“‘v‘r'i': '
go}:

’2';

 

 

._
n hf

fiWflmud.

.ft

.30, A
b ff!
IAr‘lvrl

h

I:
If ..
fizlze.

 

 

 

 

 

2T lungullij'flijlflli[lljfjlzlymll

This is to certify that the

dissertation entitled

Mineralization, Immobilization, and 15N-NMR
Spectroscopy of Organic Nitrogen in Whole
Soil and Particle-Size Fractions

presented by

Ralph John DiCosty

has been accepted towards fulfillment
of the requirements for

Ph.D Crop and Soil Sciences

 

degree in

 

M—

Major professor

Date $2 71/7?

MSU is an Affirmative Action/Equal Opportunity Institution 012771

vv

-WV 5— _4

 

 

LIBRARY

Michigan State

University

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we (#090000033651).“

MINERALIZATION, IMMOBILAZATION AND 1’N-NMR SPECTROSCOPY OF
ORGANIC NITROGEN IN WHOLE SOIL AND PARTICLE-SIZE FRACTIONS
By

Ralph John DiCosty

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Crop and Soil Sciences

1 999

MI.‘

necessz
pollutic
humific
authors.
The obj
magic-a
N fimcn‘
Stabiliza-
Was qua:
imponan
’4 montk

samples I

ABSTRACT
MINERALIZATION, IMMOBILAZATION AND l’N-NMR SPECTROSCOPY OF
ORGANIC NITROGEN 1N WHOLE SOIL AND PARTICLE-SIZE FRACTIONS
By

Ralph John DiCosty

Knowledge of the mechanisms controlling nitrogen (N) turnover in soils is
necessary to maximize agricultural production while minimizing environmental
pollution. Important mechanisms include adsorption, aggregate protection, and
humification. Humification has been linked to increases in heterocyclic N by some
authors, but this is disputed by others who consider that soil N is largely proteinaceous.
The objectives of this study were to (I) test the quantitativeness of 15N-cross-polarization
magic-angle spinning nuclear magnetic resonance (CPMAS-NMR), (2) identify organic
N functional groups during clover decomposition, and (3) evaluate mechanisms of N
stabilization. The CPMAS-NMR spectrum of a prepared, complex, soil-organic mixture
was quantitative for both heterocyclic and noncyclic N. However, peak overlap was an
important but not fatal weakness of this technique. A sandy loam soil was incubated for
14 months in the laboratory with lTN-clover (T rifolium pratense L.) with periodic
samples taken for ultrasonic particle-size fractionation followed by 15N-CPMAS-NMR
and a shaken slurry N mineralization test. In the incubation, clover-N was quickly
transferred fi'om macroorganic matter in coarse fractions to clays, and thereafter
underwent slow decomposition. One-third or less of clover-N was mineralizable in

shaken slurry tests of size fractions, and therefore the remainder was stabilized via

adsorptic
clover-N
5°‘o amin
without I
Other stu
question;
The met

oompari:

adsorption or humification, and possibly aggregate protection. The composition of
clover-N per NMR was invariably 90% amide, 5-10% guanidinium N of arginine, and
5% amino. Any humification must have involved incorporation of protein into humus
without change of functional group. Discrepancies in N composition between this and
other studies may be due to the shortness of the incubation in this study or the
questionable assumption in other studies that all proteinaceous soil N is hydrolyzable.
The mechanisms of N stabilization observed here were relevant to the field as judged by

comparison of this study to field studies.

I dedicate this dissertation to Jesus Christ, my Lord and Savior.

I thank Him for life, love, and peace.

iv

1th;
laboratory I
patience. I
encouragen
my thesis

Dr
helped me I
Sharon Ant
learning frc
for me, and
mm per Son
Wroscop

Di Natham

ACKNOWLEDGMENTS

I thank my wife Utami for her love for me, for working late into the night in the
laboratory with me, for helping me with slide presentations and formatting, and for her
patience. Ithank my parents Gene DiCosty and June Robinson for their continual
encouragement and belief in me. I thank Anna for enduring long hours spent working on
my thesis.

Dr. Eldor Paul supported me generously, encouraged me to be a deeper thinker,
helped me to appreciate the breadth of soil science, and always had time for me. Dr.
Sharon Anderson allowed me to take the risk of this research project, thus facilitating my
learning from both success and failure. She supported me generously, always took time
for me, and helped me to think more deeply. Dr. Boyd Ellis encouraged and helped me
both personally and professionally. Dr. David Weliky taught me to be an NMR
spectroscopist, allowing the use of his equipment and working with me late into the night.
Dr. Nathaniel Ostrom shared his laboratory and helpful advice with me.

Kermit Johnson troubleshot almost every imaginable NMR problem for me, and
Dr. Long Le provided my initial NMR training plus continued support. Dr. Karl Bishop
freely shared helpful ideas for my NMR work. Dr. Dave Harris helped me overcome the
challenges of C02 measurement, elemental analysis, and mass spectrometry. Bryan Pape
and Jason Franti worked hard and skillfully in many of my laboratory experiments. Dr.
Tom Willson, Dr. Sherri Morris, and Ann-Marie Ezanno freely shared ideas, data, and

expertise with me. Mike Rutter and Dr. Oliver Schabenberger gave me crucial help in

the are.

the lab

Thoma
Snyd. J

Whalon

L'nivers
I

through
Compen
T
Smite, 1
Opinions,
those oft?

A{mic-111m:

the area of statistics. Keith and Karyn Sanders gave me important help and friendship in
the laboratory.

Eugene Jones has been a faithful and encouraging friend for many years. En'c
Thomas supported and encouraged me many times. Steve F itton, Basim Aljazi, Scott
Stryd, Jack Brown, John VanderSloot, Evan Vanderwey, Dr. and Mrs. Mark and Sherre
Whalon, and Dr. Bill Metcalfe supported me with encouragement and helpfirl advice.
Thanks to the seemingly countless fiiends who prayed for and encouraged me at
University Reformed Church.

I am grateful for the financial support of the US Department of Agriculture
through a Water Sciences Fellowship and a grant within the National Research Initiative
Competitive Grants Program.

This material is based upon work supported by the Cooperative State Research
Service, US. Department of Agriculture, under Agreement No. 94-37 107-03 87. Any
opinions, findings, conclusions, or recommendations expressed in this publication are
those of the author and do not necessarily reflect the view of the US. Department of

Agriculture.

vi

LIST OF
LlST 0F

LVTROD

SLATE RI

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................... x
LIST OF FIGURES .................................................................................................. xii
INTRODUCTION ...................................................................................................... 1
Control of N Mineralization in Soils ......................................................... 1

Carbon Mineralization ................................................................... 1

Physical Location of Organic N Within

the Soil Matrix .............................................................................. 2
Chemical Structure of Organic N ................................................... 6

N Dynamics in Cover Crop Systems ........................................................ 12
Objectives ............................................................................................... 13
MATERIALS AND METHODS ............................................................................... 14
Soil Incubation Experiment .................................................................... 14
Overview ...................................................................................... 14

Soil and Plant Material .................................................................. 14
Incubation Conditions .................................................................. 18

In Situ N Mineralization ................................................................ 18

In Situ C Mineralization ............................................................... 19
Ultrasonic Fractionation of Soil ................................................... 21

N Mineralizability in Whole Soil and

Particle-Size Fractions ................................................................ 22
C, N, and 15N Analyses ............................................................... 27
Statistical Treatment of Outliers ................................................... 28

vii

REST

DISC I

CONC}

NMR Spectroscopy .............................................................................. 29

Correction for Differential Relaxation Effects ........................................ 32
RESULTS .................................................................................................................. 35
Overall N and C Mineralization ............................................................. 35
N and C Mineralization/Immobilization in Whole
Soils and Particle-Size Fractions ........................................................... 41
Recovery of Mass in Particle-Size Fractions ................................ 41
Dynamics of Clover-N, Native-N, and C ..................................... 42
Mineralization of N in Whole Soils and Particle-
Size Fractions in Aerobic Shaken Slurries ............................................. 55
Test of Quantitative NMR .................................................................... 69
NMR Spectroscopy of Incubated Soils ................................................. 73
Relationship between N Mineralization and Organic N Functional
Groups During the 14-Month Incubation .............................................. 89
DISCUSSION ............................................................................................................ 92
Dynamics of Clover-N in Particle-Size Fractions:
Agronomic Implications ........................................................................ 92
Mechanisms of Short- to Medium-Term Stabilization of
Soil Organic N ...................................................................................... 97
Chemical Structure of Soil Organic N as
Determined by NMR: A Critical Analysis .............................................. 99
CONCLUSIONS ...................................................................................................... 104
REFERENCES .......................................................................................................... 106
Appendix 1. Isotopic Tracer Equations ...................................................................... 115
Appendix 2. Modeling of Mneralization-Immobilization
of Clover-Derived N in Clay Fractions ................................................... 117
Appendix 3. Brief Overview of NMR Concepts ........................................................ l 18

viii

Appendi

 

Appendi

Appendi

Appendix

Appendix

Appendix

Appendix

Appendix 4. Detailed Table of 1’N Chemical Shifts ................................................... 120
Appendix 5. Adjustment of N and C Amounts for

Incomplete Dispersion on Day ll and for Variable

clay Recovery ........................................................................................ 124

Appendix 6. Details of analysis of variance of
shaken slurry N mineralization .............................................................. 126

Appendix 7. Computation of 30-Day Mineralization in the Undisturbed Soil (For
Comparison With Shaken Slurry Mineralization Data) .......................... 127

Appendix 8. Calculations of Aggregate Protection Factors and
Aggregate-Protected Pool Sizes ............................................................ 130

Appendix 9. Correction for Error Due to Peak Overlap
in lsN-NMR Spectrum of Clover-Uracil-Soil ........................................ 131

Appendix 10. Estimation of Uncertainty in NMR Spectra
Due to Noise and Subjectivity in Phasing/Baseline Correction ................... 134

ix

Table I.
Table 2

Table 3
ka4

Ines
Table 6
Table 7.

Table 8
Table 9.
Table II
Table I]

Table I;

Table 13
Table 14
Table 15

pi

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

Table 12.

Table 13.

Table 14.

LIST OF TABLES

Fractions of soil organic N identified by acid hydrolysis .......................... 7
Oxidative model of plant decomposition in mineral soils ........................... 11

Particle-size distribution as determined by conventional pipet
or ultrasound methods ............................................................................ 17

Description of N—free nutrient medium used in 30-day shaken-slurry N

mineralization tests ................................................................................ 26
Summary of 15N chemical shifts ............................................................... 30
Summary of kinetic modeling data ............................................................ 40
Recovery of mass in ultrasonic particle-size fractionations ...................... 43

Clover-derived N in whole soils and size fractions
(as % of clover-N initially present in whole soil) ...................................... 44

Native N in whole soils and size fractions (as % of native N

initially present in whole soil) .................................................................. 51
C (native- plus clover-derived) in whole soils and size fractions

(% of C initially present in whole soil) ..................................................... 56
Thirty-day N mineralization in shaken Slurries (as percent of

organic N in fraction) ............................................................................. 60
Soil C/N ratio (molar basis) as a function of incubation time .................... 63

Analysis of variance of 30-day N mineralization in aerobic
shaken slurries ........................................................................................ 64

Significance levels for pre-planned comparisons of
30-day N mineralization values in shaken Slurries .................................... 64

Table 15. Aggregate-protection factors for whole soil and

particle-size fractions ............................................................................... 66

Table 16. Aggregate-protected pool sizes for whole soil and
particle-size fractions (% of N in fraction) .................................................... 68

Table 17.

Table 18.

Table 19.
Table 20 '
Table 2l. '
Table 22. '.

Table 23,

Table 24
Table 25.
Table 26
Table A-l.
Table A—2

Table A-3,

Table 17. Test of quantitative 15N-NMR in prepared mixture of

(”N-uracil + lsN-plant material + unlabeled soil) .................................... 71
Table 18. Relaxation data for prepared mixture of

lsN-clover + 1SN-uracil + unlabeled soil ................................................... 72
Table 19. NMR peak area percentages for clover and incubated soil ....................... 74
Table 20. Tm values (s) for clover and incubated soils ........................................... 79
Table 21. Tun values (ms) as a function of material and firnctional group ................ 80
Table 22. T1,," values (ms) as a function of material and firnctional group ................ 81
Table 23. Correction factors for NMR peak areas due to difi‘erential

relaxation during the contact time .......................................................... 87
Table 24. 15N retention in present study compared to field studies .......................... 96
Table 25. Comparison of soil C with soil protective capacity .................................. 97
Table 26. Maximum possible mineralization of clover-N in shaken slurry ............... 100
Table A-1. lsN Chemical shifts of biologically important molecules ......................... 121
Table A-2. Chemical shifts conversion ............................................................................ 120
Table A-3. Aggregate Protection Factors And Pool Sizes ........................................ 130

F i gur
F iguri
figure

Figure

Figure
Figure
Figure I
Figure 9

Figure l
Figure l;
Flsure I:
Fliture 13
Flglue I4
figure 15.
new“ .

FIEUIe ]7_ x

LIST OF FIGURES

Figure 1. Overview of soil incubation experiment ............................................... 15.
Figure 2. Chemical structure of uracil .................................................................. 31
Figure 3. N remaining after mineralization during soil incubation ................................ 36
Figure 4. Respiration rate as a filnction of soil incubation time .................................... 38

Figure 5. C remaining in whole soil as determined by dry combustion of

whole soil and by respiration ....................................................................... 39
Figure 6. Clover-N amounts in incubated soil ..................................................... 47
Figure 7. Native N amounts in incubated soil ..................................................... 53
Figure 8. C amounts in incubated soil ................................................................. 57
Figure 9. NMIN and ln(NMrN) versus molar C:N ratio ........................................... 62

Figure 10. 15N-NMR spectra of prepared sample of
15N-uracil + 15N-clover + unlabeled soil .............................................. 70

Figure 11. NMR spectra of clover, whole soil, fine clay, coarse clay,
and fine silt .......................................................................................... 75

Figure 12. Measured and modeled Signal intensities for clover as
a function of contact time ................................................................... 82

Figure 13. Measured and modeled signal intensities for
whole soil as a function of contact time ............................................. 84

Figure 14. Measured and modeled signal intensities for fine clay
as a firnction of contact time ............................................................. 85

Figure 15. Linear regressions of 15N amounts versus NMR signal
intensity for soil samples .................................................................. 88

Figure 16. NMR detectability of fine-clay associated lsN after various
incubation times ................................................................................ 89

Figure 17. NMR spectra of ”N-labeled clover, with and without soil. ............... 91

xii

 

Figure 18.

Figure 18. Clover-derived N in fractions as a function of time
(% of remaining) ............................................................................. 94

xiii

An I.‘
(N) in soils ?
enlironmen
understandi:
independent
within the s
(e g tempel

andare dest

Control of

Cwbon Ari”

C art
agfiufltura]
microbial u;

residues wit

INTRODUCTION

An understanding of the mineralization (and immobilization) of organic nitrogen
(N) in soils is necessary to maximize agricultural production while minimizing
environmental degradation. Therefore, soil scientists have devoted considerable effort to
understanding the controls on N mineralization. These controls include four (non-
independent) categories: (1) carbon mineralization, (2) physical location of organic N
within the soil matrix, (3) chemical structure of organic N, and (4) environmental factors
(e.g. temperature, moisture, pH). The first three controls are relevant to the present study

and are described below.

Control of N Mineralization in Soils
Carbon Mineralization

Carbon mineralization is closely linked to N mineralization. Decomposition of
agricultural crop residues with C:N ratios greater than 25 is typically accompanied by net
microbial uptake of inorganic N (Paul and Clark, 1996), whereas decomposition of
residues with C:N < 25 results in net microbial production of inorganic N. Under aerobic
conditions, net C mineralization is expected to exceed net N mineralization; mineralized C
escapes as C02 but mineralized N can again be immobilized. For example, McGill (1971,
p. 77) incubated a soil with “C-aeetate + "hr—ammonium sulfate and found that the half-
life ofamino acid N was 2700 (1, whereas Sorenson and Paul (1971) found that the half-

life ofamino acid C in the same soil incubation was 1600 d.

 

Physical bx
The
important it
fractionatic
Catroux ar
mmmm
soluble, (2
pools of S
Tl
SUPPOTT n
MOM fra
decompo
On a siev.

1992). c

Physical Location of Organic N within the Soil Matrix

The physical location of organic N within the soil matrix is undoubtedly an
important factor in N mineralization. Soil scientists have established that the physical
fractionation of soil yields biologically and chemically distinct pools of organic matter (e.g.
Catroux and Schnitzer, 198 7). Although many fi'actionation schemes have been published,
the four most important pools of soil organic matter (SOM or OM) appear to be (1)
soluble, (2) macroorganic matter (MOM), (3) microbial biomass, and (4) sorbed. (These
pools of SOM have some overlap with one another.) See Christensen (1996) for details.

The soluble fraction likely represents free or weakly bound compounds that readily
support microbial growth, such as sugars or low molecular weight polypeptides. The
MOM fraction probably consists mainly of large plant fragments in the early stages of
decomposition. As defined here, this material is isolated by flotation in water or recovered
on a sieve with other sand-sized material during particle-size separation (Christensen,
1992). Generally, the MOM fraction is short-lived but it may contain resistant
components (Christensen, 1992) such as lignin or charcoal.

The third pool of SOM, microbial biomass, accounts for only 1-3 % of total soil C
and at most 5 % of total soil N (Smith and Paul, 1990). Nonetheless, the microbial
biomass is important as a transformer and short-term reservoir of soil nutrients. Direct
microscopy and chloroform firmigation are two important methods for estimating the
microbial biomass (see Paul and Clark, 1996). In the chloroform method, living microbial
cells are lysed by chloroform fumigation; the biomass of the lysed cells is then estimated
by observing the C02 or NH4 released upon re-inoculation of the firmigated soil or by

immediately extracting the fumigated soil and analyzing for elements released from the

hsed cells
The
organic nla'
biomass fra
for the rela'
the fine silt
clay ln reg
determined
Northern G
consistenrlt
Eli and com
cOrh'ersion
Several deca
C and N b0,
Rela
soil Organic
CODIUnctl'on
m“ cOnside
dill. and the
(1977)) in as:
Shaw) and f0
found in The 2

i002

lysed cells. Appropriate conversion factors are required.

The sorbed fraction is important in the medium- to long-term stabilization of soil
organic matter, in contrast to the relatively short-lived soluble, MOM, and microbial
biomass fiactions. This fraction is primarily associated with silt and clay and is responsible
for the relatively high levels of SOM found in fine-textured soils. The OM associated with
the fine silt and coarse clay appears to be more resistant than that associated with fine
clay. In regard to the long-term stability of these fi'actions, Anderson and Paul (1984)
determined the radiocarbon age of C in particle-size fi'actions from three soils in the
Northern Great Plains (North America); the 1‘C ages of fine silt and coarse clay were
consistently greater than those of fine clay. In two of the three soils, the 1‘C ages of fine
silt and coarse clay were 200 yr or greater. Tiessen and Stewart (1983) found that
conversion of grassland to agricultural land led to a net loss of soil organic C and N after
several decades; the conversion was also associated with an increase in the proportion of
C and N bound to fine silt and coarse clay.

Relatively few studies have focused on the short- or medium-term stabilization of
soil organic N in clay and silt fractions. Ladd et al. (l977a,b) amended soils with l’Nos in
conjunction with either wheat straw or glucose; alter 160 d of incubation in the laboratory
they considered that organic 15N in the fine clay was being transferred to silt and coarse
clay, and that silt was important in the long-term stabilization of OM. Paul and McGill
(1977), in association with Myers, amended field soils with “C"N-straw or ("N03 plus
straw) and followed the labeled material for 2 - 4 yr. Generally, the proportion of 15N
found in the > 0.2 pm fraction increased, whereas the proportions of 15N found in the 0.04

to 0.2 pm and < 0.04 pm fractions decreased. In a field study of the decomposition of

 

”CBS-
(and 13‘
fractior
ls‘m‘ls
of 1TN 3
analysis
was con
1 8 yr".
modeled
P
fractions
In pores l:
are dichS
adsorptior

cOnslderak

Dior

4.09

l
. Aggregale‘

In thl:

”ClsN-wheat straw during 574 d, Aita et al. (1997) found that substantial amounts of "N
(and 13C) were retained in the fi'action < 50 pm, with negligible net decomposition in said
fi'action. Balabane and Balesdent (1995) supplied 15N to maize fields in the form of either
ISNT‘IAISNO‘J or maize residues in a four-year study, and found limited net decomposition
of 15N associated with particles < 50 um. With 15N supplied in inorganic form, kinetic
analysis revealed that there were two pools of 15N within the < 50 um fraction; one pool
was considered to undergo zero decomposition and the other had a decay rate constant of
1.8 yr". Where 15N was supplied as maize residue, the 1’N in the < 50 um fraction was
modeled as a (one-pool) accumulation function without decay.

Possible mechanisms for the stabilization of soil organic N in the clay and silt
fractions include (1) adsorption per se, (2) aggregate protection’ (i.e. entrapment of OM
in pores inaccessible to microbes), and (3) chemical structure. The first two mechanisms
are discussed below, and the third mechanism is the topic of the next section. In regard to
adsorption, Hassink (1996,1997) and Hassink and Whitmore (1997) have compiled
considerable evidence that clay and silt behave as saturable adsorptive surfaces with
respect to organic matter. In a study of soils fiom five continents, Hassink (1997)

demonstrated that the sorption capacity, or protective capacity, for soil C was as follows:

protective capacity (g C kg") =

4.09 + 0.37 x (% particles < 20 um) (I)

 

' Aggregate protection has also been referred to as physical protection. The former term
is used in this study to avoid ambiguity.

4

Etidemly. till
applicable to

amilal equal

prot

le

These auth
desorption
adsorbed(
slow. The
OVetsimph

Th
explained
(1967), B
Weakly b0
ml), the I;
microaggr
Mullen:
Subscripts
ofc-p‘0;\
aggregate,

30 °C) Ofp

Evidently, the sorption of organic C by silt and clay is non-specific, as Equation 1 was
applicable to soils with difl‘erent mineralogy. Hassink and Whitrnore (1997) developed a

similar equation based on the proportion of particles < 2 am:

protective capacity (g C kg") =

21.1 + 0.375 x (% particles < 2 pm)

These authors integrated the protective capacity into a simple mechanistic adsorption-
desorption model. The model assumed that desorption must precede decomposition of
adsorbed OM, and indicated that adsorption was relatively quick whereas desorption was
slow. The assumption that desorption must precede decomposition is probably an
oversimplification, but reflects the fact that clays inhibit decomposition (Stevenson, 1994).
The aggregate protection of soil organic N in pores inaccessible to microbes is
explained by the macroaggregate—microaggregate concept of Edwards and Bremner
(1967). Based on dispersion studies, these researchers proposed that soil is composed of
weakly bound macroaggregates (> 250 um) and tightly bound microaggregates (< 250
pm), the latter being most important in aggregate protection. The structure of the
microaggregates is represented by [(C-P-OM)x]y, where C = clay mineral particle, P =
polyvalent metal such as Ca, Al, or Fe, and OM = humified organo-metallic complex. The
subscripts x and y are whole numbers, and reflect that microaggregates contain a hierarchy
of C-P-OM units. Experimentally, Edwards and Bremner (1967) investigated the
aggregate protection of N by means of an anaerobic (waterlogged) incubation (two weeks,

30 °C) of previously air-dried and ground soils. Mineralization in two soils ground to 50

um was found to be about 2.5-fold higher than mineralization in the same soils ground to
2000 um. Ifthe 50 um soil was moistened after grinding, allowed to air dry at 25 °C, and
sieved to 2000 um size, the mineralization enhancement was lost; evidently, the
microaggregates had re-fonned in the wetting-drying process. Craswell and Waring
(1972) extended the results of Edwards and Bremner (1967), and found increases in
aerobic N mineralization with grinding in fine- but not coarse-textured soils.

One might argue that the increases in aerobic N mineralization upon grinding are
the result of enhanced oxygen difi'usion to the interior of microaggregates rather than
entrapment of OM in pores inaccessible to microbes. However, Rovira and Greacen
(195 7) showed that the uptake of oxygen by an undisrupted silt loam soil was the same

under air or oxygen atmospheres.

Chemical Structure of Organic N

The chemical structure of soil organic N is controversial; therefore the relationship
between the mineralization and chemical structure of N has not been adequately
characterized. Broadly speaking, soil organic N has been studied by one of three methods:
(1) chemical extraction, (2) pyrolysis mass spectrometry, and (3) lsN-nuclear magnetic
resonance spectroscopy. One widely used method of chemical extraction is acid
hydrolysis; soils are typically treated with hot 3 to 6 M HCl for 12 to 24 h (Stevenson,
1994). The fiactions identified are summarized in Table l. The nature of the N in the
Mk, hydrolyzable unknown, and acid-insoluble fractions is not well known; Stevenson
(1994) and Schnitzer (1985) consider that about one-half of soil N remains unidentifiable

by soil hydrolysis techniques. Despite these problems, the hydrolysis fractions have been

 

obsen'ei
general!)

therefore

Table l
{.—

l .
fractlon

observed to exhibit some differences in mineralizability. For example, amino acid-N
generally decreases (as a proportion of total N) upon cultivation (Stevenson, 1994) and is

therefore a labile N form.

Table l. Fractions of soil organic N identified by acid hydrolysisf

 

 

 

 

 

 

% of
fraction explanation soil N
. 0 acid N Determlned m hydrolysate by mnhydnn 3 045
mm" methods.
Ammonia recovered upon steam distillation
“N}h”-N of Mthreated hydrolysate, sources include 20-35
amino acids destroyed by hydrolysis and
clay-fixed NHn.
Determined by either steam distillation of
hydrolysate at pH 11.2 in the presence of
amino sugar N phosphate-borate buffer (with correction for 5-10
NHz-N) or direct colorimetric analysis of
hydrolysate.
hydrol 1 le I own Hydrolyzable N minus (ammo acrd N +
N (HUN) ammo sugar N + 10-20
NHs-N).
acid-insoluble N - 20—3 5

 

 

 

 

 

T Adapted fi'om Stevenson (1994).

The chemical fi'actionation of SOM into humin, humic acid, and fillViC acid, is a
milder fractionation than acid hydrolysis and has been useful in studies of proteinaceous
soil N. A typical separation might involve treatment of soil with 0.1 to 0.5 M NaOH,

followed by separation of liquid and solid phases by centrifugation, and adjustment of the

7

 

liquid pH

the NaOH
acid remai
electropho
considered
summarize
to an arom
(1973) ext:
Polm'nylpl

materials.

liquid pH to 2 with 2 M HCl (Schnitzer, 1982). The humin is contained in the residue of
the NaOH extraction, the hunric acid is precipitate of the acidified liquid, and the fulvic
acid remains in solution. Simonart et al. (1967) applied phenol extraction and
electrophoresis to humic acid to isolate a “humoprotein” complex and protein. They
considered that the protein was hydrogen-bonded to the humic acid. Haworth (1971)
summarized numerous studies on humic acid and hypothesized that protein was attached
to an aromatic humic “core” via both covalent and hydrogen bonds. Biederbeck and Paul
(1973) extracted humic acid with phenol, and purified the phenolic extract with
polyvinylpyrrolidone (PVP, a phenol-complexing agent) to isolate relatively pure peptidic
materials. The phenol and PVP were considered to have disrupted hydrogen bonds
between peptides and an aromatic humic acid traction. These studies demonstrated that
protein is an important N-containing constituent of humic acid.

Soil scientists have applied cross-polarization magic-angle spinning 15N-nuclear
magnetic resonance spectroscopy in the solid state (CPMAS-NMR) and pyrolysis-mass
spectrometry (PyMS) to gain further information about the chemical structure of soil
organic N. The proponents of each technique claim to be able to assign all but a small
fraction of soil N to specific functional groups, but the conclusions of different
investigators are often contradictory. Based on PyMS studies of whole soils and
hydrolysis fractions, and on detailed analysis of hydrolysis fi'actions (e.g. gel
chromatography, gas chromatography, mass spectrometry), Schulten and Schnitzer (1998)
state that soil N is distributed as follows: proteins + peptides + amino acids, 40 °/o; amino
sugars, 5 - 6 %; heterocyclic N, 35 °/o; and Nth, 19 %. (NIB is an operationally defined

pool; see Table 1.) In a PyMS study of agricultural soils, Schulten and Hempfling (1992)

found a neg
authors. the
managemer
N.
In c
“/5 or more
form of ant
al, 1993, l'
in fulxr'c aci
Given that E
agricultural
Concerning 1
Botl
Organic N l
r‘E‘Imaflgeme
WantitafiOn

um “Widen

found a negative correlation between heterocyclic N and total soil N. According to these
authors, the low-N soils had a decreased capacity to immobilize N (as a result of
management practices) and thereby became enriched with relatively resistant heterocyclic
N.

In contrast to the aforementioned PyMS studies, NMR researchers have found 80
% or more of soil N in amide form (peptide bonds in proteins), with the remainder in the
form of amino acids, amino sugars, or the amino groups of nucleic acid bases (Knicker et
al., 1993, 1997; Clinton et al., 1995; Hopkins et al., 1997). Heterocych N was identified
in firlvic acid by Zhou and Wen (1992), but comprised only 9% of the spectral intensity.
Given that Schulten and Hempfling (1992) proposed that heterocych N is an indicator of
agricultural management practices (see above), it is important to reconcile the controversy
concerning the chemical nature of soil organic N.

Both PyMS and CPMAS-NMR have potential problems in quantitation of soil
organic N. In the case of PyMS, incomplete sample volatilization or thermal
rearrangement may confound the analysis (Schulten et al., 1995). In CPMAS-NMR,
quantitation may fail if differential relaxation effects during the contact and delay times are
not considered. These efl‘ects have been assessed in plant material, compost, and soluble
extracts (Knicker and Ludemann, 1995; Knicker et al., 1997), but lmve not been
thoroughly investigated in soil. (The nature of these effects is described in the Material
and Methods). Another potential problem of CPMAS-NMR is that rigid aromatic
structures may have excessively long Tm values; the delay times required to detect the N
in such structures may be too lengthy for a practical experiment. In a CPMAS-NMR

study of 13C, Wilson et al. (1984) showed that Tm values exceeded 50 s for rigid, planar,

 

aromatic m
containing
PyMS (Scl
Adi
MIR studi
insufficient
may be cur
present in 5
high freque
in soil and 1
small signal
Des
for studies t
fracrions, W
Each fraCtic
resolved N)
”Seen that
authors cultt

We: is

aromatic molecules such as naphthalene, anthracene, and phenanthrene! Similar N-
containing structures, such as those proposed by Flaig et al. (1975) or those identified by
PyMS (Schulten and Schnitzer, 1998), might also exhibit long Tm values.

Additional problems may confound quantitation of soil N by CPMAS NMR Most
NMR studies of soil N require addition of l 5N; incubation time of the 15N with soil may be
insuficient for formation of heterocyclic N (Schulten and Schnitzer, 1998). Also, NMR
may be currently unable to detect the numerous heterocyclic N structures thought to be
present in soil (Schulten and Schnitzer, 1998). The structures may be obscured by the
high frequency shoulder of the amide—N peak. If numerous heterocyclic N structures exist
in soil and have a wide range of chemical shifis, the structures may give rise to numerous
small signals that are undetectable in a typical CPMAS-NMR experiment.

Despite the limitations of CPMAS-NMR, its noninvasiveness makes it attractive
for studies of soil organic N. The technique has not been applied to soil particle-size
fi'actions, which are known to contain biologically and chemically distinct pools of OM.
Each fi'action may contain fewer firnctional groups than whole soil, and thus better
resolved NMR spectra seem possible. Further, the results of Bedrock et al. (1998)
suggest that lsN—CPMAS-NMR can serve to identify soil microbial populations. These
authors cultured soil bacteria and fungi axenically and found that the organisms had
distinct ”N-CPMAS-NMR signatures. The bacterial cells exhibited a small peak at 115-
140 ppm (referenced to NH: = 0 ppm), which was attributed to aromatic N of purine
bases. The fungal cells had a smaller peak in said region, but showed a unique peak at 10
- 25 ppm, which was attributed to N—acetyl-glucosamine. Bedrock et al. (1998) also

incubated unlabeled straw in mesh bags in a field soil fertilized with "mirror. After 12

10

 

months of c

a] (1998) s

bacteria cor

In c

organic C a

discussion '
of ”cm
relates C st
from carbo
Preservatio
cOflcentratt

resl/‘GCIlvelj

Table 2~ 0
F

\
major pee
NMR Spe
'\

Probable 1
molch 165

 

l

months of decomposition, the degraded straw had a fungal NMR signature. Bedrock et
al. ( 1998) supported their finding with other microbiological data, which indicated that
bacteria comprised less than 5% of the microbial biomass.

In contrast to nitrogen, the relationship between the chemical structure of soil
organic C and C mineralization is reasonably well characterized. This topic merits
discussion because C and N mineralization are related (see above). Based upon the results
of l3C-NMR studies, Baldock et al. (1992) proposed a simple three-stage model, which
relates C structure to mineralization (Table 2). Overall, the model describes a progression
fiom carbohydrate to aliphatic C, with an intermediate stage marked by the selective
preservation of aromatic C of lignin. The O-alkyl, aromatic, and alkyl C tend to be
concentrated in the sand (> 53 um), silt (2-53 um), and clay (< 2 pm) fiactions,

respectively.

Table 2. Oxidative model of plant decomposition in mineral soils?

 

 

 

 

stage of decomposition
early intermediate late
major peaks in 13c- O-alkyl O-alkyl alkyl
NMR spectra alkyl
aromatic
probable types of plant carbohydrate microbial peptidoglycans of
molecules carbohydrate bacterial cell walls?
selectively bacterial and plant
preserved plant lipids?
lipids?
humic polymers?
selectively
preserved plant
lignin

 

 

 

 

 

1' Adapted from Baldock et al. (1992).
ll

N Dynamics in Cover Crop Systems

Knowledge of N dynamics within cover crop systems is of fundamental importance
with the increasing emphasis on low-input sustainable agriculture. Cover crops are
beneficial because they can reduce erosion, conserve water (if kills are timed
appropriately), reduce weed biomass, and decrease fertilizer costs (Bowman et al., 1998;
Fisk, 1997). Top cover crops for the Great Lakes region of the USA include red clover,
hairy vetch, annual ryegrass, and rye (Bowman et al., 1998). Sarrantonio (1998) suggests
that from 25-50% of the N from cover crops is available to a subsequent crop, but this
estimate may be somewhat high per the 15N studies discussed below.

Harris and Hesterman (1990) evaluated the uptake of alfalfa-”N by first-year corn
and second-year barley at two sites in Michigan. After one growing season, 46% of the
15N remained in organic forms in the soil and 17 or 25% of the 13N was found in the plant.
The higher plant uptake occurred in a relatively coarse-textured soil and may have been
related to a lack of clay protection. The alfalfa-derived 15N was much less available during
the second growing season, as the barley crop recovered only 1% of the initial lsN.

Harris et al. (1994) extended the findings of Harris and Hesterman (1990) by
studying the fate of fertilizer- versus legume-”N (red clover) in conventional and low-
input farming systems in Pennsylvania. Losses of 15‘N from the two plant-soil systems
were similar after two growing seasons (39% of input). Plant uptake of 1N was higher in
the fertilized system than in the legume system (40 versus 17% of input), with nearly all
uptake in both systems occurring in the first growing season. The legume system retained
more 15\N in the soil than the fertilizer system (47 versus 17% of input), and contained a

larger microbial biomass. The high microbial biomass and soil N retention in the legume

12

 

system sug;

Objectives

The
for both he
functional

the short- 1

system suggest enhanced N-supplying capacity in long-term cover-cropped soils.

Objectives

The objectives of this study were to (1) test whether CPMAS-NMR is quantitative
for both heterocyclic and noncyclic soil organic N, (2) identify the soil organic N
functional groups derived from decomposing clover, and (3) evaluate the mechanisms for

the short- to medium-term stabilization of soil organic N.

13

 

Soil lncul
Ourmerr'
0t
esperimen
agricultura
incubated i
mineralizar
ultrasonica
were deterr

fracrions an

Shiny N mi.

Soil and pk
The
series) from
10““ de13th.
mixed by pas
The F
method The

Allardee (192

MATERIALS AND METHODS

Soil Incubation Experiment
Overview

Objectives 2 and 3 (see Introduction) were addressed via a soil incubation
experiment, as depicted in Figure 1 and summarized below. The Ap horizon of an
agricultural soil was amended with lsN-labeled red clover (T nfoliurn pratense L.) and
incubated in laboratory-scale nricrolysirneters for 14 months. In situ C and N
mineralization were measured several times during the incubation. On five dates, soil was
ultrasonically fractionated into five size classes, and the concentrations of N, 15N, and C
were determined for each size fraction and whole soil. On three of the five dates, the
fractions and whole soil were characterized by solid-state 15‘N-NMR and by a shaken-

slurry N mineralization test.

Soil and Plant Material

The soil used was a fine-loamy, mixed mesic Glossoboric Hapludalf (Marlette
series) from East Lansing, Michigan, USA. The Ap layer of soil was excavated from 0 to
10 cm depth, air-dried, sieved to remove fiagments > 2 mm in diameter, and thoroughly
mixed by passing several times through a soil splitter.

The particle-size distribution of the soil was determined by the conventional pipet
method. The method was based on reports by Gee and Bauder ( 1986), Whittig and

Allardice (1986), and Dr. Sharon Anderson (pers. comm.) Carbonates were removed and

14

nI-Situ

I Respiration C 02
+ N03 + NH 4

 

SOIL + 15N-CLOVER

 

 

 

 

 

 

 

In-Situ
N Mineralization
Ultrasound

))))> Dflifififlf; (15 WK; 1, 3, 6, 14 MO)
{nascent-"j »—-> o
1 Fine Clay 5 __,N
l Coarse Clay 5 1
E Fine Silt i ’ 5N
i Coarse Silt E ‘—"" 30-Day N Mineralization
E Sand :
““““““““““““““ —-—> lsN-NMR

 

 

 

Figure 1. Overview of soil incubation experiment. * Whole soil was not sonicated.

15

the soil cation exchange sites were saturated with sodium by treatment with 1 M sodium
acetate (adjusted to pH 5), at 70 - 80 °C. Excess salts were then removed by
centrifugation with water, after which organic matter was removed by addition of 30%
hydrogen peroxide at 70 °C. Next, appropriate dispersant (sodium hexametaphosphate)
was added followed by either 24 h or more of soaking or about 16 h of shaking; soaking
and shaking gave the same results. Sand (> 53 pm diameter) was determined by sieving,
and fine silt (2 - 10 W“) and coarse clay (0.2 — 2 urn) were determined by shaking the
remaining suspension and taking pipet aliquots at the appropriate settling times. Finally,
fine clay (< 0.2 m) was determined in a tube centrifirge based on an integrated version of
Stokes’ law. Throughout the procedure, the presence of residual salts was taken into
account by appropriate measurements and corrections. The results are shown in the
leftmost data column of Table 3. The soil was identified as a sandy loam per the USDA

textural classification system.

16

Table 3. Particle-size distribution as determined by conventional pipet or ultrasound

 

 

 

 

 

 

 

methods.
method _________________________
ultrasound
conventional pipet (38.3 k] / 28 g soil) ultrasonic yield /
fraction cm 1' 1%)1’ conventional yield
fine clay
(< 02 um) 6-50 (0.28) 3.81 (0.56) 0.59
coarse clay
02-2”“) 9-05 (0-37) 12.4 (0.7) 137
fine silt
(2-10 um) 9.94 (0.90) 12.0 (0.4) 121
coarse silt
(10-53 11m) 19-7 (0.2) 15.4 (2.9) ()73
sand
(53-2000 um) ....... 313---}??? _____________ _5,6_3_m_f3_}_) ..... 1‘03
SUM 100 100

 

 

 

 

 

T Sample standard deviations in parentheses.

The chemical characteristics of the soil prior to amendment with clover were as

follows (means :1: sample standard deviations): pH in 1:2 soil HzO = 7.20 :l: 0.03, Ctotal (g

kg") = 13.6 e 1.6, Na... (g kg") = 1.12 t 0.01, and molar C:N ratio = 14.1 e 1.8.

The clover ( T nfoliunl pratense L.) used as a soil amendment had been grown

hydroponically using 15N03 as the sole N source. The clover was freeze-dried and ball-

milled to a powdery texture prior to mixing it with soil. The characteristics of the clover

were as follows (means i sample standard deviations): Cronn (g kg") = 403 i 4, Ntotal

(g kg") = 38.9 :1: 1.0, molar C:N ratio = 12.9 r: 0.4, atom % "N =-- 91.3 i 2.0.

17

 

Incubationt
Soils
laboratory-5c
soil was ame
basis). The
adding the a
weight. All
Clark (199C
independen
rosin. T11
be obtainec
lntt
approxjma

pOteHIlaL \

Incubation Conditions

Soils were incubated at room temperature (about 23 °C) in the dark in the
laboratory-scale microlysimeters described by Nadelhofi‘er (1990). Before incubation, the
soil was amended with 34.3 g of the ”N-labeled red clover per kg amended soil (dry
basis). The total C and N contents of the soil were approximately doubled as a result of
adding the amendment, and the clover-C corresponded to about 1.4 % of the soil dry
weight. Although agricultural soils typically receive much smaller organic inputs, Paul and
Clark (1996) state that the decomposition rate of organic materials added to soil is
independent of the quantity added if the C addition does not exceed 1.5 % of the soil dry
weight. The high concentration of clover in the soil ensured that "N-NMR spectra could
be obtained in a reasonable amount of time.

Into each lysimeter, 62.1 g of amended soil (dry basis) were packed at an
approximate bulk density of 1.2 Mg m'3. Then, water was added to -33 kPa soil water
potential, with correction for moisture retained in the glass filter and glass wool
components of the lysimeters. The soil water potential was maintained by adding water

about twice per week; less than ten percent of the water evaporated between additions.

In Situ N Mineralization

In situ N mineralization was defined as the cumulative, soluble, inorganic N
leached from the lysimeters. The method was modified from Nadelhofi‘er (1990). For

leaching, soils were allowed to equilibrate with an N-free nutrient solution for 30 min, and

then vacuum-extracted at 45 - 60 kPa. Potassium chloride was added to the leachates as a

18

 

preservative

 

nutrient solut:
MgSOt. 8 33
Ml NazthC
in each lysim
the lysimeter

about ll p01

preservativez, after which the leachates were stored at -20 °C until analysis. The N-free
nutrient solution contained 1.3 mM CaClz, 0.67 mM KHzPOe, 0.33 mM K2S04, 0.33 mM
MgSOs, 8.33 "M HeBOs, 0.67 “M MnClz, 0.67 “M 211804, 0.17 ”M CuSOs, and 0.17
“M Na2M004, with final pH = 5.1. About 5 to 8 mL of excess nutrient solution remained
in each lysirneter afier leaching; the excess liquid was allowed to evaporate by uncapping
the lysimeters until they again reached -33 kPa water potential. At the end of the study,

about 11 pore volumes of leachate had been collected from each lysirneter.

In Situ C Mineralization

In situ C mineralization was measured by temporarily making the lysirneters
gastight, during which time respired C02 accumulated in the headspace (N adelhofi‘er,
1990). For a single measurement of respiration in an individual lysimeter, three to four
headspace gas samples were taken at specific times and immediately injected into an
infi'ared gas analyzer for determination of C02 concentration. Respiration in each
lysimeter was defined as the slope of a linear regression of C02 concentration versus time.
The squared correlation coefficient of the regression (Rz) exceeded 0.95 for most
measurements; measurements with R2 < 0.95 were omitted from the final data.

For the respiration measurement at incubation time = 1.3 d, the first-measured
lysimeters respired more than the last-measured lysimeters. This observation is consistent

with general principles of the decomposition of plant material in soil (see Paul and Clark,

 

2 According to Keeney and Nelson (1982), filtered 2 M KCl extracts of soil are stable
under refiigeration for several months. In the present study, KC] was added to l or 2 M
final concentration.

19

 

1996), thl
decompot
scatter die
fitting a re

for time =

1996); the rate of C loss is expected to decrease rapidly with time in the initial stages of
decomposition. Therefore, the respiration at time = 1.3 d was estimated by constructing a
scatter diagram of the natural logarithm of respiration versus exact time of measurement,
fitting a regression line through the points, and calculating a predicted value of respiration
for time = 1.3 d.

After the l4~month soil incubation, the respiration data were fitted to a two-pool
model (Paul et al., in press):

r = Crkr exp(-krt) + Czkz exp(-k2t) (2)
whae r = respiration (amount ofC per time), Cr and C2 are the amounts ofC in the two
pools, k1 and k2 are rate constants (time'l), and t = time. Respiration was defined to be a
positive number and thus equals —dC/dt. The sum of the parameters C1 and C2 was
constrained to equal initial whole soil C minus resistant C. Resistant C was estimated as
45% of the native (non-clover) C. This estimate is derived from the data of Collins et al.
(in press); these researchers determined that nonhydrolyzable C (after removal of visible
plant residues) in cultivated topsoils of the Northern Prairie and Corn Belt regions of the
USA ranged from 39 to 51% of total C. Paul et al. (1997) classify nonhydrolyzable C as
resistant based on its old radiocarbon age. For cultivated topsoils in Colorado, Nebraska,
and Arizona, Paul et al. (1997) found that nonhydrolyzable C pool had a mean “c age
ranging fiom 900 to 3300 yr whereas the 14C age of the total C in these soils ranged from
400 to 1300 yr.

For simplicity, no correction for microbial biosynthesis was made in the model.
The fitting procedure is explained below in the section N Mineralizability in Whole Soil

mid Pw'ticle-Size Fractions.

20

Ultrasonic Fractionation of Soil

On five dates, the soil from two lysirneters was composited, sonicated, and
separated into five organo-mineral size classes. The size classes were fine clay (0-0.2 pm
diameter), coarse clay (0.2-2 am), fine silt (2-10 tun), coarse silt (10-53 pm), and sand
(53-2000 pm). An ultrasonic energy of 38.3 k] per 28 g soil was used, per the results of a
preliminary study. This energy produced a particle-size distribution similar to the
conventional pipet method (Table 3); the energy was measured as described by Morra et
al. (1991) and North (1976). Sonication was carried out in a 250-mL beaker that was
resting in an ice bath, with a 1:5 soilzwater ratio, 28 g soil (dry basis), 1.27 cm (0.5 in)
ultrasonic probe diameter and 3 cm probe depth. The incubated soils required brief hand-
shaking before sonication; the shaking served to break up large clods of moist soil.

After sonication, the three finest size fractions were physically separated on the
basis of Stokes' law. The fine clay was isolated by centrifugation, the coarse clay by
gravity sedimentation at 4 °C (the low temperature served to slow microbial
transformations), and the fine silt by gravity sedimentation at room temperature. Each
separation consisted of eight centrifugation or sedimentation cycles. During the
separations, the concentration of clay was kept below 10 g/L; Elonen (1971) showed that
clay concentrations above 10 g/L significantly increase the suspension viscosity relative to
pure water and thereby confound the Stokes' law calculations. Afier the three finest size
fi'actions were isolated, the coarse silt and sand were separated by wet sieving.

When the separations were completed, the three fine size fiactions were

concentrated by flocculation with MgClz; the supernatant was removed by siphoning, with

21

centrifiigation as necessary. The two coarse size fi'actions were either flocculated as
above or allowed to settle under gravity, and were concentrated by siphoning the
supernatant. Salts were removed by dialysis at 4 °C (cellulose membrane with molecular
weight cutofi‘ 12,000 to 14,000, Spectra/For, Spectrum Medical Industries), such that the
amount of Mg remaining in the soil was small compared to Mg supplied by the liquid
medium in subsequent shaken-slurry N mineralization tests. The separation,
concentration, and dialysis were completed in about 2 weeks, with samples stored at 4 °C
between steps. The separated fiactions and a sample of whole soil were freeze-dried, and

then stored at -70 °C.

N Mineralizability in Whole Soil and Particle-Size Fractions

N mineralizability in whole soil and particle-size fractions was measured by two
complementary methods. The first method consisted of the kinetic analysis of clover-N
and nativebN in whole soils and particle-size fractions as a function of incubation time.
The amounts of clover-N and native-N were computed fi'om lsN measurements according
to the equations given in Appendix 1. As appropriate, the data were fit to one of four
mathematical models.

The first model is a one-pool exponential decay:
N = No ¢XP(-kt) (3)
where N! is the amount of N at time = t, No is the N amount in the fi'action at the
beginning of incubation, and k is a rate constant (tirne").

The second model is a two-pool exponential decay:
N. = N1 exp(-k1t) + N2 exp(-k2t) (4)

22

 

 

where the F
two pools 6
follows: i
N: = N1 expl‘

The 1
period of net

Nr=-A €Xp(-

 

where A and
depletion rate
The fo
N=M+NM
Where No is th
the Size of an I

acclll'l‘llllallon 1

where the parameters are defined analogously to Equation 3. In some cases, one of the
two pools exhibited essentially zero decay and the two-pool model was simplified as
follows:

Nt = N1 exp(-k1t) + N2 (5)

The third model is applicable where a period of net N accumulation precedes a
period of net N depletion:
N = -A exp(-kxt) + B exp(-knt) (6)
where A and B are (positive) constants, k». is the accumulation rate constant, and k1) is the
depletion rate constant. The derivation of this model is given in Appendix 2.

The fourth model simulated N accumulation without depletion:
N. = No + N1 (1 — exp(-kp.t)) (7)
where No is the N amount in the fraction at the beginning of incubation, N1 is related to
the size of an unspecified pool from which N is being transferred, and kA is the
accumulation rate constant.

For simplicity, no corrections for microbial biosynthesis were made in the models.
Parameter estimates and standard errors in the models were generated via the Levenberg-
Marquardt method within the nonlinear regression module (PROC NLIN) of the SAS
software (SAS Institute, 1989). The best model for each fiaction was usually chosen
based on visual inspection. vaisual inspection did not clearly identify one model as
superior, the model with the lowest ratios of the standard errors to the parameter
estimates was chosen.

The second method for measuring N mineralizability involved shaking the whole

soils and particle-size fractions (obtained from three dates of the l4—month soil incubation)

23

 

for 30 dfi.VS

Schnitzer (

percent N r

where X3!) =
extractable
observed fc
prior to sha
organic mar
agii’tgation
Soil:
Brim-er
5'4th to s;
was deSigne
To each flas

demed fTOrr

for 30 days in an N-free nutrient medium. The method was based on that of Catroux and

Schnitzer (1987). N mineralization was defined as shown below:

1\130 —NO - NCTRL X

 

percent N mineralized = 100 (8)

NTOTAL — NO

where N30 = extractable inorganic N in soil sample after 30 days of shaking, No =
extractable inorganic N in soil sample prior to shaking, Ncm. = net change in inorganic N
observed for a blank (no soil) sample shaken for 30 d, and Ntotal = total N in soil sample
prior to shaking. Ideally, this measurement reflects the capacity of mineral-associated
organic matter to supply inorganic N in the absence of the protective effects of soil
aggregation. Details of the measurements are described below:

Soils and 23 to 29 mL of N—free nutrient medium were added to 125-mL
Erlenmeyer flasks. Soil amounts were calculated to give 7 pmol N per mL medium,
subject to sample availability and a maximum soil:water mass ratio of 1:10. The medium
was designed to support generalized heterotrophic microbes, and is described in Table 4.
To each flask, 1 mL of inoculum was added per 29 mL of medium. (The inoculum was

derived from a moist soil that had been stored at 2 °c.3)

 

3 The inoculum served to provide microbes to the flasks without contributing a
significantamormtofN,andwasderived fromthepreviously described Mariette loam soil. A
fewgramsofthissoilwerebroughtto -33 kPawaterpotentialandincubatedeatroom
temperature. Thereafier,thesoilwasmaintainedat2°Cand-33 kPawaterpotential. Foreach
slmkakslmyNnmraahzafionexpannamaporfionoftheclfifledsoflwasmmbmedwnh
waterinthemoist soil:watermassratio of 1:10. Theresulting slurrywashand—shaken, allowed
tosettlefor30min,and1mLofsupematantwascombinedwith99mLoftheN-fieemmient
medium. Thismowhnnsuspensionwasconfimouslysfinedwlfileuansfernngappmpnate
aliquotstoeachflask.

24

immer
the 3C
extra:
shakir
enrar
conce
ahow
(In :1
ram
Syring

Preiin

Flasks were stoppered with foam plugs and inorganic N was extracted either
immediately or after 30 d of rotary shaking at 250 i 5 rpm. About twice per week during
the 30—d period, evaporated water was replaced and wall deposits were resuspended. All
extraction and shaking occurred in the dark. The average temperature for extraction and
shaking was about 23 °C, and the maximum deviation from the average was 3 °C. To
extract the soils, solid KCl was added to the suspensions to achieve a final KCl
concentration of l M. The slurries were then shaken at 250 i 5 rpm for 60 min and
allowed to flocculate (under the influence of the salts in the medium) at 2 °C for about 2 h.

(In the case of fine clay, the slurries were also centrifuged for 10 min at relative
centrifugal force = 26.) Finally, the supernatants were passed through cellulose ester
syringe-filters with pore size 0.45 um; the filtrates were stored at -20 °C until analysis.

Preliminary tests showed the filters contained no inorganic N.

25

Table 4. Description of N-free nutrient medium used in 30—day shaken-slurry N
mineralization tests?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

compound conc. (mM)
mm 22.5
mm 22.0

NaCl 8.56
Mg804 2.00

0102 0.100
macs 46.1 x 10‘3
MnClz 9.10 x 10'3
Fesot.7Hso 4.89 x 10'3
sodium tartrate 7.38 x 10'3
CuClz.4HzO 0.130 x 10'3
Znsot 0.152 x 10'3
CoClz.6H20 0.170 x 10'3
NazMOO4.2HzO 0.104 x 10‘3

 

 

 

 

1' This medium was modified from the M9 Minimal Medium (Sambrook et al., 1989, as
cited by Zuberer, 1994). The phosphate salts provided bufl‘ering near pH 7. The trace
elements were added via a stock solution described by Cote and Gherna (1994). Medium
was not sterilized

26

 

samples were

determined b}
described bel.
Solut

and shaken-s
(lachat). Im
the shaken- s]
CompOunds r
additions ( B a
Volumes of e
recovery of a
original Sam;
The ‘

Slurry emacr
(“one Scie
were diluted .‘
aqueous amm
51:“?CtrometerS

follows:

C, N, and 15N Analyses

Total C and N contents of clover, soils, and particle-size fl'actions were determined
by dry combustion (Carlo Erba NA 1500 Series 2 N/C/S Analyzer). For whole soil
samples with high levels of nitrate, both organic and inorganic N were determined. These
samples were extracted three times with 0.01 M CaClz; organic N in the solid residue was
determined by dry combustion and inorganic N in the combined extract was determined as
described below.

Soluble inorganic N (NOs' + N02' + NHI) in the leachates, whole soil extracts,
and shaken-slurry extracts was measured colorimetrically on a flow-injection analyzer
(Lachat). Interference in the colorimetric reactions occurred in many of the extracts from
the shaken-slurry mineralization tests, and was likely a result of colored organic
compounds or unflocculated colloids. The interference was overcome by standard
additions (Bader, 1980). Briefly, fixed volumes of sample were treated with fixed
volumes of either blank solution or a solution containing inorganic N. The percent
recovery of added inorganic N was computed, and the concentration of inorganic N in the
original sample was calculated accordingly.

The 15N concentrations of the clover, soils, particle-size fractions, and shaken-
slurry extracts were determined by mass spectrometry at either Michigan State University
(Europa Scientific 20—20 Stable Isotope Analyzer) or University of Georgia. Samples
were diluted as necessary with natural abundance N in the form of atropine, soil, or
aqueous ammonium sulfate to avoid introducing excessive 15N into the mass
spectrometers. The 15N concentrations in the soils and fractions were computed as

follows:

27

 

[INSMP = {A

there [INSM
”N in the mix".
concentrations

andrrmareth

 

Concentrations

In the lei
amounts of elm
“hole soil, For

Computing the 5

Statistical lrea
Dixon's

bet for Tespir:
The“? Outliers c
Clara due only It
difficuh to Corn
concentratims ‘

[ISN]SMP = {Ahr([N]er msrr) + [N]SMP mSMP) - Asm [N]srr) msrn}/(lOO mSMP)

where [”N]sMp denotes the sample 15‘N concentration, AM and Aer denote the atom %
"N in the mixture and natural abundance material, respectively, [N]sm and mats are the
concentrations of N in the natural abundance material and sample, respectively, and 111311)
and msw are the masses of natural abundance material and sample, respectively.
Concentrations were expressed on a molar basis.

In the kinetic analysis of N and C in whole soils and particle-size fractions, the
amounts of clover-N, native-N, and C were expressed as percent of initial amounts in
whole soil. For simplicity, the standard deviations of the initial amounts were ignored in

computing the standard deviations of the percentages.

Statistical Treatment of Outliers

Dixon's gap test (Bliss, 1967) was used to identify outliers at the 10 % probability
level, for respiration, elemental analysis (C and N), and shaken-slurry N mineralization.
These outliers comprised a small proportion of the data, and were excluded fi'om the final
data due only to their extreme deviations. This was deemed necessary because it is
difl'lcult to completely avoid analytical errors in samples with low masses or low
concentrations of analytes. According to Bliss (1967), ". .. the bias from rejecting a valid

observation is usually far less than that caused by retaining a contaminant."

28

NM

spinr

2004

ppm l

Appe

see As

NMR Spectroscopy4

NMR spectra of the solid state samples were acquired on a Varian VXR 400 MHz
spectrometer with a frequency of 40.5 MHz for 15N. Cross-polarization magic—angle
spinning was used to enhance the signal-to—noise ratio and hasten data acquisition. About
200-300 mg of soil were packed into a silicon nitride rotor of 7 mm diameter and spun
about the magic angle at 4500 Hz. Chemical shifts were referenced to (”NHahSOa (= 0
ppm); chemical shift assignments are summarized in Table 5 and given in detail in

Appendix 4.

 

4 See Appendix 3 for a brief overview of NMR concepts.
29

 

 

Table 5. Su

\

shift region
(ppm) ,,

 

5-34
42-76

 

 

 

158

\

l70-2ll

\

312-320

\

354
\
1 Chemical .
COWderallons‘
have a chemica

Table 5. Summary of "N chemical shifts.

 

 

 

 

 

 

 

 

 

 

shift region 1' nitrogen functionality

(PP!!!)

0 ammonium

5-34 amino groups in free and amino-terminal amino acids

42-76 amino groups of nucleic acids, guanidine nitrogens of arginine, indole-N of
mophan

83-116 amide N of proteins

120-148 heterocyclic N in certain positions in nucleic acids, in lristidine, or in flavin

158 heterocyclic N in pyrrole (chlorophyll)

170-211 heterocyclic N in nucleic acids, in flavin, or in pyrrole (chlorophyll)

312-320 heterocyclic N in oxidized flavin structures

354 nitrate

 

1‘ Chemical shifts are relative to ”NI-laNOs (= 0 ppm). Based on hydrogen bonding
considerations, the reference compound used in the present study, (lsNflahSOt, is likely to
haveachemicalshift within 1 ppmof‘5NHaNos. SeeAppendix4fordetails.

Standard samples were run to ensure optimal performance of the spectrometer
before each series of NMR experiments. The magic angle was checked and adjusted with
KBr. Then, the 90° lH pulse time, the Hartmann-Hahn match, and decoupling power
during acquisition were optimized on a sample comprised of lsN-uracil, (”NILhSOt, and
unlabeled soil. The uracil (Figure 2) served to verify that the spectrometer was able to
detect heterocyclic N, and the relatively narrow resonance range of (”NI-Iahsm allowed

for a precise determination of the Hartmann-Hahn match. The 90° lH pulses ranged from

30

Figure 2. Chemical structure of uracil

6.5 to 10.5 us, and decoupling power during acquisition ranged fi'om about 65 to 80 kHz
(as measured with an oscilloscope). With this decoupling power, linewidths (filll width at
half-height, lb = 10 Hz) were about 4 to 6 ppm for uracil and 2 ppm for ammonium
sulfate. Optimal contact times were 0.2 to 0.4 ms, delay times were 0.5 s or longer to
avoid probe overheating, and acquisition times ranged from 6 to 10 ms.

The uracil/anmronium sulfate/ soil sample used for the above optimization process
had been intimately mixed so that the molecular interactions (e.g. binding to soil particles)
would be similar to those expected in a natural soil. The uracil, ammonium sulfate, and
soil were combined in a sufficient volume of water for complete dissolution of both ls‘N—
labeled compounds; the resulting suspension was mechanically shaken for one hour,
quickly frozen in liquid nitrogen (to avoid precipitation of relatively insoluble uracil), and
lyophilized. This mixing method influenced the NMR behavior of the sample; the sample
prepared as above exhibited a lower Tm value than a sample mixed only by grinding with
mortar and pestle. This is likely due to shorter distances between 15N atoms and
paramagnetic species (e. g. Fe) in the water-treated sample.

31

Fort
intimate mixtl
technique de .
compared tot
effects ( as des<

To pro
multiplied by r
40 Hz for do
transformed s
correction we
fimctional grl

Correction OF

(affectionfi

Ditte
NMR Spectr;
differential r;

and Memmw

For testing whether NMR was quantitative (Objective 1 in Introduction), an
intimate mixture of lsN-uracil, lsN-clover, and unlabeled soil was prepared with the mixing
technique described above. The peak area proportions determined by NMR were
compared to the known sample composition, after correction for differential relaxation
efl‘ects (as described below).

To process each NMR spectrum, the raw data (free induction decay) were
multiplied by an exponential weighting function (line broadening = 100-120 Hz for soils,
40 Hz for clover), zero-filled to 16384 data points, and Fourier-transformed. The
transformed spectrum was phased and baseline-corrected. The phasing and baseline
correction were somewhat subjective and were performed twice. The peak area for each
functional group was defined as the average peak area from the two phasing/baseline

correction operations. Spinning sidebands were included in the peak area calculations.

Cw'rection for Diflerential Relaxation Effects

Difl‘erential relaxation effects must be considered for accurate quantitation of
NMR spectra. These efl'ects were evaluated via two models. The first model accounts for
difl‘erential rates of buildup or decay of magnetization during the contact time (Stej skal

and Memory, 1994):

exp (‘ tc /TlgH )- ”(1% lTNH) (9)
1- (Tan /T1pH)

 

S(tc)= So

where S(te) = measured signal intensity at contact time = to, So = theoretical signal
intensity, Tip}! = spin—lattice relaxation time in the rotating frame, and TNH is the cross-

polarization time. The model excludes Tips with the assumption that this parameter is of

32

 

negligible e1

<< 1, and 1
different con
with either t
1973, as moc
.\ll.\' of the
computed b1

group at a g

(‘13:~
expl

negligible effect (i.e. Tle >> Tlpfl). In addition, the model assumes the molar ratio lsN/‘H
<< 1, and TNH << Tlpl-l. To implement the model, NMR spectra were acquired at seven
difl‘erent contact times (ranging from 0.05 to 3.2 ms). Model parameters were generated
with either the COMPLEX nonlinear optimization program of Box (Kuester and Mize,
1973', as modified by Dr. Thomas Manetsch) or the Marquardt method within PROC
NLIN of the SAS software (SAS Institute, 1989); standard errors of the parameters were
computed by SAS. For convenience, a correction factor (CF) for a particular functional

group at a given contact time was defined by rearranging Equation 9:

l-(TNH/TIPH)

CF = ‘
°XP(‘ te/Tlplr )‘ 9XP(’ tc/TNH) (10)

 

Thus, corrected signal intensity = measured signal intensity x CF.
The second model accounts for difl‘erential decay of magnetization during the

delay time (Stejskal and Memory, 1994):
S(tb) = So [ 1 -— exp(-to/ Tm) ] (11)

where S(tn) = measured signal intensity at delay time = to, So = theoretical signal intensity,
and rm is the spin-lattice relaxation time of ‘H (in the laboratory frame). The model
assumes that spin-spin relaxation is negligible. To fit the model, NMR spectra were
acquired with three different delay times. Model parameters were generated with the

same methods used for the contact time model (Equation 9).

33

34

RESULTS

Overall N and C Mineralization

The overall N mineralization during the 14-month incubation of the clover-
amended soil is shown in Figure 3. The total N is given for the dates on which in situ N
mineralization was measured (via leaching), and was computed as initial total N minus
the cumulative leached N. Measured organic N is reported for the dates on which soils
were ultrasonically fractionateds; the basis for the modeled organic N curve will be given
later. (Gaseous losses were small as indicated by the small difference between the upper
and lower curves near the end of the incubation.) There was a substantial accumulation
of inorganic N in the middle of the incubation, as revealed by the large difference
between total and organic N at such time. This accumulation of inorganic N is likely due
to the infrequent leaching during the first half of the incubation. Such high
concentrations of inorganic N would likely not be found in a field soil and may have led
to some inhibition of N mineralization (e.g. no net change in measured organic N
between day 34 and 95). Nevertheless, Figure 3 indicates that N mineralization after 14

months was substantial and corresponded to about 30% of the initial N.

 

5 On Day 11, no whole soil sample was taken prior to ultrasonic fractionation, and
whole-soil organic N was estimated as follows:

organic N = initial whole soil N — leached inorganic N —- inorganic N recovered
immediately after sonication.

On Days 34, 95, and 190, organic N was measured on whole soil after removal of
inorganic N with 0.01 M CaClz, as described in Materials and Methods. On Day 439,
organic N was estimated as whole soil N minus inorganic N recovered immediately after
sonication.

35

 

+TOTAL N -
O ORG. N (MEAS)
-—ORG. N (MODEL)

 

    

 

 

 

 

 

N(%OF1N1T1ALTOTALN)
8

 

 

70 d
60 V T fifl I l— j l V l 7 T ‘ l r T T
0 50 100 150 200 250 300 350 400 450
INCUBATIONTIMEM)

Figure 3. N remaining after mineralization during soil incubation. (Error
bars represent sarrrple standard deviations and are invisible where error < 1%
of initial total N.)

36

The overall C mineralization is shown in Figures 4 (respiration) and 5 (whole soil
C). The initial respiration rates of the clover-amended soil were extremely high,
corresponding to 5 and 3 % of initial soil C per day after 1 and 5 days, respectively.
These rates reflect the rapid microbial attack of labile organic constituents, such as sugars
and amino acids. Respiration rates were very slow (< 0.08% of initial soil C per day)
later in the incubation, probably as a result of increased proportions of slowedegrading
compounds such as cellulose or hemicellulose. (See Paul and Clark (1996) for relative
decomposition rates of various substrates.) The respiration data were modeled with the
two-pool constrained model as described in Materials and Methods (Equation 2); the
model is appropriate as indicated by the close agreement between the measured and
modeled values (Figure 4) and the relatively low standard errors of the parameter
estimates (Table 6). The two C pools in the respiration model do not sum to 100 because
it was assumed that 45% of the native (non-clover) C was completely resistant to
mineralization throughout the incubation, as explained in Materials and Methods.

The modeled respiration data are contrasted with the whole soil C data as
obtained by dry combustion in Figure 5. As with the respiration data, the combustion
data were reasonably modeled by assuming the existence of two C pools with non-zero
decay constants and a third resistant pool (Figure 5, Table 6). The sharp difference in
slope between the combustion and respiration curves during the first month of incubation
is attributed to methodology. Respiration was measured only on days for which the soil
was at optimum water content and thus was probably overestimated: Occasionally (eight
times during the study), the soil water content substantially exceeded the optimum value

as a result of the leaching procedure for measurement of in situ N mineralization (see

37

 

 

 

 

 

 

 

 

 

6
E RESP=-dC/dl= CokoeXp (-ko t)+Ct R: 6X!) Hit 01
5
8 4 meter estimates and (standard errors)
°\° ,3 C0(%ofinitia1 soilC/d)= 40.6 (0.3)
[g a 3 . korcr‘)=0.l60 (0.002)
g :1 2 1 C, =37.5 (0.3) . MEAS
5 8 lo: 8.46e-4 (2.12e-4)
1: -—MODEL
3 r H—l
E 0 50 100 150 200 250 300 350

INCUBATION TIME ((1)

Figure 4A. Respiration rate as function of soil incubation time. About
51% of the initial soil C was derived from the clover amendment. Error
bars represent sample standard deviations.

 

 

 

 

 

 

 

 

 

:1! 0.12 ,

E . Mars

‘5

°\° A 0-03 ‘ -—M0DEL
V'U

E?)

are i r

68 0.044 I

F'- i

a J i ’ 3
g o a . . . a . . . . . . . -
33 0 50 100 150 200 250 300 350

INCUBATION TIME ((1)

Figure 4B. Expanded view of Figure 4A.

38

«mm 23 _ when 5253 mount 2.0823808 2 co v83 2e 8% corona—m8
c2308 25. dough—m8 3 n5 :8 20:3 we 33388 be 3 cognac we :8 20:3 5 méfifioa 0 .m oazwfi

 

 

 

 

 

3 $209382.
8e 8m 8n 2: o
2- . . t t t . t t on
r on

m . e a
% w
m .. m
m . E W
m S M asterowomfixo 2% +9 88min com" o cm W
a m r W
(O l ,1 CO 3

S . (

8 . Emma ammo: 9:

m: . Ewan—208 ammo:

m: . 388208982 0

 

 

 

 

 

- c:

39

130 u 3:82.00 88 .2023: :8 20.23 22202 .«0 ..\o H 2.038: .2925? :8 20:3
2.22:2 .20 e\.. u 2.5020 .0 :8 20:3 322 .20 .x. n 0 Saw com 0 220.... 2023 2225 .20 o\.. u 32 "8:22am": 5.30:8 as 0.3 3:5 .2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

grow: a 38.82%“. 8e u 3.8 u z m 2-3%: 23
8.8 a 8.8 + a 63¢ a 2 2 .8295 3.: n 38 u z m 2.3%.. an 358
as n 18 + a 823 n $5.8290 83 n 8.8 u z m 2.682 an 28
:2 mom; on 88.32.26 - a $3 an 381;» n 8 n z a 2.633: 36 .32
3 3.8.0 o 0:68.:ea - 8 ~29 n v.92 + Anne n 8 u z 9 2.623: 56 oases
E 802 .o o $8.829a - 3 an“ n 9.3 + Gem n 8 1 z a. 29:8 36 28
5232302 5.50 o :0 Enema 9 0: . 2-0330 :8 20:3
u 3.38 a mag. 8296 2 88 n 848 + 7 Good n 5 .829a so a 38 u z e 2-0526 29
88¢ n 93.8 + c 88.0 n 8: .8296 8n n :8 u z m 2.580 in 838
c Anonno n no; 82%... 8.1 a 8.8 + 20 $88 a 88.82%“. a: n a: 8 1 z e 223% an 28
c 9.22 n new“. 8293 9? 38 + a Gmmod o 33.829a 8.9 n 38. u z o 2226.6 326 .39
w some? a moon. 8290 c .m o 3.8 + a 9.088 a 88.8290 8.: a new. 1 z o 2.9% 36 ones
7 9-622 a 38.82.25 8.9. n v.0: + c a: 2 .o a 8.38298 8; n 58. u z o 2.563 36 2:0
a song o «6:823 2 .m u .38 + a :88 n 2: .829m 2a n 581 z e 223% =8 2223
SN + c 8.82 n 108ml 90 a: n :8 + c :83 a 9.08.8298 0: u 3.8 u 0 - Amowfizwmmm
o: + r 912 .N a $88298 one u 2.8 + 7 38¢ a 8:8298 23 n 381 0
c cans .N a 1882.5 Go o 28 cans .N a 2184.8 + N 8682968
8 38.0 n of .8290 So u 988 Good a CB .8 1 2e 8:292 0 =8 2&3
2200 08056 2.8 9.038230 8025.23 523 c2323 i8 :2 528%
.6983:
523:3

 

 

 

 

.2 see 2:...er 2er as 555m o 23.

40

Materials and Methods). The excess water from the leaching solution required one day
or more to evaporate and was likely inhibitory to respiration. This inhibitory efl‘ect may
have been most pronounced during the first month of the incubation; the soils were
leached twice during this period and the otherwise high respiration rate may have been
greatly decreased. Accordingly, the combustion data of Figure 5 reflect soil C dynamics
more accurately than do the respiration data.

The combustion data of Figure 5 indicate that about 40% of soil C was lost after
14 months. Thus, C mineralization (40%) exceeded N mineralization (30%, Figure 3);
the explanation is that some mineralized N was again immobilized whereas mineralized
C was permanently lost from the soil (see Introduction - C Mineralization). If it is
assumed that the loss of native C was 6%6, then about 75% of the clover C was ‘

mineralized (Figure 5, right~hand vertical axis).

N and C Mineralization/Immobilization in Whole Soils and Particle-Size Fractions
Recovery of Mass in Particle-Size Fractions

The particle-size yields following ultrasonic fractionation are shown in Table 7.
Although the particle-size distributions for the five incubation times are in general

agreement with the expected distribution, some differences are evident. Most notably,

 

6 This assumption is reasonable as judged by the data of Collins et al. These researchers
collected agricultural soils from Michigan, Minnesota, Ohio, and Wisconsin, removed
visible plant fragments, incubated the soils at 25 °C with optimum water content, and
measured respiration. The respiration data was fit to the two-pool constrained model of
Paul et al. (in press) as in the present study. Based on their model parameters, the
average C mineralization in topsoils alter 14 months would be 1 1i3% of the initial C. In
the present study, non-clover C comprised about 50% of total C, and thus the expected
mineralization of native C would correspond to only about 6% (50% x 0. l l) of total C.

41

there was incomplete dispersion on Day 11 and variable fine clay recovery on the other
dates. These discrepancies reflect difficulties in methodology and almost certainly do not
represent true differences in the particle-size distributions; N and C amounts in the size
fractions were adjusted accordingly (Appendix 5). The N and C amounts in each fraction

were also adjusted for mass losses during particle separation:

corrected N or C in fraction = measured N or C in fraction / (% recovery / 100)

The percent recovery in the above equation is equal to:

100 x soil mass recovered in all fractions / unfractionated soil mass

(See Table 7.)

Dynamics of Clover-N, Native-N, and C

The amounts of clover-derived N in the size fractions and whole soil as a function
of incubation time are given in Table 8. On Day 11, about one third of the initial clover-
N resided in the silts and sand; an additional one third of the clover-N was found in the
clays. The N found in the coarse silt and sand almost certainly is macroorganic matter
(MOM). Direct visual observation or microsc0py (50 X) revealed the presence of
relatively large pieces of organic matter in the coarse silt and sand; this type of material
was absent from the clays and fine silt. The N associated with the clays and fine silt

probably represents sorbed organic matter, and may also include small pieces of organic

42

Table 7. Recovery of mass in ultrasonic particle-size fractionations. T

 

 

 

 

 

 

 

 

 

-------_-.--._-_--------in.99.bs.t.i99_tims_(c.i) .........................
expected

fraction distribution 1 11 34 95 190 439
fine clay 3.81 i056 1.32 3.55 4.29 3.15 4.24
(< 0.2 pm)
coarse clay 12.4 i0.7 7.90 12.1 12.1 13.4 12.1
0.2 - 2 pm)
fine silt 12.0:t0.4 13.8 12.4 12.1 11.9 11.6
Q -10 m1
coarse silt 15.4 i 2.9 16.2 17.0 16.8 16.9 17.9
10 — 53 um)
sand 56.3 i 3.1 60.8 54.9 54.7 54.7 54.1
(53 — 2000 um)
SUM 100 100 100 100 100 100
recovery (%) § — 93 93 93 91 93

 

 

 

 

 

 

 

1' Data are percent of recovered mass, unless otherwise noted.

1 Means and sample standard deviations from preliminary study as described in
Materials and Methods and Table 3.

§ Recovery (%) == 100 x soil mass recovered in all fractions / unfractionated soil mass

material with entrapped mineral particles. This sorbed and fine OM likely originated
from solutes and small particles of the clover and from recent microbial products. About
15% of the initial clover-N was mineralized during particle-size separation of the Day 11
soil (Table 8). The original location of this mineralized N is unknown. This N is likely
part of an active organic matter pool; as evidence, note that the N mineralized during

particle-size separation declined throughout the incubation (Table 8).

43

 

Table 8. Clover-derived N in whole soils and size fractions (as % of clover-N initially
present in whole soil). 1'

# N o attempt was made to separate inorganic and organic N on Day 439. Thus, the
reported value probably includes a small amount of inorganic N. (Note that whole soil
organic N z sum of fractions on this day.)

'H' Whole soil organic N minus sum of fractions.

44

Table 8. Clover-derived N in whole soils and size fractions (as % of clover-N initially
present in whole soil). t

 

 

 

 

 

 

 

 

 

 

 

_--_--.irrsrrbat.i9rrtirps.(4) .................................

fraction 0 ; 11 34 95 190 439
fine clay - 6.16 7.84 11.1 6.39 7.22
(< 0,2 1"“) (0.10) (0.07) (0.8) (0.49) (0.35)
coarse clay - 26.3 38.7 40.5 33.2 25.8
(0.2 - 2 pm) (2.6) (4.1) (4.0) (0.1) (0.7)
total clay - 35.9 46.1 53.0 42.1 33.8

(2.6) (4. 1) (4. 1) (0.4) (0.8)
fine silt - 18.0 13.3 9.41 6.87 5.14
(2 - 10 11111) (0.5) (0.4) (0.27) (0.20) (0.17)
coarse silt - 5.06 0.410 0.268 0.315 0.285

10 -— 53 pm) (0.26) (0.014) (0.023) (0.024) (0.014)
sand - 7.69 0.888 0.377 0.336 0.189
53 -— 20009111) (0.38) (0.057) (0.022) (0.022) (0.015)

sum of 66.6 60.7 63.0 49.6 39.5
fi'actions

(2.8) (4.1) (4.1) (0.5) (0.9)
whole soil 100.0 82.0 § 71.3 1 70.0 1 50.5 1 41.1 #
organic N

(2.5) (1.2) (1.7) (4.9) (2.9) (1.2)

N mineralized - 15.4 10.6 6.94 0.863 1.63
during particle-
size separation
11

(3 .0) (4.4) (6.43) (2.927) (1 .48)

 

 

 

 

 

 

 

1' Data are means with sample standard devations in parentheses. Data are based on

 

isotope dilution equations, as described in Appendix 1. The data have been adjusted for
(1) loss of soil mass during particle-size separation, (2) incomplete dispersion on Day 11,
and (3) variable clay recovery. Because of these adjustments, total clay and sum of
fractions cannot be directly calculated from the data in the table. See text of Results for
details.

1 Soil was not fractionated on Day 0.

§ No whole soil sample was taken on Day 11. Whole soil organic N for this day was
estimated as described in the text.

1 On Days 34, 95, and 190, whole soil organic N was determined as described in
Materials and Methods.

45

The measured and modeled dynamics of clover-N are depicted graphically in
Figure 6; model parameters and standard errors are summarized in Table 6. The whole
soil organic clover-N was fit to a two-pool exponential decay model, constrained such
that the sum of the two pools was equal to the initial clover N (Figure 6A). Given that
the loss of native-N from the incubated soil was negligible (as will be shown later) and
that clover-N comprised 51% of the initial soil N, the model equation for loss of both

native and clover N from whole soil (Figure 3) is computed as:

N (% ofinitial) = 12.6 exp(-O.106 t) + 38.4 exp(-l .5193 t) + 49 (12)

Although one might expect C and N mineralization to follow similar models (with
perhaps slower decay constants for N), the pool sizes and rate constants in Equation 12
are substantially different from the equation for loss of C (as determined by combustion,

Figure 5). The C equation is restated below for ease of comparison:

C (% of initial) = 24 exp(-0.0303 t) + 54.1 exp(-9.08e-4 t) + 21.9
The difi‘erences probably are a result of the oversimplification in the C model; it was not
possible to model the clover- and native-C separately. The large difference between the
rate constants in the fast pool may be a result of the lack of data for whole soil C on Day
11; perhaps the initial decline in C in Figure 5 was faster than predicted by the modeled
combustion data.

Despite the aforementioned differences between the modeled C and N losses from

whole soil, Figure 6 and Table 6 indicate that the dynamics of clover-N within whole soil

46

CLOVER-N (96 0F INITIAL CLOVER-N IN WHOLE SOIL)

110

 

   
   

A. WHOLE SOIL

N = 24.7 exp(-O.106 t) + 75.3 exp (000151 1)

 

 

12 i a. FINE CLAY

 

10 J i N = -6.67 exp(-0.04101) + 10.4 exp(-0.001051)

 

 

 

  
 
   

N = -36.7 exp(-0.0661 t) + 44.6 exp(-0.00130 t)
20 . . . . . , . .

 

C. COARSE CLAY

 

 

 

1). TOTAL CLAY

  
    

N = -33.0 exp(-0.04041) + 57.6 exp(-0.00128 t)

 

 

 

I MEAS
--- MODEL

 

 

 

3O 4 I . 1 . T I r
0 100 200 300 400
INCUBATION TIME (d)

Figure 6. Clover-N amounts in incubated soil.

47

 

CLOVER-N (% OF INITIAL CLOVER-N IN WHOLE SOIL)

 

 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I MEAS
-— MODEL

 

 

 

16 j E. mm: SILT
12 . N = 11.8 exp(-0.0266 r) + 9.22 exp(-0.00137 t)
8 .4
4 w 9 . T . ‘-
6
F. COARSE SILT
4 -
. N = 27.7 exp(-0. 160 t) + 0.289
2 ..
I I
0 f 1 W —tr T 1
12
1 G. SAND
3 ‘ N = 27.2 exp(-0. 121 1) + 0.466 exp(-0.00197 r)
0 100 200 300 400
INCUBATION TIME ((1)

Figure 6 (cont'd.). Clover-N amounts in incubated soil.

4s

 

and particle-size fractions are described reasonably well by simple kinetic models. Sand-
and coarse silt-associated N were characterized primarily by a pool of fast-decaying
organic N (k = 0.12 to 0.16 d") which almost certainly represents macroorganic matter
(MOM). The fine silt contained two clover-N pools; the faster pool had an intermediate
decay rate (k = 0.03 d") and probably was MOM (with entrapped mineral particles) in an
intermediate stage of decomposition. The slow pool within the fine silt (k = 0.001 d") is
attributed to sorbed N. The clay fractions exhibited fast accumulation (k = 0.04 to 0.07
d'l) followed by slow depletion (k= 0.001 d'l) of clover-N, and were modeled with the
assumption that both processes followed an exponential trend (Appendix 2). (The large
positive error bars for fine and coarse clay on Day 11 reflect incomplete soil dispersion
(Appendix 5).) The accumulation is considered a result of the transfer of N from coarser
fractions, and the depletion reflects the slow mineralization of sorbed N. The ratios of
the standard errors to the parameter estimates are rather large for the clay fractions and
for the fast-decaying pool of whole soil N (Table 6), but this probably is due to the
inherent difiiculty in fitting models with three or four unconstrained parameters to only
five or six data points.

The overall dynamics of the clover-N in the size fractions can be summarized as
follows: Upon amendment of soil with clover, soluble and particulate N rapidly
distribute such that clover-N is found in all fractions. The coarser fractions (sand, coarse
silt, and to a lesser extent fine silt) are dominated by macroorganic nitrogen and the finer
fractions (clays, and to a lesser extent fine silt) are dominated by sorbed N. The

macroorganic N is rapidly attacked (k = 0.12 to 0.16 d‘l), and the resulting microbial

49

products accumulate as sorbed N in the clays. Thereafter, the clay-sorbed N is slowly
mineralized (k = 0.001 d"), acting as a long-term nutrient source.

The dynamics of native (non-clover) native N are presented in Table 9 and Figure
7. The distribution of native N on Day 11 is similar to that of clover-N, except that the
preportion of native N in the coarser fractions (> 2 urn) exceeded the proportion of
clover-N in said fractions (compare Tables 8 and 9). This is attributed to the possible
mineralization of relatively labile macroorganic clover-N during particle-size separation,
as well as to the grinding of the clover prior to the incubation.

Figure 7A shows that there was little or no net mineralization of native organic N
in the whole soil throughout the 14-month incubation. The large amendment of N-n'ch
clover and the high concentrations of inorganic N during much of the incubation (see
Figure 3) explain this observation The measured and modeled data for the size fractions
(Figure 78—6) are similar to that of clover-N in that a rapid depletion of N in the coarse
fractions is accompanied by an accumulation of N in the clays. The movement of
apparently labile native N from the sand— and coarse silt-associated MOM to the clays
without subsequent decomposition of the clay-associated N (in contrast to the case of
clover-N) is difficult to explain. In the N-rich (C-limited) condition of this study, native
MOM may have been attacked primarily to satisfy microbial C requirements. The
disintegration of the MOM may have released resistant N compounds, which were then
sorbed on clays. The relatively poor model fits for native N as compared to clover-N
(Table 6, Figures 6 and 7) are probably due to the higher uncertainty of the native N
amounts as well as the paucity of data points during times of native N accumulation or

depletion within the fractions.

50

Table 9. Native N in whole soils and size fractions (as % of native N initially present in
whole soil). T 1

# No attempt was made to separate inorganic and organic N on Day 439. Thus, the
reported value probably includes a small amount of inorganic N. (Note that whole soil
organic N 9' sum of fractions on this day.)

H Whole soil organic N minus sum of fractions.

51

E13313

Table 9. Native N in whole soils and size fractions (as % of native N initially present in

 

 

 

 

 

 

 

 

 

 

 

whole soil). 1' I
.............................. 11399124092999-1512.----------_----
fi‘action 0§ 11 1 34 95 190 439
fine clay - 3.73 16.6 10.0 12.7 14.8
< 0.2 pm) (0.41) (0.8) (1.5) (1.1) (2.1)
coarse clay - 25.1 42.7 41.5 38.3 45.7
(0.2 - 2 pm) (4.4) (11.6) (7.4) (1.8) (3.9)
total clay - 36.0 58.2 52.8 52.9 62.2
(4.4L 11.9 (7.6) (2.2) (4.6)
fine silt - 31.9 34.4 27.5 26.8 28.6
(2 -10 um) (2.0) (2.5) (2.3) (2.1) (1.2)
coarse silt - 13.8 6.55 6.07 6.47 5.44
(10 - 53 pm) (1.0) (0.32) (0.19) (0.31) (0.16)
sand - 22.3 7.85 9.69 8.01 3.82
(53 - 2000 pm) (1.9) (0.53) (0.26) (0.90) (0.44)
sum of 104.0 107.0 96.1 94.2 100.0
fiactions
(5.5) (11.9) (8.0) (3.2) (4.7)
whole soil 100.0 ND 103.9 1 110.2 11 104.2 1 103.2 #
organic N
(2.5) (9.5) (14.2) (10.2) (6.4)
N mineralized - ND -3.03 14.1 10.0 3.18
during particle-
size separation
II
(15.20) (16.3) (10.6) (7.96)

 

 

 

 

 

 

 

 

1' Data are means with sample standard devations in parentheses. Data are based on
isotope dilution equations, as described in Appendix 1. The data have been adjusted for
(1) loss of soil mass during particle-size separation, (2) incomplete dispersion on Day 11,
and (3) variable clay recovery. Because of these adjustments, total clay and sum of
fiactions cannot be directly calculated from the data in the table. See text of Results for
details.

I ND = not determined
§ Soil was not fractionated on Day 0.

1 On Days 34, 95, and 190, whole soil organic N was determined as described in
Materials and Methods.

52

NATIVE N (96 OF INITIAL NATIVE N IN WHOLE SOIL)

130 ‘

120
110
100

388

18

r4~
1 i

10‘

 

A. WHOLE SOIL ORGANIC N

 

%

 

 

 

B. FINECLAY

H-e-r

Ni

=..0+134(1 exp(-00651t))

 

 

 

 

30

201

C. COARSE CLAY

i

N= 0+42.-4(l-.exp(-008701))

 

 

 

 

7o—
60;
503
40:

-4
d
.4
‘
d
‘

 

1). TOTAL CLAY
i

 

I
=O+56.-8(l -.exp(-0096St))

 

 

I MEAS
— MODEL

 

 

 

30

0

I T T

200 300

INCUBATION TIME ((1)

Figure 7. Native N amounts in incubated soil.

53

 

NATIVE N (96 OF INITIAL NATIVE N IN WHOLE SOIL)

39

 

36‘
33—
30+
27‘

 

24

E. FINE SILT

N = 7.06 exp(-0.0153 1) + 27.4

 

i

 

 

 

F. COARSE SILT

N = 27.8 exp(~0.115 t) + 5.99

 

 

 

 

 

i G. SAND

  

N = 17.4 exp(-0.00507 t)

 
 

 

 

O 100 200 300 400

INCUBATION TIME (d)

Figure 7 (cont'd.). Native N amounts in incubated soil.

54

 

 

I MEAS
-- MODEL

 

 

The dynamics of C (clover plus native) within the soils and particle-size fractions
are given in Table 10 and Figure 8. As in the case of N, significant amounts of C were
mineralized during particle-size separation except on Day 439 (Table 10). The trends in
C are similar to those of N; disappearance of C in the coarse fiactions (> 2 um) is
accompanied by accumulation of C in the clays. Although no decomposition of clay-
associated C is evident (Figure 8A—C), it may be that the mineralization of labile clover-
derived C and the accumulation of resistant native C occurred simultaneously in the clay
fractions. The net decrease in sand-associated C on each measurement date (Figure 8)

suggests that POM could have provided C to the clay fractions throughout the incubation.

Mineralization of N in Whole Soils and Particle-Size Fractions in Aerobic Shaken
Slurries

An aerobic shaken slurry N mineralization test was conducted on the whole soils
and particle-size fractions from Days 34, 190, and 439 of the incubation of clover-
amended soil, as described in Materials and Methods. The test provided additional
detailed information about the N in the size fractions, and, ideally, reflected the capacity

of the mineral-associated OM to supply inorganic N in the absence of aggregate

55

Table 10. C (native- plus clover-derived) in whole soils and size fractions (% of C
initially present in whole soil). '1 t

 

 

 

 

 

 

 

 

 

 

 

................................ 19.91.1949931932219)..---------------
fraction 0 § 11 34 95 190 439
fine clay - 2.56 6.97 6.16 5.79 7.36
< 02 pm) (0.06) (1.02) (0.93) (0.86) (1.11)
coarse clay - 15.2 28.4 27.5 27.8 26.6
(0.2 - 2 pm) (0.3) (1.3) (0.6) (0.3) (0.1)
total clay - 22.8 34.9 34.4 35.6 34.8
0.3) (1 .3) (0.6) (0.4) (0.3)
fine silt - 19.9 25.3 19.2 16.2 17.6
(2 - 10 um) (0.3) (0.6) (2.0) (0.7) (0.2)
coarse silt - 9.16 5.18 3.93 4.34 4.67
(10 - 53 pm) (0.07) (0.13) (0.03) (0.08) (0.11)
sand - 18.9 8.24 7.10 5.74 4.03
(53 - 2000j1m) (0.6) (0.07) (0.18) (0.92) (0.29)
sum of 70.8 73.6 64.6 61.8 61.2
fractions
(0.8) (1.5) (2.1) (1.2) (0.5)
whole soil C 100.0 ND 83.3 72.0 68.4 58.0
(5.8) (4. 1) (2.6) (3.7) (1.6)
C mineralized - ND 9.69 7.40 6.52 -3.15
during particle-
size separation
‘II
(4.36) (3.31) (3.87) (1.69)

 

 

 

 

 

 

 

 

‘I’ Data are means with sample standard deviations in parentheses. The data have been
adjusted for (1) loss of soil mass during particle-size separation, (2) incomplete
dispersion on Day 11, and (3) variable clay recovery. Because of these adjustments, total
clay and sum of fractions cannot be directly calculated from the data in the table. See
text of Results for details.

1 ND = not determined

§ Soil was not fractionated on Day 0.

fl Whole soil C minus sum of fractions.

56

 

 

 

 

 

 

 

 

 

 

8 4
6 ..
4 u
,2 . A. FINE CLAY
d 2 ~
8 O u 7 I I I I I
LU
6
I 30-1
3 j ‘ A
.—|
E 207 B. COARSE CLAY
z
8 10 T F I I T T
65’ 40
U 4
3o -
C. TOTAL CLAY
1
20 I I I I T I
0 100 200 300 400
INCUBATION TIME (d)

Figure 8. C amounts in incubated soil. About 51% of the initial soil C

was due to clover.

 

 

3O

20‘

D. FINE SILT

 

10

 

12
10‘

E. COARSE SILT

 

 

C (% OF rNrrrAL WHOLE SOIL)
00

 

 

 

 

I

100

200

I I T

300 400

INCUBATION TIME (d)

Figure 8. (cont’d) C amounts in incubated soil. About 51% of the

initial soil C was due to clover.

 

protection. One might argue that the shaken slurry N mineralization tests were
ambiguous measures of aggregate protection because of greater diffusion of oxygen and
inorganic nutrients in the shaken Slurries relative to the undisturbed soil. However,
oxygen did not limit biological activity in an undisturbed silt loam soil (Rovira and
Greacen, 1957; see Introduction of present study), and would not be a likely limiting
factor in the sandy loam soil of the present study. The undisturbed soil of this study was
probably well-supplied with inorganic nutrients via the clover amendment, which was
grown in Hoagland’s nutrient solution. The periodic leachings of the undisturbed soil
with N-free nutrient solution supplied additional inorganic nutrients.

The shaken slurry mineralization in whole soil is considered a measure of
aggregate protection because prolonged shaking in water has been shown to disaggregate
(disperse) soils (Edwards and Bremner, 1967). Edwards and Bremner (1967) found that
shaking of soils for 10 days with 18 mL water per g soil resulted in clay yields that were,
on average, 76% of those obtained with the highly effective resin method (shaking of soil
with an Na-saturated exchange resin). No attempt was made to measure the dispersion of
whole soil in shaken slurry in the present study, and the high water: soil ratios (24 to 30
mL per g) may have resulted in less than complete dispersion. Thus the aggregate
protection results for whole soil should be viewed as semiquantitative.

Nitrogen mineralization in aerobic shaken slurry ranged from slightly negative
numbers (immobilization) to 12% of the organic N initially present in the slurry (Table
1 1). Thus, one to twelve % of the N in the whole soil, fine clay, coarse clay, and fine silt
was aggregate-protected. The non-mineralized N in these fractions was evidently

stabilized by other mechanisms, such as adsorption per se or humification. There was

59

6358.688
8: 33 Z Bum—82:8 80 8088 Z: 228 83¢sz 8 Z 038: 98 2.3.63 Sons 52388.25 89.38 02 “

335532 :2 meow—«$68 888:3.“ 29:9. .823 5 £32: 03 «:5 +

 

 

 

 

 

 

 

 

 

 

am. 8 @518 $8 5 .8 :98 3.8 on.
em; :28 a: and Id 2.»
$28 888% 58 A88 $2.8 9&8 02
mi .8 ovmd mmd 3+ 3.9 can
$2.8 38.8 98.8 58 No.8 8.8 3.
Sad. wvvd- ammo Mme». 9.: Ea
8:3 Em 838 2 xi 2:8 >26 833 a .326 new :8 2223 :8 523252888 9
. ............................................. mEWww. ............................ . ................. coca 0:5 868335

 

 

 

w .P A5888 5 2 2:38 .«o 25208 38 mots—m :33...“ :2 8:32—30:28 Z flashy—22:. .: 03am.

very limited mineralization in the coarse silt and sand, which was probably a direct result
of high C :N ratios in these fractions (Table 12). These fi'actions contained only a small
portion of the soil N recovered on Days 34, 190, and 439 (Tables 8 and 9) and will be
discussed no firrther.

The aggregate protection (N mineralization) of whole soil N between Days 34 and
439 did not change (Table 11). This is surprising, as one would expect aggregate-
protected N to be humified with time. Perhaps real differences are masked by the
variability ofthe whole soil data7. Alternatively, the high nitrate levels in the Day 34
whole soil (Figure 3) may have inhibited N mineralization.

The relationship of 30-day N mineralization in the shaken slurries to particle size,
time (i.e. days of incubation prior to fractionation), and C:N ratio was investigated by
analysis of variance (ANOVA) (GLM procedure of SAS; SAS Institute, 1989). Coarse
silt, sand, and whole soil were excluded from the AN OVA because of inhibition of
mineralization by high C:N ratios in coarse silt and sand, and high variability in the
whole soil, as discussed earlier. Both 30-day N mineralization (NMIN) and its natural
logarithm (ln(NMIN)) were subjected to analysis of variance because scatterplots and
linear regressions revealed that the relationship between NMIN and C:N ratio was more
exponential than linear (Figure 9). When only time and particle size were considered as
sources of variation, AN OVA of both NMIN and ln(NMIN) revealed a highly significant

interaction of time and particle size (Table 13). When C:N ratios were considered as well

 

7 The high uncertainties in the whole soil data stem from the relatively large values of
No (initial inorganic N) in Equation 8. As discussed previously, inorganic N levels were
high during much of the 14-month incubation (Figure 3); subtraction of the large No
value from the similarly large N30 value in Equation 8 results in considerable uncertainty.

61

 

 

 

 

 

 

 

 

 

 

 

 

 

12 4 o
10 . OFINE CLAY
E 8 ‘ ICOARSE CLAY
z 6 4
4 , AFINE SILT
2 4
o a
0 5 20 molar C:N
3 -
2.5 4
2 2 ‘ oFINE CLAY
g 1-5 “ ICOARSE CLAY
5 I ‘ AFINE SILT
0.5 ~
0 .
,0 5 6 5 20 molar C:N ratio

Figure 9. NMIN and ln(NMIN) versus molar C:N ratio. NMIN = 30-day N
mineralization in aerobic shaken slurry as percent of organic N in fraction. Data

points are averages of two or three replicates.

62

Table 12. Soil C/N ratio (molar basis) as a fiinction of incubation time.

 

 

 

 

 

 

 

 

 

 

 

incubation time (d)

fraction 0 T 1] 34 95 190 439
whole soil 13.47 ND: 11.1 9.4 8.63 11.2
(0.8) (0.5) (0.1) (0.26) (0.1)

fine clay ND 6.84 7.90 7.82 8.39 9.27
(< 0.2 pm) (0.24) (0.04) (0.26) (0.10) (0.74)
coarse clay ND 7.96 9.46 9.04 10.5 10.2
(0.2-2 pm) (0.21) (0.87) (0.12) (0.1) (0.3)
fine silt ND 11.0 14.7 14.5 13.5 14.8
(2-10 pm) (0.3) (0.2) (1.0) (0.0) (0.2)
coarse silt ND 13.5 21.4 17.9 18.5 23.5
(10-53 um) (0.3) (0.4) (0.4) (0.3) (0.5)

sand ND 17.6 27.0 20.3 19.8 29.1
(53-2000 um) (0.4) (1.5) (0.6) (1 .2) (2.7)

 

 

 

 

 

 

 

‘1 The C/N ratio on Day 0 was computed from the C and N contents of the soil and clover
amendment.

1 ND = not determined

as time and particle size, AN OVA revealed that all three factors were significant sources
of variation (Table 13). The important interpretation is that 30—day N mineralization is
significantly related to (1) the balance of C and N availabilities, (2) the extent of previous
decomposition (time), and (3) fundamental differences in the nature of each size class of
organo-mineral complexes.

Specific pre-planned comparisons among the size fractions are presented in Table
14; these were derived from the SAS least-squares means option within the ANOVA of
NMIN versus (time, particle size, time x particle size). For a given time, the data show
that NMIN decreased significantly (p < 0.05) as follows: fine clay > coarse clay > fine

silt.

63

Table 13. Analysis of variance of 30-day N mineralization in aerobic shaken slurries?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

dependent variable 1 source of variation § p—value
NMIN time 0.0001
particle size 0.0001
time x particle size 0.0001
NMIN time 0.0780
particle size 0.0116
C:N ratio 0.4708
ln(NMIN) time 0.0001
particle size 0.0001
time x particle size 0.0001
ln(NMIN) time 0.0016
particle size 0.0166
C:N ratio 0.0006

 

 

1' Only fine clay, coarse clay, and fine silt were included in the analysis of variance per
reasons stated in the text. For details of analyze of variance, see Appendix 6.

INMIN:100X(N30-No-NCTRL)/(NT-No)

§ time = incubation time prior to fractionation

(See Materials and Methods.)

Table 14. Significance levels for pre-planned comparisons of 30—day N mineralization
values in shaken slurries. t

 

 

 

 

 

 

 

 

incubation time

prior to

fractionation ((1) fine clay > coarse clay fine clay > fine silt coarse c1ay> fine silt
34 Se-S Se-S Se-S
190 0.03 0.0007 5e—5
439 0.003 5e-5 5e-5

 

1‘ Tests are for differences within rows. Comparisons were derived from pro-planned
one-tailed tests of least squares means in SAS. See text for additional details.

 

Aggregate protection factors (APF) are complementary to the above analyses of
shaken-slurry N mineralization, and are usefiil for distinguishing among N stabilization

mechanisms. The APF is defined as follows:

= 30 - d N mineralization in shaken slurry
30 - d N mineralization expected in undisturbed soil

The expected N mineralization in the undisturbed soil was computed on the basis of the
kinetic model equations for clover- and native-N given earlier; see Appendix 7 for
details.

The whole soil APF showed no change with time (Table 15). (The apparent
difference in the aggregate protection factor between Days 34 and 439 is not statistically
significant (p > 0.1 for t-test).) This finding is surprising because one might expect APFs
to decrease as humification proceeds. Possible reasons for unexpected findings for whole
soil were given earlier (high nitrate, high variability).

The highest APFs were found in the fine clay (Table 15). This is reasonable as
fine-clay associated N is expected to be less accessible to microbes compared to the N in
other fractions. The decrease in APF with time for both fine and coarse clay suggests
that humification was occurring in those fractions or that the organic N in those fractions
was becoming more strongly bound to the mineral phase. The APF for fine silt on Day
34 is less than one. This result is difficult to explain, and may be due to uncertainty in

the mineralization modeled for the undisturbed soil.

65

Table 15. Aggregate-protection factors for whole soil and particle-size fractions. '1' I

 

 

 

 

 

 

 

 

 

incubation
time prior to
fractionation
((1) whole soil fine clay coarse clay fine silt
34 3.64 :1: 0.98 10.6 :h 0.8 A a 2.67 :t 0.37 B a 0.0737 :t 0.0088 B
190 3.73i3.75 6.20i 1.01 Ab 2.44:1:015 Ba l.63:t0_31 C
439 583$ 1.59 3703:081 Ab 1.67i0.10 Ab 1.69zt0.31 A

 

1 Aggregate-protection factor = (30-day N mineralization in shaken slurry) / (30-day

mineralization in undisturbed soil); see text and Appendices 7 and 8 for details.

1 Different capital letters denote differences within rows. Different lowercase letters

denote differences within columns. (p = 0.10) Whole soil was excluded from statistical

tests due to its high variability, and no tests for time trend were performed on fine silt.
These results are based on unplanned two—tailed t-tests; the stated p-value reflects the

experimentwise error rate according to Bonferroni’s inequality (Sachs, 1982).

66

 

The estimation of aggregate-protected pool sizes in the whole soils and size
fractions constitutes another useful application of the shaken slurry N mineralization data.
The aggregate-protected pool size is estimated as:

aggregate-protected pool (% of N in fraction) = (3 0-day shaken slurry N

mineralization) - (3 O-day N mineralization in undisturbed soil)

This operational definition is similar to that used by Beare et al. (1994). The aggregate-
protected pool sizes are shown in Table 16, and range fi'om 0 to 11% of N in a given
fraction or whole soil. The data indicate that the aggregate-protected pool sizes (as a
percentage of N in fraction) decrease with increasing particle size. In the case of fine
clay and coarse clay, the pool sizes decrease with time. These observations reinforce
conclusions drawn from the APF data of Table 15. That is, aggregate protection is more
pronounced in fine fi'actions, and its decrease with time in the fine fractions indicates that
humification is occurring or that N is becoming more tightly bound to the mineral phase.
No decrease in the aggregate-protected pool size was observed in whole soil; this may

stem from the difficulty in detecting differences with highly variable data.

67

Table 16. Aggregate-protected pool sizes for whole soil and particle-size fi'actions (% of
N in fraction). 1 1

 

 

 

 

 

 

 

 

 

incubation
time prior to
fractionation
(d) whole soil fine clay coarse clay fine silt
34 6173:229 10.6i0.9 Aa 3.27:1:073 B a NC§
190 4.34 i 5.96 5.93 at 1.15 A b 2.75 i 0.29 A a 0.91 d: 0.45 B a
439 6761-222 3.022t091 Ac 1013:015 Ab 0.49i0.22 Ba

 

1' Aggregate-protected pool size (°/o of N in fraction) = (30-day N mineralization in
shaken slurry) — (30-day mineralization in undisturbed soil); see text and Appendices 7
and 8 for details.

1 Different capital letters denote differences within rows. Different lowercase letters
denote differences within columns. (p = 0.10) Whole soil was excluded from statistical
tests due to its high variability. These results are based on unplanned two-tailed t-tests;
the stated p-value reflects the experimentwise error rate according to Bonferroni’s
inequality (Sachs, 1982).

§ NC = not calculable (computed pool size is negative).

68

 

Test of Quantitative NMR

The CPMAS-”N-NMR spectrum of the prepared mixture of lSN-clover/ 15N-
uracil/unlabeled soil is shown in Figure 10. The two N atoms of uracil are evident as
well as the amide, guanidinium, and amino N atoms of clover. The peak area percentages
with and without correction for differential relaxation (per Equations 9, 10, and 11) are
given in Columns 2 and 3 of Table 17. Differential relaxation effects were small for the
sample since the relaxation-corrected peak area percentages are approximately equal to
the uncorrected percentages; this is because (1) the delay time for the spectrum of Figure
10 (1.25 s) was substantially greater than the maximum value of Tm (0.454 s, see Table
18), and (2) the correction factors for differential relaxation during the contact time were
approximately equal for all functional groups (Table 18). With or without correction for
relaxation, the NMR peak area percentages do not correspond to the known sample
composition (Table 17, column 1). Peak overlap was considered a likely source of this
error.

To correct for error due to peak overlap, the clover/uracil/soil spectrum was
approximated as a linear combination of previous spectra taken of uraciVsoil and
clover/soil. After the statistical validity of this approach was confirmed, the relative
contributions of uracil and clover to the clover/uracil/soil spectrum were computed
(Table 17, Column 4 - see Appendix 9 for details). In this case, the peak area proportions
correspond closely to the known sample composition. The important conclusion is that
l"N-NMR is quantitative in a complex soil sample for both heterocyclic and noncyclic N.
However, significant errors may result from peak overlap in samples of unknown

composition.

69

fiancee wage u .. .833 co N:2. 863% u 8 was? .6 -mz- Sagas» u 5 .m as
n 2:: >226 $8 No u 0:5 “08:00 .mom “co—092:: + eo>o_o-}_ + €55-22 mo 29:8 3835 Mo 83803 M277} _ .3 8sz

Eng
00—. cm- 0 cm on: o2 com cmm 00m cmm oov

 

 

 

Alnz-du<y5

 

70

.m 8289? a. 83.8.8.8 mm «58% ..8\8>o.o can manage: me 8:83:80 83.. .8 .885 80.888 8.88 anon. _r

5.82:. Ram? .39 .8 $2. 55 m8. 083 8.0: .8809. 28 o8.82.80
08.83 £58.... 9 26 89.8388 38 6.38 8 8.8.3883 388.80 2.... 683.82 8: 083 8335 $22 a

8882.88

8:38 2: .8 2888 883 .28 80.888380 2: £838 a. 35380:: 8 .88.. 3.2.83 083 $0858.38 .5
28.888 88.88». 0.985 03.830: 83 8.88380 2: =8 488308.358 .3 3.8.88 083 82.880280 2: .88:
88888on 89: 3 35888.8 83 8:888:00 Z: 88.0 .83: .65 88.0 5 8288380 Z: :8... .88880 N

Am 2sz 00m .uonzeva n .888 .Z: .8 8.32. ma: 9 8.8.. maromnam .r

 

 

 

 

 

 

 

 

 

 

co. co. co. 8. 35m
64 8.285%
9 9 55.8.28
Nmm 1mm Dom ode 88.0
a. 3
SN #8 mom «.3 a. £88:
a. :
Nam «.3. ..mm 63 * 2-.."..2:
waning 88.. .85 8288.8 m 9882.883 u NZ .89
8.88.2 3820......6 8.. 389:8. 3.80.5.6 8.. momma—808.. 88 .8 8383
88888.. 8.3 6.3:. £42.22 .v 38283 «owes—88.. 8.82. 542-}. .N 83.89.80 88.. 2.2
38 8.3.. g.z_ .m 29:8
.585. ..

 

 

.58 8.033: + 3.53:. “5.92: + ..88.Z:V..o 0838 3388 a. 3.2: 3.3.23.8 .8 so... .2 03a...

71

3 who 83 363 2:? 8.8 8:882 .888
5953 v8.8.2 3.3 on 8 3888 98% 382888 43288 83 08: >28 :28 36:28 Rama .8 8888828 on u <2 F

.88ch ms: 8 8.88 $8535 a

.m8 N .o u 08: 85.80 53, .3 8883.— 8a cows—3.8 88am 828280 «

8358088 8 828 c.8286 .r

 

 

 

 

 

 

 

 

8: Ga: 2 .m $3.8 $8.0 <2 853
. . . . . -fiz-
§ _ 5 8 S _ 38 8 Es o <z 825838
«.8 _ A: c 2. m 88 e 28 0 <2 825253»
«8.. am. 3 m: 38.8 886 w <2 83
MS ._ 93.8 Em A: 8.8 228 $8.8 «Sac a .2435
m2 ._ A38 Be $8.8 $8.0 63.8 3.3 a “2435
M 08: 63:8 matau AmEfiEH 88v my; ah museum
83822 380856 382883
88 888 828083

 

 

 

 

 

 

.8 :9. 8.0%? + __o§-z_ + 38.92: 8o 2.8a 8385 5 3% 88.33. .M: 23

72

N MR Spectroscopy of Incubated Soils

The CPMAS-NMR spectra of clover, whole soil, fine clay, coarse clay, and fine
silt are shown in Figure 11. (No separate spectra were acquired for coarse silt and sand
due to the low concentrations of 15N in these fractions.) The peak area percentages
(corrected for differential relaxation effects) are summarized in Table 19. Throughout
the incubation, the N functional group composition of the whole soil and size fi'actions
was not significantly different fiom that of pure clover; the organic N was approximately
90% amide, 5 - 10% guanidinium N of arginine, and 5% amino. (The identification of
the guanidinium and amino peaks is explained below.)

The spectrum of pure clover (Figure 11A) is better resolved than the spectra of the
soil samples and thus provides additional information on the organic N composition at
the beginning of (and probably throughout) the incubation. The strong amide signal (95
ppm) suggests that the clover-N is mostly proteinaceous; therefore, the other peaks in the
spectrum probably represent non-amide N within amino acids. The spectrum shows one
peak at approximately 150 ppm, two peaks in the region 30-70 ppm, and two peaks in the
region -10 to 20 ppm. The peak at 150 ppm is likely to be the heterocyclic N of histidine.
Multiple functional groups have resonances in the region 30-70 ppm (Table 5), but
stoichiometric considerations strongly suggest that the peaks at 58 and 45 ppm are the -
NH- and -NH2 units of the guanidinium group of arginine, respectively. The ratio of the
relaxation-corrected peak areas (-NH- : -NH2 = 112.5, data not shown) was close to the
expected stoichiometric value (1 :2) for the guanidinium group. The two peaks in the
region -10 to 20 ppm probably represent the amino groups of free or amino-terminal

amino acids; these peaks cannot be unambiguously assigned to specific amino acids

73

.56:an Hews :39
mo ..\o m .85 088 o: 3 mowfleooeoa 3.3 ion 2: @833 228280 82E. douse—2 .uccouobau Ba $62.50 soon 23: Sun— a

.Emoflfimmm 8: 39305 8d 53:83 12%».
:39 me o\.. E I m =2. 9:..— mo moococogv £2ng 05—83 98 mafia—E 5 $580.33 28 .86: 353% doing x8e 9 25 H

.303 anon oz: 2: «o 288 2: me. 333—08 203 303 can“. 35m 28 n933:3 8:83 + mEmfla
+ cosaeeomgwb 3955 333 9.6885 3.8 803.63: 8.58% scam 8.53% 6282823 25 no woman mm BBQ 3% seem .r

 

 

 

 

---E-----E._ ....... E._-.----.e.~. ..... E ..... BEE ...... we ....... o e ..... 92.----.03 ..... E .......... E-.. ............ Em .............
--.N.m._..----_..o..w----m_.m ...... a..m----a._.a.---a_m-.--*...M...a-----.m.% ..... a 99-x.» ..... a fisew ........ E 2%
4 2:5me0

.E ..... E ..... E ...... m ._...h.----.a.¢.wm---@...m.-.--e..k-----m_.a. ..... mm...m_--._.mm ..... a ma----.m.m.m.-------e.f ..... «MEEEE ......

-E ..... E ..... we ...... a -a.a----._-._..m.---m-w.a-.-E ..... E ..... Eéma ...... a -mm---._..e..-------.m..a.. ............. tE .............

o o o o o o o o o o o c we; 2:25: .3 Z 3:383

--.§---E ...... m ....... 3.. ..... E ...... m. ...... a. m. ...... a. _ ....... a. ...... me. ..... E ....... a.-. o -.+..@-E Egg.-.
:3 25 >20 3.38 >20 new :8 20:3 ego—o T 3:32:

 

 

 

 

 

 

 

 

m N + :8 8.385 23 56.0 .8 83:85 as been «22 .2 2.3

74

 

 

 

A CLOVER
* *
400 350 300 250 200 150 100 50 0 -50 Hun-100
B. WHOLE SOIL

     

14 MONTH INCUBATION

1 MONTH INCUBATION

 

IIUYIITfi'lIITIITTerYTTYWTIITTTIIIIIIIVTTTTTITTTf|I1

400 350 300 250 200 150 ppm 100 50 0 ~50 -100

Figure 11. NMR spectra of clover and whole soil. "‘ = spinning sideband
Contact time = 0.4 ms, delay time = l s (clover) or 0.5 s (whole soil).

75

C. FINE CLAY

    
 
    
   
 

14 MONTH INCUBATION

1 MONTH INCUBATION

 

fi—rrrjr vriv‘rrrrrrviylvv1T1T7r11.ivxiuv-yrvvvxuvru1

400 350 300 250 200 150 ppm 100 50 0 -50 -100

D. COARSE CLAY

14 MONTH INCUBATION

1 MONTH INCUBATION

 

l‘rivixxvarrvwrrvrrviiiixinzvv111vv1vvi.xirrriirvi]

400 350 300 250 200 150 ppm 100 50 0 -50 -100

Figure 11 (cont'd). NMR spectra of fine clay and coarse clay. * = spinning sideband
Contact time = 0.4 ms, delay time = 0.9 s (coarse clay, 1 month), 1 s (coarse clay, 14
month), or 0.5 s (fine clay).

76

  
  
     
 
   

E. FINE SILT

14 MONTH INCUBATION *

1 MONTH INCUBATION

 

[T Y I I lTrorTT‘rY I l I l TY V7! 1 T FTT TT 1 I l T I If‘ Tfi [ I I 1 T rr' 1] I

400 350 300 250 200 150 100 50 O -50 -100

Figure 11 (cont'd). NMR spectra of fine silt. * = spinning sideband
Contact time = 0.4 ms, delay time = 0.75 s.

77

because the a-amino groups of all the common amino acids (except proline) have
resonances in or near this region (Table 5 and Appendix 4).

As mentioned above, the NMR peak area percentages for clover and incubated
soils were presented after correction for differential relaxation efi‘ects (Table 19). The
overall corrections for differential relaxation effects were small; the maximum difference
between corrected and uncorrected peak area percentages for each sample corresponded
to no more than 3% of the total signal intensity. The nature and validity of the
corrections is discussed below.

Relaxation during the delay time was characterized by the parameter Tm
(Equation 1 1). The ratios of the standard errors to the estimated values for T; H were
generally low, indicating that the Tm model is correct (Table 20). A notable exception is
the high-frequency amino group of clover. The standard error was high for this relatively
small peak, which represented < 1% of the total spectral intensity.

The T; H values (Table 20) were generally small compared to the delay times used
in the NMR spectra of Figure 11 (t = 1 s for clover, 0.5 to 1 s for soil); thus the
differences between the measured and Tug-corrected signal intensities were small. The
Tm values for soil were in all cases lower than those for clover; this is attributed to the
relatively high concentrations of paramagnetic species in soil (e.g. Fe”) (Pfeffer et al.,
1984)

Differential relaxation during the contact time was also considered, and was
characterized by the parameters TN“ and T1,,“ (Equation 9). The relatively low ratios of
standard errors to parameter values (Tables 21 and 22) and the close correspondence

between measured and modeled values (Figures 12, 13, and 14) indicate that the contact

78

Table 20. Tm values (5) for clover and incubated soils. 1'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

functional group
................ ,----------------.--------------- ---- --------.---------------------------.
aromatic guamglrmifiifieN S Of
N of 3
material histidine amide -NH -NH2 amino
clover I 0.365 0.531 NA § 0.520 0.972 0.829
(0.425) @095) (0.103) (1.556) (0.082)
whole soil 0.187 0.523
ND 1 (0.006) NA (0.168)
fine clay 0.140
ND (0047) NA NA
coarse 0.200
clay ND (0. 05 1) NA NA
fine silt 0.175 0.210
ND (0019) NA (0.056)

 

T Standard error in parentheses. For the incubated soils, model parameters were
generated only for samples that had been incubated for 34 d. The parameters were
assumed to be equally applicable to soils that had been incubated for different times.

1 For clover, two peaks were observed in both the guanidinium and amino regions.

Values on the lefi correspond to the peaks with the higher NMR fiequency.

§ NA = no dependence of signal intensity upon delay time was observed; functional

group assumed to be fiilly relaxed. Minimum delay values tested were as follows: clover

= 1 5, whole soil = fine clay = coarse clay = fine silt = 0.5 s.

1 ND = fimctional group not detected.

79

 

Table 21. TN}; values (ms) as a function of material and functional group. '1'

 

 

 

 

 

 

 

 

 

 

 

 

.................................. fimctlonalamuv
l guanidinium N’s of
imidazole N arginine
material of histidine amide 'NH "NH2 amino
clover I 0.128 0.0588 0.0496 0.0599 0.0605 0.292
(0.053) (0.0089) (0.0181) (0.0056) (0.0583) (0.103)

. 0.0784 0.0888 0.187
Wh°le 5°“ ND § (00076) (0.0093) (0.043)
0.0757 0.0846 0.162

fine day ND (0.0089) (0.0219) (0.100)
come ch ND 0.0623 0.0553 0.130
y (0003 5) (0.0224) (0.060)

. 0.0724 0.0823 0.268
the 5‘" ND (00053) (0.0110) (0.081)

 

 

 

 

1' Standard errors in parentheses. For the incubated soils, model parameters were
generated only for samples that had been incubated for 34 d. The parameters were
assumed to be equally applicable to soils that had been incubated for different times.

t For clover, two peaks were observed in both the guanidinium and amino regions.
Values on the left correspond to the peaks with the higher NMR frequency.

§ ND = functional group not detected

80

Table 22. T1,,“ values (ms) as a function of material and functional group. 1

 

 

 

 

 

 

 

 

 

 

 

 

------_---.f11.9_<2t.i.9!1§!.sr.992 ..................................
imidazole N of 1 guanidinium N’s 1
material histidine amide of arginine amino
-NH -NH2
clover I 3.15 2.14 1.96 1.41 12.6 1.65
Q48) (0.28) (0.58) (0.11) (29.3) (0.59)

whole soil ND § 3.24 1.95 1.95
(0.33) (0.19) (0.46)

fine clay ND 2.15 1.67 0.932
(0.23) (0.40) (0.558)

coarse clay ND 2.29 1.80 2.83
(0.12) (0.60) (1.42)

fine silt ND 3.17 2.19 2.60
(0.24) (0.28) (0.88)

 

 

 

 

1’ Standard errors in parentheses. For the incubated soils, model parameters were
generated only for samples that had been incubated for 34 d. The parameters were
assumed to be equally applicable to soils that had been incubated for different times.

I For clover, two peaks were observed in both the guanidinium and amino regions.
Values on the lefi correspond to the peaks with the higher NMR fi'equency.

§ ND = functional group not detected.

81

0.03
0.07 4 I A. HISTIDINE

0.06 4

0.05 - f I
0.04 4
0.03 4

0.02 1
0.01 l

 

 

 

 

5.0
4,5 4 . B. AMIDE
4.0 4
3.5 4
3.0 4
2.5 4
2.0 4
1.5 4
1.0 4
0.5 4
0.0 . T Y , . r T

 

 

 

 

0.16
0.14 4
0.12 4
0.10 4
0.08 4
0.06 4
0.04 4
0.02 4
0.00 . , . T . , . T . . . Y .

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50

CONTACTTIME(ms)

 

 

SIGNAL INTENSITY (ARBITRARY UNITS)

 

 

 

 

 

 

 

Figure 12. Measured and modeled signal intensities for clover as a fiinction of contact
time. Error bars represent uncertainty due to spectral noise and subjectivity in
phasing/time. Error bars represent uncertainty due to spectral noise and subjectivity in
phasing/baseline correction as described in Appendix 10.

82

0.30

 

0.25 J
0.20 4
0.15 4'
0.10 4

0.05 4

 

 

 

0.00...ff.w..

 

0.07
0.06 4
0.05 4 I

 

E. AMINO (LEFT)

0.04 4

I
0.03-fl! i J

0.02 4

0.01 4f

0.00 . 1 1 r 4 . Y . v T 4 1
0.16
0.14 4
0.12 4
0.10 4
0.08 4
0.06 4
0.04 4
0.02 4
0.00 4 . . T s

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
CONTACT TIME (ms)

..

 

 

   
 
 

F. AMINO (RIGHT)

O MEAS
—MODEL

SIGNAL INTENSITY (ARBITRARY UNITS)

 

 

 

  

 

 

 

 

I T I I I T

Figure 12. (cont‘d) Measured and modeled signal intensities for clover as a fimction
of contact time. Error bars represent uncertainty due to spectral noise and
subjectivity in phasing/time. Error bars represent uncertainty due to spectral noise
and subjectivity in phasing/baseline correction as described in Appendix 10.

83

0.35
0.30 4 t
0.25 4
0.20 4

0.15 \
0.10

0.05
0.00 4 4 4 4 4 4 4

 

A. AMIDE

 

 

 

0.035
0.030 4
0.025 4

0.020 4 \

0.015 4 \
0.010 4 \{
0.005 4

0.0m jfi T T T I T

 

B. GUANIDINIUM

 

 

 

0.016
0.014 4
0012 I ”"\\i\ C. AMINO
0.010 4

0.008 4 \i

0.006 ‘ \

0.004 r \I

0.002
0.(X)O r T T 1' V 1 T
0.0 0.5 1.0 1.5 2.0 2.5 3.0 O MEAS

CONTACT TIME (ms) —— MODEL

 

SIGNAL INTENSITY (ARBITRARY UNITS)

 

 

 

 

 

 

 

Figure 13. Measured and modeled signal intensities for whole soil as a fimction
of contact time. Error bars represent uncertainty due to spectral noise and
subjectivity in phasing/baseline correction as described in Appendix 10.

84

0.60

 

0-50 ‘ ‘ A AMIDE

0.40 4 \
0.30 4 \
0.20 4 \

0.10 4

 

 

0.“) I I ' I

 

0.060

 

0050 ) B. GUANIDINIUM

0.040 4 ‘\;
0.030 4 \-

0.020 4. \I\
0.010 4 \'

0.000 4 . 4 a 4 .

 

 

 

0.040
0.035 4
0.030 - I C. AMINO
0.025 4 \

0.020 4 \

0.015 4 \-

0.010 4 N
I
0.005 1‘

0.“ I I I I I I I

 

SIGNAL INTENSITY (ARBITRARY UNITS)

 

 

 

 

0.0 0.5 1.0 1.5 2.0 2.5 3.0 . MEAS
CONTACT TIME (ms) —— MODEL

 

 

 

Figure 14. Measured and modeled signal intensities for fine clay as a function of
contact time. Error bars represent uncertainty due to spectral noise and
subjectivity in phasing/baseline correction as described in Appendix 10.

85

time model (Equation 9) was appropriate. The poorer fits for the amino group relative to
other functional groups are likely a result of the relatively small peak area of amino as
well as peak overlap (Figure 11). The correction factors for differential relaxation during
the contact time were computed fi'om TN}; and T1,,“ (Equation 10, Table 23). Within each
sample, the ratio of the maximum to the minimum correction factor among the various
functional groups varied from about 1 to 1.5. (A high maximumzminimum ratio indicates
that differential relaxation effects may be significant.) The highest correction factor
usually corresponded to the amino group, probably as a result of high TN}; values for
amino (Table 21). High TN}; values are expected in relatively mobile functional groups
such as amino (Kinchesh et al., 1995).

Correlation of the NMR spectral intensities with the amounts of ”N in the
incubated soil samples constitutes additional evidence of the quantitativeness of CPMAS-
NMR. Linear regressions were run with 15N amount and total NMR signal intensity per
scan as dependent and independent variables, respectively. Regressions were run
separately for each soil fraction with spectra that had been collected in a single NMR
session; thereby confounding effects due to soil fraction (e. g. differential Fe contents) and
variation in spectrometer performance were avoided. The correlations for whole soil,
fine clay, coarse clay, and fine silt were all significant (p s 0.05, Figure 15). The lowest
R2 value was obtained for fine clay, and a plot of detectability (NMR signal intensity per
scan per mol ”N) versus incubation time revealed an apparent decrease in the
detectability of 1’N with increasing incubation time (Figure 16). One might argue that
the decrease in detectability suggests that a form of undetectable N, such as heterocyclic

N in slow-relaxing rigid aromatic structures (or perhaps clay-fixed NI-In), may have

86

50:038.: M22 :23: 05 5:3
830: 05 o: 0:800:00 a2 05 :o 0020:» 0:23: 05:8 :5 E:_:_E:§w 0:: £3 :_ 32030 0:03 8.00: :3: .:0>o_0 :0": a

00:5 €20.30 :0: 00:33:: :00; :0: :05 0:8 9 03023:?
33:60 on 8 0083.00 0:03 0:30:33: 2:. .v E” 5.: 3:03:05 :03 :0: :2: 00383. :0: 3:: :0:0:0:0w 0:03 9:80:88:

0:08 .25.: “00:33:: 0:: :0": A3: v 5:: \ 08:: 250% 3:285: =0 :8 .0200 boifimxoaau 0:03 0:98.: 5:80:80
.: 2.0:: 003 5:00:00 02 .m :28:ch :0: 05: 83:00 2: to 5:? 00:38 880% c: “00:30 0:03 0:800: 5:00:80 +

2:: :03? :m 3203: 003 2 03.32:: H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

hm: Ev: 02.: :2: . Em 05.:
3.: 6: SN: 52 - @0838
mm: 09; CNN: 00—; - NAm6 Em
2: won: mg: NZ: . :0m 20:3
a: Sm: one: EN; 62: E...— 03: m:0>o_o
n t 1
8:00.: 5:00:50 5:: 2:5 m2 $2 0280 N 0:652 3:038
a 055w; :0 Z 3802::
:o m. Z EEEESSm
....................... 95% 3:285:

 

 

+ 0:5 88:00 0:: mEB: 5588.0: BEENE: o: 0:: 30:0 0.00: 122 5.: 0:83.: 8:00:80 .MN 030:.

87

 

60 .4
50 at WHOLE SOIL
40 R2 = 0.995 .
30

 

 

 

 

umol ”N in NMR SAMPLE
:— N
O C O

 

 

.1 L4

 

  
   

COARSE CLAY
R2 = 1.000 **

 

 

 

 

 

 

 

0.6

NMR SIGNAL INTENSITY PER SCAN ° MEAS
(ARBITRARY UNITS) “FRED

 

0.8 3.0

FINE SILT
R2 = 0.999 **

 

 

r

0.2 0.4 0.6 0.8

 

 

 

 

Figure 15. Linear regressions of 15N amounts versus NMR signal intensity for soil
samples. Each point represents a soil sample collected after 1, 6, or 14 months of
incubation. The point (0,0) is not a measured point, but was included in the input data for
the regressions since no signal would be expected for a blank sample. Error bars

represent sample standard deviations from mass spectrometric measurement of 15N. "' and
** indicate significance at the 0.05 and 0.01 probability levels, respectively.

88

 

 

 

“Z

E 0.016
ad

E 0.014 4
>‘

b 4
(I)

E 0012 «
13 0.010 ~
2

53 l
(I)

g 0.008

100

200
INCUBATION TIME (d)

V

300

400

 

500

Figure 16. NMR detectability of fine—clay associated lSN after various incubation

times.

89

slowly formed in the fine clay. If it is assumed that the lsN is 100% detectable at Day 34,
Figure 16 indicates that about 20% ofthe fine—clay l’N may have been undetectable at
Day 439; this undetectable 15N corresponds to about 4% of the remaining 15N in the
whole soil.

The above analysis of NMR data suggests that very little, if any, heterocyclic '5 N
was formed during the soil incubation. However, the broad high-frequency (left-hand)
shoulder of the amide peak observed here (Figure 11) and in other studies (e. g. Knicker et
al., 1993; Knicker and Ludemann, 1995) may be a result of heterocyclic N. Comparison
of the spectra of 15N-clover in the presence and absence of soil (Figure 17) indicates that
the shoulder is mainly a consequence of soil-induced line broadening. The shoulder does
obscure a small signal due to histidine (near 150 ppm), but this signal represents only
about 1% of the clover-”N (Table 19). The NMR peaks in these soil samples were too
broad to identify fungal glucosamine or bacterial purine as per Bedrock et al. (1998, see

Introduction).

Relationship between N Mineralization and Organic N Functional Groups During
the l4~Month Incubation

Throughout the 14-month incubation of soil amended with l’N-clover, 1SN-NMR
demonstrated that the composition of soil organic 15N in whole soils and size fractions
invariably was about 90% amide, 540% guanidinium N of arginine, and 5% amino. No
separate NMR spectra were obtained for coarse silt and sand, but it is likely that their

relatively short-lived 15N pools resembled that of the initial clover.

9O

”bl-CLOVER

 

 

 

 

1’N- CLOVER + UNLABELED SOIL

 

 

 

[IIrrrrIITITTIrTIIfTrTVTrrtrrrrTrfirrrrTTIIITITrfirrl

400 350 300 250 200 150 ppm 100 50 O -50 -100

Figure 17. NMR spectra of ISN-labeled clover, with and without soil. In both
Spectra, contact time = 0.2 ms, delay = 1 s, line broadening = 120 Hz.

91

Despite the lack of change in organic l5N filnctional groups, important differences
in mineralization behavior were observed among size fractions. Kinetic analysis of
clover-derived N (i.e. l’N) indicated the presence of sorbed N in clays and fine silt (k e
0.001 d'l), processed macroorganic matter in fine silt (k z 0.03 d'l), and relatively
unprocessed macroorganic matter in coarse silt and sand (k z 0.1 to 0.2 d"). When
particle-size fractions from the l4-month incubation were incubated separately in shaken
slurries, differences in N mineralization were observed and linked to C:N ratio, size class,
and the time of incubation prior to fractionation. However, the shaken slurry results were
measured for total N (”N + 1‘N) and cannot be assumed to be completely applicable to
15N. Nevertheless, the overall results indicate that for the entire l4-month incubation the
clover-derived N was almost entirely proteinaceous, and that differences in N
mineralization behavior were not related to the composition of organic N. Rather, clover-
N mineralization in the relatively N-rich soil of this study appeared to be driven by the
microbial requirement for C. The stability of clover-N was linked to humification or
adsorption per se, and perhaps aggregate protection, as will be presented in the

Discussion section.

92

DISCUSSION

Dynamics of Clover-N in Particle-Size Fractions: Agronomic Implications

The dynamics of clover—derived N turnover are best discussed in relation to
microbial degradation. A vigorous microbial attack ensued immediately after the
amendment of soil with ”N-clover. After 1 and 5 d of incubation, the respiration rate
corresponded to 5 and 3% of total soil C, respectively. Proteinaceous 1’N (as revealed by
l’N-NMR) was released from the original plant material and probably from the fast-
cycling microbial biomass, and accumulated in the clay fractions (and perhaps in the fine
silt) until about 95 d. (Evidently, ultrasound at 38 kJ per 28 g soil does not rupture plant
cells; ultrasonic rupture of plant cells would have resulted in greater recovery of 15N in
size fractions afier 11 d.) The slopes of the curves for clover-N loss indicate that most of
the labile clover constituents had been transformed after 34 d. (See Figure 6, especially
6F and 6G.) Microbe-on microbe attack was likely a prominent process between 34 and
95 d; the maximum clover-N accumulation in the clays occurred at about 95 d. This
interpretation of microbial dynamics is supported by McGill et al. (1975). These
researchers added l“C-acetate and l’N-ammonium sulfate to soil and observed maximal
fungal and bacterial populations at 5 and 10 d of incubation, respectively. They
considered that the fungal decline was due to either autolysis or bacterial attack.

After 95 d, the clover-N associated with the fine soil fractions exhibited a slow
degradation. The stabilization of clover-N in the fine fractions is emphasized by plotting

clover-N amounts for all fractions as percent of remaining (Figure 18), and is supported

93

 

 

k

 

+ TOTAL CLAY

 

A +COARSE CLAY
-)(-FINE SILT
+FINE CLAY
-e- SAND

 

 

 

 

 

 

—l
—X -+- COARSE SILT
o

 

 

INCUBATION TIME (d)

Figure 18. Clover-derived N in fi'actions as a function of time (% of remaining.)

94

by the mechanistic model of Hassink and Whitmore (1997). Their model characterizes
clays as saturable adsorptive surfaces, which quickly adsorb and slowly desorb OM.

The rapid transformation of clover-derived, and perhaps native, light-fiaction
material in the coarse silt and sand supports the correlation between 70- and lSO—d
nitrogen mineralization potential (NMP) and macroorganic N and C found by Willson et
al (in press). (These authors defined macroorganic N and C as the N and C in the 53 —
2000 um fraction; this fiaction was obtained by sieving a dispersed soil suspension and is
thus similar to the macroorganic fraction of the present study.) Although the correlations
of Willson et al. (in press) were statistically significant, no more than 41% of the
variability in NMP was explained by macroorganic N or C. They noted that
macroorganic matter did not increase immediately after plant residue incorporation; this
was perhaps because the residue particle size was > 2000 um. For a better prediction of
NMP, they suggested that macroorganic matter measurements could be combined with
information about recently incorporated plant residues.

Fresh plant residues may contain appreciable amounts of soluble organic N (e.g.
Aita et al., 1997). This N pool is almost certainly a significant short-term N source and
was not determined in the present study. The pool was probably lost (mineralized)
during particle-size separation (Table 8). In future studies, this pool might be preserved
by performing all fractionation steps in a chilled environment and sampling for dissolved
organic compounds.

The agronomic relevance of the apparently long-term stabilization of clover-N
observed here can be assessed by comparing the N retention in this study to that observed

in the field by others (Table 24). For these comparisons, the rate constants for clover-N

95

decomposition were adjusted to the field based on a Qto value of 2. (See Katterer et al.
(1998) for details.) No attempt was made to account for variations in moisture that would

be expected in the field, as moisture was not correlated with C02 fluxes in a field study in

southwest Michigan (Paul et al., 1999). The overall plant decomposition in the present

study closely matches that of Aita et al. (1997), but is less than that observed by Harris

and Hesterman (1990). The higher decomposition (lower retention) observed by Harris

and Hesterman (1990) is likely a result of the presence of growing plants (Cortez and

Cherqui, 1991). Roots are known to be a major source ofsoil C and lead to N

immobilization followed by mineralization (Jansson and Persson, 1982). This analysis

suggests that some but not all field mechanisms of N mineralization and immobilization

were operating in the present study.

Table 24. 15N retention in present study compared to field studies.

 

 

 

 

 

 

 

 

 

 

 

 

soil organic 15N
as % of initial
(present study,
soil organic corrected to
15N as % of environmental
agronomic soil initial (per conditions of
reference system texture conditions reference) the reference)
Aita et al. "N-wbcat straw silt 365 d, 10 °C, 62 60
(1997) decomposition loam 45 kg straw-
in absence of N per ha
growing plants
Harris and T’N-alfalfa loam 332 d, 10 °C, 38 61
Hesterman decomposition 112 kg
(1990) in presence of alfalfa-N
corn g ha
Harris and l’N-alfalfa sandy 332 d, 10 °C, 40 61
Hesterman decomposition loam 112 kg
(1990) in presence of alfalfa-N
corn per ha

 

96

 

The stabilization of considerable quantities of plant-derived N observed by Aita et
al. (1997) and Harris and Hesterman (1990) is likely due in large part to association of N
with clays via mechanisms discussed in the next section. Calculations of the degree of
saturation of the soil protective capacity per the model of Hassink and Whitmore (1997)
indicate that the soils of Aita et al. (1997), and Harris and Hesterman (1990) were
undersaturated with organic matter both before and after addition of plant residues (Table
25). The soil of the present study was undersaturated before addition of clover and at

saturation immediately after addition of clover.

Table 25. Comparison of soil C with soil protective capacity.

 

 

 

 

 

 

 

 

soil C
soil C prior immediately calculated
to addition afier addition soil
soil texture of plant of plant protective
(USDA percent clay residue residue capacity
reference , system) (< 2 pm) 1 (g C kg") 1 (g C kg") ; (g C kg‘l) §
$93.3 al. silt loam 15 9.9 11.3 27
Harris and
Hesterman loam 25-50 14 14.3 30 - 40
(1990)
Harris and
Hesterman sandy loam 0-20 8 8.3 21 - 29
(1990)
$133“ sandy loam 16 14 27 27

 

 

 

 

 

‘1' Exact values of percent clay were not given by Harris and Hesterman (1990); the range
of percent clay was deduced from their reported soil textural class.

I Exact values of native and added C were not given by Harris and Hesterman (1990) but
were estimated from OM content of soil and from C:N ratio of plant residue.

§ Soil protective capacity per Hassink and Whitmore (1997):
protective capacity (g C kg!) = 21.1 + 0.375 x (% particles < 2 pm)

97

The retention of N in cover crop systems or other systems with high organic
inputs is important in the long-term stabilization or buildup of SOM levels. High levels
of SOM may eventually lead to increases in the native N mineralized during a growing

season.

Mechanisms of Short- to Medium-Term Stabilization of Soil Organic N

Aggregate protection of organic N in this study was assessed by ( 1) the extent of
mineralization in shaken slurry, (2) the ratio of shaken slurry to undisturbed
mineralization (aggregate protection factor, or APF), and (3) the difference between
shaken slurry and undisturbed mineralization (aggregate-protected pool size). The
aggregate protection factors (APF) measured for whole soil (APF z 4, Table 15)
correspond roughly to those measured by other authors for soils with much higher clay
contents than the sandy loam soil (15% clay) of this study. Edwards and Bremner (1967)
and Craswell and Waring (1972) assessed aggregate protection by comparing two-week
N mineralization (30 °C, anaerobic or aerobic) of coarse-ground (2000 um) and fine-
ground (50 um) soils. In nine soils with clay contents ranging fi'om 26 — 72%, the APFs
ranged fi'om 1.3 to 4.3, with the exception of one observation (APF = 25). (APFs were
calculated from their data as the ratio of fine-ground to coarse-ground N mineralization.)
For soils with clay contents ranging from 10-17%, no enhancement of N mineralization
was observed (Craswell and Waring, 1972). The apparent difference between the results
of Craswell and Waring (1972) and those of the present study is explicable on the basis of
methodology. All soils studied by Craswell and Waring (1972) were air-dried prior to

grinding, and the dried clayey soils undoubtedly required considerable grinding energy

98

before passing through the 50 um sieve. The dried sandy soils, on the other hand, were
probably loosely aggregated; the small aggregates containing the protected OM probably
passed through the sieve with minimal grinding energy. In contrast, the sandy loam soil
of the present study was probably dispersed to a considerable degree during shaking as
explained in Results.

Aggregate-protected N accounted for no more than 11% of N (clover + native)
stabilization in this study. Therefore, the most prominent N stabilization mechanisms
were probably humification or clay adsorption per se. APF values and aggregate-
protected pool sizes decreased with time for fine clay and coarse clay; the explanation is
that the strength of adsorption increased with time (e.g. increased number of covalent
bonds between organic N compounds and minerals) and/or that the organic N compounds
were being humified.

Due to lack of 15N data in the shaken slurry mineralization tests, the aggregate
protection of clover-derived N cannot be directly evaluated. Nonetheless, the maximum
possible mineralization of clover-N was calculable, and the results indicated that at most
33% of the clover-N was mineralizable in shaken slurry (Table 26). Therefore,
humification or adsorption per se, and perhaps aggregate protection, were important
mechanisms for the stabilization of clover-N. The distribution of organic l5N during this
study was invariably 90% amide, 5-10% guanidinium N of arginine, and 5% amino;
incorporation of N into heterocycles was not observed. Overall, these findings support
the notion that proteinaceous N is incorporated into soil humus as reviewed in the
Introduction. In the short term, the NMR data of this study show that this incorporation

is probably not accompanied by change in the overall distribution of N functional groups.

99

Table 26. Maximum possible mineralization of clover-N in shaken slurry.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

maximum
possible
mineralization
days of prior of clover-N (%)
incubation fraction 1'
whole soil 20
fine clay 33
34
coarse clay 10
fine silt 2.4
whole soil 17
190 fine clay 19
coarse clay 9.3
fine silt 10
whole soil 25
fine clay 11
439 coarse clay 6.4
fine silt 6.9

 

 

 

 

1' (umol mineralized in shaken slurry / umol organic clover-N present at beginning of
shaken slurry) x 100

Chemical Structure of Soil Organic N as Determined by NMR: A Critical Analysis
The overall quantitativeness of NMR was demonstrated by characterization of
differential relaxation effects, analysis of a complex soil-organic mixture with known
composition, and correlation of NMR spectral intensities with 15N concentrations as
measured on a mass spectrometer. The relatively small differential relaxation effects
observed here should not be generalized; samples that contain appreciable amounts of
both high-mobility N (e. g. amino) and limited-mobility N (e. g. amide) would be expected
to exhibit relatively large differential relaxation effects. The analysis of the mixture of
known composition showed that peak overlap is an important though not fatal weakness

of the NMR analysis of soil samples.

100

Although l’N—CPMAS NMR has been used to distinguish fungal versus bacterial
biomass in pure culture and in decaying organic materials (Bedrock et al., 1998), NMR
peaks in the presence of soil are too broad to make such distinctions. Moreover, a peak
coinciding with the fungal N-acetyl glucosamine peak reported by Bedrock et al. (1998)
was observed in pure clover (Figure 11A).

Improvements in N-NMR spectroscopy of soils seem possible. Density
fiactionation leads to isolation of fractions low in specific gravity and relatively high in
organic content; such schemes were used in a 13C--NMR study by Baldock et al. (1992).
Christensen (1992) has questioned some density fractionation schemes, and considers
that density fractions of clay and silt particles may not represent truly distinct pools of
organic matter. If this is true, one might argue that the NMR spectrum of the “light”
fraction within clay and silt would be similar to that of the “heavy” fraction within said
particles. A firrther suggestion for improvement of NMR spectra is the dissolution of
mineral matter by HF (Schmidt et al., 1997). This approach would seem to forfeit the
advantage of the noninvasiveness of NMR Finally, the recently developed teclmique of
high-resolution solid-state l"’N-NMR may help resolve the current controversy
surrounding the chemical structure of native soil organic N (leschke and Jansen, 1998).
However, this approach requires high spinning speeds and firrther advances in broadband
spectral excitation, and may not be immediately applicable to soil systems.

The composition of 15N in whole soils and particle-size fractions in this study
(90% amide, 5-10% guanidinium N of arginine, and 5% amino) is in good agreement
with other ”N-CPMAS—NMR studies of soil but at variance with the composition of soil

N as presented by Schulten and Schnitzer (1998). The composition of soil N per

101

Schulten and Schnitzer is: proteins + peptides + amino acids, 40 %; amino sugars, 5 - 6
%; heterocyclic N, 35 %; and NH3, 19 %. (NH; is an operationally defined pool; see
Table l.) Schulten and Schnitzer (1998) argue that the incubation times of l’N-amended
soils in studies such as the present one may be too short for the formation (or selective
preservation) of resistant heterocyclic N. Although this caveat may certainly apply to this
study, the discrepancy in N chemical composition may stem fiom assumptions used by
Schulten and Schnitzer. Their analysis appears to rely on the assumption that no
proteinaceous N is contained in the N fraction that is resistant to hot acid hydrolysis
(Sowden et al., 1977). This assumption may be reasonable in some studies, as hot acid
hydrolysis is a traditional method for decomposing proteins into their constituent amino
acids (Creighton, 1993). However, 15N-CPMAS-NMR of the nonhydrolyzable residue of
a soil incubated for two years with 1’N-ammonium sulfate + crop residues + powdered
filter paper revealed that about 80% of the N in the residue was in the form of amide and
amino groups; about 20% could be assigned to either amide or heterocyclic N of indole
(Zhuo et al., 1992). Similarly, the ”N-NMR analysis (at natural abundance levels) ofthe
hydrolysis residue from an organic-rich lake sediment produced an amide—dominated
spectrum quite similar to the spectra acquired in the present study (Knicker and Hatcher,
1997). Further analysis of the residue (thermochemolysis followed by gas
chromatography/mass spectrometry) identified the N as proteinaceous (Knicker and
Hatcher, 1997). Based on the unchanging proteinaceous NMR signature of clover-
derived N throughout the 14-month soil incubation and the shaken slurry mineralization

tests, the considerable differences in the mineralization/immobilization behavior of

102

clover-N among the difierent particle-Size fractions appear to be linked to adsorption per

se or humification and possibly aggregate protection.

103

CONCLUSIONS

This study confirmed that soil particle-Size fractions contain biologically distinct
pools of organic nitrogen, and strongly suggested that fundamental differences in the
short- to medium-term behavior of legume-derived N among these fractions are not a
result of differences in the chemical structure of N. A transient macroorganic legume-
derived N pool was found in coarse silt and sand, with a decay rate of about 0.1 d'1 in the
laboratory. This rate corresponds to a steady-state mean residence time (MRT) of 10 d in
the laboratory or 25 d in the field (10 °C). A stable pool of legume-derived N was found
in the clays and fine silt and corresponded to about 60% of added legume-N; its decay
rate was 0.001 d’l, corresponding to an MRT of 3 years in the laboratory or 7 years in the
field (10 °C). The large Size of this stable N pool despite a high rate of legume addition
was explained in terms of the soil protective capacity of Hassink and Whitmore (1997);
the amount of native plus added C was roughly equal to the protective capacity.

Evidence fiom l’N—CPMAS-NMR suggested that both the transient and stable pools of
legume-derived N were almost entirely proteinaceous. The aggregate-protected N pool
(legumederived plus native N) was operationally defined as the 30-day shaken slurry N
mineralization minus the 30—day mineralization in undisturbed soil. This pool comprised
between 0 and 11% of the N within the size-fractions and whole soil, and increased (as a
percentage of N in the fiaction) with decreasing particle Size. These findings support the
notion that some organic N is resistant by virtue of its entrapment in pores inaccessible to

microbes (Edwards and Bremner, 1967). The generally low amounts of N mineralized in

104

shaken slurries (maximum of 12% of total N, 33% of legume-N) indicate that substantial
amounts of N are stabilized via adsorption per se or humification.

Key weaknesses of 15N—CPMAS-NMR are peak overlap and its impracticality in
establishing the chemical structure of native (unlabeled) soil organic N. Nevertheless,
lsN-CPMAS-NMR was a useful tool for the noninvasive Study of labeled soil organic N.
The technique was able to detect both heterocyclic and noncyclic N forms in a prepared
mixture of unlabeled soil and l’N-labeled organic materials, and was quantitative for
incubated soils to within 5-10% of the total spectral intensity.

Overall, the mechanistic and kinetic information of the present study are
important to both the fundamental and practical understanding of soil N availability.
Mechanistically, the short- to medium-term mineralization/unmobilization of legume-
derived N was controlled by the physical location of this N within the soil matrix rather
than by the organic filnctional group composition of this N. The kinetic rate constants
determined in the laboratory, after correction to field temperatures, reflected some but not
all of the mechanisms operating in the field. These kinetic parameters along with other
key information, such as percent clay, degree of saturation of soil protective capacity,
magnitude and timing of crop N needs, and mean annual air temperature, will allow us to
better predict N mineralization/immobilization in the field. Such prediction is important
in designing agricultural systems in which organic N is mineralized in synchrony with

plant needs.

105

REFERENCES

106

REFERENCES

Aita, C., S. Recous, and D. A. Angers. 1997. Shorts-term kinetics of residual wheat straw C
and N under field conditions: characterization by 3C 5N tracing and soil particle Size
fi'actionation. European Journal of Soil Science 48: 283- 294.

Anderson, D.W., and EA. Paul. 1984. Organo-mineral complexes and their study by
radiocarbon dating. Soil Sci. Soc. Am. J. 48:298-301.

Bader, M. 1980. A systematic approach to standard addition methods in instrumental
analysis. J. Chem. Educ. 57:703-706.

Balabane, M., and J. Balesdent. 1995. Medium-terrn transformations of organic N in a
cultivated soil. European Journal of Soil Science 46:497-505.

Baldock, J A. J. M. Oades, A.G. Waters, X Peng, A M. Vassallo, and M A. Wilson.

113‘992 Aspects of the chemical structure of soil organic materials as revealed by solid-state
13C NMR spectroscopy. Biogeochemistry 16:1-42.

Beare, M.H., M.L. Cabrera, P.F. Hendrix, and DC. Coleman. 1994. Aggregate-protected
and unprotected organic matter pools in conventional- and no-tillage soils. Soil Sci. Soc.
Am. J. 58:787-795.

Bedrock, C. N ,.M V. Cheshire, B. L. Williams, I. Solntseva, S. J. Chapman, J. A. Chudek,
and B. A. Goodman. 1998. Identification of nitrogenous components of fimgal and
bacterial origin immobilized In decomposing wheat straw by NMR spectroscopy using 5N
CPMAS. Soil Biol. Biochem. 30: 113115

Bevington, PR, and DR. Robinson. 1992. Data reduction and error analysis for the
physical sciences. 2nd ed. McGraw-Hill, New York.

Biederbeck, V.O., and EA. Paul. 1973. Fractionation of soil humate with phenolic
solvents and purification of the nitrogen-rich portion with polyvinylpyrrolidone. Soil Sci.
115:357-366.

Bliss, C.I. 1967. Statistics in biology. Vol. 1. McGraw-Hill, New York.

Blomberg, F., and H. Ruterjans. 1983. Nltrogen-IS NMR in biological systems. p. 21-73.
In L]. Berliner, and J. Reuben (ed) Biological magnetic resonance. Vol. 5. Plenum
Press, New York.

Bowman, G., C. Shirley, and C. Crarner. 1998. Managing cover crops profitably. 2nd ed.
Sustainable Agricultural Network, Handbook Series Book 3, Beltsville,MD.

107

Briggs, J ,M., and E.W. Randall. 1973. 15N chemical shifis in concentrated aqueous
solutions of ammonium salts. Mol. Phys. 26:699-703.

Catroux, G., and M. Schnitzer. 1987. Chemical, spectroscopic, and biological
characteristics of the organic matter in particle Size fractions separated from an Aquoll.
Soil Sci. Soc. Am. J. 51:1200-1207.

Christensen, HT. 1992. Physical fi'actionation of soil and organic matter in primary
particle size and density separates. p. 1-90. In B.A. Stewart (ed) Advances in Soil
Science. Vol. 20. Springer-Verlag, New York.

Christensen, RT 1996. Matching measurable soil organic matter fi'actions with
conceptual pools in simulation models of carbon turnover: Revision of model structure. p.
143-159. In D.S. Powlson, P. Smith, and J .U. Smith (eds) Evaluation of soil organic
matter models using existing long-term datasets. Vol. NATO ASI Series I 38. Springer-
Verlag, Berlin.

Clinton, P.W., RH. Newman, and RB. Allen. 1995. Immobilization of lsN in forest litter
studied by 15N CPMAS NMR Spectroscopy. European Journal of Soil Science 46:551-
556.

Collins, HP, ET. Elliott, K. Paustian, L.G. Bundy, W.A. Dick, D.R. Huggins, A.J.M.
Smucker, and EA. Paul. in press. Carbon pools and fluxes in long-term corn belt
agroecosystems. Soil Biol. Biochem.

Cortez, 1., and A. Cherqui. 1991. Plant growth and the mineralization of adsorbed ”C-
and lsN-labelled compounds. Soil Biol. Biochem. 23:261-267.

Cote, RI, and KL. Ghema. 1994. Nutrition and media. p. 155-178. In P. Gerhardt,

R. GE. Murray, W.A. Wood, and NR. Krieg (eds) Methods for general and molecular
bacteriology. American Society for Microbiology, WashingtomDC.

Craswell, ET, and SA. Waring. 1972. Efl‘ect of grinding on the decomposition of soil
organic matter: I. The mineralization of organic nitrogen in relation to soil type. Soil Biol.
Biochem. 42427-433.

Creighton, TE. 1993. Proteins: structures and molecular properties. 2nd ed. W.H.
Freeman and Company, New York.

Cross, T.A., J .A. DiVerdi, and SJ. Opella. 1982. Strategy for nitrogen NMR of
biopolymers. J. Am. Chem. Soc. 104:1759-1761.

Edwards, AP, and J .M. Bremner. 1967. Microaggregates in soils. J. Soil Sci. 18:64-73.

Elonen, P. 1971. Particle-size analysis of soil. Acta Agralia Fennica 122.

108

Fisk, J .W. (1997). Medics and clovers in wheat-no-tillage corn rotations with insights into
their use by farmers. Ph.D. Dissertation, Michigan State University, East Lansing,
Michigan.

Flaig, W., H. Beutelspacher, and E. Rietz. 197 5. Chemical composition and physical
properties of humic substances. p. 1-211. In J.E. Giesiking (ed.) Soil components: Vol. 1,
Organic Components. Springer-Verlag, New York.

Gee, G.W., and J.W. Bauder. 1986. Particle-size analysis. p. 383-411. In A Klute (ed.)
Methods of Soil Analysis, Part 1. Physical and Mineralogical Methods. Vol. . 2nd ed.
Agronomy Monograph. 9. American Society of Agronomy - Soil Science Society of
America, Madison, WI.

Harris, G.H., and 0B. Hesterman. 1990. Quantifying the nitrogen contribution fiom
alfalfa to soil and two succeeding crops using nitrogen-15. Agron. J. 82: 129-134.

Harris, G.H., O.B. Hesterman, EA. Paul, SE. Peters, and RR. Janke. 1994. Fate of
legume and fertilizer nitrogen-15 in a long-term cropping systems experiment. Agron. J.
86:910-915.

Hassink, J. 1996. Preservation of plant residues in soils difl‘ering in unsaturated protective
capacity. Soil Sci. Soc. Am. J. 60:487-491.

Hassink, J. 1997. The capacity of soils to preserve organic C and N by their association
with clay and silt particles. Plant Soil 191:77-87.

Hassink, J ., and AP. Whitmore. 1997. A model of the physical protection of organic
matter in soils. Soil Sci. Soc. Am. J. 61:131-139.

Haworth, RD. 1971. The chemical nature of humic acid. Soil Sci. 111:71-79.

Hopkins, D.W., J .A. Chudek, E.A. Webster, and D. Barraclough. 1997. Following the
decomposition of ryegrass labelled with 13C and 15N in soil by solid-state nuclear magnetic
resonance spectroscopy. European Journal of Soil Science 48:623-631.

Jansson, S.L., and J. Persson. 1982. Mineralization and immobilization of soil nitrogen. p.
229-252. In F .J . Stevenson (ed) Nitrogen in Agricultural Soils. American Society of
Agronomy - Crop Science Society of America - Soil Science Society of America,
Madison, WI.

Jeschke, G., and M. Jansen. 1998. High-resolution 14N solid-state NMR spectroscopy.
Angew. Chem. Int. Ed. Eng]. 37: 1282-1283.

Katterer, T., M. Reichstein, O. Andre’n, and A. Lomander. 1998. Temperature dependence

109

of organic matter decomposition: a critical review using literature data analyzed with
different models. Biol. Fertil. Soils 27:258-262.

Keeney, DR, and D.W. Nelson. 1982. Nitrogen - inorganic forms. p. 643-698. In AL.
Page, RH. Miller, and DR. Keeney (eds) Methods of Soil Analysis, Part 2. Chemical and
Microbiological Properties. American Society of Agronomy - Soil Science Society of
America, Madison, WI.

Kinchesh, P., D.S. Powlson, and E.W. Randall. 1995. 13C NMR studies of organic matter
in whole soils: 1. Quantitation possibilities. European Journal of Soil Science 46: 125-138.

Knicker, H., R. Fri'md, and H.-D. Li'ldemann. 1993. The chemical nature of nitrogen in
native soil organic matter. Die Naturwissenschaften 80:219-221.

Knicker, H., and PG. Hatcher. 1997. Survival of protein in an organic-rich sediment:
possible protection by encapsulation in organic matter. Die Naturwisserrschaften 84:231-
234.

Knicker, H., and H.-D. Ludemann. 1995. N-15 and C-13 CPMAS and solution NMR
studies of N-l 5 enriched plant material during 600 days of microbial degradation. Org.
Geochem. 23:329-341.

Knicker, H., F. Rudiger, and H.-D. Ludemann. 1997. Characterization of nitrogen in plant
composts and native humic material by natural-abundance ISN CPMAS and solution
NMR spectra. p. 272-294. In M.A. Nanny, R. A. Minear, and J.A. Leenheer (eds) Nuclear
magnetic resonance spectroscopy in environmental chemistry. Oxford University Press,
New York, NY.

Kuester, J .L., and J .H. Mize. 197 3. Optimization techniques with FORTRAN. McGraw-
Hill, New York.

Ladd, J .N., J.W. Parsons, and M. Amato. 1977a. Studies of nitrogen immobilization and
mineralization in calcareous soils - 1. Distribution of immobilized nitrogen amongst soil
fiactions of different particle size and density. Soil Biol. Biochem. 9:309-318.

Ladd, J .N., J .W. Parsons, and M. Amato. 1977b. Studies of nitrogen immobilization and
mineralization in calcareous soils - II. Mineralization of immobilized nitrogen from soil
fractions of different particle size and density. Soil Biol. Biochem. 92319-325.

Levy, GO, and KL. Lichter. 1979. Nitrogen-15 Nuclear Magnetic Resonance
Spectroscopy. Vol. . ed. John Wlley & Sons, New York.

McGill, W.B. (1971): Turnover of microbial metabolites during nitrogen mineralization

and immobilization in soil. Ph.D. Thesis, University of Saskatchewan, Saskatoon,
Saskatchewan.

110

McGill, W.B., J .A. Shields, and EA. Paul. 1975. Relation between carbon and nitrogen
turnover in soil organic fractions of microbial origin. Soil Biol. Biochem. 7:57-63.

Meyer, S.L. 1975. Data analysis for scientists and engineers. Wiley, New York.

Morra, M.J., RR. Blank, L.L. Freeborn, and B. Shafii. 1991. Size fractionation of soil
organo-mineral mmplexes using ultrasonic dispersion. Soil Sci. 152:294-303.

Nadelhofi‘er, K.J. 1990. Microlysirneter for measuring nitrogen mineralization and
microbial respiration in aerobic soil incubations. Soil Sci. Soc. Am. J. 54:411-415.

North, PF. 1976. Towards an absolute measurement of soil structural stability using
ultrasound. J. Soil Sci. 27:451-459.

Paul, EA, and F .E. Clark. 1996. Soil Microbiology and Biochemistry. 2nd ed. Academic
Press, San Diego.

Paul, E.A., RF. Follett, S.W. Leavitt, A. Halvorson, G.A. Peterson, and DJ. Lyon. 1997.
Radiocarbon dating for determination of soil organic matter pool sizes and dynamics. Soil
Sci. Soc. Am. J. 61:1058-1067.

Paul, E.A., D. Harris, H.P. Collins, U. Schulthess, and GP. Robertson. 1999. Evolution
of C02 and soil carbon dynamics in biologically managed, row-crop agroecosystems.
Applied Soil Ecology 11:53-65.

Paul, EA, and WE. McGill. 1977. Turnover of microbial biomass, plant residues and soil
humic constituents under field conditions. p. 149-157. In Soil Organic Matter Studies.
International Atomic Energy Agency, Vienna.

Paul, E.A., S.J. Morris, and S. Bohm. in press. The determination of soil C pool sizes and
turnover rates: biophysical fiactionation and tracers. In R. Lal, J. Kimble, and R. F ollett
(eds) Assessment methods for soil C pools. CRC Press, Boca Raton, FL.

Pfefl‘er, P.E., W.V. Gerasomowicz, and E.G. Piotrowski. 1984. Effect of paramagnetic
iron on quantitation in carbon-13 cross polarization magic angle spinning nuclear magnetic
resonance spectrometry of heterogeneous environmental matrices. Anal. Chem. 56:734-
741.

Rovira, AD, and EL. Greacen. 1957. The effect of aggregate disruption on the activity
of microorganisms in the soil. Aust. J. Agric. Res. 82659-673.

Sachs, L. 1982. Applied statistics: A handbook of techniques (Springer Series in

Statistics). Springer-Vedag, New York.
Sarrantonio, M. 1998. How much N? p. 22-23. In G. Bowman, C. Shirley, and C. Cramer

111

(eds) Managing cover crops profitably. Sustainable Agriculture Network, Handbook
Series Book 3, Beltsville,MD.

SAS Institute Inc. 1989. SAS/STAT user's guide, Version 6. Vol. 2. 4th ed. SAS Institute
Inc, Cary,NC.

Schmidti M.W.I., H. Knicker, P.G. Hatcher, and I. KOgel-Knabner. 1997. Improvement of
13C and sN CPMAS NMR spectra of bulk soils, particle-size fractions and organic
material by treatment with 10% hydrofluoric acid. European Journal of Soil Science
48:319-328.

Schnitzer, M. 1982. Organic matter characterization. p. 581-594. In AL. Page, RH.
Miller, and DR. Keeney (eds) Methods of Soil Analysis, Part 2. Chemical and
Microbiological Properties. American Society of Agronomy - Soil Science Society of
America, Madison, WI.

Schnitzer, M. 1985. Nature of nitrogen in humic substances. p. 303-325. In G.R. Aiken,
D.M. McKnight, R.L. Wershaw, and P. MacCarthy (eds) Humic Substances in Soil,
Sediment, and Water: Geochemistry, Isolation, and Characterization. John Wiley and
Sons, New York.

Schulten, H.-R., and R. Hempfling. 1992. Influence of agricultural soil management on
humus composition and dynamics: Classical and modern analytical techniques. Plant Soil
142:259-271.

Schulten, H.-R., and M. Schnitzer. 1998. The chemistry of soil organic nitrogen: a review.
Biol. Fertil. Soils 26: 1-15.

Schulten, H.-R., C. Sorge, and M. Schnitzer. 1995. Structural studies on soil nitrogen by
Curie-point pyrolysis-gas chromatography/mass spectrometry with nitrogen-selective
detection. Biol. Fertil. Soils 20: 174-184.

Sirnonart, P., L. Batistic, and J. Mayaudon. 1967. Isolation of protein fi'om humic acid
extracted from soil. Plant Soil 27:153-161.

Smith, J .L., and EA. Paul. 1990. The significance of soil microbial biomass estimations. p.
357-396. In J .-M. Bollag, and G. Stotzky (eds) Soil Biochemistry. Marcel Dekker, New
York.

Sorenson, LR, and EA. Paul. 1971. Transformation of acetate carbon into carbohydrate
and amino acid metabolites during decomposition in soil. Soil Biol. Biochem. 3: 173-180.

Sowden, F.J., Y. Chen, and M. Schnitzer. 1977. The nitrogen distribution in soils formed

under widely varying climatic conditions. Geochim. Cosmochim. Acta 41: 1524-1526.
Spiegel, MR. 1971. Schaum's outline of theory and problems of advanced mathematics

112

for engineers and scientists. McGraw-Hill, New York.

Srinivasan, PR, and RL. Lichter. 1977. Nitrogen-15 nuclear magnetic resonance
Spectroscopy: evaluation of chemical Shift references. J. Magn. Reson. 28:227-234.

Stejskal, E.O., and J .D. Memory. 1994. High resolution NMR in the solid state. Oxford
University Press, New York.

Stevenson, F.J. 1994. Humus Chemistry: Genesis, Composition, Reactions. 2nd ed. John
Wiley and Sons, New York.

Thorn, K.A., and MA. Mikita. 1992. Ammonia fixation by humic substances: a nitrogen-
15 and carbon-13 NMR study. Sci. Total Environ. 113():67—87.

Tiessen, H, and J .W.B. Stewart. 1983. Particle-size fractions and their use in studies of
soil organic matter: H. Cultivation effects on organic matter composition in size fi'actions
Soil Sci. Soc. Am. J. 47:509-514.

Whittig, L.D., and W.R. Allardice. 1986. X-ray difl‘raction techniques. p. 331-362. In A.
Klute (ed.) Methods of Soil Analysis, Part 1. Physical and Mineralogical Methods.
American Society of Agronomy - Soil Science Society of America, Madison, WI.

Wlllson, T.C., EA. Paul, and RR. Harwood. In press. Managing biologically active soil
organic matter fiactions for sustainable crop production.

Wilson, MA, AM. Vassallo, P.J. Collin, and H. Rottendorf. 1984. High-resolution
carbon-13 nuclear magnetic resonance spectroscopy and relaxation behavior of organic
solids fiom fossil fuels. Anal. Chem. 56:433-436.

Wuthrich, K. 1976. NMR in biological research: peptides and proteins. North
Holland/American Elsevier, Amsterdam/New York.

Zhuo, S.-N., and Q.-X. Wen. 1992. Nitrogen forms in hunric substances. Pedosphere
21307-3 15.

Zhuo, S.-N., Q.-X. Wen, L.-J. Du, and S.-L. Wu. 1992. The nitrogen form of
nonhydrolyzable residue of humic acid. Chinese Science Bulletin 37:508-511.

Zuberer, DA. 1994. Recovery and enumeration of viable bacteria. p. 119-144. In RW.
Weaver, J.S. Angle, and PS. Bottomley (eds) Methods of Soil Analysis, Part 2. _
Microbiological and Biochemical Properties, Soil Science Society of America Book Series
5. Soil Science Society of America, Madison, WI.

113

APPENDICES

114

Appendix 1. Isotopic Tracer Equations

Clover-derived N and native N were calculated for the various N pools (whole

soils and particle-Size fractions) throughout the incubation on the basis of the following

equations:
NPOOL=NCLOVf+NNTvg (1)
15NPOOL = ”NCLOV f+ ISNNTV g (2)

NPOOL and l5Npoor, are the amounts of N and 15N, respectively, in the pool of interest.
NCLOV and l’Ncmv are the amounts of clover-N and clover-”N, respectively, present in
the whole soil on Incubation Day 0. NW and ”Nam; are the amounts of native-N and
native-”N, respectively, present in the whole soil on Incubation Day 0. The values f and

g are defined as follows:

f = clover-derived N in pool of interest / clover N in whole soil on Day 0

g = native N in pool of interest / native N in whole soil on Day 0

All N amounts in these equations were expressed in moles. The equations assume that

isotopic fractionation effects are negligible.

115

Simultaneous solution of (l) and (2) yields:

 

 

 

 

 

H ”N
NPOOL — N NPOOL
._ INTV
f- ”N (3)
1s
NCLOV - N NCLOV
N'n'
I5
N
15 NML __ N CLOV NPOOL
s = “:0" (4)
I5N _ 15 NCLOV N
INTV NCIDV NTV

The values of f and g appear in Tables 8 and 9.

The sample standard deviations of f and g were computed with two simplifying

assumptions. First, the uncertainty in the denominators of f and g in Equations 3 and 4

were assumed to be zero. Second, errors involving differences were assumed to be

additive (a worst-case assumption). The second assumption avoided complications due

to the interdependence of 15Npoor, and Npoor, (Theses two terms are not independent

because the measurement of l’Npoor, relied on the value of Mom; see “C, N, and 15N

Analyses” in Materials and Methods.)

116

Appendix 2. Modeling of Mineralization-Immobilization of Clover-Derived N in
Clay Fractions

In the clay fractions, clover-N was observed to accumulate in the initial stages of
the incubation and to decline in the latter stages. This behavior was modeled by the

following equation:

N, = -A exp(-kAt) + B exp(-th) (1)
where A and B are (positive) constants, kA is the accumulation rate constant, and k9 is the
depletion rate constant. The derivation of the model is described below.

Assume that both the accumulation and depletion of clover-N in the clay fractions
follow first-order kinetics:

dN/dt = kn N' - kn N (2)
where N is the amount of clay-N, k, and k1) are rate constants (time") for accumulation
and depletion, respectively, and N' is the amount of N in an unspecified pool from which
N is being transferred. Accordingly,

dN'/dt = -kA N' (3)
Equations 2 and 3 constitute a system of simultaneous linear differential equations and

were solved for N as described by Spiegel (1971) to yield Equation 1.

117

Appendix 3. Brief Overview of N MR Concepts

Chemical Shift

This is a measure of the frequency at which a 15N nuclei resonates in a magnetic
field. The units are ppm. Electron-deficient N atoms (e.g. nitrate) have high chemical
Shifi (high fiequency) and appear at the left edge of the Spectrum. Electron-rich N atoms

(e.g. ammonium or amino) have low chemical shift and appear at the right edge of the

spectrum.

Cross-Polarization Magic-Angle Spinning (CPMAS)

In the solid State, 15N NMR Spectra of soils are usually acquired with cross-
polarization magic-angle spinning (CPMAS). In CPMAS, magnetization is transferred
from magnetically sensitive 1H to relatively insensitive lSN. Simultaneously, the sample
is spun 4000 - 5000 times per second about an angle of 547°. Without these methods,
solid state 15N NMR spectra would be impractically slow and the peaks obtained would

be excessively broad.

Spinning Sidebands
These are produced on both sides of a prominent peak in a CPMAS experiment.
Though they are sometimes a nuisance (they might overlap with other peaks), their peak

areas need to be added to the area of the parent peak for correct quantitation.

118

Differential Relaxation Effects

There are two times during an NMR scan in which such effects can occur. The
first time is the contact time, during which magnetization is transferred from 1H to 15N.
The rate of transfer is characterized by two relaxation (equilibration) times: TN“ and Ttpg.
The parameter Tim is a measure of the rate of magnetization transfer from 1H to l’N, and
T1,,“ is a measure of how fast the 15N nuclei loses the transferred magnetization. These
parameters are not necessarily the same for all 15N firnctional groups, and correction of
signal intensities (peak areas) may be necessary.

The second time in which differential relaxation effects can occur is the delay
time. This is the time between NMR scans in which l’N nuclei return to equilibrium;
without the delay time, the nuclei would become saturated and give no signal. (Multiple
scans are usually required to obtain an acceptable spectrum because of the low
concentration of 15N.) The rate of return to equilibrium is characterized by Tm. This
parameter may differ by functional group, and correction of signal intensities may be
necessary. Further, 15N nuclei with very long Tm values (say, > 5 5) may be undetectable

in a typical (delay time at 1 s) CPMAS-NMR experiment.

119

* Explanation of chemical Shifi scale:

The chemical shift scale here defines l’NH..NO. in 2 M HNO; as 0.00 ppm. The
solid ammonium salts 15NHaN03 and (”Nimzso. are expected to have a chemical shift
value within about 1 ppm of l5NHaNog in 2 M HNO;; thus, these solids are equivalent to
1’NHaNog in 2 M HNO; in all but the most exact studies.

Conversions from ammonium nitrate or ammonium sulfate reference to NH; (1,25
°C) reference can be made with data based on Levy and Lichter (1979):

Table A-2 Chemical Shifts conversion.

 

 

compound chemical shift (ppm)
NH3 (1,25 °C) -21.60
l’NH..NO. in aqueous HNO3 0.00

 

a

l. Blomberg and Ruterjans, 1983.

2. Cross et al., 1982. Data are for solid state.
3. Levy and Lichter, 1979.

4. Thorn and Mikita, 1992.

5. Wuthrich, 1976.

b Hawkes et al., 1977; as cited by Blomberg and Ruterjans (1983).

c

A = adenine, C = cytosine, G = guanine, T = thymine; these are the bases of DNA
Adenine also occurs in the adenosine energy-transfer molecules AMP, ADP, and ATP, as

well as in NAD. Guanine occurs as a reactant in the citric acid cycle in the form of
guanosine phosphate.

120

Appendix 4. Detailed Table or 15N Chemical Shifts

Table A-1. 1’N chemical Shifts ofbiolo 'cally important molecules

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

compound chemical shift (ppm) ref.’

ammonium sulfate or ammonium 0 -

nitrate

free amino acids (data probably 5-34 (including Gly and Pro) 5

applicable to amino-terminal

amino acids in proteins), amino

groups _________________________________________
ca. 20 (most common shift) 5
12-20 (most common Shifts in solvent 1
and pH conditions reasonable for soil)

guanidine NHZ 42-53 1,5

-NH- (epsilon) 60—65

-NH2 of nucleic acids 76 (cytosine); 59 (adenine); 53 2
(guarune)

tryptophan -NH- (indole) 61 1

peptide-N 83-116 (including Gly and Pro) 1,5
95—1 10 (common amino acids;
excluding Gly and Pro)

uracilb 111.2 (N1) 1
138.8 (N3)

histidine ring N (imidazole) 145-155 cation; 157 & 211 amphiion; 1
173 & 197 anion

aromatic N irgrucleic acidsc

................................................................. 1-----.
G3, T3, C1, G1, T1 1206-1396 2
A9, G9 1 147.5 2
---------------------------------------------------------------- ‘1 - - - - - q

C3, A3, A1, A7, G7 174-211 2

pyrrole 136 (in DMSO-ds) 4

pyrrole as part Of chlorophyll A 173, 175, 192, 233 (in acetone-d6) 1

flavin (isoalloxazine) 128-140 (N3); 142 (N10, OX); 1
170-179 (N1); 312-320 (N5, ox.)

nitrate 354 3

 

 

 

121

The reason for the chemical Shift-equivalence of these reference compounds can be
explained in terms of H-bonding. According to Briggs and Randall (1973), the chemical
Shift of the ammonium-”N in these compounds in the solution state is greatly influenced by
the H-bonding of the ammonium hydrogens to the lone pairs of electrons on nearby O
atoms. Thus, the electron density around the 15N atom should be similar regardless of
whether the H-bond donor is N03‘, 8042‘, or H20. In support of this argument, Briggs and
Randall (1973) found variation in chemical shift of 1 ppm at most when l’NH..No, or
("er.)sso. was dissolved in its parent acid with concentrations of the acid anion ranging
from 1 to 10 molal. Their results were confirmed for ”NI-ENC, in its parent acid by
Srinivasan and Lichter (1977). The invariance of the chemical Shifis with increasing
concentration suggests that the solid-state salts have the same absolute chemical shift as the
dissolved salts; this is confirmed by the very plausible peak positions detected for biological

materials in the present study.

122

123

Appendix 5. Adjustment of N and C Amounts for Incomplete Dispersion on Day 11
and for Variable Clay Recovery

Methodological difficulties led to some variability in particle-size distributions.
(See Table 7 and “Recovery of Mass in Particle-Size Fractions” in Results.) This
variability and the associated corrections of N and C amounts in the size fiactions are
explained below.

On Day 11, incomplete dispersion resulted from failure to hand-shake the
incubated soil with water prior to sonication. Clay amounts were less than expected, and
fine silt and sand amounts were greater than expected (Table 7). The N and C recoveries

in the fine silt and sand were adjusted accordingly:

adjusted N or C in fine silt or sand = (measured N or C in fraction) x

(expected mass of particles in fiaction) / (measured mass of particles in fiaction)
The total clay N and C were adjusted by adding the “excess” N or C fiom the fine silt and
sand to the N and C measured in the total clay. (The “excess” N or C in fine silt and sand
were taken as the adjusted minus the measured in said fi'actions.) The N and C measured
in the fine clay and coarse clay were not adjusted (individually) because it was felt that
there was insuflicient information for such a computation.

The variability in fine clay yields on Days 34-439 (Table 7) presented a firrther
challenge in determining the true N and C distributions among the Size fractions. This
variability was attributed to the difficulty of siphoning a centrifuged suspension without
altering the vertical distribution of fine particles. The amounts of N or C in fine clay on

these days were adjusted as follows:

124

adjusted N or C on Day X = (measured N or C on Day X) x

(average mass of fine clay for Days 34-439) / (mass of fine clay on Day X)

Finally, the coarse clay recovery for Day 190 (Table 7) was determined to be a
statistical outlier relative to Days 34, 95, and 439 per Dixon’s gap test (Bliss, 1967).
Therefore, the coarse clay on Day 190 was adjusted as follows:

adjusted N or C = (measured N or C) x

(average mass of coarse clay for Days 34, 95, 439)/

(mass of coarse clay on Day 190)

125

Appendix 6. Details of Analysis of Variance of Shaken Slurry N Mineralization

Nitrogen mineralization in shaken Slurry (N Mm) was a function of the
measurements N30, No, Nam, and NTOTAL as per Equation 8 in Materials and Methods.
These destructive measurements were performed twice or more on subsarnples derived
fiom a single bulk sample of material. For purposes of the analysis of variance

(ANOVA), NMIN was computed as:

NMINJ = 100 X (N3o,i - <No> - <NCTRL>) / (<NTOTAL> - <No>)

where i denotes replicate number and O denotes mean value. This computation of NW
reflects the fact that No and NTOTAL are pseudoreplicates (simple chemical tests repeated
on subsamples of the same material). The values N30 and New, in contrast, were
considered true replicates (since they reflected the outcome of a biological process).
However, the mean value of New was used for simplicity; Nam, values were generally
very small compared to N30. Within the ANOVA, each value of NMIN was weighted by
(VAR‘)", where VAR‘ is the variance in NMIN due to variability in No, Nam, and
N101)“, as computed by the rules of statistical error propagation (Meyer, 1975). The
ANOVA was implemented with PROC GLM within the SAS software (SAS Institute,

1939)

126

Appendix 7. Computation of 30-Day Mineralization in the Undisturbed Soil (For
Comparison With Shaken Slurry Mineralization Data)

The thirty-day N mineralization in Shaken Slurry was compared to the N
mineralization in the undisturbed clover-amended whole soil as a means of assessing
physical protection. The shaken slurry samples were derived fi'om ultrasonic
fractionation of the clover-amended whole soil after 34, 190, and 439 d of incubation;
unfractionated whole soil samples from 34, 190, and 439 d of incubation were also
subjected to the shaken slurry test. The thirty-day mineralization in the undisturbed soil
was computed on the basis of the kinetic models of clover-derived and native N within
the particle size fi'actions (Table 6).

For whole soil and fine silt, the computations of thirty-day N mineralization in

undisturbed soil were as follows:

moles native-N mineralized in undist. soil = NMIN,er = [l - (NNTv.ttsoPREr>)/(Ner,tPRED)1 X NNTV,t,MF.AS

where var+so,PRlan is the molar amount of native N in fine silt or whole soil as predicted
by the kinetic model thirty days after fractionation, Nm,r,ppgp is the molar amount of
native N in fine silt or whole soil as predicted by the kinetic model on the day of
fractionation, and ngMEAS is the molar amount of native N as measured in fine silt or
(unfi'actionated) whole soil on the day of fractionation. For whole soil, NMmer was
assumed to be zero (Table 6, Figure 7). The moles of clover-N mineralized in

undisturbed soil were computed by an analogous equation:

moles clover-N mineralized in until“. soil = NMINpLov = [1 - (Naov,t+30.PRED)/(NCLOV,1.PRED)1 X NCLov,t.MI-:AS

127

The expected thirty-day mineralization (percent) for clover- plus native-N was then

computed as follows:

NMIN.CIOV+NTV = [(NMINNTV' + Nan.CLov)/(NNTV.1.MEAS + NCLov,t.MEAs)] X 100

For fine clay and coarse clay, accumulation of N as per Equations 6 and 7 was
neglected for the sake of the comparison; this is because no transfer of material between
fractions was possible in the shaken slurry tests. Accordingly the undisturbed

mineralization was computed as follows:

NMINCLOV = [1 - (Ncwv,t+3o.PREI)/NCIOV,1.PRED)] NCLov,t.MEAS

Taking NCLOV,t+30,PRED = N' exp[- 1(1) (t +30” and NCLOV,t,PRED = N' exp(- th ) With N'
being an arbitrary constant and kn being the decay rate constant fi'om Equation 6, the

above equation simplifies to:

NMIN,CLOV= [1 - exp (-30 1(0)] Ncmv.t.MEAs

Numerically, kg was equal to 0.00105 (1'1 (fine clay) or 0.00130 (1'1 (coarse clay); see

Table 6. The value of NMIN,er (native N mineralized from undisturbed clays) was

assumed to be zero because accumulation with no depletion was observed for native N in

128

the clays (Figure 7). The expected thirty-day mineralization (percent) for clover- plus

native-N was then computed as follows:

NMIN,CIDV+NTV = [(NMIN,NTV + NMIN,CLOV)/ (NN'I'V,L1\AEAS + NCIDV,t,MEAS)] X 100

129

Appendix 8. Calculations of Aggregate Protection Factors and Aggregate-Protected
Pool Sizes

Table A-3. Aggregate Protection Factors And Pool Sizes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30—day N 30-day N
incubation mineralization mineralization as
time prior to as measured in modeled in aggregate aggregate-
fractionation shaken Slurry undisturbed soil protection protected pool
(d) (%)t (%) I: factor§ (%) ‘1
whole soil
34 8.51 :1: 2.29 2.34 3.64 a: 0.98 6.17 d: 2.29
190 5.93 a: 5.96 1.59 3.73 a 3.75 4.34 i 5.96
439 8.16:1: 2.22 1.40 5.83 21:1.59 6.76i2.22
AVERAGE = 4.40 5.76
fine clay
34 11.7m0.9 1.10 10.6:t0.8 10.6:09
190 7.07s 1.15 1.14 6.20a 1.01 5.93i1.15
439 4.14a091 1.12 3704:081 3.02i0.91
coarse clay
34 5.23 :1: 0.73 1.96 2.67 m 0.37 3.27 i 0.73
190 4.66 a 0.29 1.91 2.44 m o. 15 2.75 :1: 0.29
439 2.521015 1.51 1.67a0,10 1.01i0.10
fine silt
34 0.752 a 0.090 10.2 0.0737 a 0.0088 NC #
190 2.35 a 0.45 1.44 1.63 a 0.31 0.91 i 0.45
439 1.192: 0.22 0.704 1.69a 0.31 0.493: 0.22

 

 

 

 

 

 

'1’ (N mineralized / N present at beginning of 30-d period) x 100

1: Calculated per kinetic models as explained in Appendix 6. For simplicity, error in
modeled mineralization was assumed to be zero.

§ Aggregate protection factor = shaken slurry / undisturbed
1 Aggregate-protected pool (% of N in fraction) = shaken slurry - undisturbed

# Not calculable (computed pool size is negative).

130

Appendix 9. Correction for Error Due to Peak Overlap in 15’N-N MR Spectrum of
Clover-Uracil-Soil

To correct for error due to peak overlap, the spectrum of 15N-clover + 1SN-uracil +
unlabeled soil (CUS) was approximated as a linear combination of previous spectra taken
of (1) l’N-uracil + unlabeled soil and (2) l’N-clover + unlabeled soil. Upon showing the
linear combination to be statistically valid, the relative contributions of clover and uracil
in the CUS spectrum were estimated. The details of the computations are explained
below.

Mathematically, for each NMR frequency f in the spectrum:
SCUS,MEAS(f) *3 SCUS,SIM(f) = kus SUS,MEAS(0 + kcs SCS,MEAS(t) (1)

where Scusnmas is the measured signal intensity in the clover/uracil/soil spectrum,

Swagger is the simulated signal intensity in the clover/uracil/soil spectrum, Stigma,“ and
ScsMEas are the measured signals in the uracil/soil and clover/soil spectra, respectively,
and kits and kcs are positive constants. The statistical validity of Equation 1 was tested

by adjusting kus and kcs so as to minimize chi-square as defined below:

ism.m(r)-ssa.sn(r))

2 = r=rL (2)
(VARCUSJWEAS + VARCUSSIMM

 

X

where ft, and fa are the NMR frequencies of the left and right edges of the spectral region

of interest, df is the degrees of freedom (df = number of NMR frequencies in spectral

131

region of interest minus one), VARCUSMEAS is the variance of the measured spectrum of
clover/uracil/soil and VARcuser is the variance of the Simulated spectrum of
clover/uracil/soil. According to Bevington and Robinson (1992), a sufficiently small
value of )8 indicates that the two distributions in the numerator of Equation 2 are not
significantly different.

The variance of the measured Spectrum in Equation 2 was estimated as the
variance in signal intensity in 25 points chosen fiom a portion of the spectrum without
peaks. The variance of the simulated spectrum was based on the nrles of statistical error

propagation:

VARcuser = kiJsVAR US.MEAS + kéSVARCSJstAS (3)

where kus and kcs are the constants from Equation 1, and VARus,nmas and VARcsMp-AS
are the variances of the measured spectra of uracil/soil and clover/soil. VARusMEAS and
VARCSMEAS were computed per the aforementioned 25-point method.

The chi-square minimization per Equation 2 yielded f = 49.41, corresponding to
p = 1.000; thus the simulated and measured spectra (not shown) were not significantly
different. Therefore, the linear combination of spectra was considered valid and the peak

area percentages of the clover/uracil/soil sample were calculated as:

 

k PA _
PAPU_N3= U;A U “ms x100
TOTAL

 

1: PA _
PAPU—Nl: 112A U N1,Us x100
TOTAL

132

 

where PAPUM and PAPUM are peak area percentages in the simulated clover/uracil/soil
spectrum due to uracil N at ring positions 3 and 1, respectively. PAU-N3,Us and PAU-N1,US

are peak areas in the measured uracil/ soil spectrum of uracil-N3 and uracil-N1,

respectively. PAPCLov is the peak area percentage in the simulated spectrum due to

clover, and PAcmvps is the peak area in the measured clover/soil spectrum PATOTAL is

defined as:

PATOTAL = kus (PAH-N3,US + PAU-Nl,us) + kcs PACLOV,CS

133

Appendix 10. Estimation of Uncertainty in NMR Spectra Due to Noise and
Subjectivity in Phasing/Baseline Correction

Phasing and baseline correction are somewhat subjective (dependent upon visual
observation), and were performed twice for each spectrum. A standard deviation due to

processing (SdpRoc) was computed for each functional group in the spectrum:

 

i=1

J:(PAi —(PA))2
SdPROC =

n—l

where n = 2, PA.- is the peak area of the functional group in the ith replicate of processing,
and <PA> is the mean peak area of the functional group.

The error due to noise was also estimated for each spectrum. Twenty-five regions
of 1 ppm width in the non-peak portion of each spectrum were integrated separately; the
standard deviation of these twenty-five areas was defined to be the noise present per 1
ppm (storsa,1). From this quantity, a standard deviation due to noise (SdNOlse) was
calculated for each functional group in the spectrum based on the rules of statistical error

propagation:

 

Sd NOISE = J(WIDTH)X sdimmer = 5d NOISE-2,1 V (WIDTH)

where WIDTH is the width in ppm of the horizontal base of the peak. Finally, the overall

standard deviation was computed as:

 

SdOVERALL = V/Sdilorsr. +5d12>ROC
with the assumption that Sdmtsp, and SdpRoc are independent.

134