations with the labile organic carbon fractions.

Materials and Methods

Study site

The Greenley Memorial Research Center experimental site is located in Knox

County, Missouri, USA (40˚ 01’N, 92˚11’W) and was originally designed as a paired watershed study to investigate VFS effects on soil solution and runoff on claypan soils.

Claypan soils in the Midwestern U.S. were formed from loess over glacial till parent materials and are defined by the presence of an abrupt, argillic subsurface horizon.

Claypan soils exhibit vertic properties, such as shrink-swell and cracking behavior (Baer and Anderson, 1997), are poorly drained and have perched water tables dominated by lateral, subsurface flow (Blanco-Canqui et al., 2002). Claypan soils comprise an area of over 16,000 km2 in the central U.S., and worldwide, soils with restrictive sub-horizons cover approximately 2.9 million km2 (USDA-NRCS, 2006). Soils in the study site were formed in loess underlain by weathered glacial till (Watson, 1979) and the claypan horizon occurs at depths ranging from 20 to 62 cm (Udawatta et al., 2006). Soil series mapped in the study area follow the landscape and are classified by the USDA-NRCS as

Putnam silt loam (fine, smectitic, mesic Vertic Albaqualfs) in the summit and shoulder positions and Kilwinning silt loam (fine, smectitic, mesic Vertic Epiaqualfs) in the backslope and footslope positions. The clay mineralogy consists of 63% smectite, 33% illite and 4% kaolinite (Chu, 2011).

The study site is a small catchment (9.25 ha) that has been cultivated for more than 60 years and it has been consistently managed under a no-till, maize (Zea mays. L.) – soybean [Glycine max (L.) Merr.] rotation since 1991. The site consists of three small, gently sloping (0 – 3 %), contiguous subcatchments. Two types of upland contour VFS were planted in 1997; grass VFS were planted in the west subcatchment, agroforestry

VFS were planted in the center subcatchment and the east subcatchment was left under no-till management to serve as the control (Fig. 4.1).

The filter strip design exceeds recommendations for upland filter strips in cropped land (USDA-NRCS, 2010); each filter strip is approximately 4.5 m wide and planted on the contour with a VFS to crop width ratio ranging from 1:8 to 1:5. Overall, the filter strips account for approximately 8 to 10% of the land area and have a combined VFS width of 22.5 to 27 m in each subcatchment. The grass VFS consist of redtop (Agrostis gigantea Roth), brome (Bromus spp.), and birdsfoot trefoil (Lotus corniculatus L.). In addition to the grasses, the agroforestry VFS contain pin oak (Quercus palustris

Muenchh.), swamp white oak (Q. bicolor Willd.) and bur oak (Q. macrocarpa Michx.).

The filter strips are mowed periodically and treated with herbicides before planting and after harvest to inhibit vegetative succession. A summary description of the site and comprehensive details can be found in Veum et al. (2009) and the references therein.

Soil sampling and laboratory analyses

Surface (0 – 5 cm) and subsurface (5 – 13 cm) soil samples were collected after the fall maize harvest in November, 2007 from four landscape positions (summit, shoulder, backslope and footslope) and three conservation management practices (grass VFS,

Figure 4.1 Greenley Memorial Research Center study site in Knox County, Missouri (star). Black dots indicate sampling locations. Thick black lines are grass VFS, grey lines are agroforestry VFS and dotted lines represent soil series boundaries (adapted from Veum et al., 2011).

agroforestry VFS and no-till). In order to compare equivalent landscape positions and avoid the potentially confounding effects of VFS on adjacent soil (i.e., the potential effect of being spatially upslope or down-slope from the VFS), samples collected from the no- till subcatchment served as control samples for this study.

Three subsamples were collected from each combination of landscape position and conservation management practice, totaling 36 sampling locations (Fig 4.1). Soil cores were collected using 7.62 cm diameter aluminum rings for the surface (0 – 5 cm) and subsurface (5 – 13 cm) following the soil core method of Grossman and Reinsch (2002).

Surface residues and litter were removed prior to soil sample collection. Gravimetric water content of the soil cores was determined by oven drying at 105˚C. Soil bulk density was measured on soil core samples with an adjustment for water content. For general characterization of soil properties, one kg of bulk soil from each depth at each location was also collected using a hand trowel.

Cation exchange capacity (CEC), base saturation and particle size were determined on air-dried, ground soil samples following USDA-NRCS procedures (Burt, 2004). Total

SOC was determined as described in Veum et al. (2011). The PAO separation, detailed in

Veum et al. (2011), is a simplified version of the slaking method employed by Six et al.

(1998). In short, 100 g of air-dried soil (< 2mm) was slaked and immediately sieved in deionized water over a 23 cm diameter, 53 µm sieve for 2 min. The stable aggregates (i.e. the PAO fraction) were backwashed and dried at 50˚C. Following the aggregate size classification of Six et al. (1998), the PAO fraction consists of slaking-stable aggregates

(> 53µm) that contain free light fraction SOM, intra-aggregate particulate organic matter

(POM), and mineral-associated SOM. The discarded fraction (< 53 µm) contains mineral-

associated SOM from unaggregated clay and silt-sized particles and weakly aggregated soil. The organic carbon and total nitrogen content of the PAO fraction (PAO-C and

PAO-N) were determined on a LECO TruSpec CN® combustion analyzer and calculated on a sand-free basis. General soil characteristics are summarized by conservation management practice in Table 4.1.

Cold WEOM and WEOMP were obtained from whole soil and the PAO fraction following the general extraction method of Burford and Bremner (1975). Duplicate 20 g samples of air-dried and sieved whole soil (< 2 mm) or the PAO fraction were shaken for one hour on an end-to-end shaker in 40 ml deionized, ultrapure water at a 1:2 (w/v) ratio at 20º C. Samples were then centrifuged for 20 minutes at 3,600 rpm. The supernatant was transferred and centrifuged for an additional 30 minutes at 15,000 rpm and filtered in series through prewashed 0.45 µm and 0.2 µm Whatman Puradisc™ syringe filters to remove particulates and colloidal material. The filtrate was acidified using HCl to ca. pH

2 to remove inorganic carbonates and immediately analyzed in triplicate for non- purgeable organic carbon (WEOC from whole soil and WEOCP from the PAO fraction) and total nitrogen using a Shimadzu TOC-V™ liquid carbon and nitrogen analyzer

(Shimadzu Corp.; Kyoto, Japan) equipped with an ASI autosampler and sparging needle.

Undisturbed core samples were used to measure aggregate stability and water content, following an adapted method of Kemper and Rosenau (1986). Specifically, field-moist soil was hand-sieved using a 6 mm sieve. One-half of each sample was air-dried at 20˚C for fourteen days. Duplicate ten gram samples of soil aggregates were allowed to slowly wet via capillary action for 10 minutes in deionized water on a 250 µm sieve to reduce slaking effects, followed by wet sieving for ten minutes at approximately 30 seconds per

Table 4.1 Mean soil properties in the surface (0–70 kg m-2) layer of the study soils by conservation management practice. Values followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05 (n=12 per treatment).

Values without lowercase letters were determined using composite samples (n=4 per treatment).The standard error of the mean is stated in parentheses. Data partially obtained from Veum et al. (2011).

Surface (0–70 kg m-2 soil) Soil Property Grass Agroforestry No-Till Textural Class Silt Loam Silt Loam Silt Loam -1 † CECNaOAc (cmolc kg ) 23.3 (1.20) 23.6 (0.51) 21.0 (0.40)

Base SaturationNaOAc (%) 92.8 (2.33) 87.8 (1.58) 90.0 (2.27) Clay Content (g kg-1) 242.5 (17.59) 239.2 (9.42) 218.7 (6.22)

Soil pHw 6.6 (0.11)a 6.0 (0.16)a 6.3 (0.12)a Bulk Density (g cm-3) 1.10 (0.027)b 1.02 (0.023)b 1.29 (0.020)a

-2 ‡ WCFM (kg m ) 12.7 (0.59)b 11.7 (0.32)b 15.5 (0.58)a

-1 § WCAD (g kg ) 0.02 (0.001)a 0.02 (0.001)a 0.01 (0.001)a SOC (kg m-2)¶ 2.14 (0.060)a 2.08 (0.076)a 1.85 (0.076)a PAO-C (kg m-2)# 2.39 (0.099)a 2.57 (0.150)a 1.52 (0.063)b WEOC (g m-2)†† 11.57 (0.882)a 11.81 (0.943)a 6.16 (0.432)b

-2 ‡‡ WSAFM (kg m ) 56.9 (1.67)a 54.1 (1.26)a 56.5 (0.66)a

-2 §§ WSAAD (kg m ) 46.5 (2.23)a 46.9 (1.43)a 19.9 (2.46)b Subsurface (70–130 kg m-2 soil) Soil Property Grass Agroforestry No-Till Textural Class Silt Loam Silt Loam Silt Loam -1 † CECNaOAc (cmolc kg ) 20.8 (1.03) 19.9 (0.60) 19.62 (0.63)

Base SaturationNaOAc (%) 96.2 (0.58) 96.2 (2.9) 92.2 (2.46) Clay Content (g kg-1) 246.8 (17.49) 226.2 (9.53) 227.8 (0.47)

Soil pHw 7.0 (0.09)a 7.0 (0.05)a 6.9 (0.08)a Bulk Density (g cm-3) 1.30 (0.011)ab 1.25 (0.024)b 1.39 (0.024)a

-1 ‡ WCFM (g kg ) 10.7 (0.35)b 9.9 (0.33)b 12.9 (0.26)a

-1 § WCAD (g kg ) 0.02 (0.001)a 0.01 (0.001)a 0.02 (0.002)a SOC (kg m-2)¶ 1.26 (0.051)a 1.24 (0.044)a 1.13 (0.068)a PAO-C (kg/m-2)# 1.00 (0.035)a 0.98 (0.032)a 0.87 (0.059)a WEOC (g m-2)†† 7.06 (0.455)a 5.31 (0.618)a 4.61 (0.196)a

-2 ‡‡ WSAFM (kg m ) 45.4 (1.65)ab 41.0 (1.35)b 47.1 (0.96)a

-2 §§ WSAAD (kg m ) 30.4 (2.67)a 29.1 (1.63)a 14.6 (1.14)b † Cation exchange capacity, ‡ Field-moist water content, § Air-dried water content, ¶ Soil organic carbon, # Particulate, adsorbed and occluded (PAO) organic carbon (sand-free basis), †† Water-extractable organic carbon, ‡‡ Field-moist water stable aggregates, §§Air-dried water stable aggregates

cycle. Soil aggregate mass remaining on the sieve was dispersed in 0.5% Na- hexametaphosphate for 30 minutes on a rotary shaker to determine sand content. Initial water content prior to capillary pre-wetting was determined at the time of the aggregate stability test on both field moist and air-dried samples by drying triplicate three gram samples at 105˚C for 24 hours. The final aggregate mass was corrected for sand content and adjusted for initial water content. Sand content was not significantly different among treatments in this study, and the inference and conclusions related to the soil fractions were not affected by the adjustment for sand content. However, when separating aggregate size fractions from whole soil samples, sand is retained in certain size fractions causing a ‘dilution effect’ and the convention among soil scientists is to express these values on a sand-free basis (Elliott et al., 1991; Six et al., 1998). In this study, WSAAD,

WSAFM and PAO separations were expressed on a sand-free basis; thus, some values for the PAO fraction may appear to be greater than the whole soil counterparts. When directly comparing “whole soil” fractions with PAO fractions (i.e., to compare SOC concentration to PAO-C concentration across samples), both values were calculated on a sand-free basis. These soil properties are summarized by conservation management practice in Table 4.1.

Statistical analysis

Similar to many environmental studies, the original paired watershed used for this study design lacks replication among the different treatment factor combinations. In studies lacking plot-level replication, subsamples are often incorrectly treated as true replicates. This is sometimes referred to as pseudoreplication, which ignores the fact that the subsamples are correlated (i.e., they come from the same experimental unit) and may

lead to incorrect statistical and scientific inference (Veum et al., 2011). Therefore, testing of the main effects in this study was accomplished using a modified two-factor ANOVA

(Kuehl, 2000). The two-way interaction between conservation management practice and landscape position was used as the random error term to account for the lack of plot-level replication, and an additional random effect accounts for the subsamples. The two-factor

ANOVA analyses were conducted using SAS Proc Mixed (SAS® software, Version

9.1.3, SAS Institute Inc., Cary, NC, USA.) and Tukey's HSD procedure was used as the multiple comparison test to evaluate pair-wise differences among all treatment means.

Further details of this modified ANOVA and a full justification for this model specification can be found in Veum et al. (2011) and the Appendix. All analyses were conducted at the α=0.05 significance level. To examine the relationships among variables, linear regression and correlation analysis was considered across all subsamples.

For comparisons between soil fractions (e.g., C:N ratios or PAO-C to SOC comparisons), paired t-tests were conducted (n=36). To validate regression assumptions, residual analyses were conducted, including the Shapiro-Wilk test for normality, and no departures were found.

Results

Treatment effects on soil properties

In the surface layer, mean ρb and field-moist water content were significantly reduced under grass and agroforestry VFS relative to no-till. Mean WSAAD, WEOC, WEOCP, the

C:N ratio of WEOM and the C:N ratio of WEOMP were significantly greater under VFS relative to no-till, reflecting an approximate 2-fold increase for mean WSAAD, WEOC

and WEOCP (Table 4.1), and an approximate 3-fold increase of the WEOM C:N ratio under grass and agroforestry VFS (Table 4.2). When expressed on a gravimetric, soil volume or soil mass basis, similar trends among treatments for WEOC and WEOCP were found; therefore, only the equivalent soil mass results are reported. No significant differences were found among conservation management practices for mean pH, air-dried water content (Table 4.1), sand content (ranging from 75 g kg-1 to 85 g kg-1), or the proportion of sand-free soil retained in the PAO fraction (73.2, 74.1 and 73.8 % for grass

VFS, agroforestry VFS and no-till, respectively).

In the subsurface layer, mean ρb declined significantly under agroforestry VFS, and field moist water content declined significantly under grass and agroforestry VFS relative to no-till. Similar to the surface layer, mean WSAAD was approximately 2-fold greater under VFS than no-till (Table 4.1). In addition, the WEOM C:N ratio was significantly greater under grass VFS than agroforestry VFS and no-till, reflecting a 1.7 and 2.2-fold increase, respectively (Table 4.2). No other significant differences were found among the conservation management practices for the properties investigated.

The general trends among conservation management practices for WSAAD and

WSAFM varied slightly depending on the form of expression (gravimetric, soil volume, soil mass and organic carbon normalized values) in both soil depth layers. These results can be found in Table 4.3. Additionally, no significant differences were found among landscape positions for any variables in this study in either soil depth layer.

Table 4.2 Mean C:N ratios in whole soil, the particulate, adsorbed and occluded (PAO) fraction, water-extractable organic matter from whole soil (WEOM), and water-extractable organic matter from the particulate, adsorbed and occluded fraction (WEOMP). Values followed by a different capital letter within columns and for a given soil sampling depth and phase are significantly different. Values followed by a different lowercase letter within rows are significantly different (using

Tukey’s HSD) at α = 0.05 and the standard error of the mean is stated in parentheses.

Solid phase C:N ratios Overall Grass Agroforestry No-Till (n=36) (n=12) (n=12) (n=12) Surface (0-5 cm) Whole soil 11.4 B 10.8 (0.43) B,a 11.7 (0.36) B,a 11.7 (0.35) A,a PAO fraction 12.3 A 12.3 (0.13) A,a 12.6 (0.20) A,a 12.0 (0.17) A,a Subsurface (5-13 cm) Whole soil 11.6 A 11.3 (0.18) A,a 11.4 (0.36) A,a 12.1 (0.40) A,a PAO fraction 11.5 A 11.5 (0.17) A,a 11.2 (0.13) A,a 11.6 (0.30) A,a Water-extract C:N ratios Overall Grass Agroforestry No-Till (n=36) (n=12) (n=12) (n=12) Surface (0-5 cm) WEOM 4.2 B 5.5 (0.67) B,a 5.1 (0.48) B,a 1.8 (0.18) B,b

WEOMP 10.6 A 10.8 (0.55) A,a 12.8 (0.29) A,a 9.0 (0.27) A,b Subsurface (5-13 cm) WEOM 6.4 B 9.3 (0.92) A,a 5.6 (0.49) B,b 4.2 (0.31) B,b

WEOMP 9.1 A 8.4 (0.54) A,a 10.2 (0.24) A,a 8.8 (0.18) A,a

Table 4.3 Means of field-moist water stable aggregates (WSAFM) and air-

dried water stable aggregates (WSAAD) by conservation management

practice expressed on a gravimetric, soil volume, and soil mass basis, as

well as WSAFM and WSAAD normalized to soil organic carbon (SOC),

water-extractable organic carbon (WEOC), and particulate, adsorbed and

occluded organic carbon (PAO-C). Values followed by a different

lowercase letter within columns and for a given soil layer are significantly

different (using Tukey’s HSD) at α = 0.05. The standard error of the mean

is stated in parentheses

Comparison of whole soil to the PAO fraction

In the surface layer, aggregated across all samples (n=36), the sand-free concentration of organic carbon (OC) in the PAO fraction (37.5 g OC kg-1 PAO) was significantly different from the concentration of OC in sand-free soil (33.5 g OC kg-1 soil). However, compared within conservation management practices (n=12 per treatment), results varied.

The PAO fraction contained significantly greater concentrations of OC in the agroforestry VFS (43.2 g OC kg-1PAO and 35.8 g OC kg-1soil, respectively) and grass

VFS (41.5 g OC kg-1PAO and 36.2 g OC kg-1 soil, respectively), while no significant difference was found in the no-till soil, suggesting organic carbon was preferentially sequestered in the PAO fraction only under VFS. No significant differences were found in the concentration of OC between fractions in the subsurface layer.

In the surface layer, the solid phase C:N ratio of the PAO fraction was significantly greater than the C:N ratio of whole soil across all samples (n=36); however, results varied within conservation management practices (n=12 per treatment). Agroforestry and grass

VFS exhibited significantly greater solid phase C:N ratios in the PAO fraction relative to whole soil, while no significant difference in the C:N ratios was found under no-till. In the subsurface layer, no significant differences were found in the solid phase C:N ratios between whole soil and the PAO fraction (Table 4.2).

In the water extracts, WEOMP exhibited significantly greater OC concentrations than

WEOM (adjusted for sand content) across all samples and within the VFS soils in the surface layer. No significant differences were found in the no-till soil or in the subsurface layer. The C:N ratio was significantly greater in WEOMP than WEOM (aggregated across all samples, n=36), in the surface and subsurface layers. Within the soil surface

layer of conservation management practices (n=12 per treatment), WEOMP exhibited significantly greater C:N ratios than WEOM under grass VFS, agroforestry VFS, and no- till, representing a 2.0, 2.4 and 5.0-fold increase, respectively. In the subsurface layer, the

C:N results were similar to the surface layer, except that the C:N ratio of WEOM did not decline as dramatically under grass VFS as under agroforestry VFS or no-till (Table 4.2).

Relationships among soil properties

General regression and correlation relationships among soil properties were explored.

In the surface layer, aggregated across all samples (n=36), WSAAD demonstrated the strongest correlations with ρb (r = –0.83), WEOC (r = 0.80), PAO-C (r = 0.80) and

WEOCP (r = 0.74), while only a moderate correlation was found with SOC (r = 0.62, p =

0.002). Scatterplots and overall simple regression relationships between WSAAD and

SOC, PAO-C, WEOC and WEOCP can be found in Figure 4.2. The only variable significantly correlated with WSAFM was field-moist water content (r = 0.38, p = 0.021).

Similar, but weaker, relationships were found in the subsurface layer. Multiple regression relationships were evaluated for WSAAD and WSAFM using several soil properties. The greatest variability in WSAAD was explained using ρb and the organic carbon fractions

WEOC, WEOCP and PAO-C. The greatest variability in WSAFM was explained by field moist water content and WEOC.

Surface (0 – 70 kg soil m-2):

2 WSAAD = 84.0 – 55.6 ρb + 1.75 WEOC, (R = 0.80)

2 WSAAD = 83.0 – 56.1 ρb + 8.68 PAO-C, (R = 0.72)

2 WSAFM = 27.8 + 15.8 WCFM + 7.14 WEOC, (R = 0.35)

Figure 4.2 Scatter plots and regression relationships for air-dried water stable aggregates (WSAAD, kg m-2) and (a) total soil organic carbon (SOC, kg m-2), (b) particulate, adsorbed and occluded carbon (PAO-C, kg m-2), (c) water-extractable organic carbon (WEOC, g m-2) and (d) water

-2 extractable organic carbon from the PAO fraction (WEOCP, g m ) in the surface layer (0-70 kg soil

-2 m ) using all subsamples (n=36). Sand-free values are presented for PAO-C and WEOCP. Grass vegetated filter strips (VFS) are represented by black circles, agroforestry VFS by open circles and no-till by grey triangles. The slope parameters are statistically significant (p < 0.001).

Subsurface (70 – 130 kg soil m-2):

2 WSAAD = 67.4 – 43.9 ρb + 2.65 WEOC, (R = 0.60)

2 WSAAD = 75.3 – 46.0 ρb + 21.3 WEOCP, (R = 0.45)

2 WSAFM = 14.7 + 2.09 WCFM + 1.14 WEOC, (R = 0.35)

- where WSAAD, WSAFM, SOC, PAO-C and field-moist water content (WCFM) are in kg m

2 -2 -3 , WEOC and WEOCP are in g m , and ρb is in g cm . All parameters were significant (p

< 0.001). In addition, strong correlations were found between PAO-C and WEOCP (r =

0.86, p < 0.0001), as well as PAO-C and WEOC (r = 0.78, p < 0.0001). Overall, WEOC represented approximately 0.5% of SOC in the study soils.

Comparison of WSAFM, WSAAD and the PAO fraction

Across all surface samples (n=36), the proportion of sand-free soil remaining after sieving was significantly different among all three disaggregation methods in the order

WSAFM (mean = 81.3%) > PAO fraction (mean = 73.7%) > WSAAD (mean = 57.6%).

Within conservation management practices (n = 12 per treatment), the proportion of sand-free soil remaining after sieving was significantly greater for WSAFM than in the

PAO fraction or WSAAD under all three conservation management practices. The difference in the amount of sand-free soil remaining between the WSAAD and the PAO fraction was not significantly different under agroforestry VFS (mean = 73.5% and

74.1%, respectively) or grass VFS (mean = 73.2% and 70.2%, respectively). In contrast, the PAO fraction retained a significantly greater proportion of sand-free soil than WSAAD under no-till (mean = 73.8% and 29.0%, respectively). Across all samples, WSAFM was significantly greater than the corresponding WSAAD values, and the reduction in

aggregate stability from air-drying samples was far more dramatic under no-till compared to VFS. In the subsurface layer, the proportion of sand-free soil remaining after sieving was significantly lower for WSAAD (mean = 41.1%), but not significantly different between WSAFM (mean = 74.1%) and the PAO fraction (mean = 75.2%). Mean WSAFM was significantly greater than WSAAD within all conservation management practices, and the most dramatic difference was evident under no-till (mean = 78.4% and 24.3%, respectively).

Discussion

Previous work at this study site found VFS soils contained greater stocks of PAO-C compared to no-till, even though no significant differences in SOC stocks were found.

Additionally, the PAO fraction had a greater C:N ratio than whole soil, suggesting increased lability of SOM in the PAO fraction (Veum et al., 2011). This study demonstrated that both solids and aqueous extracts from the PAO fraction have greater concentrations of organic carbon and greater C:N ratios than whole soil. Together, these results indicate that the organic matter is preferentially sequestered in the PAO fraction, and that it is less degraded than organic matter in whole soil (Elliott, 1986). Within conservation management practices, grass and agroforestry VFS soils demonstrated preferential concentration of solid and water soluble SOM in the PAO fraction relative to whole soil as well as greater C:N ratios in the solid and aqueous PAO fractions relative to whole soil. In contrast, no-till only demonstrated a significantly greater aqueous C:N ratio in the PAO fraction. These results confirm the greater lability of the PAO fraction overall and suggest that the sequestration of labile and less decomposed organic matter is

enhanced under the perennial vegetation of VFS relative to no-till. This may be the result of differing quantity and/or quality of organic matter inputs to the soil, or the result of differences in the decomposition rate due to protection in soil aggregates (Elliott, 1986;

Landgraf et al., 2006).

The PAO fraction is separated by slaking and 2 minutes of sieving, and it consists of stable aggregates > 53µm. Following the aggregate size classification used by Six et al.

(1998), this fraction contains free light fraction SOM, intra-aggregate particulate organic matter (POM), and mineral-associated SOM. The discarded clay and silt sized fraction (<

53 µm) is presumably more humified (Six et al., 2001; Carter, 2002) and by definition, does not contain POM (Cambardella and Elliott, 1992). Typically, larger aggregates generally contain greater concentrations of total SOC (Miller and Dick, 1995; Puget et al., 1995), younger, more labile SOC (Puget et al., 1995; Angers and Giroux, 1996), and more WEOC (Gregorich et al., 2003) than smaller aggregates. Thus, through a rapid size separation technique, the PAO fraction is expected to be enriched in younger, less decomposed and more labile SOM relative to whole soil.

In contrast to the PAO fraction, the WSAFM and WSAAD fractions retain stable macroaggregates in the 250 – 2000 µm range. Therefore, in addition to removing the clay and silt-associated SOM (< 53um) removed by the PAO separation, the fine inter- aggregate POM from unstable microaggregates (53 – 250 µm) is also lost. Under VFS, there were no significant differences in the proportion of soil lost between the PAO fraction and WSAAD. However, under no-till, the WSAAD procedure lost nearly three times more soil than the PAO separation, suggesting that a large proportion of no-till soil consists of weak microaggregates. The fine inter-aggregate POM associated with this

intermediate size fraction (53-250 µm) is potentially more labile than the < 53 µm fraction, and less labile than macroaggregate POM and light fraction SOM (Cambardella and Elliott, 1992; Six et al., 2000a) remaining after sieving. Although the PAO separation subjects soil to less sieving time than the WSA separation, fast-wetting (i.e., slaking) is considered more disruptive than mechanical sieving after slow wetting (Oades and

Waters, 1991; Bouajila and Gallali, 2008). Additionally, in this study, air drying of soil resulted in significantly lower aggregate stability for all soils, although the effect was much more dramatic in the no-till soil. Under no-till, the WSAAD method led to large soil losses in the 0-53 plus 53-250 µm range; only a small portion of no-till soil contained macroaggregates strong enough to withstand the disruptive forces of the WSAAD test.

Similarly, Haynes (1993) reported that arable soils were less stable when air dried than when field-moist, and that the reverse was true for high organic matter soils.

The greater aggregation under VFS may be partly attributed to the role of roots in macroaggregate formation. Roots and fungal hyphae contribute directly to soil aggregation by exuding organic compounds that act as cementing agents, promoting microbial activity (Angers and Caron, 1998) and binding microaggegates into macroaggregates (Elliott, 1986). Roots are known to be greater under perennial vegetation than in cropped soil (DuPont et al., 2010). As a result, arable soils are more susceptible to the effects of surface slaking than soils under perennial vegetation (Tisdall and Oades, 1980). Together, these processes contribute to more frequent macroaggregate turnover and SOM mineralization (Beare et al., 1994) under no-till. Therefore, in no-till soils, microaggregates are less likely to combine into strong macroaggregtes, and labile

SOM inputs are less protected and more rapidly decomposed relative to soil under VFS.

The relationships among ρb, SOM content, porosity, water content and aggregate stability are complex. In this study, treatment effects using field-moist aggregates were contradictory to air-dried results, and overall, the field-moist soil exhibited greater mean aggregate stability values than air-dried soil, suggesting that antecedent water content may have an effect on soil cohesion (Kemper and Rosenau, 1986). The clay mineralogy at this site is dominated by smectite, and antecedent moisture may reduce the effects of slaking and swelling in smectitic soils, leading to greater aggregate stability (Reichert et al., 2009). No-till soil often has greater water content due to decreased transpiration

(Munoz-Carpena et al., 1993), increased compaction (Blanco-Canqui et al., 2009b) and decreased macroporosity (Anderson et al., 2009) relative to soil under perennial vegetation. Overall, the WSAFM results suggest that no-till may exhibit equivalent or greater aggregate stability than VFS soils under certain field conditions.

No landscape effects were detected in the variables evaluated in this study. This may be due to the gentle (0–3%) slope of the study site, the small catchment area, and the similarity of the soil series at this site. Due to soil forming processes, the soil type distribution mirrors changes in landscape position at this site, and the parent material

(loess) from which the soils form is the same. Although the proportion of illite and smectite may vary slightly across landscape positions in agroecosystems on claypan soils

(Baer and Anderson, 1997), smectitic and illitic clays exhibit similar shrink-swell characteristics; thus, differences in aggregation should result from changes in SOM, not subtle differences in mineralogy (Oades and Waters, 1991). No significant differences in

SOM quantity or quality have been observed across landscape positions at this site

(Veum et al., 2011 and this study); therefore, it is not surprising that landscape effects on aggregation were not detected.

The strong positive correlations among WSAAD, PAO-C, WEOC and WEOCP (Fig.

4.2) suggest that these organic carbon fractions play an important role in aggregate stability, that enhanced aggregate stability is protecting these fractions from degradation, or both. Although most studies have focused on the correlation between hot water extracts and aggregate stability (Haynes, 2000 and others), the strong positive correlation between cold WEOC and WSAAD in this study suggests that WEOC quantity may be a good predictor of aggregate stability in some soils. Not only does SOM act as a binding agent in the formation of aggregates, but SOM fractions may also stabilize existing aggregates by reducing slaking stress (Oades, 1984), reflecting the positive feedback mechanism between SOM and soil aggregation. Therefore, even if not directly involved in the formation of soil aggregates, some SOM fractions may play a role in stabilization of existing aggregates. Furthermore, few studies have examined the quality (e.g., biodegradability, lability) of WEOM from within aggregates to WEOM from whole soil

(Marschner and Kalbitz, 2003), and these results suggest WEOMP is more labile and less degraded than WEOM from whole soil.

Conclusions

This study confirms a shift in SOM quality toward greater stocks of labile, aggregate- associated SOM under VFS ten years after establishment in a no-till agroecosystem.

Detecting management effects on properties related to soil quality and soil organic carbon is critical for effective agroecosystem management, and in this study, the PAO

separation, aggregate stability and cold-water extracts served as rapid, cost effective indicators of short-term changes in soil quality/soil carbon change. Overall, this study provides insight into the potential role of VFS and no-till agriculture in carbon sequestration. Future investigations into SOM characterization and microbial community structure may further elucidate the ecosystem-level effects of conservation management practices in agroecosystems on claypan soils.

CHAPTER 5

MICROBIAL ENZYME ACTIVITY AND LABILE SOIL ORGANIC MATTER UNDER VEGETATIVE FILTER STRIPS AND NO-TILL AGRICULTURE

This chapter is in preparation for submission to Soil Biology and Biochemistry: Veum, K.S., Goyne, K.W., Kremer, R.J., Motavalli, P.P. Microbial enzyme activity and labile soil organic matter under vegetative filter strips and no-till agriculture.

Abstract

Despite the widespread use of conservation management practices to enhance profitability and sustainability in agroecosystems, few studies have compared soil quality among different conservation management practices. This study compared the effects of no-till, grass vegetative filter strips (VFS) and agroforestry VFS and landscape position on dehydrogenase activity (DH), phenol-oxidase activity (PO), permanganate-oxidizable organic carbon (POXC) and ultraviolet measures of water-extractable organic matter

(WEOM) chemistry. In addition, POXC content and WEOM chemistry were compared between the particulate, adsorbed and occluded (PAO) fraction and whole soil. The VFS soils contained significantly greater POXC pools than no-till. Grass VFS exhibited the greatest DH and PO, indicating enhanced overall microbial activity, potentially due to the chemical composition of WEOM. The agroforestry VFS exhibited the greatest measures of aromaticity and humification in the WEOM fraction. The PAO fraction was highly correlated with POXC and water extracts from the PAO fraction demonstrated greater aromaticity, oxygen substitution and molecular weight components relative to whole soil

extracts. Overall, this study demonstrates significant changes in SOM quality under VFS relative to no-till after 10 years, and that the PAO fraction represents a pool of partially processed, labile organic matter.

Introduction

In agroecosystems, soil organic matter (SOM) contributes to vital soil functions such as sustained biological activity, nutrient cycling, crop productivity and environmental quality (Doran and Parkin, 1994; Karlen et al., 1997). Conservation management practices demonstrate a wide range of environmental benefits (Lal et al., 1994; Holland,

2004), and have been shown to improve many aspects of SOM quality over conventional tillage (Angers et al., 1993a; Collins et al., 2000). More studies are needed, however, to elucidate the role of management practices such as agroforestry systems on SOM cycling

(Morgan et al., 2010). Total soil organic carbon (SOC) is often insensitive to short-term changes (Lal, 2004) or subtle differences among conservation management practices

(Veum et al., 2011); therefore, it is essential to identify effective and sensitive indicators of SOM quality that are relevant to the soil functions of interest (Karlen et al., 1997;

Schoenholtz et al., 2000). Studies have demonstrated the sensitivity of microbial enzyme activities (Jordan et al., 1995), labile KMnO4-oxidizable (POXC) organic matter (Culman et al., 2012), water-extractable (WEOM) organic matter (Embacher et al., 2007) and aggregate-associated SOM fractions (Cambardella and Elliott, 1994) to land use changes.

Nevertheless, the effects of land use on the composition and distribution of soil organic matter in soil aggregates is still not well understood (Helfrich et al., 2006). In particular,

little is known about the chemistry of WEOM from soil aggregates (Marschner and

Kalbitz, 2003).

The chemically oxidizable POXC fraction of SOM is influenced by agricultural practices and is used as an indicator of labile SOM (Blair et al., 1995; Tirol-Padre and

Ladha, 2004; Culman et al., 2012). Often, POXC correlates better with other indicators of

SOM quality, such as aggregate stability (Chan et al., 2001), particulate organic carbon fractions or microbial biomass carbon (Culman et al., 2012) than total SOC. A broad analysis by Culman et al. (2012) indicates that POXC is more closely related to the 53-

250 µm size fraction of particulate organic matter than the larger macroaggregate fractions or the free light fraction, suggesting that POXC is a relatively processed pool of labile SOM. The fraction represented by POXC is influenced by experimental conditions including KMnO4 concentration and reaction time (Weil et al., 2003; Tirol-Padre and

Ladha, 2004). Overall, POXC serves as a rapid, simple and inexpensive colorimetric analysis that isolates a labile and “active” organic carbon fraction when using a low concentration of KMnO4 and a short incubation time (Weil et al., 2003).

The particulate, adsorbed and occluded (PAO) fractionation technique separates an organic carbon fraction (PAO-C) that has shown promise in differentiating among management practices in an agroecosystem where total SOC differences were not detectable (Veum et al., 2011). The PAO fraction consists of stable aggregates > 53µm and, following the aggregate size classification used by Six et al. (1998), contains intra- aggregate particulate organic matter, occluded mineral-associated SOM and the free light fraction. Organic matter in the discarded clay and silt sized fraction (< 53 µm) is presumably more humified (Six et al., 2001; Carter, 2002) and by definition, does not

contain POM (Cambardella and Elliott, 1992). Both solid and aqueous extracts from the

PAO fraction have demonstrated greater concentrations of organic carbon and greater

C:N ratios relative to whole soil, suggesting preferential sequestration of more labile

SOM in the PAO fraction (Veum et al., 2011 and Chapter 4).

Microbial uptake mechanisms require an aqueous environment; therefore, WEOM is potentially the most bioavailable fraction in soils (Marschner and Kalbitz, 2003) and represents a mobile fraction of SOM that may transport pollutants (Sposito and Weber,

1986; Chantigny, 2003). Components of cold (i.e., room temperature) WEOM, such as carbohydrates, polysaccharides and organic carbon, have demonstrated positive correlations with other indicators of SOM quality including aggregate stability (Miller and Dick, 1995 and Chapter 4), soil respiration (Cook and Allan, 1992) and microbial biomass C (Zak et al., 1990; Nishiyama et al., 2001). Aliphatic components of WEOM are associated with enhanced biodegradability, whereas aromatic moieties have been associated with increased recalcitrance, reduced bioavailablity (Kalbitz et al., 2003b;

Marschner and Kalbitz, 2003) and/or inhibitory effects on microbial enzyme activity

(Toberman et al., 2008). The proportion of aliphatic and aromatic WEOM constituents, however, varies as a result of organic inputs and the extent of degradation (Kalbitz et al.,

2003a). Many studies have evaluated WEOM from whole soil, yet few studies have examined the quality (e.g., biodegradability, lability) of WEOM from aggregate size fractions (Marschner and Kalbitz, 2003).

Ultraviolet-visible absorbance spectroscopy has been used extensively to assess the degree of humification of humic substances isolated from soils and aquatic systems. The electron transfer absorption band, typically represented by 254 (A254) or 280 nm, and the

associated molar absorptivity (i.e., SUVA), have been positively correlated with aromatic content (Chin et al., 1994; Peuravuori and Pihlaja, 1997), molecular size (Chin et al.,

1994; Peuravuori and Pihlaja, 1997), hydrophobicity (Guo and Chorover, 2003) and biological recalcitrance (Kalbitz et al., 2003b). In contrast, the ratio of absorbance at 280 to 365 nm (E2:E3) has been negatively correlated with aromatic content (Peuravuori and

Pihlaja, 1997) and molecular size (Peuravuori and Pihlaja, 1997), and positively correlated with aliphatic content (Guo and Chorover, 2003). Furthermore, the ratio of absorbance in the electron transfer band at 253 nm to the benzenoid band at 203 nm

(ET:BZ) is indicative of the degree of substitution in aromatic rings with oxygen- containing functional groups (Korshin et al., 1997; Fuentes et al., 2006). In addition to the study of isolated humic substances (Chen et al., 1977 and many others), ultraviolet indices have been applied to the study of plant (Hunt and Ohno, 2007; He et al., 2009) and whole soil (Embacher et al., 2008; Takacs and Fuleky, 2010) extracts; however, little is known about the application of these measurements to WEOM from different soil aggregate sizes.

Microorganisms are key players in nutrient cycling and decomposition of SOM. They respond quickly to changes in environmental conditions and, therefore, provide an excellent early indicator of soil improvement or soil degradation in agroecosystems

(Jordan et al., 1995; Kennedy and Smith, 1995; Pankhurst et al., 1996). Microbial enzyme activities are an expression of the microbial community, reflecting metabolic factors and nutrient availability (Caldwell, 2005), and effectively demonstrating rapid responses to land use (Dick, 1997). Dehydrogenases are endocellular enzymes that facilitate the transfer of protons and electrons from substrates to acceptors in the electron

transport system and represent a broad range of microbial oxidative activities.

Measurement of dehydrogenase (DH) activity serves as a proxy for overall microbial activity and reflects carbon flux through the soil ecosystem as well as changes in SOM quantity and quality (Casida, 1977; Tabatabai and Dick, 2002).

Phenol oxidases are keystone enzymes involved in the partial to complete ligninolytic degradation and/or polymerization of lignin and lignin degradation products in the presence of oxygen (Smith and Stotz, 1949; Ander and Eriksson, 1976; Palmer and

Evans, 1983). Phenol oxidases are extracellular enzymes produced by fungi and bacteria

(i.e., Phanerochaete and Actinomyces, respectively) and they are found in plant tissues.

Phenol oxidase (PO) activity generally increases with increased oxygen levels; however, competing factors may inhibit PO activity including cooler temperatures (Freeman et al.,

2001a), acidic pH (Pind et al., 1994; Williams et al., 2000) and reduced water content

(Toberman et al., 2008). When PO activity is inhibited, phenolics accumulate in the soil and inhibit other enzymes and overall SOM degradation (Freeman et al., 2001b;

Marschner and Kalbitz, 2003).

Soil formation is the result of geomorphic and pedogenic processes, and landscape influences the spatial distribution of soil properties in agroecosystems (Yoo et al., 2006;

Hancock et al., 2010). Therefore, landscape position is an important factor in field and landscape-level experimental design. Another influential consideration in the evaluation of soil properties are the calculations used to express the results. Using gravimetric (e.g., g per kg soil) versus areal expressions (e.g., g per m2 by equivalent soil volume or equivalent soil mass) may lead to different results and inference (Ellert and Bettany,

1995; Ellert et al., 2001a; Veum et al., 2011). Although expressing results on an areal,

equivalent soil mass basis is generally accepted as the most appropriate comparison, many studies lack the requisite bulk density (ρb) data.

The objectives of this study were (1) to evaluate the effects of three conservation management practices (i.e., grass vegetative filter strips (VFS), agroforestry VFS and no- till) and four landscape positions (i.e., summit, shoulder, backslope and footslope) on enzyme activities as well as POXC content and WEOM quality from whole soil and the

PAO fraction, (2) to evaluate relationships among these variables and associated soil properties, (3) to compare POXC content and WEOM quality between whole soil and the

PAO fraction. We hypothesized that (1) increased plant residues and exudates under perennial vegetation (i.e., VFS) would result in greater enzyme activity, increased POXC and shifts in WEOM quality under VFS relative to no-till, (2) that enzyme activities would exhibit positive correlations with labile organic carbon fractions, and (3) that the

PAO fraction would demonstrate increased POXC and WEOM lability relative to whole soil.

Materials and Methods

Study site

The Greenley Memorial Research Center experimental site, located in Knox County,

Missouri, USA (40˚ 01’N, 92˚11’W), has been cultivated for more than 60 years and has been under a no-till, maize (Zea mays. L.) – soybean [Glycine max (L.) Merr.] rotation since 1994. This site was originally designed as a paired watershed study in 1991, and it is established on soils that are characterized by a claypan horizon occurring at depths ranging from 20 to 62 cm (Udawatta et al., 2006). Claypan soils have an abrupt, argillic

subsurface horizon that impedes drainage and induces lateral, subsurface water flow

(Blanco-Canqui et al., 2002; Arnold et al., 2005). Claypan soils cover over 16,000 km2 in the central U.S.; globally, soils with restrictive sub-horizons encompass approximately

2.9 million km2 (USDA-NRCS, 2006). Soils in the study area were formed in loess underlain by weathered glacial till (Watson, 1979) and are designated by the USDA-

NRCS as Putnam silt loam (fine, smectitic, mesic Vertic Albaqualfs) in the higher landscape positions (summit and shoulder) and Kilwinning silt loam (fine, smectitic, mesic Vertic Epiaqualfs) in the lower positions (backslope and footslope). Under the

Food and Agricultural Organization (FAO) soil classification system, these soils are classified as Vertic Luvisols and Gleyic Luvisols. The clay mineralogy of the soil surface horizon consists of 63% smectite, 33% illite and 4% kaolinite (Chu, 2011).

The study site is on a 9.25 ha catchment consisting of three small, gently sloping (0 –

3 %), contiguous subcatchments. In 1997, upland grass and agroforestry VFS were planted in the west and center subcatchments, respectively, and the east subcatchment was left under no-till management to serve as the control (Fig. 5.1). Each upland VFS is approximately 4.5 m wide and planted on the contour with a VFS to crop width ratio ranging from 1:8 to 1:5, exceeding recommendations for upland filter strips in cropped land (USDA-NRCS, 2010). Overall, the filter strips account for approximately 8 to 10% of the catchment area and have a combined VFS width of 22.5 to 27 m in each subcatchment. To inhibit vegetative succession, the filter strips are mowed periodically and treated with herbicides before planting and after harvest. The grass VFS consists of redtop (Agrostis gigantea Roth), brome (Bromus spp.), and birdsfoot trefoil (Lotus

Figure 5.1 Greenley Memorial Research Center study site in Knox County, Missouri (star). Black dots indicate sampling locations. Thick black lines denote grass vegetative filter strips (VFS), grey lines denote agroforestry VFS and dotted lines represent soil series boundaries (adapted from Veum et al., 2011).

corniculatus L.). The agroforestry VFS contain pin oak (Quercus palustris Muenchh.), swamp white oak (Q. bicolor Willd.) and bur oak (Q. macrocarpa Michx.) in addition to the grasses. A summary description of the site and comprehensive management details can be found in Veum et al. (2009) and the references therein.

Soil sampling and laboratory analyses

Surface (0 – 5 cm) and subsurface (5 – 13 cm) soil samples were collected after maize harvest in November, 2007 from four landscape positions (summit, shoulder, backslope and footslope) and three conservation management practices (grass VFS, agroforestry

VFS and no-till). In order avoid the potentially confounding effects of VFS on adjacent soil (i.e., the potential effect of being spatially upslope or down-slope from the VFS), samples representing no-till management were collected from equivalent landscape positions in the no-till subcatchment (i.e., subcatchment with no VFS). Three subsamples were collected from each combination of landscape position and conservation management practice, totaling 36 sampling locations (Fig 5.1). Soil cores were collected using 7.62 cm diameter aluminum rings for the surface (0 – 5 cm) and subsurface (5 – 13 cm) following the soil core method of Grossman and Reinsch (2002). Surface residues and litter were removed prior to soil sample collection. Soil gravimetric water content was determined by oven drying soil cores at 105˚C. Soil ρb was measured on soil core samples following the method of Grossman and Reinsch (2002). Soil properties were calculated on a gravimetric (i.e., g per kg soil) and an areal basis (i.e., equivalent soil volume and equivalent soil mass per m2) as outlined in Lal et al. (2001a). For general characterization of soil properties, one kg of bulk soil from each sampling depth at each

location was collected using a hand trowel. Field moist soil samples, used for determination of microbial enzyme activity, were stored in the dark at 4 ºC until analysis.

Cation exchange capacity, base saturation and particle size were determined on air- dried, ground soil samples following USDA-NRCS procedures (Burt, 2004). Total SOC and PAO-C were determined as described in Veum (2011). In brief, soil was sieved (< 2 mm), air-dried, ground and analyzed on a LECO TruSpec CN® combustion analyzer

(Leco Corp., St. Joseph, MI, USA) for SOC and total nitrogen measurements. Following a simplified aggregate separation method used by Six et al. (1998), the PAO fraction was separated by slaking and sieving 100 g of air-dried soil (< 2mm) over a 23 cm diameter,

53 µm sieve for 2 min. The organic carbon and total nitrogen content of the PAO fraction

(PAO-C and PAO-N) were determined using a LECO® combustion analyzer and the mass of PAO-C per unit mass of soil was calculated on a sand-free basis (g PAO-C kg-1 soil). Due to the calculation of PAO-C on a sand-free basis, the PAO-C values are greater than total SOC values from the same soil samples that are unfractionated. Sand content was not significantly different among treatments. General soil characteristics, summarized by conservation management practice, are shown in Table 5.1 (surface layer) and Table 5.2 (subsurface layer).

Air-dried water stable aggregates were measured on undisturbed core samples following an adapted method of Kemper and Rosenau (1986) as described in Chapter 4.

In short, field-moist soil was hand-sieved (< 6 mm) and air-dried. Aggregates were allowed to slowly rewet via capillary action for 10 minutes in deionized water on a 250

µm sieve and wet-sieved for ten minutes at 30 seconds per cycle. The remaining sample mass was corrected for sand content and initial water content.

Table 5.1 Mean soil properties in the surface (0–70 kg m-2) layer of the study soils by conservation management practice. Values followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05 (n=12 per treatment). Values without lowercase letters were determined using composite samples (n=4 per treatment).The standard error of the mean is stated in parentheses. Data partially obtained from Veum et al. (2011).

Surface (0–70 kg m-2 soil) Soil Property Grass Agroforestry No-Till Textural Class Silt Loam Silt Loam Silt Loam -1 a CECNaOAc (cmolc kg ) 23.3 (1.20) 23.6 (0.51) 21.0 (0.40)

Base SaturationNaOAc (%) 92.8 (2.33) 87.8 (1.58) 90.0 (2.27) Clay Content (g kg-1) 242.5 (17.59) 239.2 (9.42) 218.7 (6.22)

Soil pHw 6.6 (0.11)a 6.0 (0.16)a 6.3 (0.12)a Bulk Density (g cm-3) 1.10 (0.027)b 1.02 (0.023)b 1.29 (0.020)a -2 b WSAAD (kg m ) 46.5 (2.23)a 46.9 (1.43)a 19.9 (2.46)b SOC (kg m-2)c 2.14 (0.060)a 2.08 (0.076)a 1.85 (0.076)a PAO-C (kg m-2)d 2.39 (0.099)a 2.57 (0.150)a 1.52 (0.063)b POXC (g m-2)e 135.6 (2.23)a 128.2 (3.41)a 89.7 (1.41)b WEOC (g m-2)f 11.57 (0.882)a 11.81 (0.943)a 6.16 (0.432)b DH (mg m-2 hr-1)g 833.1 (76.44)a 491.5 (33.15)b 429.5 (33.16)b PO (mol m-2 h-1)h 2.08 (0.107)a 0.78 (0.053)b 1.03 (0.143)b a Cation exchange capacity, b Water stable aggregates, c Soil organic carbon, d Particulate, e adsorbed and occluded (PAO) organic carbon (sand-free basis), KMnO4-oxidizable organic carbon, f Water-extractable organic carbon, g dehydrogenase activity, h phenol oxidase activity

Table 5.2 Mean soil properties in the subsurface (70–130 kg m-2) layer of the study soils by conservation management practice. Values followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05 (n=12 per treatment). Values without lowercase letters were determined using composite samples (n=4 per treatment).The standard error of the mean is stated in parentheses. Data partially obtained from Veum et al. (2011).

Subsurface (70–130 kg m-2 soil) Soil Property Grass Agroforestry No-Till Textural Class Silt Loam Silt Loam Silt Loam -1 a CECNaOAc (cmolc kg ) 20.8 (1.03) 19.9 (0.60) 19.62 (0.63)

Base SaturationNaOAc (%) 96.2 (0.58) 96.2 (2.9) 92.2 (2.46) Clay Content (g kg-1) 246.8 (17.49) 226.2 (9.53) 227.8 (0.47)

Soil pHw 7.0 (0.09)a 7.0 (0.05)a 6.9 (0.08)a Bulk Density (g cm-3) 1.30 (0.011)ab 1.25 (0.024)b 1.39 (0.024)a -2 b WSAAD (kg m ) 30.4 (2.67)a 29.1 (1.63)a 14.6 (1.14)b SOC (kg m-2)c 1.26 (0.051)a 1.24 (0.044)a 1.13 (0.068)a PAO-C (kg/m-2)d 1.00 (0.035)a 0.98 (0.032)a 0.87 (0.059)a POXC (g m-2)e 59.7 (2.00)a 60.4 (2.38)a 52.8 (3.18)a WEOC (g m-2)f 7.06 (0.455)a 5.31 (0.618)a 4.61 (0.196)a DH (mg m-2 hr-1)g 327.1 (26.68)a 275.9 (32.86)a 116.2 (21.04)b PO (mol m-2 h-1)h 1.70 (0.214)a 0.63 (0.052)b 1.24 (0.170)ab a Cation exchange capacity, b Water stable aggregates, c Soil organic carbon, d Particulate, e adsorbed and occluded (PAO) organic carbon (sand-free basis), KMnO4-oxidizable organic carbon, f Water-extractable organic carbon, g dehydrogenase activity, h phenol oxidase activity

Water-extractable organic matter was extracted from whole soil (WEOM) and the

PAO fraction (WEOMP) following the general extraction method of Burford and

Bremner (1975) as described in Chapter 4. In brief, duplicate 20 g samples of air-dried and sieved soil (< 2 mm) were extracted on an end-to-end shaker for 1 hr using deionized, ultrapure water at a 1:2 (w/v) ratio at 20ºC. The supernatant was centrifuged and filtered in series through prewashed 0.45 µm and 0.2 µm Whatman Puradisc™ filters to remove suspended colloidal material, acidified to ca. pH 2 and analyzed for non- purgeable organic carbon from whole soil (WEOC) and the PAO fraction (WEOCP) and total nitrogen using a Shimadzu TOC-V™ liquid carbon analyzer (Shimadzu Corp.;

Kyoto, Japan).

Ultraviolet-visible absorbance measurements were collected on the same water- extractable organic matter described above by acidifying the filtrate to ca. pH 7, diluting as necessary with ultrapure deionized water, and scanning at 2 nm increments from 200 nm to 700 nm on a Genesys8™ (Thermo-Fisher Scientific, Yokohama, Japan) spectrophotometer using a 1 cm quartz cuvette. Absorbance indices were used to characterize the chemistry of WEOM and WEOMP (Chin et al., 1994; Korshin et al.,

1997; Peuravuori and Pihlaja, 1997) including the absorbance at 254 nm (A254 and

A254P, respectively), the ratio of absorbance at 280 to 365 nm (E2:E3 and E2:E3P, respectively), and the ratio of absorbance at 253 to 203 nm (ET:BZ and ET:BZP, respectively). To examine the possibility of interference from nitrate at 203 nm in the

ET:BZ ratio, the correlation between the ET:BZ ratio and two ratios not affected by nitrate were examined per Korshin et al. (1997); ratios of 253 to 220 nm and 253 to 230 nm (wavelengths not affected by nitrate) demonstrated high linear correlation coefficients

with the ET:BZ ratio (r = 0.998, p<0.0001 and r = 0.987, p<0.0001, respectively), indicating nitrate was not influential. The ratio of absorbance at 465 nm to 665 nm

(E4:E6), widely used to indicate the degree of condensation of the aromatic carbon moieties (Peuravuori and Pihlaja, 1997), was considered; however, absorbance above 500 nm generally dropped below the limit of detection and were not meaningful for samples in this study. All samples were extracted at the same ratio of soil to solution volume; therefore, results for the extracts reflect the abundance of a given variable (i.e., organic carbon, aromaticity, etc.) from an equivalent soil mass.

Potassium permanganate-oxidizable organic carbon was determined on unfractionated soil (POXC) and the PAO fraction (POXCP) using a modified method of

Weil et al. (2003) to identify the potentially labile fraction of organic matter. Specifically,

2 ml of 0.2M KMnO4 - 1 M CaCl2 stock was added to a suspension of 2.5 g of air-dried and sieved (< 2mm) soil in 18 ml of deionized, ultra-pure water (for a final reaction concentration of 0.02M KMnO4). Samples were reacted on an end-to-end shaker for 15 min and centrifuged at 3,600 rpm for 5 min. The supernatant was diluted 100x with ultrapure, deionized water, and absorbance was determined at 550 nm in a 1 cm quartz cell on a Genesys-8™ spectrophotometer.

Dehydrogenase activity was measured on unfractionated soil following an adapted method of Casida (1977). In brief, 3 ml of 2% CaCO3 and 2 ml of 3% 2,3,5- triphenyltetrazolium chloride (TTC) were added to duplicate 6 g samples of field-moist soil. Samples were mixed using a glass rod and incubated at 37 ºC in the dark for 16 hr.

The incubation was terminated by adding 10 ml of methanol, vortex mixing the sample, waiting 30 minutes and vortexing the sample again. The soil suspension was filtered

through a 7 cm diameter Whatman 42™ filter paper using methanol to quantitatively transfer and extract the colored 2,3,5-triphenyl formazan (TPF) product. Sample absorbance was immediately read at 485 nm using a spectrophotometer. Blanks (reagents without soil) and controls (soil with all reagents except substrate) were analyzed concurrently. Water content was determined at the time of the DH assay by oven-drying duplicate 3 g samples at 105ºC for 24 h. Results were expressed as µg TTC converted to

TPF per gram of equivalent oven-dried soil per hour. Results were also calculated on a soil carbon basis to explore the relationship between DH activity and different soil carbon fractions (SOC, PAO-C, WEOC, and POXC).

Phenol oxidase activity was determined on unfractionated soil following an adapted method of Pind (1994). In summary, duplicate 2 g samples of field-moist soil were homogenized for 20 s in a Stomacher 80™ lab blender (Tekmar, Cincinnati, OH) with 20 ml of 50 mM sodium acetate. Using a pipette, 3 ml of soil suspension was transferred to digestion tubes to provide 2 assay replicates and a soil control (soil with reagents except substrate). Then 3 ml of 25 mM L-3,4-dihydroxyphenylalanine (L-DOPA) was added to the assay replicates. The tubes were incubated in the dark at 37 ºC on an end-to-end shaker for 1 hr, centrifuged at low speed for 2 minutes and filtered using prewashed 0.45

µm and 0.2 µm Whatman Puradisc™ syringe filters in series to remove particulates and colloidal material. Absorbance was read immediately at 460 nm. Blanks (reagents without soil) and controls (soil with all reagents except substrate) were analyzed concurrently. Water content was determined at the time of the PO assay by oven-drying duplicate 3 g samples at 105ºC for 24 h. Results were expressed as µmol L-DOPA converted to 3-dihydroindole-5,6-quinone-2-carboxylate (DIQC) per gram of equivalent

oven-dried soil per hour. Results were also calculated on a soil carbon basis to explore the influence of different carbon fractions (SOC, PAO-C, WEOC and POXC) on PO activity.

Statistical analysis

Similar to many environmental studies, the original paired watershed used for this study design lacks replication among the different treatment factor combinations.

Treating subsamples as true replicates, sometimes referred to as pseudoreplication, ignores the fact that the subsamples are correlated (i.e., they come from the same experimental unit) and may lead to incorrect statistical and scientific inference (Veum et al., 2011). Therefore, testing of the main effects in this study was accomplished using a modified two-factor ANOVA (Kuehl, 2000). The two-way interaction between conservation management practice and landscape position was used as the random error term to account for the lack of plot-level replication, and an additional random effect accounts for the subsamples. The two-factor ANOVA analyses were conducted using

SAS Proc Mixed (SAS® software, Version 9.1.3, SAS Institute Inc., Cary, NC, USA.) and Tukey's HSD procedure was used as the multiple comparison test to evaluate pair- wise differences among all treatment means. Further details of this modified ANOVA and a full justification for this model specification can be found in Veum et al. (2011) and the Appendix. All analyses were conducted at the α=0.05 significance level.

To examine the interactive effects of conservation management practices and WEOC quantity on WEOC chemistry (i.e., A254, SUVA, E2:E3 and ET:BZ) from unfractionated soil and the PAO fraction, a form of analysis of covariance (ANCOVA) was applied

100

using SAS Proc Mixed. Data from all subsamples within each treatment were combined for the comparison (n=12 per conservation management practice). Linear correlation and regression analysis was considered across all subsamples (n=36), to examine the relationships among soil properties. To validate the regression assumptions, residual analyses were conducted, including the Shapiro-Wilk test for normality, and no departures were found. The correlation due to subsampling was not accounted for in the

ANCOVA or linear regression models. The absorbance and molar absorptivity at 254 and

280 nm were compared, with nearly identical results and inference; therefore, only the

A254 results and the associated molar absorptivity (i.e., SUVA) have been presented.

Conclusions and trends for SUVA results were identical whether analyzed on a solution or soil mass basis; thus, SUVA is presented on an equivalent soil mass basis (L cm-1 m-2) for consistency with other data in this study and to facilitate comparisons among soil properties such as WEOC (g m-2). For comparisons between soil fractions (e.g., POXC content of unfractionated soil versus the PAO fraction), paired t-tests were conducted and differences were considered significant at p < 0.05.

Results

Treatment effects

In the surface layer, mean POXC, POXCP and the proportion of POXC to total SOC were significantly greater under grass and agroforestry VFS than under no-till. When expressed on a gravimetric, soil volume or soil mass basis, similar trends among treatments for POXC and POXCP were found; therefore, only the equivalent soil mass

101

results are reported. In Table 5.1, POXC results represent values per m-2 soil, and in

Table 5.3, results are presented on a sand-free basis for the purposes of comparing unfractionated soil to the PAO fraction. Mean DH and PO were significantly greater under grass VFS than under agroforestry VFS or no-till, corresponding to a 1.6 and 1.9- fold increase in DH, and a 2.7 and 2.0-fold increase in PO, respectively (Table 5.1). The ratio of mean PO activity to mean DH activity (PO:DH) was not significantly different among conservation management practices in the surface layer. Analysis of the water extracts indicates that the agroforestry VFS exhibited the greatest A254, A254P and

SUVA relative to grass VFS and no-till, whereas no-till exhibited significantly greater

SUVAP (Table 5.2). The E2:E3 ratio was significantly different among treatments in the order grass VFS > agroforestry VFS > no-till. In contrast, the E2:E3P data demonstrate that this ratio for the PAO fraction was significantly reduced in agroforestry VFS relative to grass VFS and no-till. The ET:BZ ratio was significantly greater under grass and agroforestry VFS than no-till, and no significant differences were found among conservation management practices in ET:BZP (Table 5.3).

To further explore the effects of the interaction between conservation management practice and WEOC or WEOCP on ultraviolet absorbance indices in the surface layer, a form of ANCOVA was used. For A254, A254P, SUVA, E2:E3, E2:E3P and ET:BZ, the slopes were not significantly different among conservation management practices (p >

0.05) and the y-intercepts were significantly different (p < 0.05), demonstrating an interactive effect between treatment and the concentration of organic carbon on the ultraviolet absorbance indices. The greatest A254, A254P and SUVA per unit organic

102

Table 5.3 Mean surface layer values for whole soil and the particulate, adsorbed and occluded

fraction (subscript P) of KMnO4-oxidizable organic carbon (POXC), ultraviolet absorbance at 254

nm (A254), specific ultraviolet absorbance at 254 nm (SUVA, L cm-1 m-2), the ratio of ultraviolet

absorbance at 280 to 365nm (E2:E3), and the ratio of absorbance at 253to 203 nm (ET:BZ), POXC (g

-1 -1 POXC kg soil), POXCP (g POXC kg PAO), SUVA and SUVAP are calculated on a sand-free basis

for comparison between soil fractions. Values followed by a different capital letter within columns

for a given variable and for a given soil sampling depth are significantly different. Values followed by

a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05.

Standard errors are stated in parentheses.

Overall Grass Agroforestry No-Till POXC 2.03 B 2.37 (0.050) B,a 2.31 (0.076) B,a 1.41 (0.027) B,b

POXCP 2.51 A 2.73 (0.057) A,a 2.70 (0.078) A,a 2.10 (0.063) A,b A254 1.38 B 1.45 (0.115) B,b 1.98 (0.119) B,a 0.69 (0.045) B,c

A254P 3.66 A 3.71 (0.215) A,b 5.04 (0.346) A,a 2.23 (0.133) A,c SUVA 8.08 A 7.51 (0.264) A,b 9.44 (0.341) A,a 7.30 (0.207) A,b

SUVAP 5.26 B 4.86 (0.101) B,b 5.28 (0.122) B,ab 5.62 (0.117) B,a E2:E3 5.17 A 5.86 (0.071) A,a 4.98 (0.066) A,b 4.66 (0.079) B,c

E2:E3P 4.52 B 4.67 (0.150) B,a 3.69 (0.202) B,b 5.04 (0.104) A,a ET:BZ 0.072 B 0.094 (0.0118) B,a 0.098 (0.0094) B,a 0.025 (0.0018) B,b

ET:BZP 0.358 A 0.352 (0.0044) A,a 0.369 (0.0037) A,a 0.352 (0.0073) A,a

103

carbon was demonstrated by the agroforestry VFS (Fig. 5.2a-c), and grass VFS exhibited the least SUVAP per unit of organic carbon (Fig. 5.2d). The E2:E3 and E2:E3P ratios, reflecting the proportion of aliphatic constituents, was greatest under grass VFS (Fig.

5.3a,b). The ET:BZ ratio, reflecting the degree of substitution of oxygen-containing functional groups in aromatic structures, was greatest under grass and agroforestry VFS relative to no-till (Fig.5.3c). Significant regression relationships could not be established with ET:BZP for the ANCOVA (Fig 5.3d).

In the subsurface layer, no significant differences in POXC or POXCP content were detected among conservation management practices (Table 5.4). In contrast to the surface layer, mean DH activity was significantly greater under grass VFS and agroforestry VFS than under no-till in the subsurface layer, reflecting a 2.8 and 2.4-fold increase, respectively. Mean PO activity was significantly greater under grass VFS than under agroforestry VFS or no-till, representing a 2.7 and 1.4-fold increase, respectively (Table

5.2). However, the agroforestry VFS exhibited a significantly reduced mean PO:DH activity than no-till, corresponding to a 79% decrease. Analysis of the water extracts demonstrated significantly greater mean SUVA under agroforestry VFS, significantly greater mean E2:E3 and ET:BZ under grass VFS, and significantly greater mean E2:E3P under no-till relative to grass VFS and no-till. No significant differences in mean A254,

A254P, SUVAP or ETBZP were found among conservation management practices in this layer (Table 5.4).

The general trends among conservation management practices for DH and PO varied slightly in both soil depth layers depending on the form of expression (gravimetric, soil volume, soil mass and organic carbon normalized values). These results can be found in

104

Figure 5.2 Scatterplots and regression relationships for surface soil samples by conservation management practice for water-extractable organic carbon (WEOC) with (a) absorbance at 254 nm

-1 -2 (A254), (c) specific ultra-violet absorbance at 254 nm (SUVA, L cm m ) from unfractionated soil and from the particulate, adsorbed and occluded fraction (b and d, respectively, subscript P). Grass vegetated filter strips (VFS) are represented by black circles, agroforestry VFS by open circles and no-till by grey triangles. Slope parameters are statistically significant (p< 0.05).

105

Figure 5.3 Scatterplots and regression relationships for surface soil samples by conservation management practice for water-extractable organic carbon (WEOC) with (a) the ratio of ultra-violet absorbance at 280 nm to 365 nm (E2:E3) and (c) the ratio of ultraviolet absorbance at 253 to 203 nm

(ET:BZ) from unfractionated soil and from the particulate, adsorbed and occluded fraction (b and d, respectively, subscript P). Grass vegetated filter strips (VFS) are represented by black circles, agroforestry VFS by open circles and no-till by grey triangles. Slope parameters are statistically significant (p< 0.05).

106

Table 5.4 Mean subsurface layer values for whole soil and the particulate, adsorbed and occluded

fraction (subscript P) of KMnO4-oxidizable organic carbon (POXC), ultraviolet absorbance at 254

nm (A254), specific ultraviolet absorbance at 254 nm (SUVA, L cm-1 m-2), the ratio of ultraviolet

absorbance at 280 to 365nm (E2:E3), and the ratio of absorbance at 253 to 203 nm (ET:BZ), POXC

-1 -1 (g POXC kg soil), POXCP (g POXC kg PAO), SUVA and SUVAP are calculated on a sand-free

basis for comparison between soil fractions. Values followed by a different capital letter within

columns for a given variable and for a given soil sampling depth are significantly different. Values

followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD)

at α = 0.05. Standard errors are stated in parentheses.

Overall Grass Agroforestry No-Till POXC 1.04 B 1.07 (0.034) B,a 1.10 (0.041) B,a 0.95 (0.054) B,a

POXCP 1.18 A 1.20 (0.034) A,a 1.24 (0.045) A,a 1.09 (0.050) A,a A254 0.67 B 0.69 (0.037) B,a 0.84 (0.112) B,a 0.48 (0.032) B,a

A254P 1.45 A 1.39 (0.069) A,a 1.74 (0.183) A,a 1.22 (0.080) A,a SUVA 6.61 A 5.53 (0.204) A,b 8.63 (0.264) A,a 5.67 (0.199) A,b

SUVAP 4.88 B 4.85(0.139) B,a 4.62 (0.099) B,a 5.18 (0.182) A,a E2:E3 4.64 B 5.31 (0.117) A,a 4.05 (0.109) B,b 4.56 (0.204) B,b

E2:E3P 5.38 A 5.23(0.068) A,b 5.02 (0.085) A,b 5.89 (0.115) A,a ET:BZ 0.093 B 0.119 (0.0134) B,a 0.108 (0.0118) B,ab 0.053 (0.0056) B,b

ET:BZP 0.350 A 0.320 (0.0228) A,a 0.348 (0.0113) A,a 0.383 (0.0098) A,a

107

Table 5.5. Additionally, no significant differences were found among landscape positions for any variables in this study in either soil depth layer.

Relationships among POXC, enzyme activities and other soil properties

Pearson correlation coefficients were evaluated for relationships among soil properties in the surface layer for unfractionated soil (Table 5.6). Aggregated across conservation management practices (n=36), the strongest, positive correlations with DH and PO included the index of WEOM aliphaticity, E2:E3 (r = 0.70 and 0.71, respectively), soil pH (r = 0.60 and 0.51, respectively, data not shown), and POXC (r =

0.54 and 0.35, respectively). In addition, PO was negatively correlated with SUVA, the index of WEOM aromaticity. The labile POXC fraction was highly significantly correlated with other indicators SOM quality, including PAO-C (r = 0.83), aggregate stability (r = 0.81), and measures of WEOM chemistry (A254, E3:E3, ET:BZ and the

C:N ratio of WEOM). With the exception of SUVA, the WEOM ultraviolet indices

(A254, E2:E3, ET:BZ) and the C:N ratio of WEOM were strongly correlated with the

PAO-C, POXC and WEOC fractions; however, SOC was only moderately to weakly correlated with other soil properties. Relationships were also evaluated for the PAO fraction, and correlations among soil properties were similar to unfractionated soil with a few notable exceptions (Table 5.7). The E2:E3P ratio was highly negatively correlated with WEOCP (r = -0.86), the C:N ratio of WEOMP (r = 0.83), PAO-C (r = -0.73) and

POXCP (r = -0.64). Similar but weaker correlations were found in the subsurface (data not shown).

Overall, across conservation management practices (n=36), WEOM demonstrated an increase in total aromatic constituents (A254) as WEOC concentration increased, while

108

Table 5.5 Means of dehydrogenase activity (DH) and phenol oxidase activity (PO) by conservation management practice expressed on a gravimetric, and soil mass basis, as well as DH and PO normalized to total soil organic carbon (SOC), particulate, adsorbed and occluded organic carbon (PAO-C), KMnO4-oxidizable organic carbon (POXC), and water- extractable organic carbon (WEOC). Values followed by a different lowercase letter within columns and for a given soil layer are significantly different (using Tukey’s HSD) at α = 0.05. The standard error of the mean is stated in parentheses.

109

Table 5.6 Correlation coefficients across all surface samples (n=36) for values from unfractionated soil of absorbance at 254 nm (A254), specific ultraviolet absorbance (SUVA, L cm-1 m-2), ratio of absorbance at 280 to

365 nm (E2:E3), ratio of absorbance at 253 to 203 nm (ET:BZ), water- extractable organic carbon (WEOC, mg m-2), carbon to nitrogen ratio of water-extractable organic matter (W-C:N), KMnO4-oxidiazble organic carbon (POXC, kg m-2), particulate, adsorbed and occluded organic carbon (PAO-C, kg m-2), solid carbon to nitrogen ratio (C:N), soil organic carbon (SOC, kg m-2), water stable aggregates (WSA kg m-2), phenol oxidase activity (PO, mol m-2 hr-1), and dehydrogenase activity (DH, mg m-2 hr-1). A portion of the data from Veum et al. (2011) and Chapter 4.

110

Table 5.7 Correlation coefficients across all surface samples (n=36) for values from the particulate,

adsorbed and occluded fraction (subscript P) of absorbance at 254 nm (A254), specific ultraviolet

absorbance (SUVA, L cm-1 m-2), ratio of absorbance at 280 to 365 nm (E2:E3), ratio of absorbance at

253 to 203 nm (ET:BZ), water-extractable organic carbon (WEOC, mg m-2), carbon to nitrogen ratio

-2 of water-extractable organic matter (W-C:N), KMnO4-oxidiazble organic carbon (POXC, kg m ),

particulate, adsorbed and occluded organic carbon (PAO-C, kg m-2), solid carbon to nitrogen ratio

(C:N), soil organic carbon (SOC, kg m-2), water stable aggregates (WSA kg m-2), phenol oxidase

activity (PO, mol m-2 hr-1), and dehydrogenase activity (DH, mg m-2 hr-1). A portion of the data from

Veum et al. (2011) and Chapter 4.

PAO Fraction

SUVAP E2:E3P ET:BZP WEOCP W-C:NP POXCP PAO-C C:NP

*** * *** *** *** *** ** A254P NS -0.92 0.36 0.96 0.83 0.74 0.85 0.44

** * SUVAP NS NS NS NS -0.47 -0.37 NS

* *** *** *** *** * E2:E3P -0.35 -0.86 -0.83 -0.64 -0.73 -0.41

** ET:BZP NS 0.51 NS NS NS

*** *** *** WEOCP 0.79 0.70 0.82 NS

*** *** * W-C:NP 0.61 0.68 0.36

*** ** POXCP 0.86 0.48

PAO-C 0.57***

* Significant at p < 0.05 ** Significant at p < 0.01 *** Significant at p < 0.001

111

the proportion of aromatic constituents (SUVA), demonstrated a decreasing trend (Fig.

5.2a,c). In contrast, the proportion of aliphatic constituents (E2:E3) and the degree of oxygen substitution (ET:BZ) tended to increase as WEOC concentration increased (Fig.

5.3a,c). In the PAO fraction, A254P increased and E2:E3P declined with increasing

WEOCP (Fig. 5.2b and 5.3b), while no clear trends in SUVAP or the ET:BZP ratio were demonstrated (Fig. 5.2d and 5.3d).

Comparison of the PAO fraction to whole soil

For this comparison, sand-free values were used for unfractionated soil and the PAO fraction to avoid a dilution effect due to preferential retention of sand in the PAO fraction

(see Elliott et al., 1991 for further explication). In the surface layer, aggregated across all

-1 samples (n=36), the sand-free concentration of POXCP (2.51 g POXC kg PAO) was significantly greater than the concentration of POXC in sand-free soil (2.03 g POXC kg-1 soil). In addition, compared within conservation management practices (n=12 per treatment), similar results were found, suggesting that labile POXC was preferentially sequestered in the PAO fraction relative to whole soil (Table 5.3). Similar results and trends were seen in the subsurface layer (Table 5.4).

In the surface layer, A254P and ET:BZP were consistently greater than A254 and

ET:BZ across and within conservation management practices, indicating the concentration of water-extractable aromatic constituents and the degree of substitution of oxygen-containing constituents were greater in the PAO fraction relative to unfractionated soil. In contrast, the proportion of aromatic moieties was greater in unfractionated soil (SUVA) than the PAO fraction (SUVAP) across and within

112

conservation management practices. The index of WEOM aliphaticity was significantly greater in unfractionated soil (E2:E3) relative to the PAO fraction (E2:E3P) overall and within the VFS treatments; however, no-till demonstrated the reverse trend, where

E2:E3P was significantly greater than E2:E3 (Table 5.3). Results and trends in the subsurface were similar, except that no significant differences were found between

SUVA and SUVAP under no-till, and no significant differences were found between

E2:E3 and E2:E3P under grass VFS (Table 5.4).

Discussion

Previous work at this site indicated that one decade was not sufficient time for detectable differences in SOC stocks to emerge after grass and agroforestry VFS were installed in a no-till system. However, an increase in PAO-C (Veum et al., 2011), aggregate stability and WEOC (Chapter 4) under VFS suggested that important changes in SOM quality had occurred. Additionally, Veum et al. (2011) and Chapter 4 found that the PAO fraction represented a potentially more labile and less decomposed fraction of

SOM relative to SOM from unfractionated soil. This study examined the effects of VFS and no-till on SOM quality indicators including DH and PO activity, POXC content and

WEOM chemistry and further compared SOM quality in the PAO fraction to that from unfractionated soil.

The strong correlations among POXC, PAO-C, WEOC and WSA in this study reflect the feedback mechanism between SOM and soil aggregation; SOM fractions bond soil particles and promote aggregation while aggregates protect SOM from microbial attack

(Cambardella and Elliott, 1993; Chantigny et al., 1997; Pikul et al., 2009). In this study,

113

POXCP was significantly greater than POXC, and POXC correlated highly with PAO-C, confirming that the PAO fraction represents a more protected and less processed pool of

SOM relative to unfractionated soil. Both agroforestry and grass VFS demonstrated significantly greater stocks of POXC and POXCP relative to no-till, indicating enhanced retention and protection of partially processed, labile organic matter under VFS (Culman et al., 2012).

The WEOM fraction is potentially the most biologically available and reactive SOM pool, as well as the fraction most easily lost in runoff (Chantigny, 2003; Marschner and

Kalbitz, 2003). Quality of WEOM is influenced by the carbon source (Hobbie, 1996) as well as decomposition and transformation of SOM in the soil (Qualls, 2004). Overall, grass VFS exhibited the greatest microbial DH and PO activity, and a greater proportion of soluble, aliphatic, low molecular weight WEOM compounds, suggesting greater inputs and/or greater retention of labile SOM under grass VFS. In contrast, the agroforestry VFS contained significantly greater concentrations and proportions of aromatic WEOM and significantly less aliphatic, low molecular weight WEOM relative to grass VFS or no-till.

This suggests a difference in SOM inputs (i.e. greater aromaticity of the source carbon under agroforestry) or a reduction in ecosystem metabolism that has resulted in enrichment of recalcitrant, aromatic WEOM (Toberman et al., 2008). Decomposition has been inversely related to polyphenolic content (Qualls and Haines, 1992), and a negative correlation between PO activity and soluble phenolic content has been attributed to a combination of the biological recalcitrance of aromatic compounds as well as phenolic inhibition of microbial activity (Marschner and Kalbitz, 2003; Toberman et al., 2008).

114

Additionally, several other factors are known to inhibit PO activity including low temperature (Freeman et al., 2001a), low pH (Pind et al., 1994; Williams et al., 2000), low oxygen (Pind et al., 1994; Freeman et al., 2001b) and low water content. At this study site, decreased water content and pH have been observed under the agroforestry

VFS relative to grass VFS and no-till (Anderson et al., 2009 and Chapter 4) and these conditions may be inhibiting PO activity under agroforestry VFS, leading to the persistence and build-up of phenolics and subsequent depression of overall microbial activity. This is supported by the positive correlation between PO and E2:E3 and the negative correlation between PO and SUVA. The strong, positive correlations between

DH, PO, E2:E3 and pH illustrate the controlling influence of soil pH and the availability of soluble, low molecular weight organic compounds on overall microbial activity.

Increased aromaticity of dissolved organic matter may increase reactivity with pollutants such as polychlorinated biphenols (Uhle et al., 1999). Thus, the increased concentration and aromaticity of WEOM from agroforestry VFS suggests a greater potential for mobilization and transport of some pollutants relative to grass VFS or no-till

(Baldock, 2002). On the other hand, soil tends to preferentially retain WEOM with aromatic and carboxylic acid groups (McKnight et al., 1992), and higher molecular weight, more complex WEOM is potentially less mobile (Bu et al., 2010). Furthermore, increased infiltration capacity under VFS relative to no-till at this study site (Anderson et al., 2009) may mitigate WEOM losses, and Veum et al. (2009) found no significant differences in dissolved organic carbon losses among conservation management practices at this site.

115

The results of this study further confirm that SOM in the PAO fraction is more labile and less processed than SOM from unfractionated soil (i.e., increased POXC). Ultraviolet evaluation of the water extracts from unfractionated soil and the PAO fraction also suggests important differences in WEOM chemistry and hence quality. Plant tissues contain various aromatic compounds, and enhanced aromaticity of water extracts has been attributed to decomposition byproducts of plant residues (Qualls and Haines, 1992).

Occluded particulate organic matter represents a diverse array of fresh and partly decomposed organic matter, primarily of plant origin, while humus contains substances that stabilize soil due to reactive carboxylic and phenolic functional groups (Baldock,

2002). Therefore, the greater aromaticity (A254), molecular weight (lower E2:E3) and oxygen functionality (ET:BZ) of water extracts from the PAO fraction may represent fresh, undecomposed plant materials and the buildup of phenolic and carboxylic products of partial decomposition in this aggregate-associated fraction (Golchin et al., 1994). More studies are needed, however, to fully characterize the components of aggregate-associated

WEOM.

Conclusions

This study demonstrates that among grass VFS, agroforestry VFS and no-till, the most labile, active and aliphatic SOM was found under grass VFS. Agroforestry VFS, on the other hand, contained the most humified, aromatic WEOM and exhibited depressed

DH and PO enzyme activity relative to grass VFS. This study also found dramatic differences between water extracts of whole soil and the PAO fraction, indicating that the

PAO fraction contains a large proportion of partially processed plant residues relative to

116

whole soil. Additionally, this study suggests that organic matter in the physically protected PAO fraction may be closely related to the oxidizable POXC fraction, and that the PAO fraction may serve as an effective early indicator of soil quality change. Overall, the results of this study are important for a greater understanding of soil carbon cycling and the potential for soil carbon sequestration under conservation management practices.

117

CHAPTER 6

SPECTROSCOPIC COMPARISON OF SOIL ORGANIC MATTER AND HUMIC ACID UNDER VARYING CONSERVATION MANAGEMENT PRACTICES

This chapter is in preparation for submission to the Soil Science Society of America Journal: Veum, K.S., Goyne, K.W., Kremer, R.J., Motavalli, P.P. Spectroscopic comparison of soil organic matter and humic acid under varying conservation management practices.

Abstract

Soil organic matter (SOM) is a key determinant of soil quality and plays many roles in the biogeochemistry of agricultural soils. No-till cultivation and vegetative filter strips

(VFS), impact soil quality and potentially provide many agronomic and environmental benefits; however, the relative benefits of conservation management practices are not well understood. The objective of this study was to evaluate the effects of three conservation management practices (no-till, grass VFS and agroforestry VFS) and landscape position on the structural and compositional characteristics of SOM and humic acid (HA) using solid-state cross-polarization magic-angle-spinning 13C nuclear magnetic resonance (CP-MAS 13C NMR) and diffuse reflectance Fourier transform infrared

(DRIFT) spectroscopy. Results indicate greater content of labile and partially decomposed plant residues under grass VFS, and less biologically active SOM under agroforestry VFS and no-till. Overall, this study demonstrates the installation of VFS in a

118

no-till agroecosystem can significantly alter the characteristics of SOM after only 10 years.

Introduction

Soil organic matter (SOM) is vital to biological, chemical, and physical soil properties that influence plant growth and provide other ecosystem services (Doran and

Parkin, 1994; Karlen et al., 1997; Doran et al., 1998). Conservation management practices are known to enhance SOM relative to conventional tillage practices; however, few studies have compared the effects of varying conservation management practices.

The turnover of SOM is a balance between organic inputs and the rate of loss through decomposition or erosion, and is reflected in both the quantity and quality (i.e., composition) of SOM. Indicators of SOM quality that have demonstrated sensitivity to changes under vegetative filter strips (VFS) relative to no-till include aggregate- associated organic carbon (Veum et al., 2011), labile organic carbon fractions (Chapters 4 and 5) microbial enzyme activity (Chapter 5) and aggregate stability (Chapter 5). Overall, these studies indicate important changes in the composition and cycling of SOM under

VFS relative to no-till, even though differences in total SOC stocks were not detectable

(Veum et al., 2011).

Techniques such as 13C nuclear magnetic resonance (13C-NMR) and diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy can be used in a complementary fashion to characterize SOM or the operationally defined humic acid

(HA), fulvic acid and humin components (Stevenson, 1994). In mineral soils, the HA fraction is considered the largest SOM pool, representative of the degree of humification of organic matter (Guggenberger and Zech, 1994; Preston et al., 1994). The information

119

in 13C-NMR spectra provides information on bulk structural and compositional changes in SOM and SOM components (Kogel-Knabner, 1997; Xing et al., 1999; Mao et al.,

2000) that can be used in decomposition models (Baldock et al., 1997), carbon stabilization models (Baldock and Skemstad, 2000) and to evaluate land use effects (Ding et al., 2002). For example, alkyl-C tends to increase during degradation and O-alkyl-C tends to decrease; therefore, the ratio of alkyl to O-alkyl-C (A/A-O) indicates the relative degree of SOM decomposition (Kogel-Knabner, 1993; Baldock et al., 1997).

The presence of paramagnetic and other minerals in soil samples leads to broad signals, low resolution and low signal to noise ratios, particularly in samples with low

SOM content such as agricultural soils (Kogel-Knabner, 1997). This is commonly overcome by enriching the SOM content (i.e., extraction of inorganic materials including paramagnetic minerals) using hydrofluoric (HF) acid. The HF treatment was originally designed to obtain organic matter from sedimentary rocks (Durand and Nicaise, 1980), and has been used to purify soil HA (De Serra and Schnitzer, 1972) and SOM prior to

13C-NMR analysis (i.e., Skjemstad et al 1994, Mathers et al, 2002 and many others).

The potential drawbacks of the HF pretreatment are secondary reactions that functional groups undergo in the presence of strong acid, such as dehydration (hydrolysis of organic compounds) or recondensation of proteins and polysaccharides, as well as losses of SOM (Rumpel et al., 2002). In particular, the primary losses due to HF treatment are most likely water soluble SOM and labile proteins and carbohydrates

(Skjemstad et al., 1994) and SOM tightly sorbed to the mineral fraction (Rumpel et al.,

2002). The mineral-associated fraction is considered part of the "stable" SOM pool

(Eusterhues et al., 2003), and therefore, HF treatment could potentially affect the results

120

of SOM turnover studies (Rumpel et al., 2002). Rumpel et al (2006), however, noted that the bulk chemical composition as determined by cross-polarization, magic-angle spinning

13C-NMR (CPMAS 13C-NMR) did not appear to be sensitive to the HF treatment, although a fraction of the total C was lost (7-23%). Additionally, due to the high cost of

13C-NMR equipment and instrument time, and the time required to prepare soil samples,

13C-NMR analyses often are not feasible. Furthermore, many studies utilizing 13C-NMR lack treatment-level replication, precluding a robust statistical analysis of the spectra.

Diffuse reflectance infrared Fourier transform spectroscopy is also used to characterize SOM and SOM components using the concept of functional group frequencies (Swift, 1996). Land use, including tillage practices (Ding et al., 2002; Simon,

2007) and fertilization (Ellerbrock et al., 1999), have demonstrated shifts in relative aromaticity, aliphaticity and carboxyl functionality of SOM as identified by DRIFT techniques. Oxygen content has been equated with SOM lability, and higher ratios of O- containing functional groups to recalcitrant groups (O/R) are considered more biologically active, or labile; however, the link between SOM lability and O/R ratios is not clearly understood and may be inversely related to biological activity (Wander and

Traina, 1996). Similar to NMR spectra, removal of inorganic materials using HF acid improves the quality of DRIFT spectra, in particular the resolution of organic peaks that overlap with the Si-O region (1200 – 970 cm-1) such as carbohydrate C-O bonds (Rumpel et al., 2006). Rumpel et al. (2006) also noted that Fourier transform infrared spectra of organic horizons did not appear to be sensitive to the HF treatment.

In addition to vegetation and management practices, the occurrence and distribution of soil types and their properties are influenced by landscape position (Wilson and

121

Gallant, 2000). Pedogeomorphic processes, including physical weathering, chemical weathering, and solute and soil transport, affect soil formation in the landscape. Over time, soil is transported down-slope, accumulating in depressions (Minasny and

McBratney, 2001) and landscape position has been shown to be a factor in the spatial distribution of SOC and TN in soils with accelerated erosion and leaching (Gessler et al.,

2000; Chamran et al., 2002; Yoo et al., 2006; Hancock et al., 2010). Landscape position is also a factor affecting the rate of carbon sequestration in reclaimed prairie soils

(Guzman and Al-Kaisi, 2010) and depth to argillic horizon and associated hydraulic properties in claypan soils (Jiang et al., 2007). Therefore, landscape position is an important factor in field and landscape-level experimental design.

Previous work at this study site determined that significant changes in SOM quality have occurred under VFS after only one decade, including changes in labile and aggregate-associated fractions of SOM (Chapters 3-5). The objectives of this study were

(1) to characterize the structural and functional composition of SOM and HA under three conservation management practices (i.e., grass VFS, agroforestry VFS and no-till) and four landscape positions (i.e., summit, shoulder, backslope and footslope) using CPMAS

13C-NMR and DRIFT spectroscopic techniques and (2) to evaluate relationships between spectral characteristics and associated soil properties. We hypothesized that increased plant residues and exudates under perennial vegetation (i.e., VFS) would result in greater proportions of carbohydrate and aliphatic constituents relative to aromatic moieties, and that the peak ratios would indicate enhanced degradation of SOM under no-till relative to

SOM from VFS.

122

Materials and Methods

Study site

The Greenley Memorial Research Center experimental site, originally designed as a paired watershed study in 1991, is located in Knox County, Missouri, USA (40˚ 01’N,

92˚11’W). The study site is established on claypan soils that are characterized by an abrupt, argillic subsurface horizon that impedes drainage and induces lateral, subsurface flow (Ross et al., 1944; Blanco-Canqui et al., 2002; Arnold et al., 2005). In the central

U.S., claypan soils cover over 16,000 km2; globally, soils with restrictive sub-horizons encompass approximately 2.9 million km2 (USDA-NRCS, 2006). The study site has been cultivated for more than 60 years and has been under a no-till, maize (Zea mays. L.) – soybean [Glycine max (L.) Merr.] rotation since 1994.

The 9.25 ha catchment consists of three small, gently sloping (0 – 3 %), contiguous subcatchments. Two types of upland, contour VFS were planted in 1997; grass VFS were planted in the west subcatchment, agroforestry VFS were planted in the center subcatchment, and the east subcatchment was left under no-till management to serve as the control (Fig. 6.1). Each filter strip is approximately 4.5 m wide and planted on the contour with a VFS to crop width ratio ranging from 1:8 to 1:5, exceeding recommendations for upland filter strips in cropped land (USDA-NRCS, 2010). Overall, the filter strips account for approximately 8 to 10% of the land area and have a combined

VFS width of 22.5 to 27 m in each subcatchment. The grass VFS consists of redtop

(Agrostis gigantea Roth), brome (Bromus spp.), and birdsfoot trefoil (Lotus corniculatus

L.). The agroforestry VFS contain pin oak (Quercus palustris Muenchh.), swamp white oak (Q. bicolor Willd.) and bur oak (Q. macrocarpa Michx.) in addition to the grasses.

123

Figure 6.1 Greenley Memorial Research Center study site in Knox County, Missouri (star). Black dots indicate sampling locations. Thick black lines denote grass vegetative filter strips (VFS), grey lines denote agroforestry VFS and dotted lines represent soil series boundaries (adapted from Veum et al. 2011).

124

Soils in the study area were formed in loess underlain by weathered glacial till

(Watson, 1979) and are designated by the USDA-NRCS as Putnam silt loam (fine, smectitic, mesic Vertic Albaqualfs) in the summit and shoulder positions and Kilwinning silt loam (fine, smectitic, mesic Vertic Epiaqualfs) in the backslope and footslope positions. Under the Food and Agricultural Organization (FAO) soil classification system, these soils are classified as Vertic Luvisols and Gleyic Luvisols. The claypan horizon occurs at depths ranging from 20 to 62 cm (Udawatta et al., 2006) and the clay mineralogy consists of 63% smectite, 33% illite and 4% kaolinite (Chu, 2011). A summary description of the site and comprehensive management details can be found in

Veum et al. (2009) and the references therein.

Soil sampling and laboratory analyses

Surface (0 – 5 cm) soil samples were collected after the fall maize harvest in

November, 2007 from four landscape positions (summit, shoulder, backslope and footslope) and three conservation management practices (grass VFS, agroforestry VFS and no-till). In order avoid the potentially confounding effects of VFS on adjacent soil

(i.e., the potential effect of being spatially upslope or down-slope from the VFS), samples collected from equivalent landscape positions in the no-till subcatchment served as control samples for this study. Three subsamples were collected from each combination of landscape position and conservation management practice, totaling 36 sampling locations (Fig 6.1). Soil cores were collected using 7.62 cm diameter aluminum rings and soil bulk density was determined following the soil core method of Grossman and

Reinsch (2002). Soil gravimetric water content was determined by oven drying soil cores

125

at 105˚C. Surface residues and litter were removed prior to soil sample collection. For general characterization of soil properties, one kg of bulk soil from each sampling depth at each location was collected using a hand trowel.

Cation exchange capacity (CEC), base saturation and particle size were determined on air-dried, ground soil samples by the University of Missouri Soil Characterization

Laboratory following USDA-NRCS procedures (Burt, 2004). Total SOC was determined as described in Veum et al. (2011). The particulate, adsorbed and occluded (PAO) fraction was separated and the carbon content (PAO-C) determined as described in Veum et al. (2011). Air-dried water stable aggregates were measured on undisturbed core samples following an adapted method of Kemper and Rosenau (1986) as described in

Chapter 5. Water-extractable organic carbon from unfractionated soil (WEOC) and the

PAO fraction (WEOCP) were extracted and quantified as described in Chapter 4.

Potassium permanganate-oxidizable organic carbon (POXC), dehydrogenase activity

(DH) and phenol oxidase activity (PO), are detailed in Chapter 5. General soil characteristics, summarized by conservation management practice, are shown in Table

6.1.

Soil organic matter extraction

Subsamples (n=3) were composited from each plot at a particular landscape position for a total of 4 replicates per conservation management practice. The extraction and enrichment of SOM from whole soil samples was performed with 5% HF (v/v) following a modified procedure of Skjemstad et al. (1994) to eliminate paramagnetic materials which interfere with 13C- NMR analysis. In short, 20 g portions of air-dried and

126

Table 6.1 Mean soil properties in the surface (0–70 kg m-2) layer of the study soils by conservation management practice. Values followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05 (n=12 per treatment). Values without lowercase letters were determined using composite samples (n=4 per treatment).The standard error of the mean is stated in parentheses. Data from Veum et al. (2011) and Chapters 4 and 5.

Surface (0–70 kg m-2 soil) Soil Property Grass Agroforestry No-Till Textural Class Silt Loam Silt Loam Silt Loam -1 a CECNaOAc (cmolc kg ) 23.3 (1.20) 23.6 (0.51) 21.0 (0.40)

Base SaturationNaOAc (%) 92.8 (2.33) 87.8 (1.58) 90.0 (2.27) Clay Content (g kg-1) 242.5 (17.59) 239.2 (9.42) 218.7 (6.22)

127

sieved (< 2 mm) soil were combined with 200 ml 5% HF in a 250 ml HDPE centrifuge bottle for a soil to solution ratio of 1:10. Samples were placed on an end-to-end shaker for 2 hr then centrifuged for 20 min at 2,000 rpm. The supernatant was carefully discarded, retaining the floating light fraction for reincorporation at the completion of the purification procedure. The 2 hr HF digestion step was repeated five times, followed by a final, 12 hr overnight extraction. The SOM residue was washed with 50 ml ultrapure, deionized water 5 times, centrifuging between each wash step and decanting the wash solution. After the final wash step, the SOM residue was freeze-dried for 48 hrs, ground with an agate mortar/pestle and stored in a desiccator until further analysis.

Humic acid extraction

Subsamples (n=3) were composited from each plot for a total of 4 replicates per conservation management practice. The extraction of HA from whole soil followed the

International Humic Substances Society method as described by Gieguzynska et al.

(2009). In short, 22 g portions of air-dried and sieved (< 2mm) soil were weighed into

250 ml centrifuge tubes and 20 ml of 1N HCl was added to form a soil slurry of pH ca.

1- 2. An additional 200 ml of 0.1N HCl was added for a total soil to solution ratio of

1:10. The samples were placed on an end-to-end shaker for 1 hr, centrifuged at 3,600 rpm for 20 min. The solid residue was separated from the supernatant and neutralized with 20 ml 1 N NaOH for a pH ca. 7.0. An additional 200 ml of 0.1N NaOH was added for a soil to solution ratio of 1:10. The sample containers were sparged with N2 to prevent oxidation and placed on an end-to-end shaker for 4 hr. The samples were removed from the shaker and allowed to settle overnight. On day 2, samples were centrifuged at 3,600

128

rpm for 20 min and the HA supernatant was decanted and retained. While stirring the supernatant, 6N HCl was added drop-wise for a pH ca. 1.0. The samples were allowed to settle overnight. On day 3, samples were centrifuged at 3,600 rpm for 45 min, the supernatant was carefully decanted and discarded, retaining the HA residue.

To further purify the HA, 0.1N KOH was added to achieve a final solution volume of

220 ml to dissolve the HA residue, and 3 g of solid KCl was added to each sample. The samples were centrifuged at 3,600 for 30 min and carefully decanted to retain the HA supernatant. The residue was discarded. An additional purification step was performed by adding 6N HCl while stirring to pH ca. 1.0 to promote precipitation of HA from solution.

The samples were allowed to settle overnight. On day 4, samples were centrifuged at

3,600 rpm for 45 min and the supernatant was carefully decanted and discarded, retaining the HA residue. To remove inorganics, 0.1N HCl/0.3N HF was added to achieve a final solution volume of 220 ml and the samples were placed on an end-to-end shaker overnight. On day 5, the inorganic removal steps of day 4 were repeated. On day 6, samples were centrifuged at 3,600 rpm for 45 min and the supernatant was carefully decanted and discarded, retaining the HA residue. The HA residue was the transferred to dialysis tubing (Pierce® snakeskin 10K tubing) using ultrapure, deionized water and

- dialyzed repeatedly in ultrapure, deionized water until a negative AgNO3 test for Cl was obtained (generally 3-4 days). The HA was subsequently freeze-dried for 48 hrs, ground with an agate mortar and pestle, and stored in a desiccator until further analysis.

129

Solid state 13C nuclear magnetic resonance analysis

Approximately 200 mg of sample (SOM or HA) was packed in a 7 mm zirconia rotor with a Kel-F cap. The rotor was then spun at 5 kHz at room temperature. Solid 13C-NMR was performed on a Bruker Avance DRX300 widebore NMR spectrometer (Bruker,

Billerica, MA) equipped with a 7mm CPMAS probe. The operating frequency was

300.13 MHz for proton and 75.48 MHz for carbon, respectively. Cross-polarization with magic angle spinning and total sideband suppression (CPMAS-TOSS1) was acquired with 1 ms contact time and 1.03 s repetition delay (Dixon et al., 1982). Spectra were collected at a spin rate of 5 kHz, a recycle delay (relaxation time) of 1s and a contact time of 1ms. The 13C chemical shift was externally referenced to the carbonyl carbon signal of glycine at 176.03 ppm. The location of an NMR signal in a spectrum is reported relative to the reference signal from standard. To provide an unambiguous location unit in the spectrum, the frequency differences (Hz) are divided by the spectrometer frequency

(MHz). The resulting number is very small; thus, it is multiplied by 106, giving a locator number called the Chemical Shift, having units of parts-per-million (ppm). A minimum of 5000 scans were collected for each sample. Line broadening of 150 Hz was applied to the data before Fourier transformation. Bruker XWIN-NMR version 3.6 software was used for data collection and initial processing.

Spectral areas were calculated by integration (Xing et al., 1999) using the regions defined by Wilson (1987) and (Stevenson, 1994) as a guide (Table 6.2) and calculated as percentages of the total spectral area (0-215 ppm). Total aliphatic-C (0-110 ppm) was subdivided into alkyl-C (0-45 ppm), O-alkyl-C (45-110 ppm), methoxyl-C (45-60 ppm), carbohydrate-C (60-94 ppm), and di-O-alkyl-C (94-110 ppm). Total aromatic-C (110-160

130

Table 6.2 13C nuclear magnetic resonance (NMR) chemical shift (ppm) assignments used in this

study from Stevenson (1994) and Wilson (1987).

Assignment ppm Representative Organic Functional Groups total aliphatic-C 0-110 alkyl-C 0-45 unsubstituted, non-polar aliphatic-C O-alkyl-C 45-110 C-O, C-N bonds in carbohydrates, alcohols, esters and amino acids methoxyl-C 45-60 proteinaceous materials: amino acids, peptides, proteins carbohydrate-C 60-94 carbohydrate/polysaccharide-C di-O-alkyl-C 94-110 O-C-O anomeric-C total aromatic-C 110-160 aryl-C 110-142 aromatic-C not substituted by O or N phenolic-C 142-160 O- and N-substituted aromatic groups: phenolic OH, aromatic NH2 total carbonyl-C 160-215 carboxyl-C 160-190 carboxyl-C (COO), may overlap with phenolic-, amide- and ester-C carbonyl-C 190-215 ketone, quinone or aldehyde-C

131

ppm) was subdivided into aryl-C (110-142 ppm) and phenolic-C (142-160 ppm). Total carbonyl-C (160-215 ppm) was subdivided into carboxyl-C (160-190 ppm) and carbonyl-

C (190-215 ppm). Percent aromaticity was calculated as the total aromatic-C area (110-

160 ppm) divided by the sum of aliphatic-C and aromatic-C (0-110 ppm)+(110-160 ppm) per Chefetz et al. (2002). The alkyl to O-alkyl index (A/A-O), which increases with thedegree of decomposition, was calculated as (0-45 ppm) / (45-110 ppm) per Baldock and Preston (1995). The mass of each functional carbon group was estimated by multiplying the proportion of total organic C, as determined by integration of the NMR spectra, by the total mass of SOC in the sample (kg m-2). Due to limitations of the HA extraction procedure, the yield (i,e., kg HA kg-1 soil) could not be determined. Thus, the correlation analyses for HA proceeded using the proportional data. The organic carbon fractions and C:N ratios evaluated included POXC, PAO-C, WEOC, WEOCP, as well as the C:N ratios of the PAO fraction (C:NP), water extracts of unfractionated soil (W-C:N), and water extracts from the PAO fraction (W-C:NP).

Diffuse reflectance infrared Fourier transform analysis

The DRIFT spectra of SOM and HA were collected using a Nicolet™ 4700 mid- infrared spectrophotometer (Thermo-Fisher, Madison, WI) with a DRIFT accessory, where light output is measured as a function of infrared wavelength (or wavenumber, listed in cm-1). A mixture of 8% sample in KBr was mixed and gently ground with a mortar and pestle prior to spectral collection. The DRIFT cell was purged with CO2 and

H2O-free air for 15 min before collection of each spectrum. A background spectrum of

KBr was collected at least every 4 hrs to account for potential changes in ambient

132

environmental conditions. Exactly 400 scans were collected at a resolution of 2 cm-1 from

400 to 4000 cm-1. Spectra were converted to Kubelka-Munk units using Grams/32 v.8.0 software (Galactic Corp., Salem, NH). The HA spectra were baseline corrected as in

Chefetz et al. (1998) using 4000, 2000 and 860 cm-1 as zero absorbance points. General peak assignments and the associated wavenumbers used in this study are listed in Table

6.3. Actual peak maxima were used for the calculation of peak ratios, and the approximate location of the peaks and a description of the representative functional groups are listed in Table 6.4.

Statistical analysis

The main effects of conservation management practice and landscape position on

SOM and HA spectral characteristics were tested using a two-factor ANOVA from a completely randomized design (Kuehl, 2000). Due to the original paired-watershed study design and the lack of replication among different treatment combinations, a standard statistical analysis was not possible. The two-way interaction between conservation management practice and landscape position was used as the random error term to account for the lack of plot-level replication. Subsamples (3 per plot) were combined, resulting in n=4 for each factor (i.e., treatment) in the analysis. Further details of the modified ANOVA and a full justification for the model specification can be found in

Veum et al. (2011) and the Appendix. The two-factor ANOVA analyses were conducted using SAS Proc Mixed (SAS® software, Version 9.1.3, SAS Institute Inc., Cary, NC,

USA.) and Tukey's HSD procedure was used as the multiple comparison test to evaluate pair-wise differences in treatment means. The Tukey's HSD statistic also utilized the interaction term as the random error term in the ANOVA and accounted for the

133

Table 6.3 Diffuse reflectance infrared Fourier transform (DRIFT) spectral peak assignments and associated wavenumbers used in this study (Baes and Bloom, 1989;

Inbar et al., 1989; Stevenson, 1994).

Wavenumber (cm-1) Functional Group Assignment 3340-3280 H-bonds and OH groups, possibly water

2930-2924 CH2 asymmetric stretch

2850 CH2 symmetric stretch 1735-1713 C=O of COOH 1650 C=O, C=OH, amide H 1630-1600 aromatic ring C=C, symmetric COO- stretch 1550-1540 aromatic ring, amide 1535-1520 aromatic ring, C=C, amide

1455 C-H deformation of CH2, CH3 1425 aromatic ring C=C or carboxyl-C - 1380-1325 CH3 bending, symmetric COO stretch 1260-1240 CO, COOH, COC, phenol OH 1190-1127 aliphatic, alcoholic OH 1080-1050 CO aliphatic alcohol, C-O stretch of polysaccharides

134

Table 6.4 Diffuse reflectance Fourier transform infrared spectral peak ratios evaluated in this study and the associated functional groups (Wander and Traina, 1996; Ding et al., 2002).

Ratio Wavenumber (cm-1) Description 1 1724 / 1080 carboxyl-C/polysaccharide-C 2 1724 / 2924 carboxyl-C/aliphatic-C 3 1423 / 1648 aliphatic-C/aromatic-C 4 (1724+1648+1080)/ O-containing/recalcitrant-C (2924+1536+1510+1451+1423) 5 1724/(1451+1423) O-containing/recalcitrant-C 6 (1648+1423+1080)/3405 O-containing/recalcitrant-C 7 (1648+1423+1080)/2924 O-containing/recalcitrant-C 8 (1648+1423+1080)/1451 O-containing/recalcitrant-C 9 (1648+1423+1080)/(3405+2924+1451) O-containing/recalcitrant-C 10 (1423+1080)/3405 O-containing/recalcitrant-C 11 (1423+1080)/2924 O-containing/recalcitrant-C 12 (1423+1080)/1451 O-containing/recalcitrant-C 13 (1423+1080)/(3405+2924+1451) O-containing/recalcitrant-C 14 (1189+1129)/(1451+1423) O-containing/recalcitrant-C

135

subsamples (Littell et al., 2006). All analyses were conducted at the α=0.05 significance level. Correlation and regression analyses were conducted on data aggregated across conservation management practices (n=12).

Results

13C nuclear magnetic resonance (NMR) spectra of SOM

The CPMAS 13C NMR spectra of SOM were very similar in peak position and intensity among conservation management practices and were dominated by peaks in the aliphatic region (0-110 ppm) including a strong alkyl-C peak (110-160 ppm) and moderate peaks in the carbohydrate-C (60-94 ppm) and di-Oalkyl-C (94-110 ppm) regions. No distinct peaks were evident in the aromatic-C (110-160 ppm) or carbonyl-C

(160-215 ppm) regions (Figure 6.2). Among conservation management practices, the grass VFS soil demonstrated significantly greater proportions of aromatic, phenolic and carboxylic-C content and significantly reduced proportions of total aliphatic and alkyl-C relative to agroforestry VFS and no-till soil. The agroforestry VFS demonstrated significantly greater proportions of O-alkyl and methoxyl-C relative to grass VFS and no- till soil (Table 6.5). No landscape position effects were detected in the SOM extracts using 13C NMR.

Correlation relationships were examined between measures of SOM lability (i.e.,

SOC fractions and C:N ratios), and the CPMAS 13C NMR functional group contents.

Significant positive correlations were found between the measures of labile SOM and the functional groups aliphatic-C, O-alkyl-C, methoxyl-C and carbohydrate-C. The A/O-A ratio, an index of SOM degradation, was negatively correlated with POXC, PAO-C,

136

Figure 6.2 Examples of solid-state, cross-polarization, magic angle spinning 13C nuclear magnetic resonance (CPMAS 13C NMR) spectra of soil organic matter from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil.

137

Table 6.5 Mean proportions (%) of C structural components soil organic matter as determined by cross-polarization, magic angle spinning 13C nuclear magnetic resonance (CPMAS 13C NMR) spectroscopy from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil. Values followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05 (n=4 per treatment). The standard error is stated in parentheses.

Assignment ppm Grass Agroforestry No-Till total aliphatic-C 0-110 63.4 (3.18)b 83.0 (0.74)a 77.1 (1.45)a alkyl-C 0-45 27.6 (3.96)b 42.5 (1.01)a 43.6 (1.73)a O-alkyl-C 45-110 35.9 (1.27)ab 40.63 (0.27)a 33.8 (1.07)b methoxyl-C 45-60 8.4 (0.48)b 12.3 (0.44)a 8.5 (1.48)b

carbohydrate-C 60-94 20.7 (1.58)a 23.7 (0.41)a 19.8 (0.79)a di-O-alkyl-C 94-110 6.8 (1.04)a 4.8 (0.32)a 5.5 (0.69)a total aromatic-C 110-160 22.1 (0.60)a 10.3 (0.78)b 13.6 (0.92)b aryl-C 110-142 13.8 (3.22)a 5.5 (0.37)a 7.9 (1.24)a phenolic-C 142-160 8.9 (0.58)a 4.81 (0.46)b 5.7 (0.36)b total carbonyl-C 160-215 14.4 (1.83)a 4.6 (1.41)b 9.1 (1.71)ab carboxyl-C 160-190 10.9 (1.26)a 4.3 (0.96)b 6.2 (1.34)b carbonyl-C 190-215 3.4 (0.67)a 2.2 (0.72)a 2.9 (0.51)a % aromaticity a 25.7 (3.96)a 11.0 (0.79)b 14.93 (0.91)b A/O-A index b 0.80 (0.161)b 1.05 (0.029)ab 1.30 (0.119)a a % aromaticity calculated as 100 * (110-160) /[(0-110) + (110-160)] balkyl to O-alkyl (A/O-A) index calculated as the ratio of (0-45 ppm) / (45-110 ppm)

138

WEOC and W-C:N (Table 6.6). In addition, O-alkyl-C and carbohydrate-C were positively correlated with water stable aggregates (r = 0.67, p = 0.02 and r = 0.68, p =

0.02, respectively).

13C nuclear magnetic resonance spectra of humic acid

The CPMAS 13C NMR spectra of HA were very similar in peak position and intensity among conservation management practices and were dominated by strong peaks in the aromatic (110-160 ppm) and carboxyl (160-190 ppm) regions. The large peak at ca. 131 ppm, present in each of the HA spectra, represents ring carbons not substituted by O or

N, and is characteristic of lignin (Preston, 1996). The large carboxyl peak ca. 175 ppm is attributed to carboxylic acids, amides and esters. Moderate peaks were present in the alkyl-C (0-45 ppm), methoxyl-C (45-60 ppm) and carbohydrate-C (45-60 ppm) regions, and a small phenolic-OH /aromatic NH2 peak was present at ca. 155 ppm in all HA spectra (Figure 6.3). Among conservation management practices, the agroforestry VFS demonstrated significantly greater proportions of total aliphatic-C, O-alkyl-C, methoxyl-

C and carbohydrate-C, and significantly reduced proportions of aryl-C, total carbonyl-C and % aromaticity relative to no-till. The proportion of methoxyl-C in HA from the agroforestry VFS was also significantly greater than in HA from the grass VFS. No significant differences were detected among conservation management practices in the proportions of alkyl-C, di-O-alkyl-C, total aromatic-C, phenolic-C, carboxylic-C or carbonyl-C (Table 6.7). Representative 13C NMR spectra of HA from each conservation management practice can be found in Figure 6.3. No landscape position effects were detected in the HA extracts using 13C NMR.

139

Table 6.6 Pearson correlation coefficients (r) for carbon functional groups observed in cross polarization, magic angle spinning 13C nuclear magnetic resonance (CPMAS 13C NMR) across all samples (n=12) of soil organic matter, calculated per unit mass of soil, with soil organic carbon

(SOC), KMnO4-oxidizable carbon (POXC), particulate, adsorbed and occluded carbon (PAO-C), water extractable organic carbon from unfractionated soil (WEOC) and the PAO fraction (WEOCP), the carbon to nitrogen ratio of unfractionated soil (C:N), the PAO fraction (C:NP), water extracts of unfractionated soil (W-C:N) and the PAO fraction (W-C:NP). Functional groups not represented in this table did not present significant correlations.

Group SOC POXC PAO-C WEOC WEOCP C:N C:NP W-C:N W-C:NP aliphatic-C NS NS NS NS 0.60* NS NS NS 0.69*

O-alkyl-C 0.69* 0.62* 0.82** 0.71* 0.77** NS 0.72** 0.64* 0.62* methoxyl-C NS NS NS NS 0.68* NS 0.71** NS 0.67* carbohydrate-C 0.65* 0.58* 0.78** 0.69* 0.77** NS 0.73** 0.65* 0.64*

A/O-A index NS -0.61* -0.60* -0.68* NS NS NS -0.68* NS

* Significant at p < 0.05 ** Significant at p < 0.01 *** Significant at p < 0.001

140

Figure 6.3 Examples of solid-state, cross-polarization, magic angle spinning 13C nuclear magnetic resonance (CPMAS 13C NMR) spectra of humic acids extracted from grass vegetative filter strip

(VFS), agroforestry VFS and no-till soil.

141

Table 6.7 Mean proportions (%) of humic acid (HA) C structural components as determined by cross-polarization, magic angle spinning 13C nuclear magnetic resonance (CPMAS 13C NMR) spectroscopy from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil. Values followed by a different lowercase letter within rows are significantly different (using Tukey’s HSD) at α = 0.05 (n=4 per treatment). The standard error is stated in parentheses.

Assignment ppm Grass Agroforestry No-Till total aliphatic-C 0-110 41.5 (1.16)ab 45.2 (1.60)a 40.1 (0.96)b alkyl-C 0-45 17.2 (1.00)a 19.1 (0.72)a 17.0 (0.50)a O-alkyl-C 45-110 24.4 (0.34)ab 26.2 (0.89)a 23.2 (0.53)b methoxyl-C 45-60 7.7 (0.16)b 8.3 (0.29)a 7.2 (0.19)b carbohydrate-C 60-94 11.3 (0.16)ab 12.4 (0.39)a 10.9 (0.25)b

di-O-alkyl-C 94-110 5.4 (0.30)a 5.5 (0.31)a 5.1 (0.13)a total aromatic-C 110-160 35.4 (0.86)a 32.4 (1.10)a 36.4 (0.71)a aryl-C 110-142 28.1 (0.58)ab 25.3 (0.92)b 29.0 (0.55)a phenolic-C 142-160 7.3 (0.34)a 7.12 (0.20)a 8.7 (1.28)a total carbonyl-C 160-215 22.9 (0.67)ab 22.2 (0.48)b 23.4 (0.39)a

carboxyl-C 160-190 19.4 (0.29)a 18.9 (0.40)a 19.8 (0.27)a carbonyl-C 190-215 3.6 (0.41)a 3.5 (0.19)a 3.6 (0.19)a % aromaticity a 46.0 (1.25)ab 41.8 (1.67)b 47.6 (1.08)a A/O-A index b 0.71 (0.039)a 0.73 (0.004)a 0.73 (0.009)a a % aromaticity calculated as 100 * (110-160) /[(0-110) + (110-160)] b alkyl to O-alkyl (A/O-A) index calculated as the ratio of (0-45 ppm) / (45-110 ppm)

142

Correlation analysis of 13C NMR spectra of HA proceeded using the functional group proportions directly. The only significant positive correlations were between variables associated with the PAO fraction (PAO-C, WEOCP, C:NP, W-C:NP) and the proportions of aliphatic-C, O-alkyl-C, methoxyl-C, carbohydrate-C and di-O-alkyl-C.

Additionally,the proportion of aromatic-C and aryl-C were significantly negatively correlated with the variables associated with the PAO fraction (Table 6.8).

Diffuse reflectance infrared Fourier transform spectra of SOM

The major peaks in the SOM spectra (Figure 6.4) included the hydroxyl peak at 3350 cm-1, the alkyl peaks at 2912 and 2847 cm-1, the carboxylic peak at 1723 cm-1, the carbonyl peak at 1645 cm-1, the aromatic and phenolic peaks at 1511 and 1240 cm-1, the

-1 CH2, CH3 peaks at 1451, the carboxyl peak at 1423 cm , and the polysaccharide C-O peak at 1080 cm-1. The aromatic peak at 1600 cm-1 was a slight shoulder on the carbonyl peak. Of all the ratios examined (see Table 6.4 in Materials and Methods for a comprehensive list), only two ratios demonstrated significant differences among treatments. No-till demonstrated a significantly greater carboxyl-C to polysaccharide-C ratio than agroforestry VFS, and the agroforestry VFS exhibited a significantly greater carboxyl-C to recalcitrant-C ratio relative to no-till (Table 6.9). No significant differences were found among conservation management practices using the other ratios, and no significant differences were found among landscape positions.

143

Table 6.8 Pearson correlation coefficients (r) across all samples (n=12) for carbon functional groups, identified using 13C nuclear magnetic resonance (NMR), carbon functional groups of humic acids with soil organic carbon (SOC), KMnO4-oxidizable carbon (POXC), particulate, adsorbed and occluded carbon (PAO-C), water extractable organic carbon from unfractionated soil (WEOC) and the PAO fraction (WEOCP), the carbon to nitrogen ratio of unfractionated soil (C:N), the PAO fraction (C:NP), water extracts of unfractionated soil (W-C:N) and the PAO fraction (W-C:NP).

Functional groups not represented in this table did not present significant correlations.

NMR Group SOC POXC PAO-C WEOC WEOCP C:N C:NP W- W- aromatic-C NS NS -0.57* NS -0.62* NS NS NS -0.64* C:N C:NP aliphatic-C NS NS NS NS 0.60* NS NS NS NS

O-alkyl-C NS NS 0.72** NS 0.68* NS NS NS 0.63* methoxyl-C NS NS 0.68* NS 0.67* NS 0.58* NS 0.63* carbohydrate-C NS NS 0.58* NS 0.66* NS NS NS 0.67* di-O-alkyl-C NS NS 0.69* NS NS NS NS NS NS aryl-C NS NS -0.64* NS -0.68* NS NS NS -0.70*

* Significant at p < 0.05 ** Significant at p < 0.01 *** Significant at p < 0.001

144

Figure 6.4 Examples of diffuse reflectance infrared Fourier transform infrared (DRIFT) spectra of soil organic matter extracted from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil.

145

Table 6.9 Diffuse reflectance infrared Fourier transform (DRIFT) spectral peak ratios of soil organic matter from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil. Means within rows with different lowercase letters are significantly different (p < 0.05). The standard error is stated in parentheses.

Wavenumbers (cm-1) Description Grass Agroforestry No-Till 1724 carboxyl-C 3.35 ab 2.89 b 4.03 a 1080 polysaccharide-C (0.228) (0.229) (0.236)

1423+1080 carboxyl-C 1.51 ab 1.54 a 1.46 b 1451 recalcitrant-C (0.011) (0.008) (0.020)

146

Diffuse reflectance Fourier transform infrared spectra of humic acid

The major peaks in the HA spectra were the hydroxyl peak at 3340 cm-1, the alkyl peaks at 2935 and 2850 cm-1, the carboxylic peak at 1725 cm-1, the carbonyl peak at 1670

-1 -1 cm , the aromatic and phenolic peaks at 1600, 1510 and 1240 cm , the CH2, CH3 peak at

1451, the carboxyl peak at 1423 cm-1, and the polysaccharide C-O peak at 1080 cm-1

(Figure 6.5). Of all the ratios examined (see Table 6.4 in Materials and Methods for a comprehensive list), no-till demonstrated significantly greater carboxyl-C to polysaccharide-C, carboxyl-C to aliphatic-C, carboxyl-C to recalcitrant-C and reactive-C

(i.e., O-containing) to recalcitrant-C ratios relative to agroforestry VFS (Table 6.10). No significant differences were found among conservation management practices using the other ratios, and no significant differences were found among landscape positions.

Discussion

This study examined the composition of SOM and HA using CPMAS 13C NMR and

DRIFT spectroscopy and found significant differences among grass VFS, agroforestry

VFS and no-till conservation management practices. The bulk chemical composition of

SOM reflects both the source plant material as well as the extent of degradation (Baldock and Preston, 1995). Generally, as decomposition progresses, the O-alkyl content decreases due to preferential consumption by microbes; thus the A/O-A ratio increases

(Kogel-Knabner, 1993; Baldock et al., 1997). Grass VFS had a significantly smaller A/O-

A index, followed by agroforestry VFS and no-till. This result suggests reduced degradation and retention of partially decomposed organic residues under VFS, in particular grass VFS. In addition, the A/O-A index may be inversely related to the quality of SOM as a microbial substrate (Webster et al., 2000), and these results are supported by

147

Figure 6.5 Examples of diffuse reflectance infrared Fourier transform (DRIFT) spectra of humic acid from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil.

148

Table 6.10 Diffuse reflectance infrared Fourier transform (DRIFT) spectral peak ratios of humic acids from grass vegetative filter strip (VFS), agroforestry VFS and no-till soil. Means within rows with different lowercase letters are significantly different (p < 0.05). The standard error is stated in parentheses.

Wavenumbers (cm-1) Description Grass Agroforestry No-Till

1724 carboxyl-C 2.27 ab 2.11b 2.34 a 1080 polysaccharide-C (0.041) (0.032) (0.080)

1724 carboxyl-C 1.09 ab 1.07 b 1.18 a 2924 aliphatic-C (0.026) (0.035) (0.031)

1724 carboxyl-C 0.81 a 0.77 b 0.82 a 1451+1423 recalcitrant-C (0.010) (0.012) (0.008)

1724+1648+1080 reactive-C 0.69 ab 0.67 b 0.70 a 2924+1536+1510+1451+1423 recalcitrant-C (0.007) (0.011) (0.004)

149

enhanced microbial enzyme activity under grass VFS (Chapter 5). In contrast, microbial enzyme activities in the agroforestry VFS suggest depressed biological activity relative to grass VFS. This inference is supported by the higher O/R DRIFT peak ratio under agroforestry VFS, which is thought to be inversely related to biological activity (Wander and Traina, 1996). The depressed microbial activity may be the result of differing organic inputs under agroforestry VFS, and/or the inhibition of microbial activity due to the build-up of phenolic compounds (Chapter 5).

Grass VFS also exhibited greater proportions of aromatic, phenolic and carboxyl groups in SOM relative agroforestry VFS or no-till, indicating greater pools of fresh and partially decomposed residues. Protection of SOM in soil aggregates leads to an accumulation of partial decomposition products such as aromatic and phenolic structures.

For example, tannins and lignins contribute to the aromatic and phenolic regions of 13C-

NMR spectra (Preston, 1996), and reduced decay rates of recalcitrant materials such as lignin, may lead to increased aromatic content (Ding et al., 2006). The enhanced accumulation of partial decomposition products under grass VFS is likely related to increased aggregate stability (Chapter 4). There is a strong feedback mechanism between soil aggregation and SOM turnover. The physicochemical protection afforded by aggregation inhibits the biochemical attack of otherwise decomposable organic matter, sequestering carbon in the soil (Jastrow et al., 2007). The preferential retention of residues under VFS is supported by the greater C:N ratios of the aggregate-associated

PAO fraction and the increased PAO-C content under VFS (Chapter 3), the increased

150

water stable aggregates and WEOC under VFS (Chapter 4) and the greater content of labile, POXC under VFS (Chapter 5). In addition, the reduced decomposition of SOM under VFS was demonstrated by greater aromatic and O-containing moieties in the water extracts (Chapter 5). Overall, arable soils are subject to more frequent macroaggregate turnover and increased SOM mineralization (Beare et al., 1994). Therefore, in no-till soils, aggregation is reduced, and labile SOM inputs are less protected and more rapidly decomposed relative to soil under VFS.

The 13C-NMR carbohydrate-C and O-alkyl-C content of SOM, representing fresh, undecomposed materials (Kogel-Knabner, 1993; Kogel-Knabner, 1997), demonstrated strong, positive correlations with several indicators of SOM quality, such as aggregate stability, SOC, PAO-C, POXC, WEOC and WEOCP, as well as C:N ratios of the solid and water fractions. The bulk composition of HA, as determined by NMR, demonstrated significant correlations with variables relating to the PAO fraction only. The PAO-C content, WEOCP content and the carbon to nitrogen ratio of the PAO water extracts were negatively correlated with the aromatic and aryl-C content of the HA, and positively correlated with the carbohydrate, O-alkyl, methoxyl and aliphatic-C. This suggests that as the HA becomes increasingly humified, it is less related to the PAO fraction, and that the

PAO fraction serves as a useful proxy of SOM protection and degradation.

Conclusions

This study confirms previous work indicating that changes in the bulk chemical composition of SOM have occurred among conservation management practices after 10 years. The grass VFS contains the greatest proportions of aromatic, phenolic and

151

carboxylic SOM, representing the most protected and least degraded SOM. Overall, these results suggest SOM is preferentially protected under VFS relative to no-till, leading to a reduced degradation rate. The enhanced soil aggregation under VFS provides protection from microbial degradation, reduces the SOM turnover rate, and enhances carbon sequestration potential.

152

CHAPTER 7

CONCLUSIONS

This study examined the effects of three conservation management practices (i.e., no- till, grass vegetative filter strips (VFS) and agroforestry VFS) and four landscape positions (i.e., summit, shoulder, backslope and footslope) on many aspects of soil organic matter (SOM) quantity and quality. Initial work indicated that losses of dissolved organic carbon (DOC) in runoff (Chapter 2) and stocks of soil organic carbon (SOC)

(Chapter 3) were not significantly different among conservation management practices 10 years after installation; therefore more sensitive indicators of SOM quality were required to detect changes in SOM at this spatial and temporal scale.

Using multiple physical, chemical and biological indicators, this study demonstrated that grass VFS enhance aggregate-associated organic carbon in the particulate, adsorbed and occluded fraction (PAO-C) (Chapter 3), aggregate stability and water-extractable organic carbon (WEOC) (Chapter 4) and labile KMnO4-oxidizable organic carbon

(POXC) and microbial enzyme activity (Chapter 5).

Few studies have evaluated the chemistry of water extracts from soil aggregates, and in Chapter 5, water-extractable organic matter from the PAO fraction (WEOMP) was characterized using ultraviolet spectrophotometry, finding that WEOMP had greater aromatic and phenolic content relative to unfractionated soil extracts (WEOM), indicating reduced degradation of the parent SOM material. In addition, this study found that organic carbon in cold water-extracts from unfractionated soil (WEOC) and the PAO fraction (WEOCP), were highly correlated with aggregate stability (Chapter 4).

153

Spectroscopic analysis of SOM and humic acids (HA) confirmed the greater proportion of organic functional groups generally associated with partial degradation of plant residues (e.g., aromatic, carboxyl and phenolic-C), and a lower index of degradation, especially under grass VFS (Chapter 6). This may be the result of differing quantity and/or quality of organic matter inputs to the soil, or the result of differences in the decomposition rate due to protection in soil aggregates (Elliott, 1986; Landgraf et al.,

2006).

This study also demonstrated the efficacy of a new labile SOM separation technique.

The PAO-C fraction, described in Chapter 3, effectively differentiated VFS from no-till and was highly correlated with many other aspects of SOM quality including aggregate stability and WEOC (Chapter 4), POXC (Chapter 5) and spectroscopic indicators of

SOM composition (Chapter 6). Furthermore, this technique is rapid and inexpensive, making it an ideal candidate for analysis of management effects in agroecosystems. The

PAO fraction elucidated changes in the distribution and quality of SOM under VFS and served as an effective early indicator of management effects.

Overall, this study contributes to a greater understanding of conservation management practices on a field scale, and has implications for the role of management practices in the global soil carbon cycle. Perennial vegetation (i.e., VFS) enhances stocks of labile, partially decomposed SOM by slowing the decomposition rate through physical protection. In contrast, the reduced aggregate stability under no-till leads to increased rates of SOM turnover. Therefore, although no differences in SOC stocks were found among conservation management practices after 10 years, VFS soils may have increased

C sequestration potential over no-till in the long-term.

154

APPENDIX

Two-Factor ANOVA with Subsamples

The study consisted of two factors, landscape position and conservation management practice. In order to test for the effects of these factors, an analysis of variance (ANOVA) was used. In particular, due to the original study design, the analysis should be treated as a two-factor ANOVA with subsamples, from a completely randomized design (CRD).

However, due to lack of replication, there are zero error degrees of freedom under this model specification and thus proper hypothesis tests cannot be formulated for the landscape by conservation management practice interaction. For a comprehensive discussion see Kuehl (2000).

Fortunately, it is still possible to conduct a proper test for the main effect due to conservation management practice and, in principal, for the main effect due to landscape position. In this case, due to the lack of replication, the interaction term becomes the error term (i.e. the replication). As a result, the model becomes

where i = 1,…,a, j =1,…,b, k = 1,…,n, with ~ iid N (0, ) and ~ iid N (0, ) mutually independent. For our specific case a = 3, b = 4, and n = 3; therefore, the correct

ANOVA table is given by:

155

Source Degrees of Mean Square Expected Mean Square Freedom

subsampling Total

where and respectively.

In matrix notation the model can be expressed as

where

, ,

and

Note, is the identity matrix, denotes the Kronecker product and denotes the vector of ones of length k; e.g.,

Further, the defining matrix, , associated with each (quadratic form) sum of squares

can be expressed as

156

where is equal to the matrix of ones. Finally, the covariance matrix of Y is given

In the above model it is important to point out that we are not asserting that the interaction of two fixed effects is being accounted for by a random variable. The issue is that if one includes the interaction term in this model, there are no error degrees of freedom due to lack of replication. However, if one assumes that the model has no significant interaction term, then a constrained model can be imposed (i.e., a model with no interaction term). In this case, the sum-of-squares for the error term becomes the sum- of-squares for the interaction (which is no longer in the model) plus the sum-of-squares for the error. Due to the lack of true replication, this reduces to the sum-of-squares for the interaction. In other words, the interaction serves as the replication and is thus a random effect in the constrained model. Under the constrained model, proper tests for main effects can be formulated.

The effects model, as stated above, is necessary for deriving the relevant test statistics and their associated distributions. However, as currently stated, this model is over- parameterized and would lead to estimation problems without imposing some constraints.

The estimation constraints, imposed by SAS Proc Mixed (SAS® software, Version 9.1.3,

SAS Institute Inc., Cary, NC, USA.), consist of setting the last factor level to zero (e.g., the corresponding row in the design matrix X is set to zero). It is important to note that this constraint is not unique; instead one could use a sum-to-zero constraint (i.e., the sum over all of the levels of a given factor equals zero). However, the sum-to-zero constraint becomes intractable with unbalanced designs and is thus less general and not imposed in

SAS (Littell et al., 2006).

157

Let E[MSα] and E[MSe] denote the expected mean square for conservation management practice and expected mean square error respectively. Under the null hypothesis, H0, of no main effect due to conservation management practice, it follows that Qα = 0. Thus, under the null hypothesis, E[MSα] = E[MSe] and the appropriate test statistic for testing no main effect due to conservation management practice is

Now, under H0, it is straightforward to show that is idempotent and symmetric with rank equal to two and that the noncentrality parameter is equal to zero.

Therefore, it follows that (Theorem 2, Page 57, Searle, 1997), follows a central chi-squared distribution with two degrees of freedom. Analogously,

follows a central chi-square distribution with six degrees of freedom. With the distributions of the sum of squares for conservation management practice and sum of squares for error established, we can combine the two to form an F- statistic. However, the quadratic forms need to be independent. Now, since

and are independent (Theorem

3, Page 59, Searle 1997). Additionally, note that

Hence, F follows a central F distribution with two numerator degrees and six denominator degrees of freedom (Searle, 1997; Page 101). The test statistic for landscape position can be derived analogously. It is important to emphasize that treating the subsamples as true replication would result in an incorrect test statistic. In fact, the test

158

statistic derived in this case would have the same numerator degrees of freedom.

However, the denominator degrees of freedom would now be 24 and could thus result in incorrect scientific inference (Table A1.1).

159

Table A.1.1 Comparison of p-

values using a standard two-

factor ANOVA with

subsamples as replicates

(pseudoreplication) and a

modified two-factor ANOVA

for soil organic carbon

(SOC), and total nitrogen

(TN) calculated on a

concentration, soil volume

and soil mass basis in the

surface layer (0 – 5 cm) for

no-till (NT), agroforestry

(AGF) and grass (GR) soils.

160

BIBLIOGRAPHY

Amezketa, E., 1999. Soil aggregate stability: a review. J. Sustainable Agric. 14(2-3), 83- 151.

Ander, P., Eriksson, K.-E., 1976. The importance of phenol oxidase activity in lignin degradation by the white-rot fungus Sporotrichum pulverulentum. Arch. Microbiol. 109(1), 1-8.

Anderson, S.H., Udawatta, R.P., Seobi, T., Garrett, H.E., 2009. Soil water content and infiltration in agroforestry buffer strips. Agroforestry Systems 75, 5-16.

Angers, D.A., Caron, J., 1998. Plant-induced changes in soil structure: processes and feedbacks. Biogeochemistry 42(1/2), 55-72.

Angers, D.A., Eriksen-Hamel, N.S., 2008. Full-inversion tillage and organic carbon distribution in soil profiles: a meta-analysis. Soil Sci. Soc. Am. J. 72(5), 1370- 1374.

Angers, D.A., Giroux, M., 1996. Recently deposited organic matter in soil water-stable aggregates. Soil Sci. Soc. Am. J. 60(5), 1547-1551.

Angers, D.A., Kay, B.D., Groenevelt, P.H., 1987. Compaction characteristics of a soil cropped to corn and bromegrass. Soil Sci. Soc. Am. J. 51(3), 779-783.

Angers, D.A., Mehuys, G.R., 1989. Effects of cropping on carbohydrate content and water-stable aggregation of a clay soil. Can. J. Soil Sci. 69, 373-380.

Angers, D.A., N'dayegamiye, A., Côté, D., 1993a. Tillage-induced differences in organic matter of particle-size fractions and microbial biomass. Soil Sci. Soc. Am. J. 57(2), 512-516.

Angers, D.A., Samson, N., Legere, A., 1993b. Early changes in water-stable aggregation induced by rotation and tillage in a soil under barley production. Can. J. Soil Sci. 73, 51-59.

Arnold, J.G., Potter, K.N., King, K.W., Allen, P.M., 2005. Estimation of soil cracking and the effect on surface runoff in a Texas Blackland Prairie watershed. Hydrological Processes 19, 589-603.

Baer, J.U., Anderson, S.H., 1997. Landscape effects on dessication cracking in an Aqualf. Soil Sci. Soc. Am. J. 61, 1497-1502.

Baes, A.U., Bloom, P.R., 1989. Diffuse reflectance and transmission fouier transform infrared (DRIFT) spectroscopy of humic and fulvic acids. Soil Sci. Soc. Am. J. 53, 695-700.

161

Baldock, J.A., 2002. Interactions of organic materials and microorganisms with minerals in the stabilization of structure. In: P.M. Huang, J.M. Bollag, N. Denesi (Eds.), Interactions Between Soil Particles and Microorganisms: Impact on the Terrestrial Ecosystem. John Wiley and Sons, West Sussex, England, pp. 85-131.

Baldock, J.A., Oades, J.M., Nelson, P.N., Skene, T.M., Golchin, A., Clarke, P., 1997. Assessing the extent of decompostition of natural organic materials using solid- state 13C NMR spectroscopy. Aust. J. Soil Res. 35, 1061-1083.

Baldock, J.A., Preston, C.M., 1995. Chemistry of carbon decomposition processes in forests as revealed by solid-state carbon-13 nuclear magnetic resonance. In: W.W. McFee, J.M. Kelly (Eds.), Carbon Forms and Functions in Forest Soils. SSSA, Madison, WI, pp. 89-117.

Baldock, J.A., Skemstad, J.O., 2000. Role of the soil matrix and minerals in protecting natural organic materials against biological attack. Org. Geochem. 31, 697-710.

Ball, B.C., Cheshire, M.V., Robertson, E.A.G., Hunter, E.A., 1996. Carbohydrate composition in relation to structural stability, compactibility and plasticity of two soils in a long-term experiment. Soil Till. Res. 39(3–4), 143-160.

Barling, R.D., Moore, I.D., 1994. Role of buffer strips in management of waterway pollution: a review. Environ. Manage. 18(4), 543-558.

Barthes, B., Roose, E., 2002. Aggregate stability as an indicator of soil susceptibility to runoff and erosion; validation at several levels. Catena 47, 133-149.

Beare, M.H., Bruce, R.R., 1993. A comparison of methods for measuring water-stable aggregates: implications for determining environmental effects on soil structure. Geoderma 56(1-4), 87-104.

Beare, M.H., Hendrix, P.F., Cabrera, M.L., Coleman, D.C., 1994. Aggregate-Protected and Unprotected Organic Matter Pools in Conventional- and No-Tillage Soils. Soil Sci. Soc. Am. J. 58(3), 787-795.

Beke, G.J., 1989. Sediment and runoff water characteristics as influenced by cropping and tillage practices. Can. J. Soil Sci. 69, 639-647.

Blair, G.J., Lefroy, R.D.B., Lisle, L., 1995. Soil carbon fractions based on their degree of oxidation, and the development of a carbon management index for agricultural systems. Aust. J. Agric. Res. 46, 1459-1466.

Blanco-Canqui, H., Gantzer, C.J., Anderson, S.H., Alberts, E.E., Ghidey, F., 2002. Saturated hydraulic conductivity and its impact on simulated runoff for claypan soils. Soil Sci. Soc. Am. J. 66(5), 1596-1602.

Blanco-Canqui, H., Lal, R., 2008. No-tillage and soil-profile carbon sequestration: An on-farm assessment. Soil Sci. Soc. Am. J. 72(3), 693-701.

162

Blanco-Canqui, H., Mikha, M.M., Benjamin, J.G., Stone, L.R., Schlegel, A.J., Lyon, D.J., Vigil, M.F., Stahlman, P.W., 2009a. Regional study of no-till impacts on near- surface aggregate properties that influence soil erodibility. Soil Sci. Soc. Am. J. 73(4), 1361-1368.

Blanco-Canqui, H., Stone, L.R., Schlegel, A.J., Lyon, D.J., Vigil, M.F., Mikha, M.M., Stahlman, P.W., Rice, C.W., 2009b. No-till induced increase in organic carbon reduces maximum bulk density of soils. Soil Sci. Soc. Am. J. 73(6), 1871-1879.

Bouajila, A., Gallali, T., 2008. Soil organic carbon fractions and aggregate stability in carbonated and no carbonated soils in Tunisia. Journal of Agronomy 7, 127-137.

Bricklemyer, R.S., Miller, P.R., Paustian, K., Keck, T., Nielsen, G.A., Antle, J.M., 2005. Soil organic carbon variability and sampling optimization in Montana dryland wheat fields. J. Soil Water Conserv. 60(1), 42-51.

Bu, X., Wang, L., Ma, W., Yu, X., McDowell, W.H., Ruan, H., 2010. Spectroscopic characterization of hot-water extractable organic matter from soils under four different vegetation types along an elevation gradient int he Wuyi Mountains. Geoderma 159, 139-146.

Bullock, M.S., Kemper, W.D., Nelson, S.D., 1988. Soil cohesion as affected by freezing, water content, time and tillage. Soil Sci. Soc. Am. J. 52, 770-776.

Burford, J.R., Bremner, J.M., 1975. Relationships between the denitrification capacities of soils and total, water-soluble and readily decomposable soil organic matter. Soil Biol. Biochem. 7(6), 389-394.

Burt, R., 2004. Soil Survey Laboratory Methods Manual. Soil Survey Investigations Report No.42, Version 4.0, USDA-NRCS, Lincoln, NE.

Buyanovsky, G.A., Wagner, G.H., 1998. Carbon cycling in cultivated land and its global significance. Global Change Biol. 4, 131-141.

Caldwell, B.A., 2005. Enzyme activities as a component of soil biodiversity: a review. Pedobiologia 49, 637-644.

Cambardella, C.A., Elliott, E.T., 1992. Particulate soil organic-matter changes across a grassland cultivation sequence. Soil Sci. Soc. Am. J. 56(3), 777-783.

Cambardella, C.A., Elliott, E.T., 1993. Carbon and nitrogen distribution in aggregates from cultivated and native grassland soils. Soil Sci. Soc. Am. J. 57(4), 1071-1076.

Cambardella, C.A., Elliott, E.T., 1994. Carbon and nitrogen dynamics of soil organic matter fractions from cultivated grassland soils. Soil Sci. Soc. Am. J. 58(1), 123- 130.

163

Cantón, Y., Solé-Benet, A., Asensio, C., Chamizo, S., Puigdefábregas, J., 2009. Aggregate stability in range sandy loam soils: Relationships with runoff and erosion. Catena 77(3), 192-199.

Caron, J., Espindola, C.R., Angers, D.A., 1996. Soil structural stability during rapid wetting: Influence of land use on some aggregate properties. Soil Sci. Soc. Am. J. 60, 901-908.

Carter, M.R., 2002. Soil quality for sustainable land management: Organic matter and aggregation interactions that maintain soil functions. Agron. J. 94, 38-47.

Carter, M.R., Kunelius, H.T., Angers, D.A., 1994. Soil structural form and stability, and organic matter under cool-season perennial grasses. Soil Sci. Soc. Am. J. 58(4), 1194-1199.

Casida, L.E., Jr, 1977. Microbial metabolic activity in soil as measured by dehydrogenase determinations. Appl. Environ. Microbiol. 34(6), 630-636.

Chamran, F., Gessler, P.E., Chadwick, O.A., 2002. Spatially explicit treatment of soil- water dynamics along a semiarid catena. Soil Sci. Soc. Am. J. 66, 1571-1583.

Chan, K.Y., Bowman, A., Oates, A., 2001. Oxidizible organic carbon fractions and soil quality changes in an oxic paleustalf under different pasture leys. Soil Sci. 166(1), 61-67.

Chan, K.Y., Heenan, D.P., 1999. Microbial-induced soil aggregate stability under different crop rotations. Biol. Fertility Soils 30(1), 29-32.

Chantigny, M.H., 2003. Dissolved and water-extractable organic matter in soils: a review on the influence of land use and management practices. Geoderma 113, 357-380.

Chantigny, M.H., Angers, D.A., Prévost, D., Vézina, L.-P., Chalifour, F.-P., 1997. Soil aggregation and fungal and bacterial biomass under annual and perennial cropping systems. Soil Sci. Soc. Am. J. 61(1), 262-267.

Chefetz, B., Tarchitzky, J., Benny, N., Hatcher, P.G., Bortiatynski, J., Chen, Y., 1998. Characterization and properties of humic substances orininating from an activated sludge wastewater treatment plant. In: G. Davies, E.A. Ghabbour (Eds.), Humic Substances: Structures, Properties and Uses. The Royal Society of Chemistry, pp. 69-78.

Chen, C.R., Senesi, N., Schnitzer, M., 1977. Information provided on humic substances by E4/E5 ratios. Soil Sci. Soc. Am. J. 41, 352-358.

Chen, C.R., Xu, Z.H., Mathers, N.J., 2004. Soil carbon pools in adjacent natural and plantation forests of subtropical Australia. Soil Sci. Soc. Am. J. 68, 282-291.

164

Chin, Y.P., Aiken, G., O'Loughlin, E., 1994. Molecular weight, polydispersity and spectroscopic properties of aquatic humic substances. Environmental Science and Technology 28, 1853-1858.

Christopher, S.F., Lal, R., Mishra, U., 2009. Regional study of no-till effects on carbon sequestration in the Midwestern United States. Soil Sci. Soc. Am. J. 73(1), 207- 216.

Chu, B., 2011. Sorption and Transport of Veterinary Antibiotics in Agroforestry Buffer, Grass Buffer and Cropland Soils. Department of Soil, Atmospheric and Enivronmental Sciences, University of Missouri.

Chu, B., Goyne, K., Anderson, S., Lin, C.-H., Udawatta, R., 2010. Veterinary antibiotic sorption to agroforestry buffer, grass buffer and cropland soils. Agroforestry Systems 79(1), 67-80.

Clapp, C.E., Stark, S.A., Clay, D.E., Larson, W.E., 1986. Sewage sludge organic matter and soil properties. In: Y. Chen, Y. Avnimelech (Eds.), The Role of Organic Matter in Modern Agriculture. Martinus Nijhoff Publishers, Dordrecht, The Netherlands, pp. 209-253.

Collins, H.P., Elliott, E.T., Paustian, K., Bundy, L.G., Dick, W.A., Huggins, D.R., Smucker, A.J.M., Paul, E.A., 2000. Soil carbon pools and fluxes in long-term corn belt agroecosystems. Soil Biol. Biochem. 32(2), 157-168.

Conant, R.T., Ogle, S.M., Paul, E.A., Paustian, K., 2011. Measuring and monitoring soil organic carbon stocks in agricultural lands for climate mitigation. Front. Ecol. Environ. 9(3), 169-173.

Cook, B.D., Allan, D.L., 1992. Dissolved organic carbon in old field soils: total amounts as a measure of available resources for soil mineralization. Soil Biol. Biochem. 24(6), 585-594.

Cousen, S.M., Farres, J.P., 1984. The role of moisture content in the stability of soil aggregates from a temperate silty soil to raindrop impact. Catena 11, 313-320.

Culman, S.W., Snapp, S.S., Freeman, M.A., Schipanski, M.E., Beniston, J., Lal, R., Drinkwater, L.E., Franzluebbers, A.J., Glover, J.D., Grandy, A.S., Lee, J., Six, J., Maul, J.E., Mirksy, S.B., Spargo, J.T., Wander, M.M., 2012. Permanganate oxidizable carbon reflects a processed soil fraction that is sensitive to management. Soil Sci. Soc. Am. J. 76(2), 494-504.

De Serra, M.O., Schnitzer, M., 1972. Extraction of humic acid by alkali and chelating resin. Can. J. Soil Sci. 52(3), 365-374.

Dick, R.P., 1997. Soil enzyme activities as integrative indicators of soil health. In: C. Pankhurst, B.M. Doube, V.V.S.R. Gupta (Eds.), Biological Indicators of Soil Health. CABI Publishing, New York, NY.

165

Dillaha, T.A., Sherrard, J.H., Lee, D., 1986. Bulletin 153: Long-Term Effectiveness and Maintenance of Vegetative Filter Strips, Virginia Water Resources Research Center, Blacksburg, VA.

Ding, G., Liu, X., Herbert, S., Novak, J., Amarasiriwardena, D., Xing, B., 2006. Effect of cover crop management on soil organic matter. Geoderma 130, 229-239.

Ding, G., Novak, J.M., Amarasiriwardena, D., Hunt, P.G., Xing, B., 2002. Soil organic matter characteristics as affected by tillage management Soil Sci. Soc. Am. J. 66, 421-429

Dixon, W.T., Schaefer, J., Sefcik, M.D., Stejskal, E.O., McKay, R.A., 1982. Total suppression of sidebands in CPMAS C-13 NMR. J. Magn. Reson. Imaging 49(2), 341-345.

Doran, J.W., Jones, A.J., Arshad, M.A., Gilley, J.E., 1998. Determinants of soil quality and health. In: R. Lal (Ed.), Soil Quality and Soil Erosion. CRC Press, Boca Raton, FL, pp. 17-39.

Doran, J.W., Parkin, T.B., 1994. Defining and assessing soil quality. p. 3-21. In: J.W. Doran, D.C. Coleman, D.F. Bezdicek, B.A. Stewart (Eds.), Defining Soil Quality for a Sustainable Environment. SSSA Spec. Publ. No. 35, Madison, WI, pp. 410.

Dou, F., Wright, A.L., Hons, F.M., 2008. Dissolved and soil organic carbon after long- term conventional and no-tillage sorghum cropping. Commun. Soil Sci. Plant Anal. 39, 667-679.

DuPont, S.T., Culman, S.W., Ferris, H., Buckley, D.H., Glover, J.D., 2010. No-tillage conversion of harvested perennial grassland to annual cropland reduces root biomass, decreases active carbon stocks, and impacts soil biota. Agric., Ecosyst. Environ. 137, 25-32.

Durand, B., Nicaise, G., 1980. Procedures for kerogen isolation. In: B. Durand (Ed.), Kerogen: Insoluble Organic Matter from Sedimentary Rocks. Editions Technip, Paris, France, pp. 35-54.

Ellerbrock, R.H., Höhn, A., Rogasik, J., 1999. Functional analysis of soil organic matter as affected by long-term manurial treatment. Eur. J. Soil Sci. 50(1), 65-71.

Ellert, B.H., Bettany, J.R., 1995. Calculation of organic matter and nutrients stored in soils under contrasting management regimes. Can. J. Soil Sci. 75, 529-538.

Ellert, B.H., Janzen, H.H., McConkey, B.G., 2001a. Measuring and comparing soil carbon storage. In: R. Lal, J.M. Kimble, R.F. Follett, B.A. Stewart (Eds.), Assessment Methods for Soil Carbon. CRC Press, Boca Raton, FL, pp. 131-146.

166

Ellert, B.H., Janzen, H.H., McConkey, B.G., 2001b. Measuring and comparing soil carbon storage. In: R. Lal, J.M. Kimble, R.F. Follett, A.J. Stewart (Eds.), Assessment Methods for Soil Carbon. CRC Press, Boca Raton, FL.

Elliott, E.T., 1986. Aggregate structure and carbon, nitrogen, and phosphorus in native and cultivated soils. Soil Sci. Soc. Am. J. 50(3), 627-633.

Elliott, E.T., Palm, C.A., Reuss, D.E., Monz, C.A., 1991. Organic matter contained in soil aggregates from a tropical chronosequence: correction for sand and light fraction. Agriculture, Ecosystems & Environment 34(1–4), 443-451.

Embacher, A., Zsolnay, A., Gattinger, A., Munch, J.C., 2007. The dynamics of water extractable organic matter (WEOM) in common arable topsoils: I. Quantity, quality and function over a three year period. Geoderma 139(1-2), 11-22.

Embacher, A., Zsolnay, A., Gattinger, A., Munch, J.C., 2008. The dynamics of water extractable organic matter (WEOM) in common arable topsoils: II. Influence of mineral and combined mineral and manure fertilization in a Haplic Chernozem. Geoderma 148(1), 63-69.

Essington, M.E., 2004. Soil and water chemistry: an integrative approach. CRC Press, Boca Raton, FL.

Eusterhues, K., Rumpel, C., Kleber, M., Kögel-Knabner, I., 2003. Stabilisation of soil organic matter by interactions with minerals as revealed by mineral dissolution and oxidative degradation. Org. Geochem. 34(12), 1591-1600.

Eynard, A., Schumacher, T.E., Lindstrom, M.J., Malo, D.D., 2005. Effects of agricultural management systems on soil organic carbon in aggregates of Ustolls and Usterts. Soil Till. Res. 81(2), 253-263.

Freeman, C., Evans, C.D., Monteith, D.T., Reynolds, B., Fenner, N., 2001a. Export of organic carbon from peat soils. Nature 412(6849), 785-785.

Freeman, C., Ostle, N., Kang, H., 2001b. An enzymic 'latch' on a global carbon store. Nature 409(6817), 149-149.

Fu, X., Shao, M., Wei, X., Horton, R., 2010. Soil organic carbon and total nitrogen as affected by vegetation types in Northern Loess Plateau of China. Geoderma 155(1-2), 31-35.

Fuentes, M., González-Gaitano, G., García-Mina, J.M., 2006. The usefulness of UV– visible and fluorescence spectroscopies to study the chemical nature of humic substances from soils and composts. Org. Geochem. 37(12), 1949-1959.

García-Orenes, F., Guerrero, C., Mataix-Solera, J., Navarro-Pedreño, J., Gómez, I., Mataix-Beneyto, J., 2005. Factors controlling the aggregate stability and bulk

167

density in two different degraded soils amended with biosolids. Soil Till. Res. 82(1), 65-76.

Garten, C.T., Wullschleger, S.D., 1999. Soil carbon inventories under a bioenergy crop (switchgrass): measurement limitations. J. Environ. Qual. 28(4), 1359-1365.

Gessler, P.E., Chadwick, O.A., Chamran, F., Althouse, L., Holmes, K., 2000. Modeling soil - landscape and ecosystem properties using terrain attributes. Soil Sci. Soc. Am. J. 64, 2046-2056.

Gieguzynska, E., Amine-Khodja, A., Trubetskoj, O.A., Trubetskaya, O.E., Guyot, G., ter Halle, A., Golebiowska, D., Richard, C., 2009. Compositional differences between soil humic acids extracted by various methods as evidenced by photosensitizing and electrophoretic properties. Chemosphere 75(8), 1082-1088.

Gilmour, D., 1961. The Biochemistry of Insects. Academic Press, New Youk.

Golchin, A., Oades, J.M., Skemstad, J.O., Clarke, P., 1994. Study of free and occluded particulate organic matter in soils by solid state 13C CP/MAS NMR spectroscopy and scanning electron microscopy. Aust. J. Soil Res. 32, 285-309.

Gregorich, E.G., Beare, M.H., Stoklas, U., St-Georges, P., 2003. Biodegradability of soluble organic matter in maize-cropped soils. Geoderma 113, 237-252.

Gregorich, E.G., Carter, M.R., Angers, D.A., Monreal, C.M., Ellert, B.H., 1994. Towards a minimum data set to assess soil organic matter quality in agricultural soils. Can. J. Soil Sci. 74, 367–385.

Grossman, R.B., Reinsch, T.G., 2002. Bulk density and linear extensibility. In: J.H. Dane, G.C. Topp (Eds.), Methods of Soil Analysis. Part 4. Physical Methods. Soil Science Society of America, Madison, WI, pp. 201-254.

Guggenberger, G., Zech, W., 1994. Dissolved organic carbon in forest floor leachates: simple degradation products or humic substances? Sci. Total Environ. 152(1), 37- 47.

Guo, L.B., Gifford, R.M., 2002. Soil carbon stocks and land use change: a meta analysis. Global Change Biology 8(4), 345-360.

Guo, M., Chorover, J., 2003. Transport and fractionation of dissolved organic matter in soil columns. Soil Sci. 168(2), 108-118.

Guzman, J.G., Al-Kaisi, M., 2010. Landscape position and age of reconstructed prairies effect on soil organic carbon sequestration rate and aggregate associated carbon. J. Soil Water Conserv. 65(1), 9-21.

168

Guzman, J.G., Al-Kaisi, M., 2011. Landscape position effect on selected soil physical properties of reconstructed prairies in southcentral Iowa. J. Soil Water Conserv. 66(3), 183-191.

Hancock, G.R., Murphy, D., Evans, K.G., 2010. Hillslope and catchment scale soil organic carbon concentration: An assessment of the role of geomorphology and soil erosion in an undisturbed environment. Geoderma 155(1-2), 36-45.

Hansen, E.A., 1993. Soil carbon sequestration beneath hybrid poplar plantations in the North Central United States. Biomass Bioenergy 5(6), 431-436.

Haynes, R.J., 1993. Effect of sample pretreatment on aggregate stability measured by wet sieving or turbidimetry on soils of different cropping history. Eur. J. Soil Sci. 44(2), 261-270.

Haynes, R.J., 2000. Interactions between soil organic matter status, cropping history, method of quantification and sample pretreatment and their effects on measured aggregate stability. Biol. Fertility Soils 30(4), 270-275.

Haynes, R.J., 2005. Labile organic matter fractions as central components of the quality of agricultural soils: a review. Advances in Agronomy 85, 221-268.

Haynes, R.J., Beare, M.H., 1996. Aggregation and organic matter storage in meso- thermal, humid soils. In: M.R. Carter, B.A. Stewart (Eds.), Structure and Organic Matter Storage in Agricultural Soils. CRC Lewis Publishers, New York, NY, pp. 213-262.

Haynes, R.J., Swift, R.S., 1990. Stability of soil aggregates in relation to organic constituents and soil water content. Eur. J. Soil Sci. 41(1), 73-83.

He, Z., Mao, J., Honeycutt, C., Ohno, T., Hunt, J., Cade-Menun, B., 2009. Characterization of plant-derived water extractable organic matter by multiple spectroscopic techniques. Biol. Fertility Soils 45(6), 609-616.

Helfrich, M., Ludwig, B., Buurman, P., Flessa, H., 2006. Effect of land use on the composition of soil organic matter in density and aggregate fractions as revealed by solid-state 13C NMR spectroscopy. Geoderma 136, 331-341.

Hobbie, S.E., 1996. Temperature and plant species control over litter decomposition in Alaskan tundra. Ecol. Monogr. 66(4), 503-522.

Holland, J.M., 2004. The environmental consequences of adopting conservation tillage in Europe: reviewing the evidence. Agriculture, Ecosystems and Environment 103, 1-25.

Horowitz, J., Ebel, R., Ueda, K., 2010. "No-Till" Farming is a Growing Practice. Economic Information Bulletin 70, U.S. Department of Agriculture.

169

Hunt, J.F., Ohno, T., 2007. Characterization of fresh and decomposed dissolved organic matter using excitation−emission matrix fluorescence spectroscopy and multiway analysis. J. Agric. Food Chem. 55(6), 2121-2128.

Idowu, O.J., 2003. Relationships between aggregate stability and selected soil properties in humid tropical environment. Commun. Soil Sci. Plant Anal. 34(5), 695 - 708.

Inbar, Y., Chen, Y., Hadar, Y., 1989. Solid-state carbon-13 nuclear magnetic resonance and infrared spectroscopy of composted organic matter. Soil Sci. Soc. Am. J. 53(6), 1695-1701.

Izaurralde, R., Williams, J., Post, W., Thomson, A., McGill, W., Owens, L., Lal, R., 2007. Long-term modeling of soil C erosion and sequestration at the small watershed scale. Clim. Change 80(1), 73-90.

Jastrow, J., Amonette, J., Bailey, V., 2007. Mechanisms controlling soil carbon turnover and their potential application for enhancing carbon sequestration. Clim. Change 80(1), 5-23.

Jenkinson, D.S., 1988. Soil organic matter and its dynamics. In: A. Wild (Ed.), Russell's Soil Conditions and Plant Growth. John Wiley and Sons, Inc., New York, pp. 564-607.

Jiang, P., Anderson, S.H., Kitchen, N.R., Sadler, E.J., Sudduth, K.A., 2007. Landscape and conservation management effects on hydraulic properties of a claypan-soil toposequence. Soil Sci. Soc. Am. J. 71(3), 803-811.

Jones, M.B., Donnelly, A., 2004. Carbon sequestration in temperate grassland ecosystems and the influence of management, climate and elevated CO2. New Phytol. 164(3), 423-439.

Jordan, D., Kremer, R.J., Bergfield, W.A., Kim, K.Y., Cacnio, V.N., 1995. Evaluation of microbial methods as potential indicators of soil quality in historical agricultural fields Biol. Fertility Soils 19(4), 297-302.

Kalbitz, K., Schmerwitz, J., Schwesig, D., Matzner, E., 2003a. Biodegradation of soil- derived dissolved organic matter as related to its properties. Geoderma 113, 273- 291.

Kalbitz, K., Schwesig, D., Schmerwitz, J., Kaiser, K., Haumaier, L., Glaser, B., Ellerbrock, R., Leinweber, P., 2003b. Changes in properties of soil-derived dissolved organic matter induced by biodegradation. Soil Biol. Biochem. 35(8), 1129-1142.

Karlen, D.L., Mausbach, M.J., Doran, J.W., Cline, R.G., Harris, R.F., Schuman, G.E., 1997. Soil quality: a concept, definition, and framework for evaluation. Soil Sci. Soc. Am. J. 61, 4-10.

170

Kay, B.D., 1998. Soil structure and organic carbon: a review. In: R. Lal, J.M. Kimble, R.F. Follett, B.A. Stewart (Eds.), Soil Processes and the Carbon Cycle. CRC Press, Boca Raton, FL, pp. 169-197.

Kemper, W.D., Koch, E.J., 1966a. Aggregate stability of soils from western United Sates and Canada. Technical Bulletin No. 1355. U.S. Department of Agriculture- Agriculture Research Service, pp. 52.

Kemper, W.D., Koch, E.J., 1966b. Aggregate stability of soils from Western United States and Canada Technical Bulletin no. 1355. U.S. Dept. of Agriculture Agricultural Research Service.

Kemper, W.D., Rosenau, R.C., 1984. Soil cohesion as affected by time and water content. Soil Sci. Soc. Am. J. 48(5), 1001-1006.

Kemper, W.D., Rosenau, R.C., 1986. Aggregate stability and size distribution. In: A. Klute (Ed.), Methods of Soil Analysis, Part I. Physical and Mineralogical Methods - Agronomy Monograph no. 9. American Society of Agronomy - Soil Science Society of America, Madison, WI, pp. 425-442.

Kennedy, A., Smith, K., 1995. Soil microbial diversity and the sustainability of agricultural soils. Plant Soil 170(1), 75-86.

Kitchen, N.R., Sudduth, K.A., Drummond, S.T., 1999. Soil electrical conductivity as a crop productivity measure for claypan soils. J. Prod. Agric. 12, 607-617.

Kitchen, N.R., Sudduth, K.A., Kremer, R.J., Myers, D.B., 2010. Sensor-based mapping of soil quality on degraded claypan landscapes of the Central United States. In: R.A. Viscarra Rossel (Ed.), Proximal Soil Sensing: Progress in Soil Science 1. Springer Science, pp. 363-373.

Knops, J.M.H., Tilman, D., 2000. Dynamics of soil nitrogen and carbon accumulation for 61 years after agricultural abandonment. Ecology 81(1), 88-98.

Kogel-Knabner, I., 1993. Biodegradation and humification processes in forest soils. In: J.M. Bollag, G. Stotzky (Eds.), Soil Biogeochemistry. Marcel Dekker, New York, NY, pp. 101-137.

Kogel-Knabner, I., 1997. 13C and 15N NMR spectroscopy as a tool in soil organic matter studies. Geoderma 80, 243-270.

Korshin, G.V., Li, C.W., Benjamin, M.M., 1997. Monitoring the properties of natural organic matter through UV spectroscopy. Evaluation of a consistent theory. Water Res. 31(7), 1787-1795.

Kravchenko, A.N., Robertson, G.P., 2011. Whole-profile soil carbon stocks: the danger of assuming too much from analyses of too little. Soil Sci. Soc. Am. J. 75(1), 235- 240.

171

Krug, E.C., Hollinger, S.E., 2003. Identification of Factors that Aid Carbon Sequestration in Illinois Agricultural Systems. Contract Report 2003-02, Illinois Department of Natural Resources, Illinois State Water Survey, Atmospheric Environment Section. 103 p.

Kuehl, R., 2000. Design of Experiments: Statistical Principles of Research Design and Analysis, 2nd ed. Duxbury Press, Pacific Grove, CA.

Lal, R., 2004. Soil carbon sequestration to mitigate climate change. Geoderma 123(1-2), 1-22.

Lal, R., Kimble, J.M., Follett, R.F., Stewart, B.A., 1998. Soil Processes and the Carbon Cycle. CRC Press, Boca Raton, FL.

Lal, R., Logan, T.J., Eckert, D.J., Dick, W., Shipitalo, M.J., 1994. Conservation tillage in the corn belt of the United States. In: M.R. Carter (Ed.), Conservation Tillage in Temperate Agroecosystems: Development and Adaptation to Soil, Climatic, and Biological Constraints. CRC Press, Boca Raton, FL, pp. 73-114.

Lal, R., Reicosky, D.C., Hanson, J.D., 2007. Evolution of the plow over 10,000 years and the rationale for no-till farming. Soil Till. Res. 93, 1-12.

Landgraf, D., Leinweber, P., Makeschin, F., 2006. Cold and hot water–extractable organic matter as indicators of litter decomposition in forest soils. J. Plant Nutr. Soil Sci. 169(1), 76-82.

Li, K., Xing, B., Torello, W.A., 2005. Effect of organic fertilizers derived dissolved organic matter on pesticide sorption and leaching. Environ. Pollut. 134, 187-194.

Li, Z., Shuman, L.M., 1997. Mobility of Zn, Cd and Pb in soils as affected by poultry litter extract-I. Leaching in soil columns. Environ. Pollut. 95(2), 219-226.

Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D., Schabenberger, O., 2006. SAS for Mixed Models, 2nd edition. SAS Institute Inc., Cary, NC.

Liu, X., Zhang, X., Zhang, M., 2008. Major factors influencing the efficacy of vegetated buffers on sediment trapping: a review and analysis. J. Environ. Qual. 37, 1667- 1674.

MacDicken, K.G., 1997. A Guide to Monitoring Carbon Storage in Forestry and Agroforestry Projects Winrock International Institute for Agricultural Development: Forest Carbon Monitoring Program. 91p.

Mann, L.K., 1986. Changes in soil carbon storage after cultivation. Soil Sci. 142(5), 279- 288.

172

Mao, J.-D., Hu, W.-G., Schmidt-Rohr, K., Davies, G., Ghabbour, E.A., Xing, B., 2000. Quantitative characterization of humic substances by solid-state carbon-13 nuclear magnetic resonance. Soil Sci. Soc. Am. J. 64, 873-884.

Marschner, B., Kalbitz, K., 2003. Controls of bioavailability and biodegradability of dissolved organic matter in soils. Geoderma 113, 211-235.

Mason, R., Gunst, R., Hess, J., 1989. Statistical Design and Analysis of Experiments. Wiley, New York.

McKnight, D.M., Bencala, K.E., Zellweger, G.W., Aiken, G.R., Feder, G.L., Thorn, K.A., 1992. Sorption of dissolved organic carbon by hydrous aluminum and iron oxides occurring at the confluence of Deer Creek with the Snake River, Summit County, Colorado. Environ. Sci. Technol. 26(7), 1388-1396.

Miller, M., Dick, R.P., 1995. Dynamics of soil C and microbial biomass in whole soil and aggregates in two cropping systems. Appl Soil Ecol 2(4), 253-261.

Minasny, B., McBratney, A.B., 2001. A rudimentary mechanistic model for soil formation and landscape development: II. A two-dimensional model incorporating chemical weathering. Geoderma 103(1-2), 161-179.

Morgan, J.A., Follett, R.F., Allen, L.H., Del Grosso, S., Derner, J.D., Dijkstra, F., Franzluebbers, A.J., Fry, R., Paustain, K., Schoeneberger, M.M., 2010. Carbon sequestration in agricultural lands of the United States. J. Soil Water Conserv. 65(1), 6A-13A.

Munoz-Carpena, R., Parsons, J.E., Gilliam, J.W., 1993. Numerical approach to the overland flow process in vegetative filter strips. Transactions of the American Society of Agricultural Engineers 36(3), 761-770.

Nishiyama, M., Sumikawa, Y., Guan, G., Marumoto, T., 2001. Relationship between microbial biomass and extractable organic carbon content in volcanic and non- volcanic ash soil. Appl Soil Ecol 17(2), 183-187.

Oades, J.M., 1984. Soil organic matter and structural stability: mechanisms and implications for management. Plant Soil 76, 319-337.

Oades, J.M., Waters, A.G., 1991. Aggregate hierarchy in soils. Aust. J. Soil Res. 29, 815- 828.

Oelbermann, M., Paul Voroney, R., Gordon, A.M., 2004. Carbon sequestration in tropical and temperate agroforestry systems: a review with examples from Costa Rica and southern Canada. Agric., Ecosyst. Environ. 104(3), 359-377.

Palmer, J.M., Evans, C.S., 1983. The enzymatic degradation of lignin by white-rot fungi, Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences. The Royal Society, pp. 293-303.

173

Panabokke, C., Quirk, J., 1957. Effect of initial water content on stability of soil aggregates in water. Soil Sci. 83(3), 185-196.

Pankhurst, C.E., Ophel-Keller, K., Doube, B.M., Gupta, V.V.S.R., 1996. Biodiversity of soil microbial communities in agricultural systems. Biodivers. Conserv. 5(2), 197- 209.

Parsons, J.E., Gilliam, J.W., Munoz-Carpena, R., Daniels, R.B., Dillaha, T.A., 1994. Nutrient and sediment removal by grass and riparian buffers. In: K.L. Campbell, W.D. Graham, A.B. Bottcher (Eds.), Environmentaly Sound Agriculture: Proceedings of the Second Conference. American Society of Agricultural Engineers, Orlando, FL, pp. 147-154.

Parsons, J.W., Tinsley, J., 1975. Nitrogenous substances. In: J.E. Gieseking (Ed.), Soil Components Volume 1. Springer-Verlag, New York, pp. 263-304.

Paul, A., Paustian, K., Elliott, E.T., Cole, C.V., 1997. Soil Organic Matter in Temperate Ecosystems: Long-Term Experiments in North America. CRC Press, NY, NY.

Paz-González, A., Vieira, S.R., Taboada Castro, M.T., 2000. The effect of cultivation on the spatial variability of selected properties of an umbric horizon. Geoderma 97(3- 4), 273-292.

Peuravuori, J., Pihlaja, K., 1997. Molecular size distribution and spectroscopic properties of aquatic humic substances. Anal. Chim. Acta 337, 133-149.

Pikul, J.L., Jr., Chilom, G., Rice, J., Eynard, A., Schumacher, T.E., Nichols, K., Johnson, J.M.F., Wright, S., Caesar, T., Ellsbury, M., 2009. Organic matter and water stability of field aggregates affected by tillage in South Dakota. Soil Sci. Soc. Am. J. 73(1), 197-206.

Pind, A., Freeman, C., Lock, M., 1994. Enzymic degradation of phenolic materials in peatlands — measurement of phenol oxidase activity. Plant Soil 159(2), 227-231.

Post, W.M., Pastor, J., Zinke, P.J., Stangenberger, A.G., 1985. Global patterns of soil nitrogen storage. Nature 317(6038), 613-616.

Preston, C.M., 1996. Application of NMR to soil organic matter analysis: history and prospects. Soil Sci. 16, 829-852.

Preston, C.M., Newman, R.H., Rother, P., 1994. Using 13C CPMAS NMR to assess effects of cultivation on the organic matter of particle-size fractions in a grassland soil. Soil Sci. 157, 26-35.

Puget, P., Chenu, C., Balesdent, J., 1995. Total and young organic matter distributions in aggregates of silty cultivated soils. Eur. J. Soil Sci. 46, 449-459.

174

Puget, P., Lal, R., 2005. Soil organic carbon and nitrogen in a Mollisol in central Ohio as affected by tillage and land use. Soil Till. Res. 80(1-2), 201-213.

Qualls, R.G., 2004. Biodegradability of humic substances and other fractions of decomposing leaf litter. Soil Sci. Soc. Am. J. 68, 1705-1712.

Qualls, R.G., Haines, B.L., 1992. Biodegradability of dissolved organic matter in forest throughfall, soil solution, and stream water. Soil Sci. Soc. Am. J. 56(2), 578-586.

Reichert, J., Norton, D., Favaretto, N., Huang, C.-H., Blume, E., 2009. Settling velocity, aggregate stability, and interrill erodibility of soils varying in clay mineralogy. Soil Sci. Soc. Am. J. 73(4), 1369-1377.

Rietveld, B., Irwin, K., 1996. Agroforestry Notes: Agroforestry in the United States, U.S. Department of Agriculture Forest Service.

Ross, R.C., Fielder, V.B., Walter, G.H., 1944. Farming in Cumberland County: A Study of the Effect of Differences in Soil and Topography, University of Illinois Agricultural Experiment Station & U.S. Department of Agriculture Bulletin 506.

Ruehlmann, J., Korschens, M., 2009. Calculating the effect of soil organic matter concentration on soil bulk density. Soil Sci. Soc. Am. J. 73(3), 876-885.

Rumpel, C., Kögel-Knabner, I., Bruhn, F., 2002. Vertical distribution, age, and chemical composition of organic carbon in two forest soils of different pedogenesis. Org. Geochem. 33(10), 1131-1142.

Rumpel, C., Rabia, N., Derenne, S., Quenea, K., Eusterhues, K., Kogel-Knabner, I., Mariotti, A., 2006. Alteration of soil organic matter following treatment with hydrofluoric acid (HF). Org. Geochem. 37, 1437-1451.

Schlesinger, W.H., 1986. Changes in soil carbon storage and associated properties with disturbance and recovery. In: J.R. Trabalka, D.E. Reichle (Eds.), The Changing Carbon Cycle. Springer, New York, pp. 194-220.

Schoenholtz, S.H., Van Miegroet, H., Burger, J.A., 2000. A review of chemical and physical properties as indicators of forest soil quality: challenges and opportunities. For. Ecol. Manage. 138, 335-356.

Searle, S.R., 1997. Linear Models. Wiley, New York.

Semlitsch, R.D., Jensen, J.B., 2001. Core habitat, not buffer zone. National Wetlands Newsletter 23(4), 5-11.

Shepherd, T.G., Saggar, S., Newman, R.H., Ross, C.W., J.L., D., 2001. Tillage-induced changes to soil structure and organic carbon fractions in New Zealand soils. Aust. J. Soil Res. 39, 465-489.

175

Shi, W., Dell, E., Bowman, D., Iyyemperumal, K., 2006. Soil enzyme activities and organic matter composition in a turfgrass chronosequence. Plant Soil 288(1), 285- 296.

Simon, T., 2007. Quantitative and qualitative characterization of soil organic matter in the long-term fallow experiment with different fertilization and tillage. Archives of Agronomy and Soil Science 53(3), 241-251.

Six, J., Bossuyt, H., Degryze, S., Denef, K., 2004. A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil Till. Res. 79, 7-31.

Six, J., Conant, R.T., Paul, E.A., Paustian, K., 2002. Stabilization mechanisms of soil organic matter: Implications for C-saturation of soils. Plant Soil 241(2), 155-176.

Six, J., Elliott, E.T., Paustian, K., 1999. Aggregate and soil organic matter dynamics under conventional and no-tillage systems. Soil Sci. Soc. Am. J. 63(5), 1350- 1358.

Six, J., Elliott, E.T., Paustian, K., 2000a. Soil macroaggregate turnover and microaggregate formation: a mechanism for C sequestration under no-tillage agriculture. Soil Biol. Biochem. 32, 2099-2103.

Six, J., Elliott, E.T., Paustian, K., Doran, J.W., 1998. Aggregation and soil organic matter accumulation in cultivated and native grassland soils. Soil Sci. Soc. Am. J. 62(5), 1367-1377.

Six, J., Guggenberger, G., Paustain, K., Haumaier, L., Elliott, E.T., Zech, W., 2001. Sources and composition of soil organic matter fractions between and within soil aggregates. Eur. J. Soil Sci. 52, 607-618.

Six, J., Paustain, K., Elliott, E.T., Combrink, C., 2000b. Soil structure and organic matter: I. Distribution of aggregate-size classes and aggregate-associated carbon. Soil Sci. Soc. Am. J. 64, 681-689.

Skjemstad, J.O., Clarke, P., Taylor, J.A., Oades, J.M., Newman, R.H., 1994. The removal of magnetic materials from surface soils. A solid state 13C CP/MAS n.m.r. study. Aust. J. Soil Res. 32, 1215-1229.

Smith, F.G., Stotz, E., 1949. A colorimetric method for the determination of phenol oxidase in plant material. Journal Paper No. 773, New York State Agricultural Experiment Station, Cornell University, Ames, IA.

Sollins, P., Homann, P., Caldwell, B.A., 1996. Stabilization and destabilization of soil organic matter: Mechanisms and controls. Geoderma 74(1-2), 65-105.

Sposito, G., Weber, J.H., 1986. Sorption of trace metals by humic materials in soils and natural waters. Critical Reviews in Environmental Control 16(2), 193-229.

176

Staddon, W.J., Locke, M.A., Zablotowicz, R.M., 2001. Microbiological characteristics of a vegetative buffer strip soil and degradation and sorption of Metolachlor. Soil Sci. Soc. Am. J. 65(4), 1136-1142.

Stevenson, F.J., 1994. Humus Chemistry: Genesis, Composition, Reactions. John Wiley & Sons, New York.

Swanston, C.W., Torn, M.S., Hanson, P.J., Southon, J.R., Garten, C.T., Hanlon, E.M., Ganio, L., 2005. Initial characterization of processes of soil carbon stabilization using forest stand-level radiocarbon enrichment. Geoderma 128(1-2), 52-62.

Swift, R.S., 1996. Organic matter characterization. In: D.L. Sparks (Ed.), Methods of Soil Analysis. Part 3. Chemical Methods. SSSA, Madison, WI, USA, pp. 1011-1069.

Tabatabai, M.A., Dick, W.A., 2002. Enzymes in Soil: Research and developments in measuring activities. In: R.G. Burns, R.P. Dick (Eds.), Enzymes in the Environment: Activity, Ecology, and Applications. Marcel Dekker, Inc., New York, pp. 567-596.

Takacs, M., Fuleky, G., 2010. Characterization of dissolved organic matter (DOM) extracted from soils by hot water percolation (HWP). Agrokemia es Talajtan 59(1), 99-108.

Tate, K.W., van Kessel, C., Atwill, E.R., Dahlgren, R.A., 2004. Evaluating the effectiveness of vegetated buffers to remove nutrients, pathogens, and sediment transported in runoff from grazed, irrigated pastures, University of California Water Resources Center.

Tatzber, M., Stemmer, M., Spiegel, H., Katzlberger, C., Haberhauer, G., Gerzabek, M.H., 2008. Impact of different tillage practices on molecular characteristics of humic acids in a long-term field experiment -- An application of three different spectroscopic methods. Sci. Total Environ. 406(1-2), 256-268.

Tirol-Padre, A., Ladha, J.K., 2004. Assessing the reliability of permanganate-oxidizable carbon as an index of soil labile carbon. Soil Sci. Soc. Am. J. 68(3), 969-978.

Tisdall, J.M., Oades, J.M., 1980. The effect of crop rotation on aggregation in a red- brown earth. Aust. J. Soil Res. 18, 423-433.

Tisdall, J.M., Oades, J.M., 1982. Organic matter and water-stable aggregates in soils. J. Soil Sci. 33, 141-163.

Toberman, H., Evans, C.D., Freeman, C., Fenner, N., White, M., Emmett, B.A., Artz, R.R.E., 2008. Summer drought effects upon soil and litter extracellular phenol oxidase activity and soluble carbon release in an upland Calluna heathland. Soil Biol. Biochem. 40(6), 1519-1532.

177

Tufekcioglu, A., Raich, J.W., Isenhart, T.M., Schultz, R.C., 2003. Biomass, carbon and nitrogen dynamics of multi-species riparian buffers within an agricultural watershed in Iowa, USA. Agroforestry Systems 57, 187-198.

Udawatta, R.P., Henderson, G.S., Jones, J.R., Hammer, R.D., 2006. Runoff and sediment from row-crop, row-crop with grass strips, pasture, and forest watersheds. Rev Sci Eau 19(2), 137-149.

Udawatta, R.P., Motavalli, P.P., Garrett, H.E., 2004. Phosphorus loss and runoff characteristics in three adjacent agricultural watersheds with claypan soils. J. Environ. Qual. 33, 1709-1719.

Uhle, M.E., Chin, Y.-P., Aiken, G.R., McKnight, D.M., 1999. Binding of polychlorinated biphenyls to aquatic humic substances: the role of substrate and sorbate properties on partitioning. Environ. Sci. Technol. 33(16), 2715-2718.

Unklesbay, A.G., Vineyard, J.D., 1992. Missouri Geology. Univ. of Missouri Press, Columbia, MO.

USDA-NRCS, 1998. World Soil Resources: Global Major Land Resource Stresses Map, Washington, DC.

USDA-NRCS, 2006. Land resource regions and major land resource areas of the United States, the Caribbean and the Pacific Basin. Agriculture Handbook 296, USDA- NRCS, Government Printing Office, Washington, DC.

USDA-NRCS, 2010. Conservation Practice Standard: Contour Buffer Strips. Code 332, USDA-NRCS, Government Printing Office, Washington, DC.

VandenBygaart, A.J., Angers, D.A., 2006. Towards accurate measurements of soil organic carbon stock change in agroecosystems. Can. J. Soil Sci. 86(3), 465−471.

VandenBygaart, A.J., Bremer, E., McConkey, B.G., Ellert, B.H., Janzen, H.H., Angers, D.A., Carter, M.R., Drury, C.F., Lafond, G.P., McKenzie, R.H., 2011. Impact of sampling depth on differences in soil carbon stocks in long-term agroecosystem experiments. Soil Sci. Soc. Am. J. 75(1).

VandenBygaart, A.J., Gregorich, E.G., Angers, D.A., McConkey, B.G., 2007. Assessment of the lateral and vertical variability of soil organic carbon. Can. J. Soil Sci. 87(4), 433-444.

Veum, K.S., Goyne, K.W., Holan, S.H., Motavalli, P.P., 2011. Assessment of soil organic carbon and total nitrogen under conservation management practices in the Central Claypan Region, Missouri, USA. Geoderma 167/168, 188-196.

Veum, K.S., Goyne, K.W., Motavalli, P.P., Udawatta, R.P., 2009. Runoff and dissolved organic carbon loss from a paired-watershed study of three adjacent agricultural watersheds. Agric., Ecosyst. Environ. 130(3-4), 115-122.

178

Virto, I., Moni, C., Swanston, C., Chenu, C., 2009. Turnover of intra- and extra-aggregate organic matter at the silt-size scale. Geoderma 156(1-2), 1-10.

Wander, M., Nissen, T., 2004. Value of soil organic carbon in agricultural lands. Mitigation and Adaptation Strategies for Global Change 9(4), 417-431.

Wander, M.M., Bidart, M.G., Aref, S., 1998. Tillage impacts on depth distribution of total and particulate organic matter in three Illinois soils. Soil Sci. Soc. Am. J. 62(6), 1704-1711.

Wander, M.M., Traina, S.J., 1996. Organic matter fractions from organically and conventionally managed soils: II. Characterization of compostion. Soil Sci. Soc. Am. J. 60, 1087-1094.

Watson, F.C., 1979. Soil Survey of Knox, Monroe and Shelby Counties, Missouri, Soil Conservation Service , Missouri Agric. Exp. Stn.

Watts, C.W., Dexter, A.R., 1997. The influence of organic matter in reducing the destabilization of soil by simulated tillage. Soil Till. Res. 42(4), 253-275.

Webster, E.A., Chudek, J.A., Hopkins, D.W., 2000. Carbon transformations during decompositions of different components of plant leaves in soil. Soil Biol. Biochem. 32, 301-314.

Weil, R.R., Islam, K.R., Stine, M.A., Gruver, J.B., Samson-Liebig, S.E., 2003. Estimating active carbon for soil quality assessment: a simplified method for laboratory and field use. Am. J. Alternative Agric. 18(1), 3-17.

West, T.O., Post, W.M., 2002. Soil organic carbon sequestration rates by tillage and crop rotation: a global data analysis. Soil Sci. Soc. Am. J. 66, 1930-1946.

Wiebe, K., Gollehon, N., 2006. Agricultural Resources and Environmental Indicators. EIB-16, U.S. Department of Agriculture - Economic Research Service.

Williams, C.J., Shingara, E.A., Yavitt, J.B., 2000. Phenol oxidase activity in peatlands in New York state: response to summer drought and peat type. Wetlands 20(2), 416- 421.

Wilson, J.P., Gallant, J.C., 2000. Digital Terrain Analysis. In: J.P. Wilson, J.C. Gallant (Eds.), Terrain Analysis: Principles and Applications. John Wiley and Sons, NY, NY, pp. 1-27.

Wilson, M.A., 1987. NMR Techniques and Applications in Geochemistry and Soil Chemistry. Pergamon Press, Oxford.

Xing, B., Mao, J.D., Hu, W.-G., Schmidt-Rohr, K., Davies, G., Ghabbour, E.A., 1999. Evaluation of different solid-state 13C NMR techniques for charaterizing humic acids. In: G. Davies, E.A. Ghabbour (Eds.), Understanding Humic Substances:

179

Advanced Methods, Properties and Applications. The Royal Society of Chemistry, Cambridge, pp. 49-61.

Yang, X.M., Drury, C.F., Wander, M.M., Kay, B.D., 2008. Evaluating the effect of tillage on carbon sequestration using the minimum detectable difference concept. Pedosphere 18(4), 421-430.

Yimer, F., Ledin, S., Abdelkadir, A., 2007. Changes in soil organic carbon and total nitrogen contents in three adjacent land use types in the Bale Mountains, south- eastern highlands of Ethiopia. For. Ecol. Manage. 242(2-3), 337-342.

Yoder, R.E., 1936. A direct method of aggregate analysis of soils and a study of the physical nature of erosion losses. Journal of the American Society of Agronomy 28, 337-351.

Yoo, K., Amundson, R., Heimsath, A.M., Dietrich, W.E., 2006. Spatial patterns of soil organic carbon on hillslopes: Integrating geomorphic processes and the biological C cycle. Geoderma 130(1-2), 47-65.

Young, A., Menz, K., Muraya, P., Smith, C., 1998. SCUAF Version 4: A Model to Estimate Soil Changes Under Agriculture, Agroforestry and Forestry. ACIAR Technical Report Series No. 41. 49 p, The Australian Centre for International Research.

Zak, D.R., Grigal, D.F., Gleeson, S., Tilman, D., 1990. Carbon and nitrogen cycling during old-field succession: constraints on plant and microbial biomass. Biogeochemistry 11(2), 111-129.

Zsolnay, A., 2003. Dissolved organic matter: artefacts, definitions and functions. Geoderma 113, 187-209.

180

VITA

KRISTEN SLOAN VEUM

EDUCATION

2012 Ph.D., Soil Science: Emphasis in Environmental Soil Chemistry – University of Missouri. Advisor: K.W. Goyne. Dissertation: Chemical and spectroscopic characterization of soil organic matter under varying conservation management practices.

2006 M.S., Fisheries and Wildlife: Emphasis in Chemical Limnology – University of Missouri. Advisor: J. R. Jones. Thesis: Disinfection by- product precursors and formation potentials of Missouri reservoirs.

2003 B.S., Geological Sciences: Emphasis in Hydrogeology and Chemistry – University of Missouri.

REFEREED PUBLICATIONS

Veum, K.S., Goyne, K.W., Kremer, R., Motavalli, P.P. Water stable aggregates and organic matter fractions under agricultural conservation management practices. (under review, 2012, Soil Science Society of America Journal).

Veum, K.S., Goyne, K.W., Holan, S.H., Motavalli, P.P. (2011). Assessment of soil organic carbon and total nitrogen under conservation management practices in the Central Claypan Region, Missouri, USA. Geoderma 167/168: 188-196.

Hua, B., Veum, K.., Yang, J., Jones, J., Deng, B. (2010). Parallel factor analysis of fluorescence EEM spectra for identification of THM precursors in lake waters. Environmental Monitoring and Assessment 161(1-4): 71-81.

Veum, K.S., Goyne, K.W., Motavalli, P. Udawatta, R.P. (2009). Runoff and dissolved organic carbon loss from a paired-watershed study of three adjacent agricultural watersheds. Agriculture, Ecosystems, and Environment 130(3-4): 115-122.

Hua, B., Veum, K., Koirala, A., Jones, J., Clevenger, T., Deng, B. (2006). Fluorescence fingerprints to monitor total trihalomethanes and N-nitrosodimethylamine formation potentials in water. Environmental Chemistry Letters 5(2): 73-77.

181