L5"...
. Ll,

 

 

 

 

 

Hum-u

 

 

 

 

 
 

 

 

:2. llllllll‘lllllll llllll llllll lllllllll

( 171M ) LIBRARY 3 1293 017
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

MANAGING NITROGEN MINERALIZATION AND BIOLOGICALLY ACTIVE
ORGANIC MATTER FRACTIONS IN AGRICULTURAL SOIL

presented by

Thomas C. Willson

has been accepted towards fulfillment
of the requirements for

Ph.D. Crop and

Soil Sciences

Ma 5 rofes or
(KW flz/

degree in

 

 

Date December 11, 1998

 

MSUL'S an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to
TO AVOID FINE re

MAY BE RECALLED with ear

remove this checkout from your record.
turn on or before date due.
lier due date if requested.

 

DATE DUE_

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1]” Wm“

MANAGING NITROGEN MINERALIZATION AND
BIOLOGICALLY ACTIVE ORGANIC MATTER
FRACTIONS IN AGRICULTURAL SOIL

By
Thomas C. Willson

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
1998

ABSTRACT

Managing Nitrogen Mineralization and Biologically Active Organic
Matter Fractions in Agricultural Soil

By
Thomas C. Willson

Integrated cropping systems use biological resources such as organic fertilizers,
nitrogen fixing plants, and crop and cover crop residues to reduce the need for
chemical inputs and improve sustainability. This research explores the effect of
integrated management on nitrogen mineralization in a variety of com-based
agricultural rotations and early successional treatments at the Kellogg Biological
Station (KBS) in Southwestern MI. Part 1 defines the effect of management on the
seasonal dynamics of nitrogen mineralization potential (NMP), the intrinsic ability
of a soil to mineralize N over time (ideally a growmg season). Measurements of
NMP, based on the accumulation of inorganic N in soils incubated under aerobic
conditions for 10, 30, 70, 150 and 310 days at 25 °C, are evaluated using 1st order
mineralization kinetics and repeated measures analysis of variance. Nitrogen
mineralization potential was greatest in the spring and lowest in the fall in all
treatments. It increased in response to compost additions, was greater in low-input
rotations with cover crops than in conventional rotations, and was greater in

successional treatments without tillage than in tilled succession or agronomic

treatments. During a com-corn-soybean-wheat rotation, NMP was lowest under 211‘1

year corn and highest under wheat and 1st year corn, reflecting both legume inputs
and the lack of spring tillage in wheat. Part 2 of the research found that
macroorganic matter was correlated with NMP in most treatments, but that
microbial biomass was not. The strongest correlation (r2=.41) was between
macroorganic C in the 53-1000 um size class and inorganic N accumulation in 150
day incubations. Macroorganic matter increased more rapidly after compost
additions than NMP, but was less affected by residue inputs. Part 3 of the research
found that conventional fertilization produced higher corn and wheat yields and
greater leaching loss than additions of leaf and dairy compost. Net annual N
mineralization (calculated from plant uptake and leaching loss) was greater in the

compost treatments, but N availability was low because of lower than expected
compost mineralization (9% yrl). Nitrogen mineralization was greatest in years with
corn production and lowest in years with soybean and wheat production, reflecting
(in part) differences in residue input from the previous crop. Much of the variation
in N mineralization could be predicted by laboratory measurements of NMP and
macroorganic matter, particularly total macroorganic N (53 - 1000 um), potentially
mineralizable organic matter (N o), and N mineralization predicted in 70 and 150
day incubations based on the lst order kinetics. Leaching loss (October through
September) was best predicted by initial inorganic N and total inorganic N (initial
plus mineralized) in incubations performed after sidedress fertilization in the
preceding season. This research shows that N mineralization varies appreciably

within and between growing seasons, and responds predictably to integrated

agricultural management.

AKNOWLEDGEMENTS

I want to thank the many people at MSU and elsewhere who have made this
dissertation possible. My Co-Advisors, Dr. Richard Harwood and Dr. Eldor Paul
have offered invaluable assistance and advice throughout this process, as have the

other members of my Guidance Committee, Dr. George Bird, Dr. Francis Pierce,
and Dr. Alvin Smucker. Many members of Dr. Harwood’s laboratory have helped
me in sampling and processing these data, including my fellow graduate students:
Puma Chetri, Neva Dehne, Tim Eisenbeiss, Ann Marie Ezzano, John Fisk, Sue
Ellen Johnson, Marcus Jones, Anna Kareem, Jose Sanchez, and Jeff Smeenk, our
technicians Elaine Parker and Tim Pruden, Dr. Ernesto Franco-Vizcaino, and the
many students who have helped me in the lab, including: J oonu Andrews, Curtis
Brown, Marti Johnson, Todd Martin, Christy McGrath, Hugh Smeltekopf, Brad
Sneed,. Phil Weinstein, and Gary Zehr,. In Dr. Paul’s lab I am particularly indebted

to Dr. Dave Ham's, who was extremely helpful in getting me started with this

project.

Most of all I want to acknowledge the help and support of my family, particularly
my wife Yvette, without whom this would not have been possible. Thanks for

putting up with me. I promise that our best days are yet to come.

iv

Table of Contents

LIST OF TABLES ............................................................................................ VIII
LIST OF FIGURES .............................................................................................. X
LIST OF ABBREVIATIONS ............................................................................. XII
INTRODUCTION ................................................................................................ 1
CHAPTER 1
SEASONAL CHANGES IN NITROGEN MINERALIZATION POTENTIAL IN
AGRICULTURAL SOILS: AN OVA AND N ON-LINEAR REGRESSION
ANALYSIS OF MANAGEMENT EFFECTS ..................................................... 10
ABSTRACT .............................................................................................. 10
INTRODUCTION ................................................................................... 12
Mineralization Kinetics as an indicator of NMP .............................. 13
Statistical Analysis of N Mineralization Kinetics ............................. l3
Repeated Measures AN OVA as an alternative to Kinetic Analysis .. 15
Objectives: ...................................................................................... 16
MATERIALS AND METHODS: ............................................................. 17
Sample Handling and N mineralization .......................................... 19
Data Analysis ................................................................................. 19
RESULTS ................................................................................................. 22
LT ER 1994 ..................................................................................... 22
LTER 1995-96 ................................................................................ 25
LF L 1994 ....................................................................................... 28
LFL 1995-96 ................................................................................... 32
LFL 1995-96 Rotation Etfect ........................................................... 33
DISCUSSION: .......................................................................................... 34

The Seasonality of N Mineralization Potential ................................ 34

Treatment effects on N mineralization potential .............................. 34
Evaluation of the Methods used in this study .................................. 37
REFERENCES ......................................................................................... 39
CHAPTER 2
MANAGING BIOLOGICALLY ACTIVE SOIL ORGANIC MATTER
FRACTIONS FOR SUSTAINABLE CROP PRODUCTION .............................. 41
ABSTRACT .............................................................................................. 41
INTRODUCTION ................................................................................... 43
MATERIALS AND METHODS .............................................................. 47
RESULTS ................................................................................................. 52
DISCUSSION ........................................................................................... 62
Pool sizes in relation to plant needs. ................................................ 62
Pool sizes in relation to N mineralization Potential .......................... 64
Seasonal Changes in Microbial Biomass ......................................... 66
REFERENCES ......................................................................................... 68
CHAPTER 3
BIOLOGICAL INDICATORS OF CROP PERFORMANCE, NITROGEN
MINERALIZATION, AND LEACHIN G LOSS IN INTEGRATED CROPPIN G
SYSTEMS ........................................................................................................... 72
ABSTRACT .............................................................................................. 72
INTRODUCTION ................................................................................... 74
MATERIALS AND METHODS: ............................................................. 77
Site Description and Treatments Sampled ....................................... 77
Field Measurements ....................................................................... 78
Calculation of Field N Mineralization ............................................. 81
Laboratory Measurements .............................................................. 85

vi

Correlation Analysis ....................................................................... 87

RESULTS ................................................................................................. 89

Crap Performance, N Leaching Loss and Apparent N Mineralizatiorgl9

Correlation analysis ........................................................................ 93
DISCUSSION: .......................................................................................... 99
Limitations in the Analysis ........................................................... 103
REFERENCES ....................................................................................... 105
SUMMARY ...................................................................................................... 109
APPENDIX ....................................................................................................... 112
REFERENCES ....................................................................................... 115

Table
Table I.1

Table 1.1

Table 1.2

Table 1.3

Table 1.4

Table 1.5

Table 2.1

Table 2.2

Table 2.3

Table 3.1

LIST OF TABLES

Caption
Timing of sample dates relative to field operations

Summary of ANOVA for the 1994 LTER data set

Kinetic parameters, predicted values, and their

asymptotic standard errors (in parentheses) for
treatment and date effects in the 1994 LTER data set

Summary of ANOVA for the 1995-96 LTER data set

Kinetic parameters, predicted values and asymptotic
standard errors (in parentheses) for the 1995-96

LTER data set

Average accumulation of inorganic N in 10, 30, 70,
and 150 day incubations for the LFL Integrated
Fertilizer rotation in the 1995*96 data set

Date and treatment effects on microbial C:N ratios in
the 1995-96 data set

Correlations (r) between biologically active organic
matter fractions and cumulative mineralization
measurements across all treatments and dates

Selected N pools in 1995-96 relative to one another
and to soil N

History of N inputs in all entry points in the Integrated
Compost and Integrated Fertilizer management types

at the LFL

Page
*9

23

24

26

27

33

55

61

63

79

Table 3.2

Table 3.3

Table 3.4

Table 3.5

Table 3.6

Table 3.7

Table 3.8

Table A.1

Table A2

Laboratory measurements and kinetic parameters
used in the correlation analysis

LFL crop yields averaged between cover and no-cover
treatments.

Annual N budget for the LFL Integrated Fertilizer and
Compost treatments with cover crops in 1994 — 1996.
Parentheses indicate the standard error of the mean.

Annual N balance (not including gaseous losses) for
the Compost and Fertilizer treatments with cover
crops

Laboratory measurements with the highest correlation
to field mineralization

Laboratory measurements with the highest correlation
to corn yield

Laboratory measurements with the highest correlation
to leaching loss

Literature values for the percent of soybean N derived
from atmospheric (%Ndfa)

Literature values for the percent of red clover N
derived from atmosphere (%Ndfa)

86

90

91

93

94

96

97

113

114

Figure

LIST OF FIGURES

Caption

 

Page

 

 

 

 

 

 

 

 

 

Figure l.1

Figure l.2

Figure l.3

Figure 1.1

Figure 1.2

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Plot map of the K88 LTER main-site experiment.
Blocks 1 through 4 of treatments 1 (CT), 2 (NT), 3 (LI), 4
(2|), 7 (HTS) and 8 (NTS) were sampled along with a
tilled micro-plot in treatment 7 (ATS).

Plot map of the LFL experiment as it existed in 1994.
Data for this project comes from the Integrated Fertilizer
(Int. Fert.) and Integrated Compost (Int. Comp.)
management types.

Diagrammatic representation of the crop and cover crop
sequences at the Living Field Laboratory. Bars
represent the time in which living crops and cover crops
are in the field. Field preparation for wheat occurs
immediately after soybean harvest.

Accumulation of inorganic N during laboratory
incubations at 250 in the 1994 LFL data set. Error bars
represent standard errors for the mean of the 10, 30,
70, and 150 day sub-samples. Units are based on a
depth of 25 cm.

Accumulation of inorganic N during laboratory
incubations at 25C for the 199596 LFL data set. Error
bars represent standard errors for the mean of the 10,
30, 70, and 150 day sub-samples. Units are based on a
20 cm sample depth.

Microbial biomass C in the top 25 cm in the 1995 LTER
Soybean Treatments. Bars represent the standard
error of the mean.

Microbial biomass C and N content in the top 20 cm in
the 1995-96 LFL Samples. Bars represent the standard
error of the mean.

Two size classes of macroorganic Matter C and N in the
top 25 cm of the1994 LFL Treatments. Bars represent
the standard error of the mean for each treatment

Macroorganic C and N in two size fractions in the top 25
cm of LTER wheat and successional treatments In

29

31

52

56

58

Figure 2.5

Figure 3.1

Figure 3.2

1995-96. Bars represent standard errors of the mean.

Macroorganic Matter C and N in two size fractions in the
top 25 cm of LFL treatments in 1995-96. Bars
represent standard errors of the mean.

Corn yield as a function of the supply of inorganic N
from mineralization (Nmin*) and fertilization

Late Fall whole profile NO3-N as a predictor of leaching
in the following field season

xi

60

100

102

LIST OF ABBREVIATIONS

 

 

 

 

 

 

 

_*

Abbreviation Definition

Locafions:
LFL Living Field Laboratory
LTER Long Term Ecological Research Experiment
KBS Kellogg Biological Station

LTER Treatments:
CT High-input rotation with conventional tillage
NT High-input rotation with no-till management
Ll Low chemical input rotation with cover crops
Zl Zero chemical input rotation with cover crops
HTS Historically tilled succession
NTS Never-tilled succession
ATS Annually tilled succession

Measurements: 1'
NMP
MM
MB

Nitrogen mineralization potential
Macroorganic matter

Microbial biomass

T See Table 3.1 (page 75) for additional abbreviations used in Chapter 3.

xii

INTRODUCTION

Integrated cropping systems use biological resources such as organic fertilizers,
nitrogen fixing plants, and cover crop residues to reduce the need for chemical
inputs and improve sustainability. If we are to manage integrated cropping systems
for high productivity and minimal NOs-N loss, we must understand how our

management affects the timing and extent of nitrogen mineralization.

This dissertation explores the effect of integrated management on N mineralization
from three perspectives. Chapter 1, “Seasonal Changes in Nitrogen Mineralization
Potential in Agricultural Soils: AN OVA and Non-Linear Regression Analysis of
Management Effects,” explores the effect of management on the seasonal dynamics
of nitrogen mineralization potential (NMP), the intrinsic N supplying capacity of the
soil. We evaluate NMP based on the accumulation of inorganic N in long term
aerobic incubations at 25 °C, and compare two methods of assessing it statistically.
In chapter 2, “Managing Biologically Active Soil Organic Matter Fractions for
Sustainable Crop Production,” we explore the effect of management on the C and N
content of the soil microbial biomass (determined by the chloroform fumigation,
incubation procedure) and of macroorganic matter, which we have defined as the
organic material in the sand-sized (53-1000 um) fraction of dispersed soil. Both of
these “biologically active” organic matter fractions are involved in the supply of N in
agroecosystems: microorganisms as the primary agents of decomposition, and
macroorganic matter as an important storage pool of actively decomposing and

potentially decomposable organic matter particles. The methods used to measure

these pools are much less time consuming than long term incubations, and might
therefore be more practical as routine indicators of N mineralization potential. In
chapter 3, “Biological Indicators of Crop Performance, Nitrogen Mineralization,
and Leaching Loss in Integrated Cropping Systems,” we evaluate laboratory
measurements of NMP, microbial biomass and macroorganic matter for their ability
to predict crop performance, N mineralization, and leaching loss in the field.
Leaching loss is measured using subsurface leaching lysirneters, and net annual N

mineralization is calculated based on an annual N budget derived for each plot.

The data presented in this dissertation are based on measurements taken over a
three year period (1994 through 1996) at the Long Term Ecological Research
(LTER) main-site experiment and the Living Field Laboratory (LFL) experiment at
the Kellogg Biological Station in Hickory Corners, MI (42°24’N, 85°23’W). These
experiments are located adjacent to one another on a gently rolling glacial outwash
plain (2 to 4 % slope) consisting of Kalamazoo loam and Oshtemo sandy loam
(Typic Hapludalfs). Precipitation averages 880 mm yrl at this site(l988-l997), with
precipitation exceeding potential evapotranspiration from October through March.
The actual precipitation during the years of this study was slightly lower than
normal, 821, 820 and 651 mm based on the October through September crop
production year. There was a short period of insufficient moisture in the late spring

of 1994, and prolonged drought in the summer of 1996.

The KBS LTER site was established in 1988 to investigate the ecological

interactions within and between the major managed and unmanaged ecosystems in

the US corn belt. The LTER main-site experiment (Figure 1.1) consists of four corn-
based field-crop rotations, high input conventional tillage (CT), high input no-till
(NT), low input conventional tillage (LI), and zero input conventional tillage (Z1),
two perennial monoculture productions systems, alfalfa and poplar, and two early
successional systems, historically tilled (HTS), and never rifled (NT 8). The CT and
NT treatments were managed as conventional com-soybean rotations from 1989
through 1994 and wheat was added to the rotation in 1995. The L1 and 21
treatments are corn-soybean-wheat rotations receiving either reduced chemical
inputs (L1) or zero chemical inputs (21), and featuring hairy vetch as a green manure
cover crop frost seeded into wheat and interseeded into corn. Prior to 1993 the L1
and 21 treatments had been ridge-tilled, but since that time they have been
moldboard plowed, as is the CT treatment. The HTS was last tilled in 1989, while
the NTS treatment has been maintained by armual (winter) mowing. The HTS
treatment contains a micro-plot that is tilled annually with a moldboard plow (ATS).
The NTS treatment is located just south of the other treatments on a remnant of the
pre-modern-agn'cultural soil profile. We sampled the CT, NT, LI and Z1 treatments

in 1994, and the CT, HTS, ATS, and NTS treatments in 1995 and 1996.

The LFL experiment (Figure 1.2) was established in 1993 to investigate the
sustainability of corn production strategies ranging from a low diversity,
conventionally managed, corn monoculture, to a high diversity, organically
managed, com-corn-soybean-wheat rotation with cover crops. The experiment is

laid out in a split-plot, randomized complete block design, with main plots for each

 

 

Ancillary Plots

 

 

 

 

Treatment Key

 

 

 

 

. High Input corn/soybean/wheat,
conventional till (CT) 6 2 3
2. High Input com/soybean/wheat,
no till (NT)
3. Low Input com/soybean/wheat, 5 1 7 4
CT, legume cover (Ll) Block 5

 

 

 

 

 

4. Zero Input com/soybeanlwheat,
CT. legume cover (2|) 1 2

5. Populus (perennial woody)
6. Alfalfa (perennial herbaceous)

7. Successional, historically tilled 7 3 4 2
(HTS) and annually tilled (ATS)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Successional. never tilled (NTS) 1 2 km
(located off-site)
5 4 7 1
I Ancillary Plots j
5 6 2 3 5 6 4
7 4 4 I 3 6
Cl
ws
2 6
3 1 5 7
Block 6
. H 6 7 1 s
ADCIIIBW 0L5 BIOCIt 1 Block 2 Block 3 BIock 4 L
100 m

Figure l.1 Plot map of the KBS LTER main-site experiment. Blocks 1
through 4 of treatments 1 (CT), 2 (NT), 3 (LI), 4 (2|), 7
(HTS) and 8 (NTS) were sampled along with a tilled
micro-plot in treatment 7 (ATS).

Living Field Lab. 1994

Cover
Treatment

I
I]
11

Red
Clover

Annual
Rygrass

Rygrass+
Clover

no cover
planted

Rotation sequence: Corn. Corn. Soybeans. Wheat

Emu--

 

Compost

Organic

rm
W‘

 

its}! 107

-_-m-

Fertilizer

 

Conventional

Plot size: 30’ x 50' 2nd Year of rotation

Figure l.2. Partial plot map of the LFL experiment. (Block 1 of 4 in
1994.) The Integrated Fertilizer (Fertilizer) and Integrated
Compost (Compost) management types were sampled.

of four management types, Organic, Integrated Compost (Compost), Integrated
Fertilizer (Fertilizer) and Conventional, split for each entry point in a com com-
soybean-wheat rotation plus continuous corn so that each crop in the rotation is
represented each year. The Compost and Fertilizer management types use a
combination of banded herbicides and mechanical tillage for weed control, while the
Conventional treatment uses broadcast herbicides and the Organic treatment uses
no chemicals for weed or pest control. The Conventional and Fertilizer treatments
utilize urea and ammonium nitrate to maintain N fertility, whereas the Compost
and Organic management types use compost made from leaves and dairy manure.
All treatments use a chisel plow for primary tillage. Each entry point subplot in the
Fertilizer, Compost, and Organic treatments are further split based on the presence
or absence of cover crops. During this project, cover crops included red clover frost
seeded into wheat and interseeded into lst year corn, annual ryegrass interseeded
into 2nd year corn, and a mixture of red clover and annual ryegrass interseeded into
continuous corn (Figure 1.3). We sampled the lst year corn soybean and wheat
entry points of the integrated fertilizer and integrated compost management types
(cover and no-cover) in 1994 and all of the integrated fertilizer and integrated
compost treatments except soybean and 2nd year corn in integrated compost in 1995

and 1996.

Soil Samples were taken to a depth of 20 cm on April 11, June 5, August 8, October
4 and November 20 in 1994, and to a depth of 25 cm on April 29, July 3, September

17, and November 29 in 1995, and on April 14 in 1996. The timing of these

Crop and Cover Crop Sequence
Rotation

1st yearoom

   
  

2ndyearoom

Soybean

Monoculture

 

’7:;:':': f, ' I“ "W“ ‘7‘"; 2 3:21in

October January April July

 

 

 

 

Figure 1.3 Diagrammatic representation of the crop and cover crop
sequences at the Living Field Laboratory. Bars represent
the time in which living crops and cover crops are in the
field. Field preparation for wheat occurs immediately after

soybean harvest.

samples relative to field operations in the agronomic treatments is shown in Table
1.1. Each soil sample consisted of 12 or 15, 2 cm diameter cores, bulked together
and passed through a 6 mm sieve to remove large roots and residue fragments. Soil
samples were cooled immediately and stored at 4 °C before processing for NMP,
microbial biomass and macroorganic matter measurements. Detailed descriptions
of these procedures are provided in Chapters 1 and 2. The procedures used to
measure yield, crop and cover crop biomass, soil inorganic N, and leached inorganic
N are discussed in Chapter 3. With only a few exceptions, these field measurements
were obtained by co-workers associated with the Mott Chair for Sustainable

Agriculture.

The overall objective of this research is to identify specific management practices
that influence the availability and loss of nitrogen in cropping systems, and to
identify specific laboratory measurements at specific sample dates that can be used

to estimate the supply of N from soil under diverse management conditions.

Emu 038mm 928 86668 mm; 65 53980 Ea: a 3:865 ..x.. comm +

 

 

 

 

 

 

 

x x 583
x x x 93805 xmon.
x x x x x x x 8:53 .050
x x x x x x x x x chmE no.0
x x x x x x x x x venom 5an00
x x x x x x x x x mom=F
“was;
x x x x “moan:
x x x x x $253 xmon.
x x x x x 82.62. .050
x x x x x x AEooV mmofiogw
x x x x x x x nausea no.0
x x x x x x x wanna 33500
x x x x x x x mama...
cman>om new :50
_. .. .2 ->. FEE..- _ . IE. 1 r3 . a? . -. .c..< 2mm
08.. mm? vmmv

 

E. .chzmaaao Ba: 9 823.9 mount maEmm Co 058:. 9.. min»

CHAPTER 1

Seasonal Changes in Nitrogen Mineralization Potential in Agricultural
Soils: AN OVA and Non-Linear Regression Analysis of Management
Efiecm

ABSTRACT

A soil’s nitrogen mineralization potential (NMP) changes over time in response to
the addition and decomposition of organic material. We used non-linear regression
and repeated measures analysis of variance to explore seasonal changes in NMP in
field experiments in Southwestern MI. In the analysis of variance, NMP was based
on the average accumulation of inorganic N in 10, 30, 70 and 150 day aerobic
incubations at 25 °C. This was 29% greater in agronomic treatments receiving
compost than in inorganically fertilized treatments after 3 years of management, and
32% greater in com-soybean-wheat rotations with cover craps than in com-soybean
rotations without cover crops after 5 years. A 6th year old-field succession had 50%
greater NMP than either an annually tilled succession or an agronomic rotation.
There were strong seasonal changes in NMP in all treatments with 65% greater
accumulation in June than October samples in 1994 and 39% greater accumulation
in April than September samples in 1995. The kinetics of N accumulation generally
fit the single-pool exponential model N; = Ni + N o (at/MRT), where N! is the
inorganic N at time t, Ni is the initial inorganic N, No is the pool of mineralizable
organic N and MRT is the steady-state mean residence time of N 0. However, lower

than expected mineralization at 10 and 30 days had the effect of inflating estimates

10

of No and MRT in some incubations. Although we present several improvements in
non-linear regression analysis, neither No nor MRT were themselves reliable

indicators of NMP.

11

INTRODUCTION

One of the major challenges facing agriculturists is to develop management
strategies that better synchronize mineral N availability with crop demand. The
supply of inorganic N must be suflicient during the growmg season to achieve high
levels of productivity, but soil nitrate levels must remain low at all other times to
avoid loss and environmental contamination through leaching and denitrification
(Oberle and Keeney, 1990, Smith et al., 1990). Such losses are particularly likely
when rainfall exceeds potential evapotranspiration, as occurs in the cooler months in
temperate climates. Our ability to develop truly sustainable management systems --
whether these are based on organic inputs or chemical fertilizers -- will depend on
the ability to predict and manage the seasonal patterns of mineralization from soil

organic matter and plant residues.

Nitrogen mineralization potential (NMP) refers to the intrinsic nitrogen supplying
capacity of the soil. Measurements of NMP provide relative estimates of field N
mineralization, and can be used to model N availability and fertilizer N
requirements if environmental conditions are known (Smith et al., 1977, Campbell
et al., 1997). Nitrogen mineralization potential can be useful as an indicator of soil
quality (Doran et a1. 1987, Gregorich et al., 1994), and has been correlated with
microbial biomass, microbial respiration, total organic matter and macroorganic

matter content, and various physical indices of soil quality.

12

The NMP of a soil changes over time in response to the addition and decomposition
of organic materials. By repeatedly measuring NMP over time, we can generate a
seasonal profile that reflects the storage and release of organic-N in a soil under field

conditions.

Mineralization Kinetics as an indicator of NMP

Stanford and Smith (1972) used first order kinetics to define a pool of potentially
mineralizable N, No, which decomposes exponentially according to a rate constant,
k based on the equation:

Eq. [I] Nmin=No (l-e'k‘)

where Nmin is the cumulative amount of N mineralized at time t in the incubation.

In the original Stanford and Smith method, soil is mixed with sand and
mineralization measurements are obtained by periodic leaching. Maynard et al.
(1983) tested an alternative method in which inorganic N is allowed to accumulate
in replicate soil samples, which are then destructively sampled at various times
during the incubation. While this method produced slightly difl°erent kinetic
parameters than the leaching method, the quantity of N mineralized over time was
the same, a finding confirmed by Motavalli et al. (1995). We have preferred this

method in our lab because it can be more easily paired to C mineralization studies.

Statistical Analysis of N Mineralization Kinetics

Statistical methods for choosing between regression models are well developed (e. g.

Robinson, 1985). However, methods for identifying significant treatment effects on

13

mineralization kinetics have received less attention. Most non-linear regression
procedures provide asymptotic standard error estimates for each parameter
estimated, and some, such as the NLIN procedure of SAS (SAS Institute, 1988),
also provide asymptotic standard error estimates for the predicted values of the
kinetic equation. One can determine whether two treatments produce different
estimates for a parameter or predicted value by fitting a regression model with
separate parameter estimates for each treatment, and using the asymptotic SE
estimates to construct independent sample t-tests between the treatments.
Alternatively, one can determine whether a given treatment or set of treatments has
a significant effect on the regression as a whole by comparing the residual sum of
squares of a model with unique parameter estimates for those treatments to a
simplified model that does not include them. The appropriate test for the hypothesis
that the simplified model is sufficient to describe the system is the sum of squares
reduction test (Feber and Wild, 1989):
Eq. [2] [8813952 -— SSResi)/(deesz — deesi)]/M8Re51

~F(a, (dflfesz Men ), dflfesz)
where SSRes/ and 8819632 refer to the residual sum of squares of the full and reduced
models, respectively, de651 and deesz refer to the residual degrees of freedom of
the two models, and MSResi denotes the residual mean square error for the full
model. This test is also used to compare alternative kinetic models where one

model can be derived fiom the other by simplification.

Inferences based on NLR statistics are only valid if the same kinetic equation

provides a reasonably good fit to all treatments, and if each mineralization

l4

measurement can be considered an independent estimate of the overall curve for
that treatment. This independence criterion is met in the N accumulation method
because each measurement is made on a unique subsample, but it is not met in the
leaching method because repeated measurements are made on the same soil
sample. The inherent autocorrelation in that method can be overcome by using
mixed-model NLR techniques (Littell et al., 1997, Gregoire and Schabenberger

1996a, 1996b) such as the SAS NLINMIX macro (SAS Inst, 1996).

Repeated Measures ANOVA as an alternative to Kinetic Analysis

Non-linear regression techniques do not require a specific strategy for sampling
inorganic N during an incubation. If incubations are sampled at discrete intervals, it
may be preferable to define NMP based on repeated measures analysis of variance
(AN OVA). Using this approach, each measurement is considered a repeated
measure of the mineralization potential of the original soil sample, and the effect of
a given treatment (i.e. its NMP) is based on the average inorganic N measurement
for that treatment across all replications and incubation times. The AN OVA
approach has two important advantages over non-linear regression. First, it does not
assume that mineralization follows a particular pattern with time, and treatment
comparisons can be made regardless of whether all treatments exhibit similar
kinetics. Second, it facilitates the explicit specification of an experimental design,
and greatly simplifies the interpretation of main effects and interactions in a factorial

setting.

15

Objectives:

The objectives of this study are:

1. to determine the seasonality of N mineralization potential and the factors that
affect it,

2. to quantify the effect of agricultural management on N mineralization potential,

3. to compare repeated measures AN OVA with non-linear regression for its ability
to interpret treatment effects on both the overall level of mineralization and the

kinetics of mineralization curves.

16

MATERIALS AND METHODS:

The soils used in this study were obtained from the Long Term Ecological Research
Site (LTER) and the Living Field Laboratory (LFL) at Kellogg Biological Station
(KBS) in Hickory Corners, MI. These experiments are located adjacent to one
another on a gently rolling glacial outwash plain (2 to 4 % slope) consisting of
Kalamazoo loam and Oshtemo sandy loam (Typic Hapludalfs). The Ap horizon
reaches a depth of 20 to 30 cm and has an organic C content of 10 g kg1 to 20 cm.
(Robertson et al., 1997). Rainfall has averaged 880 mm per year from 1988-1997.
Precipitation exceeds potential evapotranspiration from October through March.
The never-tilled treatment at the LTER has an organic C content of approximately

l7gkg1.

The KBS-LTER main site experiment, initiated in 1989, consists of 7 treatments in a
randomized complete block design with the never tilled treatment located just off-
site. In April, June, and November 1994 we sampled two com-soybean rotations
with conventional fertilization; conventional till (CT) and no-till (NT), and two com-
soybean-wheat rotations with cover crops; low chemical input (LI) and zero
chemical input (ZI). All plots were planted to soybean in June, 1994 and to wheat in
November, 1994. (This was the first wheat crop in CT and NT.) The L1 and 21
treatments were interseeded with hairy vetch when in corn and frost seeded with
either hairy vetch or red clover when in wheat. In April, July, September and

November 1995, and April 1996, we sampled the CT treatment and three

17

successional treatments, historically tilled succession (HTS), a 6th year old field,
never tilled succession (NT 8), a mowed grass treatment on the original soil profile,
and annually tilled succession (ATS), a tilled micro-plot within the HTS treatment.

Additional information is available on the intemet (KBS-LTER, 1998 [online]).

The Living Field Laboratory was established in 1992 to explore the range of corn
production alternatives from a conventionally managed corn monoculture, to an
organically managed corn-com-soybean-wheat rotation with cover crops. The
factorial design of the LFL facilitates the separation of each component of these
systems: rotation vs. monoculture, cover vs. no-cover, chemical vs. organic
fertilization, and chemical vs. organic pest and weed management (c.f. Jones, et al.,

1998).

In 1994, we sampled 1st year corn, soybean and wheat, in the Integrated Fertilizer
(Fertilizer) and Integrated Compost (Compost) treatments with and without cover
crops, and in 1995-96 we sampled 1st year corn, wheat and continuous corn
(monoculture) in the same treatments. In addition to the April, June, and
November 1994 sample dates described for the LTER treatments, we sampled 1st
year corn and soybean in August and October 1994. In 1995 we also sampled 2nd
year corn and soybeans were in the Fertilizer treatment. The Fertilizer and
Compost rotations are identical except for the use of urea and ammonium nitrate

fertilizer or dairy manure -leaf compost to maintain fertility.

18

Sample Handling and N mineralization

One bulk soil sample consisting of 12, or more, 2 cm diameter cores was obtained
from each plot to a depth of 25 cm in 1994, and 20 cm in 1995-96. After sampling,
the soils were cooled immediately and stored at 4 0C until processing. Each soil
sample was thoroughly mixed in the lab and passed through a 6 mm sieve at field
moisture, and 20 g dry-weight—equivalent sub-samples were produced and adjusted
to a gravimetric moisture content of 0. l9. Sub-samples were pre-incubated for five

days.

Extractable N03 and NH4 were measured at the start of each incubation and at day
10, 30, 70, and 150. An additional incubation to approximately 310 days was
performed for the 1995-96 samples. All of the sub-samples for a given plot and date
were placed in a single, glass, quart canning jar with a small pool of free water to
prevent dehydration. Base traps containing 2 ml of 3M NaOH were used to absorb
COZ-C for measurement during the incubation period. The inorganic N content
was determined by 5:1 extraction with 1N KCl. The filtrate was analyzed at the
Michigan State University Soil Testing lab using a Lachat automated analyzer

(Lachat Instruments Inc. Milwaukee, WI).
Data Analysis

The data are divided into 5 discrete sets for analysis: 1) the 1994 LTER data set
consists of four agronomic treatments examined at three sample dates, 2) the 1995-

96 LTER data set consists of one agronomic and three successional treatments

19

sampled at 5 dates, 3) the 1994 LFL consists of a 2 x 3 x 2 factorial for fertilizer
source, crop, and cover over 3 dates, but will also be analyzed as a 2 x 2 x 2 factorial
over 5 dates, 4) the 1995-96 LFL factorial data, is a 2 x 3 x 2 factorial at five dates,
and 5) the 1995—96 Fertilizer rotation, which is a 5 x 2 factorial consisting of all entry
points in Fertilizer rotation plus continuous corn, split by cover, over 5 dates. Each
field treatment consists of four replications (blocks), and inorganic N was measured

six or seven times during each incubation.

We performed an AN OVA for each data set, using the SAS MIXED procedure
(SAS Inst, 1997) in which each sample date and incubation time was treated as a
repeated measurement of the overall NMP of the corresponding field plot. The SAS
MIXED procedure estimates the correlation structure within experimental units and
uses it to construct an accurate univariate analysis of treatment means. This
technique overcomes many of the inherent limitations of multivariate repeated
measures analyses (Cole and Gissel, 1966). The optimal covariance structure was

determined using Akaike’s Information (Littell et al., 1997).

Having used AN OVA to identify effects that produce significantly different
mineralization curves, we used the NLIN procedure of SAS (SAS Inst, 1988) to
parameterize these curves and identify significant differences in parameter estimates
and predicted values. We used the SSRT (Eq. [2]) to compare the overall fit of three
models: a single pool 1st order exponential model, Eq. [3], a linear model, Eq. [4],
and a two-pool 1st order exponential model Eq. [5]. In these models MRT, MRTI

and MRT2 are the steady state mean residence times of the mineralizable organic N

20

pools N0, N1 and N2 respectively. The parameter Ni represents the initial inorganic

N content of the soil, and R is a constant mineralization rate.

Eq. I3] ”1 = Ni+ Na(1 - e‘m'”)
Eq. [4] N1, = Ni+ kt
Eq. [5] N1: NI+N1(1_e’“”RTa)+(N2X1_e-t/mtrz)

21

RESULTS
LTER 1 994

The results of the AN OVA of the 1994 LTER data set (4 treatments x 3 dates x 5
incubation times) are summarized in Table 1.1. We observed less mineralization in
the CT and NT treatments (49.7 and 53 kg N ha'l) than in either the Low-input,(66.3
kg N ha'l) or Zero Input (69.9 kg N ha'l) treatment. There was also a significant
date effect, with the April and November sample dates resulting in significantly
greater overall mineralization (64 and 67.6 kg N ha-l) than the June sample (47.5 kg

N ha!) None of the interaction terms was statistically significant.

The key factor in increasing NMP in the L1 and Z1 treatments may have been the
use of leguminous cover crops as green manure. The use of wheat in the rotation
provides excellent conditions for the establishment and growth of the cover crop.
The production of hairy vetch following wheat in the L1 and 21 treatments averaged
421 g rrr2 in 1990 whereas only 50 g In2 of hairy vetch was produced following corn

in 1991 (Robertson et al., 1997 [online]).

The accumulation of inorganic N generally fit the single-pool, 1st order exponential
model (Eq. [3]), with estimates of 145 kg N Mgl soil for No, 17 kg N Mgl soil for
Ni, and a mean residence time of 96 days across all treatments and sample dates

(Table 1.2). The predicted values for the date and treatment effects generally

22

59.. 32.85 93 u .N 59.. _8_ano 33 u 3 $32 u #2 82.: 3:88.80 u

k0 ”Science/x A

 

 

 

 

 

co 88E.
I .535 2a 20892:. new 38:0 £9: 58536 =< .me. $3 9:335 omxi w<m
a SE 0.3 59562
a 8i 9: 95.. . . 2mm
n 6: Se E? 83 Rod cu m 8 v
n 8.x QB _N
am 8i 08 3
8 8.3 98 E . . a me
o 6.3 6% n 5 Sue Rod m m a v E .5 »
modxa c8 m3 865E800 mm:
: Bum. 2:5 7 as 2 9. 5.9952 lb: 8 m
cm; 8829a 5:325:02 .96.. .28“. 980 5:33... magi do ofinfim «0 n.
co “comma 9.90 2 £599... 53> c8022.... -n.
aoconcQw Co Gm. :82 co. o:_m>-n_
mach cozEwaam cam—2 Becki <>Oz< .
a an 5% mm: 82 as 8. <>oz< Lo 5258 . . 2%..

23

modue mgzgmggbfi_§gz§ 82n> .uflmnmgaoczi,

BEBE»

 

 

 

 

" u fl @. . H5.§§&_
Emozhaofi n SEE-races... : E9. 5 figacwxmwegnmumncfinfia
.OSSE§z§_BnBaueE§sa§n§“832585828828885
.QBBanzflmSuoamiw
E a E 8 E : e: 8 a: o: E m. and .2
n E B a E a E a as 8 c: we e: an 3552
a E k a E 8 E 2 ea 5 a: 8 e: a 9:.
n E 8 n E 8 E : ea a: ea 3 e: 8 =2
Eases .8 Emmy 3.8.8. mean 88
n E 8? n E e. E 9 ea 8 e: a: E t n
n E 8 n E 8 E 2 as as e: E E a a
a E 8 a E 8 E 3 E B e: 8? E : .2
a E R an E 8 E m 69 3: E E E 5 a. 8
cosdvaEmE esgrmzne»
.22 e. .229. 4229. e F.Eze. a 7.229.
89 8n 8? E2 02 z
22 cansfing 853a 8E 0.22

 

H .nmnSmEn§.§c_neanmEnn§fieE
A8853". .5 Se wanna 88% is an 83.9 Enema SEER 2.8: 3 new»

24

immobilization during the April CT incubations and this had the effect of inflating
the estimates of No and MRT in April across treatments and in the CT treatment
across dates. An SSRT analysis shows that there is a significant improvement in
overall fit when the treatments are fit individually, but little improvement when

parameters are fit for each date.
LTER 1995-96

Nitrogen mineralization potential was an average of 47% greater in the HTS and
NT 8 treatments than the ATS and CT treatments, in the 1995-96 LTER data set
(Table 1.3) despite greater net primary productivity and residue return in the tilled
systems (Robertson, et al., 1997 [online]). Mineralization potential was significantly
greater in April 1994 than in any of the other sample dates. There were significant
interactions between incubation time and each of the other terms. This indicates
that the relative mineralization at each point of the incubation curve varies with both
treatment and date. All of the curves were well described by single pool, 1st order
kinetics (Table 1.4). The CT and ATS treatments produced nearly identical
parameter estimates (No = 112 kg N ha-1 soil, MRT=86 days for CT, and N o = 110
kg N ha“1 and MRT= 77 days for ATS). The HTS treatment produced a similar
MRT (84 days), but a higher No (183 kg N ha-l), while the NTS treatment had the

highest estimates for

25

cgwmmooam 00:: 50>GZ H WFZ

8.3808 8:: 2.32 n m2 87.883 85 2.8293: n we: $9.: 38:82.8 n 5 “283.659.. #
.96. $3 2:8er omxi m<m :0 Baum ,,

 

 

 

 

 

mood mnmd wed? mm; 28
x .5662...
n E B 8 .34.
on g No 89552
o 5 mm 59533
8 E 8 2%
m i R mm E94 wood wood 36 36 23
m E E 92
a E 8 m?
m E E m»:
n 3v 5 a .5 oood Sod ad EN? 2958;
3.03 :5 omv .BmEEocmo mm:
z 8:2 92% 7 m; z 9. .coumcoE:z \ km:
990 ecumSEBo< .96.. 86mm QED cozmnsoc. m2m>d do unmamfii .98“.
8:85:05 2 2599.. 55, 5399:.
do mm. :82 .2 3.95
98k 5:833 :82 293i <>Oz<
.58 .28 $5 8&8. 9.. .2 <>oz< Co $5.25 3 2an

26

5388.5 3...... .962 u 9.2

56885 no.3. >=m=c< u wh< :o_mmooo:m BE. 38:29: u 9.1 39:... .m:o=:m>:oo u ._.0 ”239552
.o :6: :0 32:8 2 2590:. 88:65 2.55 525.85 .o :3 cc: 2 .3 .86 :0 22:8 2 2:88; 33.8.5 a.

3305 E53: E283: 328.23% 8: 2: 5:2 08mm 92 53> 8:33 .808. m_ 5%: :8 .2389: 2:2 w<m :o :83 ._.

 

 

 

 

E 8. E 8 E 2 c 5 2: E 9: SE 3 98 __<
a E o: n E 8 a E N: a an. 3; pm a: 9: m 8.9 2: 08. .__a<
an E m3 3 E 2. n E 2 m :E at m as own 5 NE 3 .35962
a E N: a E R an E 3 am 33 8 a a: 2. a GE 3 352:8
an E «N: a E 8 an E 9 am cm. 2: n 3: m9 n 6.9 an 22.
m c 5 9: m E 8. m E «N a 3: E a s: 9.: a at a... 82 .__a<
33 v a Emma $5.58.. $98 8.8
a E 9: m E 8 am E m: a GE 2: m 8: 8w 3 SE 3 92
o E 8 a E 8 a E 2 n .v: t. o E c: on EE «d m?
a E «2 m E 8. a E a a E 3 a E 9: 5 2E mm 9:
u E mm a E No a E 2 a a: 8 o E N: a E: w: 5
:80... v a Ema 3% 988 28,58:
_. 2 z 9. .. 2 z 9. c 2 z 9. as: .. 2 z 9. v 2 z 9.
82 2:. 2: Es. oz _2
a :2 9,83 82m> 8.2.85 «565%: 2.25.

 

h. .3 Em: mm: 8.39 9: :2 333.5ch .5 Echo 23:5» 89:83: ecu 82? 3862: £53.55: 0:32 vé 22m...

27

both parameters (181 days and 290 kg N ha-l). The predicted values follow the same
pattern as the AN OVA means, with NTS and HTS predicted to mineralize at a

greater rate than CT and ATS.

Among sample dates, there was more consistency in No than in MRT. The April
1995 samples had the lowest MRT (61 days) and a relatively low No (159 kg ha'1
soil), while November and April 1996 have the highest (220 kg ha-1 and 170 days for
November and 180 kg ha'1 and 144 days for April 1996). The relatively high MRT
values in November 1995 and April 1996 can be explained in part by greater
immobilization at those dates. Treatments with high predicted mineralization have
high values for No and MRT whereas dates with high mineralization have low
values of No and MRT. This suggests that-No is relatively more important in
establishing treatment effects, while MRT is relatively more important in
establishing date effects. Management may have a greater effect on the quantity of
mineralizable N (N 0), whereas its quality and the availability (MRT) may be more

strongly controlled by seasonal differences.

LFL 1994

Figure 1.1 shows the accumulation of N03 at 10, 30, 70, and 150 days in the 1994
LFL incubations. The top of each bar segment represents the accumulation of
inorganic N from the start of the incubation to the corresponding incubation time.
The error bars represent the mean N accumulation across incubation times (the

mean used in the AN OVA) plus and minus one standard error.

28

1994 LFL ist Year Corn
[:1 150 day incubation
300 w [2:] 70 day incubation

        

 

 

     

   

 

       

- 30 day incubation
- 10 day incubation
c o Incubation Mean (SE)
8 200 T
in, i
to = I i“
o 2 l a I I
.E a: I
2 x 100 1 -
z i: “i a? H ’ ' i '3' a
.. ‘“‘ .4 .
d Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover
1994 LFL Soybean
00 - C: 150 day incubation

[:3 70 day incubation
- 30 day incubation
- 10 day incubation
200 .1 o Incubation Mean (SE)

N Mineralized
kg ha.1

   
        

 

     

! I I III E [El
5 a a H a .955
0 J - z. -.
Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover
1994 LF L Wheat

CZ] 150 day incubation

 

 

 

   

 

 

 

300 'i :21 70 day incubation
_ 30 day incubation
- 10 day incubation
.0 o lncubationMean (SE)
0 F 200 -+
g .
s 2
.2 3
2 x
Z 100 . I I
- E
0 d Apr Jun Nov Apr Jun Nov Apr Jun Nov Apr Jun Nov
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover

Figure 1.1 Accumulation of inorganic N during laboratory incubations
at 250 in the 1994 LFL data set. Error bars represent
standard errors for the mean of the 10, 30, 70, and 150
day sub-samples. Units are based on a depth of 25 cm.

29

Both AN OVA models (3 crop x 3 date and 2 crop x 5 date) detected significant
fertilization, crop, and date effects and significant fertilization x date and crop x date
interactions. There was 20% greater NMP in the Compost treatment (86 kg N ha'l)
than in the Fertilizer treatment (72 kg N ha‘l), and there was 13% greater NMP in
the 1st year corn plots (83 kg N ha'l) and 9% greater NMP in the wheat plots (80 kg
N ha'l) than in the soybean plots (72 kg N ha'l) based on the April, June, and
November samples. The increased mineralization associated with the Compost
treatment was greatest immediately following compost applications. Compost was
not applied in the Soybean plots until November (before wheat), and this accounts
for part of the difference between crops. However, differences in residue quality
were also important. Soybean follows corn or annual ryegrass, whereas wheat
follows soybean, and 1st year corn follows either red clover (cover) or weedy fallow

(no cover), both of which have a high N content when incorporated in the spring.

The mineralization curves for the 1994 LFL data set fit a single-pool, 1st order
exponential model and produced reasonable parameter estimates, except in April
1994, when immobilization tended to produce inflated estimates of No and MRT.
As with the LTER data, these regression models produced 70 and 150 day predicted
values (not tabulated) that would provide the same interpretation of treatment and

date effects as the AN OVA means.

30

[:3 150 day incubation
300 1 1st year Corn :3 70 day incubation
- 30 day incubation
- 10 day incubation
o Incubation Mean (SE)
200 4

       

 

   

E
2:; I
0
E a I I I II
5 100i ll... gal—I a I a 5'5 3
Eggs -559 a! an
d Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover

aoo - Continuous Corn CI] 150 day incubation

[:21 70 day incubation
- 30 day incubation
- 10 day incubation

o Incubation Mean (SE)

to
8

100-4

N Mineralized
kg ha ‘

 

.m-

   

Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr
Fertilizer No Cover Fertilizer Cover Compost No Cover

Apr Jul Sep Nov Apr
Compost Cover

[2 150 day incubation
30° 1 Wheat [:3 70 day incubation
- 30 day incubation
- 10 day incubation

o Incubation Mean (SE)

§

     
   

N Mineralized
kg ha-1

‘2’

[III-

   

I
H 5

Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr
Fertilizer No Cover Fertilizer Cover Compost No Cover

  

 

 

,q-

 

Apr Jul Sep Nov Apr
Compost Cover

Figure 1.2 Accumulation of inorganic N during laboratory incubations
at 25C for the 1995-96 LF L data set. Error bars represent
standard errors for the mean of the 10, 30, 70, and 150

day sub-samples. Units are based on a 20 cm sample
depth.

31

LFL 1995-96

In the 1995-96 data set (Figure 1.2) there was 25% greater NMP in the Compost
treatments (82 kg N ha'l) than in the Fertilizer treatments (74 kg N ha'l), and 29%
greater N accumulation in the two April samples (89 and 80 kg N ha'l) than in July,
September, and November 1995 samples (68, 66, and 64 kg N ha'l, respectively).
There was greater NMP in the plots with cover crops (75 kg N ha-‘) than in the plots
without cover crops (72 kg N ha'l) and greater NMP in the wheat plots (77 kg N ha-
1) than in the continuous corn plots (71 kg N ha'l), although 1st year com (73 kg N
ha-l) was not significantly different from either. The highest mineralization
potentials were achieved in wheat following soybean (97.5 kg N Ha'l in April 1995)
and in 1st year corn with cover following wheat (93 kg N Ha'l in April 1995, 90.5 in
April 1996). These figures compare to an average of 78.3 kg N ha'lfor the other

treatments in April 1995 and 78.8 in April 1996.

The first order model provided a good fit to the data except in November, when
immobilization resulted in inflated parameter estimates. The Compost treatment
produced a significantly greater No estimate (165 kg N ha-l) and significantly lower
Ni and MRT estimates (20 kg N ha'1 and 70 days) then the Fertilizer treatment (No
= 146 kg N ha-l, Ni = 23 kg N ha'l, and MRT == 90 days). Wheat was associated
with a lower Ni estimate (15 kg N ha'l) than 1st year ( 31 kg N ha-l) or continuous
com (27 kg N ha-l), but there were no significant differences between the crop or
cover treatments in No or MRT. The strong immobilization experienced in

November may have been the result of root decomposition. Measurements of root

32

biomass taken in a related study (W illson etal., 1997) showed a dramatic decrease

in identifiable roots between September and November.
LFL 1995-96 Rotation Effect

When we examine the Fertilizer rotation as a whole (Table 1.5), we find that 2nd
year com, has a significantly lower mineralization rate than 1st year corn, or wheat
on average (58.9 kg ha-1 compared to 65.3 and 67.3 kg ha'1 respectively). The
average mineralization potential associated with the continuous corn treatment
(62.4 kg ha'l) is approximately equal to the average mineralization potential across
the rotation as a whole (62.4 kg ha’l). The use of cover crops significantly increased
NMP overall, but not in any particular crop. There is a significant crop x date
interaction in the rotations (not tabulated) that seems to be related to the high
November NMP relative to April 1996 in the soybean treatments. These plots were
tilled and planted to wheat in November, which may have stimulated the early

decomposition of the soybean residue.

Table 1.5 Average accumalation of inorganic N in 10, 30, 70, and 150 day
incubations for the LFL Integrated Fertilizer rotation in the
1995-96 data set.

 

 

Cover No Cover Combined
Rotation kg ha '1 kg ha '1 kg ha '1
soybean 64.3 ab T 59.4 ab 61.9 ab
wheat 69.1 a 65.5 a 67.3 a
lst year corn 68.4 a 63.2 ab 65.8 a
2nd year com 59.5 b 58.2 b 58.9 b
Rotaiton Average 63.9 A 60.7 B 62.3
Comtinuous corn 64.7 ab 60.0 ab 62.4 ab

 

T Based on a 20 cm sample depth

:t Values with the same letter are not significantly different at o. = 0.05.
Capital letters designate differences within rows, non-caps
designate differences within columns

33

DISCUSSION:

The Seasonality of N Mineralization Potential

Nitrogen mineralization potential was greatest in spring and smallest in the early fall
in almost all of the treatments observed in this study. Although Bonde and Rosswall
(1987), El-Harris et al. (1983), and Franzluebbers et al. (1994) examined different
cr0pping systems at different sample dates, they also found that NMP was highest
before planting and after harvest and lowest near physiological maturity. Thus as a
general rule we can say that NMP increases when plant residues and other material
is made available through tillage or senescence, but decreases as N is mineralized

and sequestered in crop tissue.

Treatment effects on N mineralization potential

The 1994 LTER data set presents a contrast between two reduced chemical input
treatments with comparatively high mineralization potentials and two conventional
treatments with low mineralization potential. We identified several factors that
could have contributed to this difference, including the recent use of leguminous
cover crops and the historical presence of wheat followed by hairy vetch in the
rotation. The reduced levels of chemical fertilization and weed control in these
treatments may also have had some effect on mineralization potential by selecting
for more N efficient plant and microbial communities. This explanation has been

proposed to describe the greater N-efficiency of a low input cash grain cropping

34

system relative to a conventional cash grain cropping system at the Rodale Institute

(Doran et al., 1987, Harris 1993, and Wander et al., 1994)..

The HTS and ATS treatments in the 1995-96 LTER data set demonstrate that a
dramatic increase in mineralization potential can be achieved after only a few years
of undisturbed fallow in this climate, and that this increase can be eliminated just as
quickly after the resumption of tillage. These changes in NMP appear to involve
both the direct effect of disturbance in promoting decomposition, and its indirect
effect on the plant community. The diverse, perennial, plant community found in
the HTS treatment seems to enhance the storage of labile sources of organic N,
probably in the form of fine roots and rhizosphere organisms, which increases the

potential of this soil to mineralize N when disturbed.

A similar effect could be involved in increasing the supply of N following wheat
production. There is a 20 month period between tillage for wheat and tillage for
corn in the com (com) soybean wheat rotations that we examined. During this time
there is nearly continuous plant production, even in systems without cover crops.
Additional study will be required to determine how important this fallow period is in
increasing N availability to corn, although it has the obvious effect of increasing the

time for red clover production in systems with cover crops.

We recorded an average increase in inorganic N accumulation of 22% after two
years of compost addition and 29% after three years. The amount of compost N

applied differed each year. In 1993, all of the plots received 43 kg N ha". In 1994,

35

the wheat plots received 209 kg N ha'land the corn plots received 697 kg N hat-1. In
1995, the wheat plots received 489 kg N ha'1 but the corn plots received only 67 kg N
ha'l. The total input of N in the Compost plots was approximately double the input

in the fertilized plots over the first three years of the experiment. .

The only significant cover crop effect was a temporary increase in NMP associated
with the red clover following wheat. It is possible that there will be greater

divergence between the cover crop and no cover treatments over time.

The Fertilizer rotation provides a useful window on the impact of crop sequence and
rotation on mineralization potential. Mineralization potential is highest in 1st year
corn and wheat and lowest in 2nd year com and soybean, with continuous corn
somewhere in the middle. While small, these differences may be important in
explaining the functioning of the com-com-soybean-wheat rotation. Seasonal levels
of inorganic N are at their highest in 1st year corn and decrease each year until they
reach their lowest point under wheat. First year corn provides the greatest yield, and
the yield of 1st year corn following red clover is equally high in the fertilized and

compost amended plots (Jones et al, 1998).

The LFL and LTER data presented in this study strongly suggests that the type of

vegetation and the time without disturbance are important determinants of NMP

36

Evaluation of the Methods used in this study

The repeated measures analysis of variance provided a more robust analysis of
management effects than the linear regression. Because they are based on the
average accumulation in 10, 30, 70 and 150 day incubations, the least square means
are inherently weighted toward long term results, but also provide some information
about the extent of immobilization. It would be easy to re-weight the analysis

mathematically by choosing different incubation lengths.

The non linear regression analysis was weakened by immobilization, the effect of
which were most apparent in the April, 1994 and November 1995 sample dates.
Immobilization can be a problem in any kinetic analysis of Net N mineralization,
and we echo the suggestion of Smith (Ref) that samples early in the incubation
should be avoided. We observed immobilization well past the 4th week in these
incubations. Although kinetic models have been proposed that allow for a lag in
mineralization (Bond and Lindberg, 1988 White and Marinakas, 1991), use of the
more conventional exponential models to study short term N mineralization will

‘ require the measurement of gross N mineralization.

The regression parameters No and MRT were extremely sensitive to immobilization,
and were often difficult to interpret. The 70 and 150 day predicted values were less
effected by immobilization and produced a similar separation of the treatments to
that provided by the AN OVA least square means even when the parameter

estimates were distorted. It is not clear, however, what advantage can be gained by

37

performing regression analyses when the parameter estimates themselves are not

interpretable.

38

REFERENCES

Bonde, TA. and T. Lindberg. 1988. Nitrogen mineralization kinetics in soil during
long-term aerobic incubations, a case study.

Bonde, TA. and T. Roswall. 1987. Seasonal variation of potentially mineralizable
nitrogen in four cropping systems. Soil Sci. Soc. Am. J. 51:1508-1514.

Campbell, C.A., Y.W. Jame, A. J alil, and J Schoenau. 1997. Use of hot KCl-NH4-N
to estimate fertilizer N requirements. Can. J, Soil Sci. 77: 161-166.

Cole, J .W.L. and J .E.Gissel. 1966. Applications of unuvariate analysis of variance
to repeated measures experiments. Biometrics 22:810-828.

Doran, J .W., D.G. Fraser, M.W. Culik, and WC. Liebhardt. 1987.1nfluence of
alternative and conventional management on soil microbial processes and
nitrogen availability. J Alt. Agn'c. 2299-106.

El-Harris, M.K., Cochran, V.L, Elliot, LR, and Bezdicek, DP. 1983. Effect of
tillage, cropping, and fertilizer management of soil nitrogen mineralization
potential. Soil Sci. Am. J. 47:1157-1161.

Franzluebbers, A.J., F .M. Hons, D.A. Zubuerer. 1994. Seasonal changes in soil
microbial biomass and mineralizable C and N in wheat management systems.
Soil Biol. Biochem. 26: 1469-1475.

Gregoire, TC. and O. Schabenberger. 1996a. A non-linear mixed effects model to
predict cumulative bole volume of standing trees. J. Appl. Stat. 23:25 7-271.

Gregoire, TC. and O Schabenberger. 1996b. Nonlinear mixed-effects modeling of
cumulative bole volume with spatially correlated within-tree data. J. Agric. Biol.
Environ. Stat. 12107-119.

Gregorich E.G., M.R. Carter, D.A. Angers, C.M. Monreal, and B.H. Ellert. 1994.
Towards a minimum data set to assess organic matter quality in agricultural soil.
Can. J. Soil. Sci.74:367-385.

Hess, T.F., and SK. Schmidt. 1994. Improved procedure for obtaining statistically
valid parameter estimates from soil respiration data. Soil Biol. Biochem. 27:1-7.

Jones, M.E., R. Harwood, N. Dehne, and J Smeenk. (in press) Cr0p yield and net
economic return of corn, soybean and wheat as affected by position in the
rotation, fertility source and cover crop. J. Prod. Agric.

Littell, R.C., G.A. Milliken, W.W. Stroup, and RD. Wolfinger. 1997. SAS system
for mixed models. SAS Inst. Inc. Cary, N .C.

KBS LTER. 1998 [Online]. KBS Long-Term Ecological Research (LTER) Project.
Available at http://lter.kbs.msu.edu/, Dec. 10, 1998.

39

Maynard, D.G., J .W.B. Stewart, and J .R. Brittany. 1983. Sulfirr and nitrogen
mineralization in soils compared using two incubation techniques. Soil Biol
Biochem. 15:251-256.

Motavalli, P.P, Frey, S.D., Scott, NA. 1995. Effects of filter type and extraction
efliciency on nitrogen mineralization measurements using the aerobic leaching
soil incubation method. Biol. F ert. Soils 20: 197-204.

Oberle, S.L. and DR. Keeney. 1990. Factors influencing corn fertilizer N
requirements in the northern corn belt. J Prod Agric. 3:52 7-534.

Peters, S., R. J anke, and M. Bohlke. 1992. Rodale’s farming system trial 1986-1990.
Rodale Inst., Kutztown, PA.

Robertson, G.P., K.L. Gross, and RR. Harwood, 1997 [online]. KBS LTER Site
Data Catalog Aboveground Net Primary Production (KBSOl9). Available at
h_ttp://lter.kbs.msu.edu/Npp/Kbscore/KbSOI9.html. Updated Dec. 20, 1997

Robertson, G.P., K.M. Klingensmith, M.J. Klug, E.A. Paul, J .R. Crum, and B.G.
Ellis. 1997. Soil Resources, microbial activity, and primary production across an
acricultural ecosystem. Ecological Applications 7: 158-170.

Robinson, J .A. 1985. Determining microbial kinetic parameters using nonlinear
regression analysis: advantages and limitations in microbial ecology. Adv.

Microb. Ecol. 8:61-114.

SAS Inst. Inc. 1988. SAS/ STAT user’s guide Vol. 2, version 6, fourth edition. SAS
Inst. Inc., Cary, NC.

SAS Inst. Inc. 1997. SAS/ STAT software changes and enhancements through
release 6.12. SAS Inst. Inc., Cary NC.

Smith, S.J., L.B Young, and GE. Miller. 1977. Evaluation of soil nitrogen
mineralization potential under modified field conditions. 8011 Scr. Soc.Am J
41 :74-76.

Stanford, G. and SJ. Smith. 1972. Nitrogen mineralization potentials of soils. Soil
Sci. Am. Proc. 36:465-472.

Wander, M.M., S.J. Triana, B.R. Stinner, and SE. Peters. 1994. Organic and
conventional management effects on biologically actrve sorl orgamc matter

pools. Soil Sci. Am. J. 58:1130—1139.

White, CS. and Y.D. Marinakis.1990. A flexible model for quantitative
comparisons of nitrogen mineralization patterns. Biol. Fert. Sorls 11:239-244

Willson, T.C., E.A. Paul, R.R. Harwood, A.J.M. Smucker, EM. Parker and D.
Harris. 1997. Carbon 13 analysis of below ground productrvrty and $011
respiration. ASA Abstracts, Anaheim CA October 1997.

40

CHAPTER 2

Managing Biologically Active Soil Organic Matter Fractions for
Sustainable Crop Production

ABSTRACT

We sampled agricultural and plant successional experiments in Southwest MI over
a two year period to study the effects of alternative management practices on
microbial biomass and macroorganic matter. Microbial biomass was determined by
the chloroform fumigation, incubation method, and macroorganic matter was
defined as the C and N content in sand-size soil separates. Microbial C decreased
during a drought in 1994, and was greater in treatments receiving compost than
fertilizer in 1995. Microbial C:N ratio was lower in July and November (6.0) than in
April and September (7 .3) , and lower in successional treatments without tillage
(5.2) than in agronomic treatments (6.7). Microbial biomass accounted for 2.6% of
soil C and 4.9% of soil N in 1995. Macroorganic matter accounted for 19.7 % of C
and 14.8% of N. Its C:N ratio was 20.8 in a never tilled successional treatment and
16.0 in other soils, 17.0, on average, in the 250 — 1000 um fraction and 15.5 in the 53
- 250 um fraction. Macroorganic C increased after compost additions, and was
greater in successional plots without tillage than in tilled treatments. It was strongly
correlated with N mineralized in 70 and 150 day aerobic incubations. Microbial
biomass was not. There was greater N mineralization per unit macroorganic matter

in April and November than in September and October. This suggests that

41

macroorganic matter could be used to estimate N mineralization if combined with

information about recently deposited plant residues.

INTRODUCTION

The need to develop cropping systems that better synchronize the availability of
inorganic N with crop demand is a major challenge facing farmers and agricultural
researchers. Sufficient inorganic N must be available during the growing season to
facilitate high yields, but an accumulation of excess soil N03' can lead to increased
N-loss and environmental contamination, particularly when precipitation exceeds

evapotranspiration.

The use of inorganic N fertilizer provides a partial solution to the problem of
synchrony in that the amount and timing of fertilizer applications can be controlled.
Logistical and economic constraints often limit farmers to a single application, and
even when split applications are possible, losses are inevitable. J okela and Randall
(1997) applied l5N labeled fertilizers at the time of planting and at the V8-stage in
continuous corn in the Northern US Corn belt and found that losses from the plant
soil system were 13 - 27 % (depending on application rate) regardless of the time of
application. Similar losses have been reported across a range of crops and
conditions (Russelle et al., 1981, Kitur et al., 1984, Meisinger et al. 1985, Reddy and

Reddy, 1993, Tirnmons and Baker, 1992).

Fertilizer N is often applied at unnecessarily high rates, leading to increased N loss.
Keeney (1987) and Meisinger (1984) note that fertilizer recommendations are often
made on the basis of insufficient information about the supply of N from plant and

soil resources. Libra et al. (1987a, 1987b) and Hallberg (1986) estimate that 50-80 kg

43

N ha'1 arable land is leached into groundwater each year in the Big Springs
watershed in Iowa, and attribute much of this flux to a failure to account for N
mineralized from legumes and animal wastes in estimating crop needs. Existing
recommendation systems are often ignored or misused. In a study of irrigated com
production in Nebraska, Schepers et al. (1986) observed that farmers overestimated
their yield by an average of 32 bu acre-1, resulting in an average over-fertilization of
50 kg N ha-l. Although Hallberg (1987) notes that farmers are becoming increasingly
aware of the issue of groundwater contamination, there is usually much more

incentive to prevent under- than over-fertilization.

Biologically based cropping systems use a combination of organic amendments and
biological N2 fixation to increase the N supplying capacity of the soil. Such systems
have the potential to provide better synchronization with crop demand than
conventionally fertilized systems, if only because inorganic N is released more
gradually over the grong season. Mineralization is affected by temperature and
moisture availability, so the supply of inorganic N can be expected to decrease
under conditions not favorable to plant growth. However, management practices
that increase mineralization during periods of high demand may also increase
mineralization when demand is low and the potential for loss is high. Dou et al.
(1995) found that the incorporation of hairy vetch as a green manure before corn
production resulted in elevated N03 levels throughout the season, and a higher N03
level at the last fall sample date than was observed for fertilized corn following a

non-legume fallow. In a similar study, Brown et al. (1993) found that late fall N03

levels can increase in both tilled and no-till systems if the decomposition of vetch is

delayed by lack of soil moisture.

There is a great potential to use crop and cover crop sequences to enhance the
synchrony of both biologically and chemically fertilized systems (McGill and Myers,
1987, Pierce and Rice, 1988, Myers et al. 1994). Winter cover crops have been
shown to reduce late fall N03 levels in a variety of systems (Smith et al., 1987,
Solberg, 1995, Philips and Stopes, 1995), while the incorporation of these residues
may either increase or decrease spring N03 levels depending on the extent of
immobilization (W egger and Mengal, 1988). Crop sequences also have
consequences on N availability, although increased N availability may account for
only a small portion of the increased yields commonly associated with crop

rotations that include legumes (Baldock et al., 1981 , Hesterman et al., 1987).

We define nitrogen mineralization potential (NMP) as the intrinsic ability of the soil
to supply inorganic N through mineralization over time. Nitrogen mineralization
potential is constantly changing in response to the accumulation and decomposition
of organic materials, and can be altered by managing the inputs of organic material,

and by changing the conditions under which decomposition occurs.

In a previous paper, (W illson et al., 1999) we measured NMP based on the
accumulation of inorganic N in 150 and 310 day aerobic incubations of soil from
agricultural and extended fallow treatments at the Kellogg Biological Station in

Hickory Comets, MI. We found that NMP was highest in the spring and lowest in

45

fall, greater in unfilled perennial systems than in tilled annual systems, increased
with the addition of compost, and was greater in rotations including wheat and
legume cover crops than in com-soybean rotations without cover crops. These
findings confirm that NMP is sensitive to short term changes in agricultural

management.

In this paper, we use data from the same field experiments to examine the
relationship between NMP and the C and N content of microbial biomass and
macroorganic matter. Both of these pools are closely associated with N
mineralization: microbial biomass because mineralization is a microbially-mediated
process (Paul and Voroney 1984, Smith and Paul 1989), and macroorganic matter
because it contains much of the partially decomposed plant material that firels
mineralization (Gregorich et al., 1994, Hassink, 1995a, 1995b). The objective of this
paper is to determine how alternative management practices alter the size and
composition of these fractions and to determine whether these fractions can be
managed, or used to predict changes in NMP. Measurements of organic matter
fractions are relatively simple, and might be useful as a basis for fertilizer

recommendations.

MATERIALS AND METHODS

Soil samples were obtained from experimental treatments at the Living Field
Laboratory (LFL) and the Long Term Ecological Research (LTER) experiment at
Kellogg Biological Station in Hickory Comers, MI (42°24’N, 85°23’W). These
experiments are located adjacent to one another on a mixture of Kalamazoo loam
and Oshtemo sandy loam (Typic Hapludalfs). The climate is temperate, with warm
humid summers and cool moist winters. Precipitation on site has averaged 880 mm

yrl (1988-1997) evenly distributed throughout the year.

Composite soil samples consisting of 12-15 two cm diameter cores were obtained to
a depth of 25 cm in April, June, August, October and November, 1994 and to a
depth of 20 cm in April, July, September and November, 1995 and in April 1996.
The samples were cooled immediately, sieved through a 4mm screen and stored at
4°C during processing. Because of the change in depth, and because different

treatments were sampled, the 1994 and 1995-96 data sets are analyzed separately.

The KBS LTER site was established in 1989 to facilitate cross-disciplinary research
on the ecological interactions inherent in row-crop agriculture and related
ecosystems in the Midwestern United States. Treatments examined in this study
include four com-based rotations and three successional treatments. The agronomic
treatments included conventionally fertilized, com-soybean rotations with and
without tillage (conventional tillage (CT) and no-till (NT )), and com-soybean-wheat

rotations with hairy vetch and red clover as green manure and either reduced

47

fertilization (low-input (LI)) or no chemical inputs of any kind (zero input (ZI)). The
successional treatments include a mowed, but never-tilled, successional meadow
(NTS), a historically tilled, old-field succession (HTS), and an annually tilled
successional treatment (ATS). All four com-based treatments were sampled in
1994, while CT and the three successional treatments were sampled in 1995 and
April 1996. The agronomic treatments were planted to soybean in June, 1994 and
wheat in October 1994 after having been in corn with or without hairy vetch in 1993.
The CT and NT treatments had not been planted to wheat prior to 1994.
Management protocols for the LTER main site experiment can be found on the

intemet (KBS LTER, 1998 [online])

The LP L was established in 1993 to provide additional information about the effects
of alternative management strategies in sustainable agriculture. The LFL has a
factorial design that allows the separation of important management alternatives
such as rotation vs. monoculture, inorganic vs. organic fertilizer (compost),
broadcast vs. banded vs. non-chemical weed control, and the presence or absence of
cover crops. Each crop in the com-com-soybean-wheat rotation is grown each year.
This prevents the confounding of crop sequence effects with year to year variations

and facilitates the comparison of rotation vs. continuous com(J ones et al., 1998).

We sampled 1St year corn, soybean and wheat in 1994, from the Integrated Compost
(Compost) and Integrated Fertilizer (Fertilizer) management types with and without
cover crops. In 1995-96 we sampled the complete com-com-soybean-wheat rotation

along with continuous corn in the integrated fertilizer management type but only 1st

48

year corn, continuous corn, and wheat in the Compost management type. The
integrated fertilizer and Compost management types differ only in N source;
ammonium nitrate fertilizer and dairy manure - leaf compost. Both management
types utilize reduced tillage (chisel plow) and a combination of banded herbicides
and mechanical cultivation for weed control. Cover crops include red clover frost
seeded into wheat, red clover interseeded into 15t year corn and continuous com,

and annual ryegrass interseeded into second year corn.

Microbial biomass was determined using the chloroform fumigation, incubation
method described by Harris et al. (1997), including the partial control subtraction

equation for microbial C (equation [1]) developed empirically by Horwath et a1.

(1996),

[1] MBC = 1.73 Cf— 0.53 Cc

and the corresponding microbial N equation (equation [2]) of Harris et al. (1997),

[2] MBN = MBC (0.56 (Nf-ch)/(Cf-pCc) +.095)

where Cf and Nf refer to the net C02-C and NH4-N respectively produced in
fumigated soil over a 10 day incubation at 25°C, Cc and Ne refer to the net C and N

mineralization in a non-fumigated control under the same conditions and p and q
are the proportion of the non-microbial C and N in the control incubation that is

consumed in the fumigated sample. The value of q and p is assumed to be equal for

49

the purpose of this analysis, and is derived using equation [3] from Horwath et a1.

(1996).
[3] p = q = 0.29 (Cf/Cc) +0.23

All COz-C measurements were made by'titration of NaOH traps, and NOa-N and
NHo-N measurements were made on 1M KCl extracts using a Lachat automated

colorometric analyzer (Lachat Instruments Inc. Milwaukee, WI).

Macroorganic matter is defined as the total C and N content of sand-sized (53 —
1000 um) soil separates. Air dried soil was dispersed in 5% sodium polyphosphate
solution, shaken for 8 hours, and passed through nested sieves of 1000, 250, and 53
um using a flow of distilled water to ensure separation. The material remaining on
the 250 and 53 um sieves was analyzed for total C and N using a Carlo Erba N A
1500 Series 2 N/ C/ S analyzer (CE Instruments Milan, Italy) and defined as coarse

and fine macroorganic matter respectively.

Treatment and date effects on both microbial biomass and macroorganic matter
were analyzed using Proc Mixed of SAS (SAS Institute, 1997) given a randomized
complete block design with 4 blocks at the LTER and a split-split-plot randomized
complete block design at the LFL with management type (compost or fertilizer) as
the main plot, each crop as a subplot and the presence of cover cr0ps as a sub
subplot treatment. The location of the cover and no cover sub-subplots in the LFL is

not randomized within each crop subplot, so cover crop effects could be confounded

50

by positional effects. The correlations between treatment means for microbial
biomass C and N and each macroorganic matter fraction, and between these indices
and the N mineralization data of Willson et al. (1999) are generated using SAS

procedures (SAS Institute, 1988).

51

RESULTS

There were no significant differences in microbial biomass C due to management in
the 1994 data set, but there was a pronounced decrease in biomass in the June
sample relative to other sample dates (Figure 2.1). The relatively low June biomass
C coincides with a prolonged drought: the LTER weather station recorded only 1.4
cm of precipitation between May 1 and June 6. The greatest decrease in biomass C
from April to June (963 kg ha'1 or 70%) occurred in the LTER NT treatment.
Additional data taken at the time of sampling (not shown) shows that this treatment
retained more soil moisture to 25cm than the other agronomic treatments, but also
had the highest proportion of microbial biomass near the surface, where it is more

likely to encounter temperature and moisture stress.

1994 Microbial Biomass
1800 l Mcrobial c

.2...- e 0]) r
E] D [1

AprJun Aug Oct Nov AprJunAugOctNov AprJun AugOctNov AprJunAugOctNov
Conventional (CT) No-Till (NT) Low Input (Ll) Zero Input (Zl)

Microbial C
Kg ha'
J—l
J-i
j
J
J—i
J-i

J

§

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0

Figure 2.1 Microbial biomass C in the top 25 cm in the 1995 LTER
Soybean Treatments. Bars represent the standard error
of the mean.

52

The lack of a significant difference in microbial C between the compost (833 kg ha")
and fertilizer (754 kg ha") treatments is surprising given the large amounts of
compost applied. The 1994 wheat plots received 3150 kg C ha:1 in October 1993, the
com plots received 9000 kg C ha'1 in May 1994, and the soybean plots received 6500
kg C ha:1 in October 1994. (All plots also received 500 kg C ha'1 of compost in April,
1993.). These additions are larger than the above ground net primary productivity in
these treatments, and contain 2 to 7 times as much N as was applied in the

corresponding fertilizer treatment.

The effect of compost was much more pronounced in the 1995 data set (Figure 2.2).
There was an average of 728 kg C ha'l for the compost plots and 598 kg ha:1 in the
fertilizer plots for tlrel5t year corn, continuous corn and wheat treatments. The
difference was greatest in the 15t year corn plots (789 kg C ha-1 for compost, 504 kg
hal for fertilizer) even though these plots received the least compost over the life of
the experiment (4950 kg C ha'l, compared to 6960 kg C ha-1 for wheat and 10800 kg
C ha-l for continuous corn). Microbial C tended to increase in the warmer months.
It was significantly greater in July or September 1995 than in April 1995 or April

1996 in this data set.

Microbial C and N were significantly greater in the never tilled succession (NT S)
than in the other LTER treatments in the 1995-96 data set, but this treatment also
had a greater total C content (1.7%) than the others (1.0%). Expressed per unit C,

there are no significant differences in microbial C or N. The HTS and NTS

53

Microbial C and N

Microbial C and N
kg ha'1

Microbial C and N
kg ha‘

kg ha"

1 600
I

1200 ~

8

§

1 st Year Corn

 

1600 '1

1200 .1

§

L

400

 

0..

1600 'l
1200 1
800 4

ml

 

Apr Jul Sep Nov Apr
Fertilizer No Cover

:21 Microbial C
- Microbial N

 

Continuous Corn

Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr
Fertilizer No Cover

Apr Jul Sep Nov Apr
Fertilizer Cover

 

 

[:1 Microbial C
- Microbial N

LFL Wheat

Apr Jul Sep Nov Apr
Fertilizer No Cover

 

% Ir

 

 

 

 

 

Fertilizer Cover

:I Microbial C
- Microbial N

 

Apr Jul Sep Nov Apr
Fertilizer Cover

 

 

 

 

 

 

Apr Jul Sep Nov Apr
Compost No Cover

 

Compost No Cover

 

 

Apr Jul Sep Nov Apr
Compost No Cover

 

11

 

II IIIII

Apr Jul Sep Nov Apr
Compost Cover

 

 

 

 

 

 

 

 

 

Apt Jul Sep Nov Apr
Compost Cover

 

Apr Jul Sep Nov Apr
Compost Cover

Figure 2.2 Microbial biomass C and N content in the top 20 cm in the
1995-96 LFL Samples. Bars represent the standard error
of the mean.

54

treatments have lower microbial C:N ratios (5.0 and 5.4) than any of the other
treatments (6.3 to 7.2) in the 1995-96 data set (Table 2.1). The C:N ratios in the
HTS and NTS treatments were relatively stable whereas those in the tilled
treatments were significantly lower after tillage in July and after senescence in
November than in the other sample dates. These C:N ratios suggest that the
microbial community is shifted toward bacterial production when fresh residues are

available (July and November), whereas there are relatively more fungi at other

sample dates.

The greatest changes in macroorganic matter in the 1995-96 data set (Figure 2.3) are
associated with the large compost inputs following soybean (October 1994), wheat
(October 1993) and 1st year corn (May 1994). Macroorganic C increased by 1300kg

ha-1 and macroorganic N increased by 50 kg ha'1 in the November LFL soybean

 

 

 

Table 2.1 Date and treatment effects on microbial C:N ratios in the 1995-96 data set
Month

Treatment April 1995 July September November April 1996 Average

LTER NTS 1* 6.4 b: 5.2 4.8 c 3.8 b 4.5 b 5.0 b
HTS 5.4 b 5.4 6.0 b ' 5.3 ab 4.9 ab 5.4 b
ATS 7.0 ab 5.6 7.7 a 6.4 ab 5.0 ab 6.3 ab
CT 8.5 a 5.6 7.4 a 6.8 a 7.6 a 7.2 a

LFL Compost 7.4 6.6 6.5 b 6.0 5.9 b 6.5
Fertilizer 7.9 5.7 7.9 a 6.0 7.3 a 6.9

All Data 7.4 A§ 5.9 B 7.0 A 5.9 B 6.3 B 6.6

 

1‘ Abbreviations: NTS = Never Tilled Succession HTS = Historically Tilled Succession
ATS = Annually Tilled Succession CT = Conventionally Tilled Wheat

1: Values with the same letter are not significantly different (a=0.05).
§ Capital letters indicate comparisons between dates.

55

1994 LF L 1st Year Corn

 

6000 .I

  

 

4000 4

Macroorganic C
kg ha '

    
 

    

2000 l

 

    

 

10000 1 :3 Coarse Macroorganic C 500
8°00 ' - Fine Macroorganic N l
Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov

[:1 Fine Macroorganic C
- Coarse Macroorganic N
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover
1994 LF L Soybean

 

     

         

           

 

   

  

 

     

10000 . C23 Coarse Macroorganic C 500
[2:1 Fine Macroorganic C
0 3°00 ‘ - Coarse Macroorganic N 400
.9 - Fine Macroorganic N z
E '7 6000 - 300 1:27
a 2 a
g .9 4000 - I l l l l l l 200 3
g l l -
°°°I'I'l"l'I' I'fffl' I'ffl'I' I'I'I'I'I' "’°
0 Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov Apr Jun Aug Oct Nov 0
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover
1994 LFL Wheat
10000 - 500
1: Coarse Macroorganic C
:3 Fine Macroorganic C
800° ‘ - Coarse Macroorgainc N ‘00

 

    

 

Macroorganic C
kg ha"
3

      

   

 

°°°°r “(m I ll r ,
.‘rr ”r ”r M fr

AprJunAugOctNov AprJunAugOctNov AprJunAugOctNov AprJunAugOctNov
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover

       

Figure 2.3 Two size classes of macroorganic Matter C and N in the
top 25 cm of the1994 LFL Treatments. Bars represent the
standard error of the mean for each treatment

56

with compost samples relative to the April through October samples. This
difference is equivalent to 23% of the applied compost C and 26% of the applied
compost N. Similarly, there is an increase of 1400 kg ha'I C and 115 kg ha:1 N
between the April and June samples in the lSt year corn with compost plots, which
representsl7% and 19% of the applied compost C and N in those treatments. Much
of this increase can be explained by the macroorganic matter content of the compost
itself. We performed a size fractionation procedure directly on a sample of compost
from the same source as applied in 1994 and found that 46% of compost C and 33%

of compost N separated in the 250-1000 urn fraction and an additional 10% of

compost C and 8% of compost N separated in the 53-250 um fraction. The
treatments that did not receive compost in 1994 averaged 4900 kg ha-1 macroorganic
matter C and showed only small changes during the season. There were no

treatment or date effects on macroorganic matter C or N in the 1994 LTER data set.

Among the four LTER treatments sampled in 1995, NTS had the highest amount of
macroorganic C, and the widest C:N ratio (Figure 2.4). It had about the same
amount of macroorganic N as the HTS treatment. The HTS treatment had
significantly greater macroorganic C and N than either of the tilled treatments (ATS
and CT), but the difference was much greater in the coarse fraction (48% more C

and 53% more N) than in the fine fiaction (15% and 18% respectively). The

difference in macroorganic C between the NT 8 and HTS treatments is primarily in

the fine fraction. The coarse fraction may be more sensitive to recent changes in

management, such as the conversion from tilled agriculture to succession without

57

tillage in the HTS treatment, whereas the fine fraction accumulates more slowly and

is thus more indicative of long term management.

LTER Wheat and Successional Treatments

 

 

 

 

 

 

 

10000 - :3 Coarse Macroorganic C _. 500
D Fine Macroorganic C ll 1,
- Coarse MacroorganicN %
0 300° ‘ - Fine Macroorganic ill " ‘00 z
-‘-’ L .2
c .— 300 =—
(a I (0'
92 92
o c» 0 or
go: r 200 88
g or
2
r- 100
l- 0

 

 

 

 

 

o .
Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr
Conventional Historically Tilled Annually Tilled Nerver Tilled
CT HTS ATS NTS

Figure 2.4 Macroorganic C and N in two size fractions in the top 25
cm of LTER wheat and successional treatments in 1995-
96. Bars represent standard errors of the mean.

The NTS treatment had significantly wider C:N ratios for both coarse (23.4) and fine
(19.0) macroorganic matter than the other LTER treatments. The NT soybean
treatment and the other successional treatments (HTS and ATS) also had higher
coarse fraction C:N ratios (17.9, 17.2, 18.5) than the tilled agronomic treatments
(16.3 for CT soybean, 16.9 for CT wheat, 15.7 for L1, and 16.8 for 21). The two
low-input treatments, LI and 21 had the lowest fine fraction C:N ratios (15.9, 15.6)
and the lowest total macroorganic matter C:N ratios (15.8, 16.1) at the LTER. It
would appear that the C:N ratio of plant residues play a role in generating the

observed treatment differences. The ATS and NTS treatments are dominated by

58

relatively low diversity of grass species whereas vegetation in the HTS treatment is

more diverse and has a higher proportion of legumes.

The LFL treatments (Figure 2.5) continued to show significant effects of compost on
macroorganic matter in 1995-1996. The compost treatments contained an average
of 24% more macroorganic C and N than the fertilizer treatments. The coarse
fraction accounts for 2/ 3 of this difference. All of the treatments in the 1995-96 data
set show a sharp decrease in macroorganic matter in November relative to the other
sample dates. While macroorganic matter is expected to reach a minimum in the
fall, it is possible that some of this decrease may have been caused by the necessity
of removing snow from the surface of certain treatments before sampling, which

may have removed some of the surface soil.

The C:N ratios at the living field laboratory were extremely consistent throughout
the study. While there were no significant differences between treatments or dates,
there were significant differences between the size fractions. The C:N ratio of the
fine fraction (15.28) was significantly lower than that of the coarse fraction (16.52)
and was less variable (SD: 0.82 for fine macroorganic matter compared to
SD=1.73 for the coarse fraction). These differences are consistent with the idea that
newer macroorganic matter fragments may be larger, and have a higher C:N ratio,

than older, finer, macroorganic matter.

Table 2.2 shows the coefficient, r, for the correlations between the organic matter

pools discussed in this paper and the N mineralization measurements performed in

59

LF L 1st Year Corn

    

 

 

 

 

 

 

 

 

 

 

 

10000 I ‘ — 500
C21 Coarse Macroorganic C
1:: Fine Macroorganic C
0 300° 1 - Coarse Macroorganic N I] - 400
2 - Fine Macroorganic N z
c " 600“ ~ 300 '2-
8 g g2
g 3’ 4000 i i— 200 b 3’
m 0
2 2
2000 i - 100
o- = to
Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr ADV JU' SOP NOV Apr Apr Jul Sep Nov Apr
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover

LFL Continuous Com

 

 

 

 

 

 

 

 

 

 

 

 

10°00 ‘ 1:] Coarse Macroorganic C r- 500
:1 Fine Macroorganic C
3000 .l - Coarse Macroorganic N 1 FL 1- 400
(3 - Fine Macroorganic N F9 2
'E ,_ 2
g in 5000 4 ~ 300 5,“
h S g o:
g 2 4000 -1 — 200 g 9
(I
2 2
2000 a _ 100
0 j - 0
Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr JUI Sep NOV Apr Apr Jul Sep Nov Apr
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover
LFL Wheat
10°00 ‘ [:1 Coarse MacroorganicC F 500
11: Fine Macroorganic C
8000 - - Coarse Macroorganic N L 400

- Fine Macroorganic N

    

 

 

 

 

 

 

 

o
.9 g
E .. 6000 4 h 300 F
N '.'
8 f:- 32
t0
2 2
2000 J L 100
0 .1 - 0
Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr Apr Jul Sep Nov Apr
Fertilizer No Cover Fertilizer Cover Compost No Cover Compost Cover

Figure 2.5 Macroorganic Matter C and N in two size fractions in the
top 25 cm of LFL treatments in 1995-96. Bars represent
standard errors of the mean.

60

Willson et al. 1999. The strongest correlations are those between macroorganic
matter C (coarse and total) and the accumulation of inorganic N after 150 days of
incubation. Macroorganic N is also a good predictor of NMP, but microbial
biomass is not. The 10 and 30day N mineralization measurements are poorly
correlated with all other indices, due to the influence of immobilization early in the
incubation. The relationship between macroorganic C and 150 day N
mineralization changes over the season as the availability of residue N changes
(data not shown). Nitrogen mineralization per unit macroorganic matter was lowest
in October and September when most plants are near peak biomass and highest in
November and April when most plants have senesced. This is similar to the overall
pattern found for NMP in Willson et al. (1999), and is consistent with the idea that
macroorganic matter and fresh plant residues are the two most important sources of
mineralizable N in these soils.

Table 2.2 Correlations (r) between biologically active organic matter

fractions and cumulative mineralization measurements across
all treatments and dates.

 

 

 

Organic Matter Inorganic N accumulation
Fraction 10 days 30 days 70 days 150 days
Macroorganic C
Combined 0.12 NS 0.39 *” 0.58 m 0.64 't'
250- 1000 um 0.17 * 0.41 m 0.59 '“ 0.64 "’*
53 - 250 um 0.03 NS 0.30 m 0.47 “* 0.54 *"
Macroorganic N
Combined 0.16 ‘ 0.44 "* 0.57 m 0.59 ‘"
250-1000 um 0.19 " 0.45 m 0.57 "* 0.57 '"
53-250 um 0.08 NS 0.36 ”* 0.47 *“ 0.51 ”"
Microbial Biomass
C -0.13 NS 0.12 NS 0.21 “ 0.28 *“
N -0.17 " -0.12 NS 0.14 NS 0.41”"

 

‘. ”, and "* represent significance at the .05. .01 and .001 probability levels.

61

DISCUSSION
Pool sizes in relation to plant needs.

The N contents of the microbial biomass and macroorganic matter fractions relative
to above ground plant biomass, yield, and residue N is calculated in Table 2.3 The
minimum inorganic N throughput is an estimate of the quantity of inorganic N that
must be made available during the growing season to supply plant needs. It is based
on an overall uptake efficiency of 50% of available inorganic N (50% of fertilizer plus
mineralized N is captured in plant tissue). This is near the upper end of the range of
recovery efficiencies reported in fertilization studies (Keeney, 1982, J okela and

Randall, 1997).

Microbial biomass accounts for only 4% of Soil N in the fertilized 1st year corn
treatment and 8% in the compost amended and a native succession soils. The data
presented in Figure 3 shows that microbial N changed substantially between dates in
each treatment, with a maximum increase of 120 kg N ha'1 between April and July
in the C1 compost- plots, more than doubling the initial N content. Given that the
inorganic N pool never exceeded 45 PPM (117 kg N ha'l) in the top 20 cm during
this study, microbial immobilization would seem to be very important in modifying
inorganic N availability. The change in microbial N from April to July was smaller
following the incorporation of red clover in the 1st year corn, cover plots in 1995 than
following the incorporation of a higher C:N ratio “weedy fallow” in the no cover

plots. This suggests that low C:N ratio cover crops could be used to

62

Table 2.3. Selected N pools in 1995-96 relative to one another and to soil N.

 

 

 

 

 

kg N ha-1 % of Soil N
LFL 1“t year corn LTER LFL 15‘ year corn LTER
without cover without cover

Fertilizer Compost NTS Fertilizer Compost NTS
SOM 1- 2340 2470 2990 100 100 100
Microbial biomass 95 136 239 4.1 5.5 8.0
macroorganic matter 53-250 135 239 225 5.8 9.7 7.5
macroorganic matter 250-2000 154 265 170 6.6 10.7 5.7
macroorganic matter Total 289 504 395 12.4 20.4 13.2
Above ground biomass :t 248 209 68 10.6 8.5 2.3
Yield Removed 137 125 5.9 5.0
Residue Returned (1995) 1 1 1 85 68 4.7 3.4 2.3
Fertilizer / Compost 134 67 5.7 2.7
Residue Input (from 1994) 51 46 79 2.2 1.8 2.6
Minimum N Required § 496 418 136 21.2 16.9 4.5

 

1' Soil parameters are averaged across sample dates.
1 Plant parameters were measured at time of peak biomass (late summer) except yield (late fall).
§ The minimum amount of inorganic N required for uptake, assuming 50% efficiency.

63

decrease the extent of early season immobilization, thereby freeing more inorganic
N for plant uptake. A microbial turnover rate of 2-3 generations per year (Harris and
Paul, 1994) would result in an average of 250 —~ 370 kg N ha'1 cycled through the
microbial biomass in the LFL agronomic treatments, which is similar to the
minimum throughput of 240-500kg N ha'1 estimated to be needed to produce the

observed crop biomass.

Macroorganic matter represents 20% of soil N in the 1st year corn, compost
treatment, and 12 and 13% in the fertilizer and NTS treatments respectively. As
discussed previously, there were large increases in Macroorganic N associated with
the large compost additions in 1994, and there was a temporary decrease in most
treatments in November 1995. Date to date changes were relatively small,
otherwise, indicating that the decomposition of macroorganic matter is balanced by
the formation of new macroorganic matter for much of the season. The decrease in
Macroorganic N in November 1995 was equivalent to 100 kg N ha-1 on average and
a maximum of 214 kg N ha-l in the 1St year corn treatment with compost and cover.
It has been estimated that the steady state turnover time for macroorganic matter is
20-40 years. This would imply an annual turnover of 7-25 kg N ha'1 in the
treatments described in Table 2.3, which is 6-30% of the total amount of plant and

compost N added to the soil in these treatments.

Pool sizes in relation to N mineralization Potential

Willson et a1. (1999) found that nitrogen mineralization potential (NMP) was

enhanced by the addition of compost, the incorporation of legume residues, and the

64

conversion from annual to perennial vegetation as in the HTS and N T S treatments.
Nitrogen mineralization potential was greatest in the late spring (April - June), and
lowest in the early fall (September - October), although it often recovered by

November.

Like NMP, macroorganic C and N are enhanced by compost additions and are
much greater in the NTS and HTS treatments than in the tilled treatments.
However, the incorporation of residues did not significantly change macroorganic
matter C and N during in this study, whereas the incorporation of legume residues
increased spring and fall NMP in Willson et al. (1999), particularly in contrast to the

low NMP observed in September and October.

The correlation between macroorganic C and NMP found in this study, and in
Hassink (1995), suggests that macroorganic C may be useful as an aid to fertilizer
recommendations, particularly if combined with information about recently
incorporated residues. The measurement of macroorganic C should be sufficiently
simple to be included in routine testing, and would provide an objective method for
modifying fertilizer recommendations in cases where past management has resulted
in unusually high or low N supplying capacity. Macroorganic matter measurements
that combine size and density fractionation (Meijbloom et al., 1995, Gadisch et al.,
1996, Barios et al., 1996) and the anaerobic incubation of Waring and Bremner
(1964) have also proven to be excellent indicators of N mineralization, but would be

more expensive to implement in routine testing.

65

Microbial biomass was a less reliable indicator of changes in substrate availability
than either macroorganic matter or NMP. Microbial biomass not enhanced by cover
crops, increased with compost only in the 1995-1996 season and experienced a

decline in June 1994 that was not related to substrate availability.

Seasonal Changes in Microbial Biomass

There was a trend toward greater microbial biomass in the warmer months (J uly-
October) than the cooler months (April and November). This is consistent with the
data at other moist temperate sites (Kaiser and Heinmeyer, 1993, Kirchner et al.,
1993, Jorgensen et al., 1994, and Paul and Harris, 1997 [online]). The opposite
pattern was found for cool-season production systems in semi arid

environments(V an Gestel and Ladd, 1992, Bremer and van Kessel, 1992, Collins et
al., 1992, and Franzluebbers et al.,1994). Maxwell and Coleman (1995) also report
lower microbial biomass in the summer in a moist riparian habitat, but in that study
summer microbial biomass seems to have been limited by microfaunal predation.
While substrate availability is usually assumed to be the most important
determinant of microbial biomass, it is clear from these studies that other factors

have a greater influence on patterns of seasonal change.

We detected significant differences in microbial C:N in the 1995-96 data set, with
lower C:N ratios after tillage and plant senescence (July and November, 1995) in the
tilled treatments and lower C:N ratios at all dates in the untilled successional
treatments. The C:N ratio of the microbial biomass should reflect the species

composition, particularly the ratio of bacteriazfungi (J enkinson, 1976, Shen et al.,

66

1984, Sparling and Zhu, 1993). The C:N ratios of Fungi range from 4.5:1 to 15:1
while the C:N ratios of bacteria tend to be between 3:1 and 5:1 (Paul and Clark,
1996). Ifone assumes a C:N ratio of 10 for firngi and 4 for bacteria, the
bacteriazfungi ratio of the July and November samples overall would be
approximately 2:1 whereas the ratio in April and September would be 1:1 and the
ratio in the NTS treatment would be approximately 6:1 across all dates. Direct
microscopy of LTER treatments in 1991 and 1993 and of the LFL in 1994 resulted
in bacteriazfungi ratios ranging from 0.321 to 5:1 (Paul and Harris, 1997 [online]).
The high relative abundance of bacteria at this site contradicts the assumption of

Anderson and Domsch (1975) that agricultural soils are dominated by fungi.

67

REFERENCES

Anderson, TH. and KB. Domsch. 1975. Measurement of bacterial and fungal
contributions to respiration of selected agricultural and forest soils. Can. J.
Microbiol. 21:314-322.

Baldock, J .O., R.L. Higgs, W.H. Paulson, J. A. Jacobs, and W.D. Schrader. 1981.
Legume and mineral N effects on crop yields in several cr0p sequences in the
Upper Mississippi Valley. Agron. J. 73:885-890.

Barios, E. R.J. Buresh, and J .I. Sprent. 1996. Organic matter in soil particle size and
density fractions from maize and legume cropping systems. Soil Biol. Biochem.
28:185-193.

Bremer, E. and C. van Kessel. 1992. Seasonal microbial biomass after addition of
lentil and wheat residues. Soil Sci. Soc. Am. J. 56:1141-1146.

Brown, R. E., G.E. Varvel, and CA. Schapiro. 1993. Residual effects of interseded
hairy vetch on soil nitrate-nitrogen levels. Soil Sci. Soc. Am. J. 57:121-124.

Collins, H.P., P.E. Rasmussen, and CL. Douglas, Jr.. 1992. Crop rotation and
residue management effects on soil carbon and microbial dynamics. Soil Soc.
Am. J. 56:783-788.

Dou, Z., R. H. Fox, J. D. Totlr. 1995. Seasonal soil nitrate dynamics in corn as
affected by tillage and nitrogen source. Soil Sci. Soc. Am. J. 59:585-864.

Franzluebbers, A.J., F.M. Hons, and DA. Zubuerer. 1994b. Seasonal changes in
soil microbial biomass and mineralizable C and N in wheat management
systems. Soil Biol. Biochem. 26:1469-1475.

Gadisch, G, H. Imhof, S. Urquiaga, R.M. Boddey, and KB. Giller, 1996. Carbon
turnover and nitrogen mineralization potential of particulate light soil organic
matter after rainforest clearing. Soil Biol. Biochem. 28: 1555-1567.

Gregorich, E.G., M.R. Carter, D.A. Angers, C.M. Momeal, and B.H. Ellert. 1994.
Towards a minimum data set to assess soil organic matter quality in agricultural
soils. Can. J. Soil Sci. 75:367-385.

Hallberg, GR. 1986. From hoes to herbicides: Agriculture and groundwater
quality. J. Soil and water conserv. 412357-364

Hallberg, GR. 1987. Agricultural chemicals in ground water: extent and
implications. J. Altem. Agric. 223-15

Harris, D. R.P. Varoney, and EA. Paul. 1997. Measurement of microbial biomass
N:C by chloroform fumigation-incubation. Can. J. Soil Sci. 77:507-514.

Hassink, J. 1995. Density fractions of soil macroorganic matter and microbial
biomass as predictors of C and N mineralization. Soil Biol. Biochem. 27:1099-
1 108

68

Hesterman, O.B., M.P. Russelle, C.C. Schaeffer, and GR Heichel. 1987. Nitrogen
utilization from fertilizer and legume residues in legume-com rotations. Agron.
J. 79:726-731.

Horwath W.R., E.A. Paul, D. Harris, J, Norton, L. J agger, and K.A. Horton. 1996.
Defining a realistic control for the chlorogorm fumigation incubation method
using microscopic counting and l4C-substrates. Can. J. Soil Sci. 76:459-467.

J enkinson, D. S., 1976. The effects of biocidal treatments on metabolism in soil-IV.
The decomposition of fumigated organisirns in soil.Soil Biol. Biochem. 8:203-
208.

J okela, W. E. and G. W. Randall. 1997. Fate of fertilizer nitrogen as affected by
time and rate of fertilizer application. Soil Sci. Soc. Am. J. 61:1695-1703.

Jones, M.E., R. Harwood, N. Dehne, and J Smeenk. 1998 (in press) Crop yield and
net economic return of corn, soybean and wheat as affected by position in the
rotation, fertility source and cover crop. J. Prod. Agric.

J oergensen, R. G., B. Meyer, and T. Mueller, 1994. Time-course of the soil
microbial biomass under wheat: A one year field study. Soil Biol. Biochem.
26:987-994.

Kaiser, E. A. and O. Heinemeyer. 1993. Seasonal variations of soil microbial
biomass carbon within the plow layer. Soil Biol. Biochem. 25: 1649-1655.

KBS LTER. 1998 [online]. The KBS LTER Site Agronomic Protocols for Main
Cropping Systems Study. Available at http://1ter.kbs.msu.edu/Agronomics/
protocolhtm, Dec. 10, 1998.

Keeney DR. 1982. Nitrogen availability indices. pp. 711-733 Dz'AL. Page et al.
(ed.) Methods of Soil Analysis, Part 2, 2nd Ed. Chemical and microbiological
properties. Agron. Monogr. 9, ASA and SSSA, Madison, WI.

Keeney, DR. 1987. Nitrate in ground water - Agricultural contribution and control.
p. 329-351. InProc. Conf. On Agricultural Imapcts on Ground Wate, Omaha,
NE. 11-13 Aug. 1986. National Water Well Assoc, Dublin, OH.

Kirchner, M. J ., A. G. Wollum II, and L. D. King, 1993. Soil microbial pOpulations
and activities in reduced chemical input agroecosystems. Soil Sci. Soc. Am. J.
57:1289-1295.

Kitur, B.K., M.S. Smith, R.L. Blevins, and W.W. Frye. 1984. Fate of l5N-depleted
ammonium nitrate applied to no-tillage and conventional corn. Agron. J. 76:
240-242.

Libra, R.D., G.R. Hallberg, and BE. Hoyer. 1987a. Impacts of Agricultural
chemicals on groundwater quality in Iowa. p. 185-217 .0) Fairchild, D.M. (ed.)
Ground Water Quality and Agricultural Practices. Lewis Publ. Inc. Chelsea, MI.

Libra, R.D., G.R. Hallberg, B.E. Hoyer, and LG. Johnson. 1987b. Agricultural
impacts on groundwater quality: The Big Spring study, Iowa. InProc. Conf. On

69

Agricultural Imapcts on Ground Wate, Omaha, NE. 11-13 Aug. 1986. National
Water Well Assoc, Dublin, OH

Maxwell, R. A. and D. C. Coleman, 1995. Seasonal dynamics of nematode and
microbial biomass in soils of riparian-zone forests of the Southern Appalachians.
Soil Biol. Biochem. 27:79-84.

McGill, W.B. and R.J.K. Myers. 1987. Controls on dynamics of soil and fertilizer
nitrogen. pp. 73-99. 0) Soil Fertility and Organic Matter as Critical Components
of Production Systems. SSSA. Special Publication 19. SSSA, ASA. Madison,
Wisconsin.

Meijbloom, F.W., J. Hassink, and M Van Noordwijk. 1995. Density Fractionation

of soil macroorganic matter using silica suspensions. Soil Biol. Biochem
27:1109-1111.

Meisinger, J .J ., 1984. Evaluating plant available nitrogen in soil-crop systems. p.
391-416 InNitrogen in Crop Production. ASA, CSSA, SSSA, Madison, WI.

Meisinger, J .J ., V.A. Bandel, G. Stanford, and J .O. Legg. 1985. Nitrogen utilization
of corn under minimal tillage and moldboard plow tillage. I. Four-year results
using labeled N fertilizer on an Atlantic Coastal Plain soil.Agron. J. 772602-611.

Myers, R. J. K., C. A. Palm, E. Cuevas, I. U. N. Gunatilleke and M. Brossard,
1994. The Synchronization of nutrient mineralization and plant nutrient
demand. In: Woomer, P. L. and M. J. Swift (eds) The Biological Management
of Tropical Soil Fertility. Tropical Soil Biology and Fertility Programme (TSBF)
and Sayce Publishing with John Wiley & Sons. 81-116

Paul, E.A. 1994. Dynamics of organic matter in soils. Plant and Soil 76:275-285.

Paul, E.A. and RE. Clark. 1996. Soil Microbiology and Biochemistry. Academic
Press. San Diego

Paul, E.A. and D. Harris. 1997 [online] KBS LTER Site Data Catalog Microbial
Biomass (KBSOl 1). Available, at http:/ /lter.kbs.msu.edu/
Microbial%20Biomass/index.htm, updated October 22, 1997

Paul, E. A. and R. P. Voroney, 1984. Field interpretation of microbial biomass
activity measurements. In: Current Perspectives in Microbial Ecology. M. J.
Klug and C. A. Reddy (eds). American Society for Microbiology. 509-514.

Philips, L. and C. Stopes. 1995. The impacts of rotational practice on nitrate
leaching losses in organic farming systems in the United Kingdom. pp. 123-134.
firNitrogen Leaching in Ecological Agriculture. A B Academic Publishers,
Great Britain.

Pierce, El. and CW. Rice. 1988. Crop rotation and its impact on efficiency of water
and nitrogen use, pp. 21-42, 02, Cropping Strategies for Efficient Use of Water
and Nitrogen. ASA Special Publication Number 51. ASA, CSSA, SSSA.
Madison, Wisconsin

70

Reddy, GB, and KR. Reddy. 1993. Fate of nitrogen-15 enriched ammonium
nitrate applied to corn. Soil Sci. Soc. Am. J. 572111-115.

Russelle, M.P., E.J. Deibert, R.D. Hauck, M. Stevanovic, and RA. Olson. 1981.
Effects of water and nitrogen management of yield and 15N-depleted fertilizer-
use efficiency of irrigated corn. Soil Sci. Soc. Am. J. 45:553-558.

SAS Inst. Inc. 1988. SAS/ STAT user’s guide Vol. 2, version 6, fourth edition. SAS
Inst. Inc., Cary, N .C.

SAS Inst. Inc. 1997. SAS/ STAT software changes and enhancements through
release 6.12. SAS Inst. Inc., Cary NC.

Schepers, J .S., K.D. Frank, and C. Bourg. 1986. Effect of yield goal and residual soil
nitrogen considerations on fertilizer recommendations for irrigated maize in

Nebraska. J. fert. Issues 32133-139.

Schepers, J .S. and D. Martin. 1987. Public perception of ground water quality and
the producers dilemma. pp. 399-411 firProc. Conf. On Agricultural Imapcts on
Ground Wate, Omaha, NE. 11-13 Aug. 1986. National Water Well Assoc.,

Dublin, OH.

Shen, S.M., G. Pruden, and BS. J enkenson.1984. Minerilization and
Immobilization of Nitrogen in fumigated soil and the measurement of microbial

biomass nitrogen. Soil Biol. Biochem. 16:43 7444.

Smith, J. L. and E. A. Paul, 1989. The significance of soil microbial biomass
estimations in soil. In: Soil Biochemistry. Vol. 6. Marcel Decker, New York.

Solberg, S. O. 1995. Influence of crops and cultivation management on the nitrogen
leaching potential on ecological farms in east norway. In: Nitrogen Leaching in
Ecological Agriculture, pp. 115-121. A. B. Academic Publishers, UK.

Sparling, G. and C. Zhu. 1993. Evaluation and calibration of biochemical methods
to measure microbial biomass C and N in soils from Western Australia. Soil

Biol. Biochem. 25:1793-1801.

Tirnmons, DR, and J .L. Baker. 1992. Fertilizer management effect on recovery of
labeled nitrogen by continuous no-till. Agron. J. 84:490-496.

Waring, S.A. andJ.M. Bremner. 1964. Ammonium production in soil under
waterlogged conditions as an index of N availability. Nature 2012951-952.

Wegger, M.G.and D.B. Mengal. 1988. The role of nonleguminous cover crops in the
efficient use of water and nitrogen, pp. 115-127. .02, Cropping Strategies for
Efficient Use of Water and Nitrogen. ASA Special Publication Number 51. ASA,

CSSA, SSSA. Madison, Wisconsin.

Willson, T.C., R.R. Harwood, E.A. Paul, D. Harris, and O. Schabenberger. 1999 (in
process). Seasonal changes in nitrogen mineralization potential in agricultural
soils: AN OVA and non-linear regression analysis of management effects.

71

CHAPTER 3

Biological Indicators of Crop Performance, Nitrogen Mineralization,
and Leaching Loss in Integrated Cropping Systems

ABSTRACT

Integrated cropping systems use biological resources such as organic fertilizers,
nitrogen fixing plants, and crop and cover crop residues to reduce the need for
chemical inputs and improve sustainability. We present crop performance, and N
cycling data for corn (Zea mays L.) monoculture and com com soybean ( Glfl'r'ne
maXL.) wheat (Tnb'cum aesa'mm L.) rotations receiving compost or inorganic
fertilizer, with and without cover crops. Laboratory measurements of microbial
biomass, macroorganic matter, and N mineralization potential (NMP) were
evaluated for their ability to predict yield, leaching loss, and N mineralization under
field conditions. Nitrogen mineralization was calculated using the equation: Nmin*
= Nharvestcd + Nresidue + Nleached - Nfertilizer - Next-o,- which does not include losses due
to denitrification or volatilization. Leaching was measured using 0.3 m diameter,
intact core lysirneters installed 0.3 m below the surface to facilitate tillage. The
rotation produced about the same mineralization and leaching loss as the corn
monoculture. The use of compost increased mineralization and decreased leaching
loss. It produced equal yields in 1st year corn and soybean, but lower yields in 2nd
year corn, continuous corn and wheat. Field N mineralization was correlated with

predicted N mineralization in 70 and 150 day aerobic incubations and with

72

macroorganic N in the 53-1000 um size class across all dates. Correlations between

NMP and both leaching loss and yield were often negative because of lower N
availability in the compost treatment. However leaching loss was positively

correlated with inorganic N and 10 day N mineralization in the previous summer.

73

INTRODUCTION

Nitrate contamination of groundwater is becoming prevalent in many agricultural
regions of North America (Hallberg 1987, Power and Schepers, 1989, Hamilton and
Helsel, 1995, Goss and Goorahoo, 1995 Richards et al., 1995, Franco and Cady,
1997). As this contamination spreads, there will be increasing pressure to develop
cropping systems that minimize leaching loss while producing high economic
returns. A key strategy in the development of sustainable crop production systems
has been to increase the use of biological resources such as crop residues, Nz-fixing
plants, and organic fertilizers to reduce the need for conventional N fertilization.
While these innovations can occasionally be implemented without decreasing
profitability, (c.f. Dobbs et al., 1988, for small grains, Kelly et al., 1995, and Creamer
et al., 1996, for tomatoes, Chase and Duffy, 1991, and Lichtenberg et al., 1994, for
corn rotations), it has yet to be shown that biologically fertilized systems are

inherently better for the environment.

To manage biological N resources successfully, we must be able to predict their
effect on inorganic N availability and loss. Specifically, we need to understand how
management changes the timing and extent of N mineralization relative to plant
uptake, and we must identify accurate predictors of N mineralization that can be

included in our calculation of fertilizer N requirements (Keeney, 1982).

In a previous paper (W illson et al., 1999b), we presented measurements of N

mineralization potential (NMP) based on the accumulation of inorganic N in

74

aerobic laboratory incubations of soils obtained from field experiments in
Southwestern, MI. At the Living Field Laboratory (LFL) site we sampled
continuous corn and selected crops in a com-com-soybean-wheat rotation, with and
without cover crops, and receiving either inorganic N fertilization or leaf-dairy
manure compost. We found that NMP was highest in the spring and lowest in the
early fall in all treatments, that compost additions increased NMP throughout the
season, and that legume cover crops increased NMP primarily in the early spring, In
a companion study (W illson et al., 1999a) we measured the C and N content of the
soil microbial biomass and two macroorganic matter size fractions in the same soil
samples to determine whether these biologically active organic matter fractions
respond to management in the same way as NMP. Microbial biomass increased
slightly with compost additions, but was otherwise insensitive to management, and
was not strongly correlated to NMP. A11 macroorganic matter fractions were
correlated with NMP and followed a similar seasonal pattern. The strongest
correlation was between the C content of macroorganic matter in the 250-1000 um
size class and change in inorganic N in the first 150 d of the incubation. The $10pe
of this relationship was dependent on the time of sampling, with the greatest
mineralization potential per unit macroorganic C in the spring and the lowest in the
early fall. In general, macroorganic matter was more sensitive to compost additions

and less sensitive to crop residue additions than was NMP.

In this paper we explore whether the laboratory measurements presented in Willson
et a1. (1999 a and b) can be used as predictors of N availability, crop performance

and leaching loss in the field. We present yield, N uptake and N leaching data for

75

the Integrated Fertilizer and Integrated Compost treatments at the LFL and use
these data to estimate net annual N mineralization. We then perform a correlation
analysis to determine which laboratory measurements at which sampling dates
provided the best predictions of yield, leached N and field N mineralization both in

the current and in the following grong season.

76

MATERIALS AND METHODS:

Site Description and Treatments Sampled

The data used in this analysis is from the 2nd, 3rd and 4th year of the Living Field
Laboratory (LFL) experiment in Hickory Comers, MI (42°24’N, 85°23’W). This
experiment is located on a gently sloping glacial outwash plain (2 to 4 % slope) with
underlying material consisting primarily of coarse sand. The soil is a mixture of
Kalamazoo loam and Oshtemo sandy loam (Typic Hapludalfs). The depth of the
Ap horizon is 20 to 25 cm at this site with an average organic C content of
approximately 10 g kgl. Average precipitation is 880 mm yr-l (1988-1997), with

precipitation exceeding potential evapotranspiration from October through March.

The LP L was established in 1993 to explore the sustainability of corn production
strategies ranging from a low diversity, conventionally managed, corn monoculture,
to a high diversity, organically managed, com-com-soybean-wheat rotation with
cover crops. The experimental design is split-split-plot, randomized complete block,
with main plots for each of four management types, subplots for each entry point in
a com—com-soybean-wheat rotation plus continuous corn, and sub-subplots based
on the presence or absence of cover crops. Cover crops included red clover frost
seeded into wheat and interseeded into lst year corn, annual ryegrass interseeded
into 2nd year corn, and a mixture of red clover and annual ryegrass interseeded into

continuous corn. Sub-subplots measure 4.6 x 15.2 m, corresponding to a single pass

77

with a 6-row equipment (0.76 m row spacing). Reduced tillage (chisel plow) is used

throughout the experiment, and the design is replicated in four blocks.

We present data from the Integrated Fertilizer (Fertilizer) and Integrated Compost
(Compost) management types at the LFL. These systems are managed identically
except for the use of inorganic fertilizers or leaf - dairy manure compost to maintain
N fertility. Fertilizer N inputs (Table 3.1) are based on soil test recommendations
for wheat and the pre-sidedress nitrate test (PSNT) for corn, with a yield goal of 9.4
Mg ha:1 (150 bu ac‘). Compost additions were intended to provide adequate N for
cr0p production based on an assumed mineralization rate of 25% per year. The
small applications in 1993 (based on an estimated crop demand of only 12 kg N ha-l)
were judged insuflicient, and starting in 1994, applications were based on an
expected crop demand of 125 kg N ha'1 at an assumed mineralization rate of 20%.
The 1993 application of compost in the soybean plots was intended to provide
differentiation between the compost and fertilizer treatments in the first year of the
study. Over the first 4 years of the experiment, twice as much total N was added to

the compost plots as to the fertilizer plots.
Field Measurements

Yield was determined in July for wheat, in September for soybean, and in the mid to
late fall for corn, depending on the dryness of the grain. Corn was hand harvested in
1993-1995, but all other yield data was obtained using 2 or 3-row harvesters. Crop

biomass samples were taken at physiological maturity for each

78

 

Eco maozczcoo n 00 60:3 u >> cmon>0m u w Eco be» new n «0 Eco 60> .3 n _.0 85530594 .F

 

 

 

 

3. N2 00 «NF 00 dd 00 2: 00
non o w o? No om 5 E 3
t.» e: «o F: 5 om 3 o w
vwm wa F0 mm >> 0 m no? No Legion.
NAN mm >> o m cm «0 m9. _.0 3352:.
mom mm? 00 3 00 So 00 m... 00
«R o m 8 No one 8 me >>
am mom «0 E 8 mom 2, mu m
wow mm? 5 omv >> o m no NO 38280
a: B 3 o m one No me 5 8.98.5
To; 2 9. 7o; 2 9. 8.0 to; z 9. 85 To; 2 9. 85 to; z 9. .590
IoIIim. End 89 39 39. 89.

 

4n... 9: a 8%.

EoEommcmE 55:5“. 3.9035 new «89:00 3562:. 9: Co 358 Eco =m E $qu 2 ho bore... mm ass...

79

crop, and cover crop biomass was measured at approximately the same time as corn

biomass.

Nitrogen leaching loss was measured using subsurface leaching lysimeters. These
were installed in the cover crop sub-subplot of each entry point in the integrated
management types and in two entry points in the Conventional and Organic
treatments. (We have used the lysimeters in the Conventional treatment to estimate
leaching loss in the Fertilizer, no cover treatment for the purpose of the correlation
analysis. These are treatments are identical except for weed control methodology,
banded herbicides in the Fertilizer treatment and broadcast in the Conventional.)
Each lysirneter was constructed of a 0.6 m section of 0.3 m diameter PVC pipe
buried 0.3 m below the surface to facilitate tillage, and contains an intact soil profile
0.3 - 0.9 m in depth. Leachate was pumped from a 201 carboy attached to the base
of the lysirneter at approximately one month intervals or as needed to prevent
overflow of the reservoir. Leachate was stored at 4 °C until analyzed for NOr-N and

NH4-N using a Lachat automated analyzer(Lachat Instruments Inc. Milwaukee,

WI).

Soil nitrate was sampled to a depth of 0.3 m, 7-12 times a year and to 0.9 m in
November. The soil samples were dried at 35 °C, ground, shaken for 30 min. in a
5:1 dilution of IMO, and gravity filtered using pre-leached filter paper. Plant
tissue analysis was performed using a Carlo Erba NA 1500 Series 2 N/ C/ S analyzer

(CE Instruments Milan, Italy) and NO3-N was measured using a Lachat automated

80

analyzer. The reader is referred to Dehne (1995), Jones (1996), and Jones et al.,

(1998) for more a more detailed description of these protocols.
Calculation of Field N Mineralization

The problem of estimating the nutrient supplying capacity of the soil to prevent both
over- and under-fertilization has been addressed by numerous authors (c.f. Cabrera
et al., 1994, Rice and Havlin, 1994, Campbell et al., 1997). Meissirrger (1984)
suggested that fertilization recommendations should ultimately be based on an N-
balance approach. He proposed the equation:

[I] NFertilizer + NMisc. — NHarvested - NLeached - NGas - NErosion = A Nsorvr + A
Nlnorganic.

to described the annual N budget of a given soil under agricultural management,
where NMisc. includes all external inputs other than fertilization and A NSOM includes
all of the organic material in the soil. In systems without legumes, NMisc. consists
primarily of atmospheric deposition, and both NMisc. and NErosiorr are usually small.
Atmospheric deposition at KBS was measured at 6.3 kg N ha'1 yr-l from 1994-1996
(National Atmospheric Deposition Project, 1998 [online]), which is probably greater
than erosion and much less than the other fluxes in equation [1]. Under these
conditions, Meisinger suggests that NMisc. and NErosion can be removed from the
analysis leaving the equation:

[2] NFertilizer " NHarvest ' NLcached " NGas = A NSOM '1' A Nlnorganic.

We can generalize equation [2] for use with integrated cropping systems by adding

specific input terms for N2 -fixation (N Fixed) and organic fertilization (N or), and by

81

redefining the system boundaries to include plant residues (N Res) as a storage pool of

N. The resulting equation would be:

[3] NFert-+ NOF + NFixed ' NLeached ' NHarvest ‘ NGas =7 A NSOM + A Nlnorganic + ANRcs,

Ifwe define NResl and NRcSZ as the non-harvested plant biomass N at the beginning
and the end of the year such that

[4] A NRes = NResZ - NResl,

and we time the endpoints of our crop production year such that all of the plant
material added to the soil during the year comes from Nresl, we can say that

[5] A SOM = NResl + Nor - NMin

where NMin is the annual net N mineralization in the field. By substituting [4] and
[5] into [3] we can calculate Nmin as:

[6] NMin 2' NHarvested + NLeached + NGas +NResZ+ ANlnorganic - NFertilizer' NFixed-

Equation [6] should be applicable whenever the assumptions of eq. [5] are valid.
This can be achieved in annual cropping systems by setting the endpoint of the crop
production year after the physiological maturity of summer crops, but before those
crops are incorporated into the soil. We base our analysis on a crop production year
of October 1 to September 30, which coincides with the transition from soybean to

wheat in the com-com-soybean-wheat rotation.

In order to satisfy equation [5], measurements of NRes should include root
production. We assume that com root N is 50% of corn stover N based on LFL data

reported in Paul et al. (1998), and that wheat and soybean root N are 72% and 58%

82

 

 

of wheat straw and corn stover N respectively (Bouyanovsky and Wagner, 1986).
(These calculations assume that root C:N equals straw or stover C:N at
physiological maturity.) Estimation of cover crop root biomass is complicated by
the fact that these plants are in an intermediate stage of deveIOpment when sampled.
We base our analysis on direct measurements of root and shoot N provided by
Kunelius et al. (1992) who reported a root to shoot ratio of 1.15 for red clover and
0.29 for annual ryegrass underseeded into barley. The “weedy fallow” biomass

following wheat is assumed to have the same root:shoot N ratio as red clover.

We have calculated Nfixcd based on the assumption that the proportion of soybean
and red clover N derived from the atmosphere (N dfa) is constant for each plant.
Percent Ndfa in soybean depends on the variety and maturity group (Henderson and
Zapata, 1984, Newhausen et al., 1998), and the type and abundance of bacterial
symbionts (V asilas and Fuhrman, 1993, Galal, 1997), but is particularly sensitive to
soil N03-N content. Soil NOa-N is generally low at the LFL, rarely exceeding 10
mg N kgl soil during soybean or red clover production. We assume a value of 60%
Ndfa for soybean, which is greater than the 33-49% reported for a similar soil N
content by Henidge et al. (1990), but less than the 69% reported by George and
Singleton (1992) or the 71-77% reported by Newhausen et. al. (1988). Values
reported for red clover Ndfa (e.g. Heichel et al., 1985, Mallarino et al., 1990,
Fumham and George, 1993) tend to be slightly higher but they are just as variable,
and appear to be just as sensitive to soil NOs-N levels. We assume a value of 70%

Ndfa for red clover in this study.

83

Robertson (1997) notes that accurate data for denitrification and volatilization

(N Gas) are rarely available for upland farming systems. Ammonia volatilization is
greatest after the application of ammonium and urea based fertilizers, and can range
from < 1% to >60% of applied N depending on the soil acidity, the method and rate
of application, the moisture content of the soil, and the pattern of rainfall after
application (Shankaracharya and Mehta, 1969, Gezgin and Bayrakil, 1995, Fox et
al., 1996). Significant N-volatilization has been measured in maturing plants
(Hoopker et al., 1980, Francis et al., 1993) and in decaying plant tissues (J anzen and
McGinn, 1991), but the overall volatile-N balance for the plant - soil system is
unknown due to the complexity of sources and sinks (Francis et al., 1997). Point
measurements of denitrification rate can be made using either the Csz inhibition
technique (Y oshinari et al., 197 7) or the 15N enrichment technique (Buresh and
Austin, 1988), but the extreme spatial and temporal variability of denitrification
makes it difficult to derive accurate field scale values from these measurements

(Robertson, 1977).

The best estimates of Ngas come from N balance studies in which gaseous loss is
determined by difference. Robertson (1997) suggests that an average of 20-40 kg N
ha-1 is typical for upland agricultural soils, but data from the studies reviewed by
Allison (1955), von Reinbaban (1990) and Peoples et al. (1995) range from near 0 to
>80 kg N ha'l. Faced with so much uncertainty, and in the absence of any direct
data, we prefer to base our analysis on a simplified version of equation [6]

[7] NMin* = NHarvest +NRes2 +NLeacoed +A Nlnorganic - NFertilizer - NFixed

where Nmin* underestimates Nmin by the quantity of Ngas.

84

Laboratory Measurements

The laboratory measurements used in this paper (Table 3.2) are identical to those
used in Willson et al. (1999a and b) except that separate kinetic analyses were
performed for each combination of treatments (management type, entry point and
cover crop) at each sample date. Soil samples were obtained in April, June, August,
October and November 1994, in April, July, September and November 1995, and in
April 1996. Each sample consists of a composite of 12, 2 cm diameter soil cores
taken to a depth of 25 cm in 1994 and 20 cm in 1995-96. The lst year corn, soybean
and wheat entry points were sampled in the Compost and Fertilizer rotations in
1994 (cover and no cover), while in 1995, all of the Fertilizer plots were sampled
along with lst year corn, continuous corn and wheat in the Compost plots.
Microbial biomass N was determined only in 1995-96 and data for N mineralization
and macroorganic matter C and N were not obtained in the wheat plots in August or

October 1994.

The kinetic parameters No, MRT and Ni in Table 3.2 are based on the single pool
exponential model N = Ni + No (l-e-VMRT) where N is the inorganic N content of a
replicate soil sample at time t (days) in an aerobic incubation at ideal moisture and
25 °C. (The measurements of N are designated N_0, N10, N70, and N150 in Table
3.2.) The values p10, p30, p70, and p150 represent the predicted accumulation at
time t, which is equivalent to No (Lei/MRT). The measured increase in N above
N_0 is designated d10, d30, d70, and d150. All of these parameters can be
considered measurements of NMP except N_0 and Ni , which measure the

inorganic N at the time of sampling. All should be positively correlated with

85

.28 82 c. 88.8 8.. 28888:. 28.2 m
.220 8.00 0.950 89 =54 2.0 89 828052 new 888.com .23.. 05 E 8:22.80 08:33:. 200 05 A

.28. 8880 .98 882 e. 8... .82; 8 8o... o8; 98888:. 02.. o8 .20 .812 9.. 2:0 .2

88...

8288... 8.8.8. 2888.8 a 82.88 no 2 8n 0 «.88... 8.882
.092 .292 .o 8.8.8.8 9.. e. 8.5. .8288... .8 2 e. 8.88. 980
.88 28 E: 82 - 8N 2. z 28 0 88... 2:88.82

.88 one .8 8m - 8 e. z .8 0 88:. 28.8.82

.88 88 E: 82 - 8 e. 2 98 0 .88. 88.8.82

.28 2.” 8o .8. .2 .8 .2 8.8 z ".288... e. 888... 88.8...

.28 22 o8 .8. .2 .8 .2 88 z 888... e. 888... 85.8.2
8.88... e. 2.. o8 8. .2 .8 .2 8 .n z 2.88... 888.2

.0 on... .o 2 2.88... 8.882

a 2m: .02..

20

2.2.20 .0220

2.2.2.. .022“.

2.2.2 0.2.2

8:. .28 .88 .28
8:. .2o .08 .28

H .82 .22 .82 .22
2 orz

8.2.88 88... 2. 8:8 8. o on... 8 z 2.88:. .z

... 2 u E2 8.8... .88 .2 8 89.8 on on... .8. 8o... 28 88.0 82
88:... .88 .2 a 85.8 on 2 o. 8 88.88.... 8 8n. oz

8...:80 8.8.882

 

 

0.02.25 820.928 05 5 com... 2082889082.... 2.0 28.5508... 2.12883

NM 030...

86

inorganic N availability except MRT, which is negatively correlated. All
incubations were performed at a constant 25 °C and gravimetric moisture content of
19%, (about 2/ 3 of water holding capacity in KBS soils). Extractions of
accumulated inorganic N (NH4 plus N03) were performed as described above for

soil nitrate.

Macroorganic matter C and N fractions are measured by shaking each soil sample
overnight in 5% (N aPO3)n solution (3:1 dilution), passing the diSpersed soil through
nested 1mm, 250 um and 53 um sieves using a flow of water to transfer the particles,
and measuring the C and N content of the material retained on the 250 and 53 um
sieves (coarse and fine macroorganic matter respectively) on the Carlo Erba CNS
analyzer. Total C and N were also measured using this instrument. Microbial
biomass C and N were measured by the chloroform fumigation, incubation
technique (J enkinson and Pawlson, 1976) using the calculations for C and N content

described by Horwath et al. (1996) and Harris et al. (1997) respectively.

Correlation Analysis

Each of the parameters in Table 3.2 was analyzed for its ability to predict yield (bu
acl), leaching loss (kg N ha'1 yrl) and apparent field N mineralization (kg N ha-1 yr
1) using the SAS CORR procedure (SAS Inst., 1986). The correlation analyses were
performed on a plot by plot basis within each month. Laboratory measurements
taken in the 1994 season were analyzed for correlation with the 1994 and 1995 field
data, and laboratory measurements taken in 1995-96 were analyzed for correlation

with the 1995 and 1996 field seasons. Because of the inherent differences between

87

corn, soybean and wheat yield, separate analyses were performed for each cr0p.

Leached N and apparent N mineralization were analyzed across crops.

88

RESULTS
Crop Performance, N Leaching Loss and Apparent N Mineralization

Crop yields were strongly affected by moisture, crop sequence, and fertility source in
the three years of this study (Table 3.3). Total precipitation (based on a crop
production year of October 1 through September 30) was 821 mm in 1994, 820 mm
in 1995, and 651 mm in 1996, compared to a 10 year average of 880 mm (1988-
1997). A prolonged drought in the summer of 1996 dramatically reduced corn and
soybean yields. Other than a late spring drought in 1994, moisture levels were
generally adequate for wheat production. First year corn (following wheat)
consistently outperformed 2nd year and continuous corn in all treatments. The use
of compost rather than inorganic fertilizers decreased the yield of 2nd year corn,
continuous corn, and wheat, but did not decrease the yield of lst year corn or
soybean. The use of annual ryegrass as an interseeded cover crop decreased 2“‘1 year
corn yield in 1994, and the use of red clover as a preceding as well as interseeded

cover crop increased 1St year corn yield in 1995, but there were no other significant

cover crop effects.

The annual N budget of the LF L cropping system (Table 3.4) suggests that many of
the observed differences in yield can be explained by differences in inorganic N
availability. In each case that yield is lower in the Compost treatment, plant N
(harvested plus residue) and leached N were also significantly lower, indicating

lower N availability. By contrast, leaching loss and plant N in 1st year corn with

89

cover were not significantly higher in the Fertilizer treatment, indicating that the
supply of N from mineralization was nearly as great in the Compost treatment as the
supply of N from mineralization and fertilization in the Fertilizer treatment. An
average of 2.5 kg N was removed as yield for each kg N leached in the Fertilizer
rotation, compared to 2.9 in the Fertilizer monoculture, 4.1 in the Compost rotation
and 6.1 in the Compost monoculture. Although leaching loss was not significantly
greater in the fertilized rotation than in the monoculture, leaching loss was
equivalent to 60% of fertilizer N applied in the rotation, compared to only 30% in the
monoculture. Because of greater N2 fixation in the rotation, harvested yield

accounted for 133% of fertilizer N applied (113% for 1st and 2nd year corn),

 

 

 

Table 3.3 LFL crop_yields avegged between cover and no-cover treatments 1'
Crop Management 1994 1995 1996 Average
Monoculture --------- Mg ha'1
Corn Fertilizer 8.7 ' 8.7 " 4.1 ' 7.2 "
Compost 7.6 8.0 3.3 6.3
Rotation
1st year corn Fertilizer 11.8 9.7 4.5 8.7
Compost 11.7 9.7 T 4.1 7.9
2nd year corn Fertilizer 10.5 * ’r 8.6 4.3 7.8 ‘
Compost 7.4 1- 8.0 4.2 6.1
Soybean Fertilizer 2.4 3.4 1 .2 2.3
Compost 2.9 3.4 1.2 2.5
Wheat Fertilizer 3.1 4.1 " 4.4 ‘ 3.9 *
Compost 3.2 3.6 3.3 3.4

 

1‘ Cover crop eTiECTwas not significant except as noted
“ Significant difference between the Compost and Fertilizer treatments (a=0.05)

compared to 85% in the monoculture.

90

 

 

 

 

 

3:855:58» .838 95 55:5“. 2.. 80.28 885% Emoscgw .
bodice“. - 2 85m - z 858.. + 2 9.28m + 2 8.82m: + z 2590:. c u .c.Er. a.
58 s 299.8: 8; :8me .z .966 8m .6 $2 .2 cmoaow a $8 8 838.8 85.... «z .m
new 30.5 .98 wounBE z 2633. H
82 .8 .2853 2 82 .F .820
N9 8 E 8 av 02 J8 8 39:9, 5858 o .P
3.— mm 8. m: t. 8: mm P 6. N: 96:9 55:5... cozfiom .3“.
.2 o: E 9 a: at he an .220 8. 7.858
2: o: 8 . E 31. :3 8w . 8v 8 $5.0 8. sues“. “8:2,
mm 3 F 6. 9 A8 .3 8 5 Now 9.0: «89:00
mm. P .. e . 8. mm CL mm a: mo oco: emu-Eon. cmoaom
m! o a: mm 8V 3 6V 2 828». .955 6858
mm: o 8. . a: 8 . 6: mm . a: a: $98: .855 co~___eou :8 52 EN
SN 2 Q on a: N2 8: mm. 55.0 8.
mm: 2 x: a: ow as a: a: a? 88.0 B. .858”. 58 .8» a.
cos—20m
R: m E 3 8: 8 .8 mm 898.: + .920 7.858
m? m a: . é mm 6: 8 . a: .2 $98». + .920 55:5“. 58
757m; 2 9. 93.3952
a. .522 a :8me f z co~___eou__ z 858.. a 2 2.28m 2 8.85: .260 $55". no.0

 

was; 36.350

 

«dag 3.382

.59: 05 Co cote Emncmfi 9: 2365 mowonacmcmd.
.82 - 32 c. 890 .28 5? 252.8: 8858 95 533.3 8.935 .r: 2. 3. 583 z .852

in 03m...

91

The estimates of Nmin* (net mineralization minus gaseous loss) range from 93 kg N
ha‘1 yr'1 for soybean to 247 kg N ha'1 yr1 for lst year corn in the Compost rotation,
and from 109 kg N ha-1 yrl in wheat to 168 kg N ha-1 yrl in lst year corn in the
Fertilizer rotation. These differences correspond to a decrease in N mineralization
potential from 1St to 21“‘1 year corn reported in Willson et al. (1999a), and can be
explained in part by the lack of compost additions in the soybean treatment, and in
part by differences in residue input from the previous crop. The residue input before
lst year corn in the Compost treatment with cover crops was 178 kg N ha1 yrl while
the input before soybean was only 52 kg N ha'1 yrl. In the Fertilizer rotation, the
input before lst year corn was 203 kg N ha-1 yr1 compared to 88 kg N ha'1 yrl before
wheat. Nitrogen mineralization was an average of 11 kg N ha-1 yr-1 greater in the
Compost rotation and 30 kg N ha'1 yr1 in the Compost monoculture than the
corresponding Fertilizer treatments. Given that residue inputs in the Fertilizer
system were 29 kg N ha'1 yrl greater in the rotation and 23 kg N ha'1 yr1 greater in
the monoculture, the decomposition of compost must have accounted for 53 kg ha'1
yr1 in the monoculture and 40 kg N ha-1 yrl in the rotation. This implies an average
mineralization rate for compost of only 9% per year based on the history of

applications, which is about half of the expected mineralization rate.

Our analysis of the total inputs and outputs for the Fertilizer and Compost systems
(Table 3.5), indicates a net loss of soil N in the Fertilizer system and a net gain in
soil N in the Compost system. Including losses for denitn'fication and volatilization
would decrease the balance in both systems (typically, by 20 - 40 kg N ha-l yr-l) but

would not alter this interpretation. Current LFL protocols call for an annual input of

92

 

120 --140 kg N ha-l yr-l for each treatment, depending on the N content of the
compost. At this rate, soil N will increase by another 180 — 200 kg N ha-l yr-l and

mineralization by 15 —- 20 kg N ha-l yr-l between 1996 to 2000.

Correlation analysis

 

 

 

 

 

 

Table 3.5 Average annual N balance (not including gaseous loss) for the Compost
and Fertilizer treatments with cover crops
Inputs Outputs
Management Rotation External N2 fixation Total Hanest Leaching Total Change
kg N l’ia‘1yr’1
Fertilizer Monoculture 1 19 9 128 101 35 136 -8
Rotation 78 59 138 1 12 47 159 .21
Compost Monoculture 288 9 297 85 14 99 198
Rotation 239 60 299 90 23 1 13 187

 

Laboratory measurements of N mineralization potential (NMP) and macroorganic
matter taken in 1994 and 1995 were reasonably good predictors of field N
mineralization (N min*) in 1994 and 1995 (Table 3.6). The correlations between
laboratory measurements in1995 and 1996 Nmin* were weaker, probably because of
the 1996 drought. The best indicators of Nmin* across the range of treatments
tested were the predicted values for 70 and 150 day mineralizations (based on 1st
order kinetics), and the C and N content of macroorganic matter in the 53-1000 and
250-1000 um size classes. The kinetic parameters No (potentially mineralizable
organic matter) and Ni (initial inorganic N) were also good predictors of Nmin* at
certain dates. Measurements taken in August and October 1994 tended to have the
highest correlation with Nmin* in 1994, although this may have been influenced by
the fact that the wheat plots were not sampled at those dates. Measurements taken

in November 1994 and April and September 1995 were the best predictors of Nmin*

93

Table 3.6 Laboratory measurements with the highest correlation to field mineralization.

 

 

 

Laboratory Field Field Predictors of N Mineralization
Data Set Season Treatments 1' Correlation :1: Date Parameter n

7994 1994 (:13er 0.83 75mg. p70 20 *"
(IC and IF) 0.77 Jun. NI 28 '“
0.76 Aug. MMN 20 m
0.73 Oct. No 20 "'
0.72 Oct. p150 20 m
1995 C2. W, C1 0.86 Nov. p150 27 “"
(IC and IF) 0.82 Nov. p70 27 “'
0.77 Nov. No 27 '“
0.76 Nov. p30 27 m
0.73 Nov. p10 27 “'
1995-96 1995 C1, W, CC (IC) 0.72 Apr. 95 No 39 "'
C1, C2. S, W, CC (IF) 0.71 Sep. p70 39 *"
0.70 Sep. p30 38 m
0.71 Apr. 95 p150 39 “'
0.70 Sep. p10 40 “*

1996 C2, C1. CC (IC) 0.47 Sep. NI 38 “

C1, C2. S, W, CC (IF) 0.41 Apr. 95 CMMN 38 "

0.40 Apr. 95 MMN 38 ‘

0.40 Sep. No 38 "

0.40 Nov. N10 38 ‘

 

T Abbreviations: 01- 1st year corn, CZ - 2nd year corn. CC - continuous com. S - soybean,
W - wheat, IC - integreated compost, IF - integrated fertilizer.
1: If there are more than 5 significant correlations, the 5 highest are shown.

94

in 1995 and 1996. The highest levels of NMP were found in the 1994 wheat plots
with cover, which had deve10ped an excellent stand of red clover in the winter of
1994-95. These plots produced the greatest Nmin* estimates in both 1994(1st year

corn) and 1996 (an year corn).

Laboratory measurements of NMP and macroorganic matter were also significantly
correlated with corn yield (Table 3.7), although many of these correlations are
negative because yields were higher in the Fertilizer treatment than in the Compost
treatment. Yield tended to be positively correlated with NMP within each
management type, which accounts for the positive correlations observed in 1995. As
in the previous analysis, high NMP in the November 1994 wheat samples was
correlated with high performance in 19951St year corn. Wheat yield (not shown)
tended to be negatively correlated with NMP because of the N deficit in Compost
wheat, whereas there was little correlation between NMP and soybean yield. The
measurements N30, N70 and N150 include both initial inorganic N and N
mineralized during laboratory incubations. In July 1995 they reflect the supply of

inorganic N from fertilization and mineralization.

Leaching loss tended to be negatively correlated with NMP measurements taken
during the same year, but positively correlated with measurements taken during the
previous year (Table 3.8). The best predictors of leaching loss in the following year
were inorganic N and short term N mineralization measurements taken after

sidedress fertilization in the previous year (August 1994 and July 1995). These

95

 

 

 

Table 3.7 Laboratory measurements with the highest correlation to corn yield.
Laboratory Field Field Corn
Data Set Season Treatmentst Correlation i Date§ Parameter n
1994 1994 C1 063 Oct. 310 12 T
(IC and IF)
1995 01.02 0.71 Nov. d70 24 ‘“
(IC and IF) 0.64 Nov. N70 26 m
0.59 Nov. d30 28 m
0.57 Nov. p10 32 m
0.56 Nov. p30 32 '“
1995-96 1995 C‘I.CC(IC) 0.58 Jul N70 37 ""
C1,CZ, CC (IF) 0.47 Jul N30 40 “
0.46 Jul. N150 40 “
0.46 Apr. 95 No 40 "'
0.43 Nov. NI 40 “
1996 C1,02, CC -0.52 Sep. CMMC 48 "*
(IC and IF) -0.50 Sep. 070 47 m
-0.49 Sep. CMMN 48 ‘“
-0.47 Sep. N70 47 m
045 Nov. p70 48 "'

 

T Abbreviations: 01- 1st year corn, CZ - 2nd year corn, CC - continuous corn
IC - integreated compost. lF - integrated fertilizer.

:I: If there are more than 5 significant correlations, the 5 highest are shown.

§ This is the sample date for the laboratory measurement.

96

 

 

 

Table 3.8 Laboratoereasurements with the west correlation to N leaching loss.
Laboratory Field Field Predictors of Leached N
Data Set Season Treatments L Correlation ; Date Parameter n
1994 1994 C1, S, W -0.56 Nov. d30 23 “
(IC and IF) -0.54 Aug. C10 28 “
-0.54 Nov. p30 28 "
0.54 Jun. N_0 28 "
-0.54 Nov. p10 28 “
1995 C2, W. CI 0.71 Aug. N_0 27 '“
(IC and IF) 0.57 Aug. p10 20 ”
-0.57 Nov. N__0 22 “
0.49 Aug. p30 20 '
-0.48 Aug. d30 18 '
1995-96 1995 C1, W, CC (lC) -0.39 Sep. p150 39 "
C1, CZ, S, W, CC (IF) -0.39 Sep. p70 39 ‘
-0.39 Nov. CMMC 39 "
-0.37 Sep. p30 39 '
-0.37 Sep. d150 39 '
1996 CZ, C1, CC (lC) 0.68 Jul. N10 38 *“
C1, CZ. S, W, CC (IF) 0.66 Jul. N_0 39 “‘
0.62 Jul. N150 39 m
0.60 Jul. N70 36 m
0.59 Sep. N_0 39 "'

 

’r Abbreviations: Cl- lst year com, CZ - 2nd year corn, CC - continuous com. S - soybean,

W - wheat, lC - integreated compost, lF - integrated fertilizer.

I If there are more than 5 significant correlations, the 5 highest are shown.

97

measurements would be expected to reflect the supply of inorganic N from fertilizers

as well as from mineralization.

98

DISCUSSION:

In the first four years of the LFL experiment we have seen a gradual decrease in
field N mineralization (N min") in the Fertilizer treatment, which has been
accompanied by increased fertilizer N requirements and decreased corn yields.
Although the drought conditions in 1996 exaggerated this apparent decline,
inorganic N levels early in the experiment may have been supplemented by the
decomposition of root material from an 8-year stand of grassy alfalfa plowed down

in October 1992.

Although twice as much total N was added to the Compost plots as the Fertilizer
plots between 1993 and 1996, these additions were not sufficient to overcome the
lower N status of the Compost treatments. The combination of compost and red
clover has provided sufficient N for lst year corn production, but there has been
insufficient N to maintain high yields in 2nd year and continuous corn. The use of
compost rather than inorganic fertilizers has resulted in lower leaching loss and
greater production per unit leached, but it is difl'icult to determine whether the same
results could have been obtained by simply reducing the amount of fertilizer applied
in the Fertilizer system. Because the compost used at the LFL is a waste product of
nearby dairy and leaf removal facilities, its cost is relatively low. An economic
analysis performed by Jones (1996) found that that the economic return was greater
for the compost system than the fertilizer system because of the lower cost of inputs.
The same study found that there was no economic advantage in choosing the 4-year

rotation, with or without cover crops, over the corn monoculture in the first 3 years

99

of the experiment. This conclusion may change if the productivity of corn in the

monoculture continues to decline relative to corn in the rotation.

We can gain a reasonable estimate of inorganic N availability to each crop by
adding the estimated N mineralization (N min*) to the amount of fertilizer N
applied. This analysis confirms that N availability was nearly as high for 1st year
corn with compost and cover (245 kg N ha'1 yrl) as for fertilized com (272, 288 and
245 kg N ha'1 yr1 for 1st year, 2nd year, and continuous corn, respectively), but was
much lower for 2nd year and continuous corn in the Compost treatment (145 and
166 kg N ha'1 yr-l respectively). There was a good correlation between this
availability index and corn yield in 1994 and 1995, but not in 1996 (Figure 3.1), and
not for soybean or wheat (not shown). This supports the conclusion that com yields

have tended to be N limited, particularly in the Compost treatment.

Corn yield vs. inorganic N availability

 

 

15
l , .
Vv o v... o .0 o .
Corn 10 1 vv V O .V
. '8? c
Yleld .vzgzvgyéy v v v 0
Mg ha-1 ' :7. fl 7 . 1994
5‘ _::u-l.,- : v 1995
I. I 1996
o . . T . L
O 100 200 300 400 500

Inorganic N Availability
(Mineralized N + Fertilizer N)
kg N ha'1

Figure 3.1 Corn yield as a function of the supply of inorganic N from
mineralization (Nmin*) and fertilization

100

The pre-sidedress nitrate test (PSNT) for corn uses the concentration of soil nitrate
under V4-V6 stage corn to determine the additional fertilizer N required to achieve a
given yield goal. It is recognized as an effective method of reducing over-
fertilization (Musser et al., 1995), and N-leaching (Durieu et al., 1996), in humid
and sub-humid agricultural systems. One reason that PSNT is so effective is that late
spring soil NOa-N levels reflect both the residual inorganic N in the soil and the rate
of early spring N mineralization. From this perspective, PSNT represents one of the
few instances where an actual measurement of mineralization potential (albeit an

indirect one) is used to make fertilizer recommendations (Keeney, 1982).

In this study there was no relationship between early June NOs-N levels and
estimated N mineralization in years with corn, or between PSNT-N and the yield of
corn in the Compost treatment. It would appear that the factors that determine
spring NOs-N levels are too complex (or too variable) for this to be an adequate

index of N mineralization.

The installation of leaching lysimeters at the LFL allowed us to measure the loss of
soil NOa-N directly, but similar measurements are rarely available in cropping
system experiments. Many authors assume that leaching loss is closely related to
the soil NOa-N level in the late fall (e.g. Brown et al. (1993), Solberg, 1995, Philips

and Stopes, 1995), which is sometimes referred to as nitrate leaching potential.

We found that there was no correlation between fall NOs-N levels and leaching loss

in three of the four years examined (Figure 3.2). Soil NOs-N levels remained

101

November inorganic N (090 cm)
vs. N leached in the following year

300 1 1994

1995

200 J

100 ~

 

 

 

 

 

 

 

1 e
o'. o'
e
F ii.
2".“ 04—. T —fi
8.: 0 100 200 0
Z
3 m 300. 1996 ,
2x
0
2001
o
100‘ . 1
-os
‘ o
I
oi—h— . a '
0 100 200 0

 

 

November Soil NOa-N,
kg N ha-1

0 Corn Plots
Soybean Plots

I Wheat Plots

Overall Trend

 

Figure 3.2 Late Fall whole profile NOa-N as a predictor of leaching in
the following field season.

102

relatively low at the end of the 1993, 1994, and 1995 field seasons, and leaching loss
in the following years (1994-1996) was more closely related to N mineralization than
inorganic N content. However, reduced crop N uptake and delayed N
mineralization caused by the 1996 drought dramatically increased fall inorganic N
levels and produced high leaching losses in 1997. Much of the observed variation in
leaching loss in 1994-1996 may be an artifact of preferential flow patterns in the soil
above the lysimeters. We have found much greater variation in leachate volume

than in nitrate concentration between replicate lysimeters.

Limitations in the Analysis

Although the idea of using a correlation analysis to link laboratory measurements to
field data is sound, its application in this study was hampered by the fact that not all
treatments were sampled at each sample date, and by the influence of specific
treatments on individual measurements, both in the laboratory and in the field. It
would be more appropriate to use this technique to compare a larger number of
treatments or a larger number of replicate data points within a smaller range of
treatments. These limitations notwithstanding, it is clear from our analysis that
laboratory measurements of NMP are correlated to N mineralization and leaching
loss in the field, although the highest correlations to leaching loss are achieved by

mid-season measurements that include both soil inorganic N content and NMP.

While the simplified formula for N mineralization was successful in demonstrating
differences between treatments and years, our failure to measure Ngas, Nfixed, and

the root component of Nres may have reduced the accuracy and integrity of the

103

analysis. The failure to measure Ngas probably resulted in N mineralization rates
that are underestimated by 5-25% in all treatments — about the same as leaching
loss. Gaseous losses should be greater in the Fertilizer treatment than the Compost
treatment because of the generally higher soils NOs-N content and the use of urea
and ammonium nitrate in wheat and corn respectively. Accurate measurements of
soil N would allow this loss to be calculated by difference in treatments without

legumes or where inputs by N2 fixation are also measured. 1

Indirect evidence that gaseous loss is no greater than leaching loss comes from a

 

15N fertilizer study performed on site in 1995-1996. At harvest in 1996, Willson et
al., (1996) were able to account for 65% of the sidedress fertilizer N applied in 1995
in the top 20 cm of soil, in plant residues, and in the harvested components from
1995 and 1996. This total loss of 45% over 18 months (including loss to soil below
20 cm) compares to an average annual leaching rate of 22% of available N for the
same treatments (fertilizer lst year corn in 1995 and fertilizer 2nd year corn in 1996)

in this study.

The generally lower NOs-N content of soils in the Compost treatments may have
increased the root to shoot ratio for all crops, and increased %Ndfa for soybean and
red clover, relative to the Fertilizer treatments. Lower gaseous loss and greater
%NDFA would decrease our calculation of Nmin* for the Compost treatments,

while increased root to shoot ratio would have the opposite effect.

104

REFERENCES

Allison, RE. 1955. The enigma of soil nitrogen balance sheets. Adv. Agron. 7 :213-
250.

Bouyanovsky, G.A., and G.H. Wagner. 1986. Post-harvest residue input into
agriculture. Plant and Soil. 93:57-65.

Buresh, R.J., and FR. Austin. 1988. Direct measuremetn of dinitrogen and nitrous
oxide flux in flooded rice fields. Soil Soc. Sci. Am. J. 52:681-688.

Cabrera, M.L., D.E. Kissel, and M.F. Vigil. 1994. Potential nitrogen mineralization:
laboratory and field evaluation. p. 15-30 in Soil testing: pospects for improving
nutrient recommendatons SSSA special publication (40),. J .L. Havlin and J .S.

J acobsen ed., SSSA Madison, WI.

Campbell, C.A., Y.W. Jame, A. J alil, and J Schoenau. 1997. Use of hot KCl-NH4-N
to estimate fertilizer N requirements. Can. J, Soil Sci. 77:161-166.

Chase, C. and M. Duffy. 1991. An economic comparison of conventional and
reduced chemical farming systems in Iowa. Am. J. Altem. Agric. 6:160-173

Creamer, N.G., M.A. Bennett, B.R. Stinner, and J. Cardina. 1996. A comparison of

four processing tomato production systems differing in cover crop and chemical
inputs. J. Am. Soc. Hort. Sci. 121:559-568.

Dehne, NC. 1995. Compost and fertilizer effects on nitrogen movement within a
cropping system during transition Ph. D. Dissertaion. Michigan State
University, Dept. of Crop and Soil Sciences

Dobbs, T.L., M.G. Leddy, and J .D. Smolik. 1988. Factors influencing the economic
potential for alternative farming systems: case analyses in South Dakota. Am. J.
Altern. Acric. 3:26-34.

Durieux,-R.P.; Brown,-H.J.; Stewart,-E.J.; Zhao,-J.Q.; J okela,-W.E.; Magdoff,-F.R.
Implications of nitrogen management strategies for nitrate leaching potential:

roles of nitrogen source and fertilizer recommendation system. Agron.J. 87: 884-
887.

Fahmam, DE, and J .R. George. 1993. Dinitrogen fixation and nitrogen transfer
among red clover cultivars. Can. J. Plant Sci. 73:1047-1054.

Fox, R.H., W.P Piekielek, and KB. Macneal. 1996. Estimating ammonia
volatilization losses from urea fertilizer using a simplified micrometeorological
sampler. Soil Sci. Soc. Am. J. 60:596-601.

Francis, D.D., J .S. Shepers, and AL. Sims. 1997. Ammonia exchange from corn
foliage during reproductive growth. Agron. J. 89:941-946.

Francis, D.D., J .S. Shepers, and M.F. Vigil. 1993. Post anthesis nitrogen loss from
corn. Agron. J. 85:659-663.

105

Franco, J. and CW. Cady. 1997 . Preventing nitrate groundwater contamination in
California: a nonregulatory approach. J. Prod. Agric. 10: 52-57.

Galal, Y.G.M. 1997. Dual inoculation of Bradyrhizobium japonicm
andAzospirillum brsilense to improve growth and biological nitrogen fixation of
soybean (Glycine max L.). Biol. Fert. Soils 24:317-322.

George, J. and PW. Singleton. 1992. Nitrogen assimilation traits and dinitrogen
fixation in soybean and common bean. Agron. J. 84:1020-1028.

Gezgin, S. and F. Bayrakil. 1995. Ammonia volatilization from ammonium sulfate,
ammonium nitrate, and urea surface applied to winter wheat on calcarious soil.
J. Plant Nutr. 18:2483-2494.

Goss, M.J. and D. Goorahoo. 1995. Nitrate contamination of groundwater:
measurement and prediction. Fertilizer Research 422331-338

Hallberg, GR. 1987 . Agricultural chemicals in ground water: extent and
implications. J. Altem. Agric. 223-15

Hamilton,-P.A. and D. Helsel. 1995. Effects of agriculture on ground-water quality
in five regions of the United States. Ground water. 332217-226

Harris, D., R.P. Varoney, and E.A. Paul. 1997. Measurement of microbial biomass
N:C by chloroform fumigation-incubation. Can. J. Soil Sci. 77:507-514.

Heichel, G.H., C. P. Vance, D.K. Barnes, and K. I. Henjum. 1985. Dinitrogen
fixation and N and dry matter distribution during four year stands of birdsfoot

trefoil and red clover. Crop Sci. 25:101-105.

Herridge, D.F., F.J. Bergersen, and MB. PeOples. 1990. Measurement of nirtogen
fixation in soybean in the field using the ureide and natural 15N abundance

methods. Plant Physiol.

Hooker, M.L., D.H. Sander, G.A. Petersen, and LA. Duigger, 1980. Gaseous N
loss from winter wheat. Agron. J. 72:789-792.

Horwath W.R., E.A. Paul, D. Harris, J, Norton, L. Jagger, and K.A. Horton. 1996.
Defining a realistic control for the chlorogorm fumigation incubation method
using microscopic counting and l4C-substrates. Can. J. Soil Sci. 76:459-467.

J anzen, H.H. and SM. McGinn. (1991). Volatile loss of nitrogen during
decomposition of legume green manure. Soil Biol. Biochem. 23:291-297..

J enkinson, D. S. and D. S. Powlson, 1976. The effects of biocidal treatments on
metabolism in soil-V. A measure for measuring soil biomass. Soil Biol. Biochem.

82209-213.

Jones, ME. 1996. Enhancing yield, profitability and nitrogen mineralization in
com-based integrated cropping systems. Ph.D.Dissertation, Michigan State

University, Dept. of Crop and Soil Sciences.

Jones, M.E., R.R Harwood, N.C. Dehne, J. Smeenk, and B. Parker. 1998.
Enhancing soil nitrogen mineralization and crop yield with overseeded cover
crops. J. Soil Water Conserv. 53:253-257.

106

Keeney DR. 1982. Nitrogen availability indecies. pp. 711-733 In: AL. Page et al.
(ed.) Methods of Soil Analysis, Part 2, 2nd Ed. Chemical and microbiological
properties. Agron. Monogr. 9, ASA and SSSA, Madison, WI.

Keeney, DR. 1987. Nitrate in ground water - Agricultural contribution and control.
p. 329-351. In Proc. Conf. On Agricultural Imapcts on Ground Wate, Omaha,
NE. 11-13 Aug. 1986. National Water Well Assoc., Dublin, OH.

Kelly, T.C., Y.C. Lu, A.A. Abdul-Baki, J .R. Teasdale. 1995. Economics of a hairy
vetch multch system for producing fresh market tomatoes in the mid-Atlantic
region. J. Am. Soc. Hort. Sci. 1201854860.

Kunelius, H.T., H.W. Johnston, and J .A. MacLeod. 1992. Effect of undersowing
barley with Italian ryegrass or red clover on yield, crop composition and root
biomass. Agric. Ecosyst. Environ. 38:127-137.

Lichtenberg, E., J .C. Hanson, A.M. Decker, A.J., Clark. 1994. Profitability of
legume cover crops in the mid Atlantic region. J. Soil Water Conserv. 49:582-
585.

Mallarino, A.P., W.P. Wedin, R.S. Goyenola, C.H. Perdomo, and GP. West. 1990.
Legume species and proportion effects on symbiotic dinitrogen fixation in
legume-grass mixtures.

Meisinger, J .J., 1984. Evaluating plant available nitrogen in soil-crop systems. p.
391-416 In Nitrogen in Crop Production. ASA, CSSA, SSSA, Madison, WI.

Musser,-W.N.; Shortle,-J.S.; Kreahling,-K.; Roach,-B.; Huang,-W.C.; Beegle,-D.B.;
F ox,-R.H. An economic analysis of the pre-sideress nitrogen test for
Pennsylvania corn production. Rev. agric. econ. 17:25-35.

National Atmospheric Deposition Project. 1998 [online]. NADP/NTN Monitoring
Location M126. Available at: http:/ /nadp.sws.uiuc.edu/nadpdata/
siteinfo.asp?id=M126&net=NADP, December 10, 1998.

Paul, E.A., H.P. Collins, D. Harris, and U. Schultes. 1998. Soil Organic Carbon
Dynamics in Com-based Agroecosystems of the Central USA: Results from 13C
natural abundance.

Philips, L. and C. Stopes. 1995. The impacts of rotational practice on nitrate
leaching losses in organic farming systems in the United Kingdom. pp. 123-134.
In Nitrogen Leaching in Ecological Agriculture. A B Academic Publishers,
Great Britain.

Peoples, M.B., A.R. Mosier, and J .R. Freney. 1995. Minimizing gaseous losses of
nitrogen. p.565-602 In: P.E. Bacon (ed.) Nitrogen Fertilization in the
Environment. Marcel Dekker, New York.

Power, J .F., and J .S. Scheppers. 1989. Nitrate contamination of groundwater in
North America. Agric-Ecosyst-Environ. 26:165-187.

von Rheinbaben, W. 1990. Nitrogen losses from agricultural soils through
denitrification — a critical evaluation. Z. Pflanzemahr. Bodenk. 153:157-166.

107

Rice, C.W., and J .L.Hav1in, 1994 Integrating mineralizable nitrogen indecies into
fertilizer nitrogen recommendations. p. 1-15.in Soil testing: pOSpects for
improving nutrient recommendatons SSSA special publication (40),. J .L. Havlin
and J .S. J acobsen ed., SSSA Madison, WI

Richards, R.P.; D.B. Baker, N.L.Creamer, J .W. Kramer, D.E. Ewing, B.J.
Merryfield, and L.K. Wallrabenstein. 1995. Well water quality, well
vulnerability, and agricultural contamination in the midwestem United States.
J. Environ. Qual. 25:389-402.

Robertson, GP. 1997. Nitrogen Use Efficiency in Row-Crop Agriculture: Crop
Nitrogen Use and Soil Nitrogen Loss. In: Jackson, L.E. (ed.) Ecology in
Agriculture. Academic Press, San Diego. 1997.

SAS Inst. Inc. 1988. SAS/ STAT user’s guide Vol. 2, version 6, fourth edition. SAS
Inst. Inc., Cary, NC.

Shanaracharyna, N.B., and B.V. Mehta. 1969. Evaluation of loss of nitrogen by
ammonia volatilization from soil fertilized with urea. J. Indian Soc. Soil Sci.
17:423-430.

Solberg, S. O. 1995. Influence of crops and cultivation management on the nitrogen
leaching potential on ecological farms in east norway. In: Nitrogen Leaching in
Ecological Agriculture, pp. 115-121. A. B. Academic Publishers, UK.

Willson, T.C., M.E. Jones, E.M. Parker, and R.R. Harwood. 1996. Is particulate
organic matter a sink for fertilizer nitrogen? Agron. Abst. ASA-CSSA-SSSA
Annual Meeting,

Willson, T.C., E.A. Paul, and R.R. Harwood. 1999a (in process). Managing
biologically active soil organic matter fractions for sustainable crop production.
(This will be submitted to Appl. Soil Ecol. in December, 1998.)

Willson, T.C., E.A. Paul, O. Schabenberger, and R.R. Harwood. 1999b (in
process). Seasonal changes in nitrogen mineralization potential in agricultural
soils: AN OVA and non-linear regression analysis of management effects. (This
will be submitted to Soil Sci. Soc. Am J. in January, 1999.)

Yoshinari, T., R. Hynes, and R. Knowles. 1977. Acetylene inhibition of nitrous
oxide reduction and measurement of denitrification and nitrogen fixation in soil.
Soil Biol. Biochem. 9:177-183.

108

SUMMARY

In these course textured soils potential and actual N mineralization are closely
linked to the supply of organic materials from above and below ground plant
residues. Nitrogen mineralization potential (NMP) was greatest following residue
incorporation in the spring and fall and lowest at peak plant biomass (late summer).
It significantly increased in systems with untilled fallow including the wheat-clover,
wheat-fallow, and historically tilled succession (HTS) treatments. Nitrogen budget
analysis suggests that much of the difference in N mineralization over the course of
the com-com-soybean-wheat rotation at the LF L can be explained by differences in

the return of residue-N from the previous crop.

The use of external organic inputs in the form of composted leaves and dairy
manure increased mineralization (NMP and Nmin*) at the LFL. Application rates
were sufficient to provide high productivity in 1st year corn when combined with
inorganic N from the decomposition of red clover following wheat, but insufficient
to produce the same productivity in 2nd year corn, continuous com or wheat as was
found in the fertilized system. Compost appears to have provided only about 50 kg

N ha-1 yrl between 1994 and 1996, consistent with a mineralization rate of

approximately 9% per year.

Tillage affects these systems directly by incorporating residues and disrupting soil
aggregates and indirectly by altering the plant community. The incorporation of

residues resulted in a temporary reduction in the C:N ratio of microbial biomass and

109

an increase in macroorganic matter and NMP. In unmanaged systems, the lack of
tillage promotes the growth of perennial vegetation that appears to favor the storage
of N below ground. In tilled systems, the longer period between tillage events in the
wheat-fallow and wheat-clover systems promotes the robust growth of winter cover
species and therefore greater supply of NMP within the system and a greater supply
of N to the following corn crop. The lack of spring tillage in wheat increases the
temporal continuity of the rhizosphere and reduces the disruption of soil aggregates,

which may increase the storage of labile organic N in the soil.

Our attempts to find practical indicators of N mineralization in the field have met
with mixed success. In general, mineralization in 70 and 150 day aerobic
incubations provide the best indication of mineralization potential and are similar to
the average of 10, 30, 70, and 150 day incubations used in the AN OVA in chapter
one. However these measurements are time consuming and may be impractical in
many cases. As defined in this paper, measurements of macroorganic matter are
quick and easy and provide more information about NMP than do total C and N or
microbial biomass. However, macroorganic matter measurements may
underestimate the contribution of recently incorporated residues, particularly
legume residues, which may release large amounts of N that does not contribute
directly to macroorganic matter. Macroorganic matter measurements could be used
to estimate N mineralization for the purposes of making fertilizer recommendations
if they are combined with information about the productivity of the previous crop or

cover crop. Additional testing will be required to determine whether macroorganic

110

 

matter provides a useful index of N mineralization potential across soils and

indifferent climatic conditions.

111

 

APPENDIX

The following tables provide supplementary references for the nitrogen fixation

estimates in Chapter 3.

112

 

 

 

 

58: EN... :0 88m ._.
E83 8:8. 626.880 8-8 or
89 .8 8 8..8> Emo. Em £882 8.3 o.. 5.88.5 oz :88
8.8:? m N .3 91 mm or
:8. ._: .: 88> 8.8.8.. m P e. m: - t :P 88.52:. :2 88.8
858.08 80:50 E ad :. mm om T9. 9 md
:8. .m .: 88-8.0 89: .988 8 :N :9. a m: :88:
.8888. :8. a: .58 .8 N: :9. 9. :3
:88 3:38E oo 28 >Emo. 83.08 E. :9. 9. mp
.888. :8. .8 .58 .8 m: :9. 9. :3
mm? .8 8 cmmamcamz Q35 3::8E o 9.8 >Eoo. .88m 2. + :9. 9. mp c838
828 26.8. 500 mm .258. 8......
:8. 3 3.....8. :9...
:8. .a .o :9. 3:. 8 .:__.:o. ::.x. :88:
8.8 o :2 9. on
:8 : :9. 9. :3
:8. .a .o 82.8: no: : .. .8: 9. 8% :88:
8.8.8> m 8.3. cow
8.8:m> m N E. 8-8 cm
8.8:? m ~79 Q?
.89 .m8amN :8 5288... 8.8:? n F .5 8-: on 5.88:... 02 88.2%
8:. 58.8.5. N: 88>: :8; 888.... :888
2 :8 :9. a 5.:
3 our :9. a 6.:
8.88.: E on? a: m .9. 0 Ed
R cm? :9. a to
Nmmw 590.05% Dan @9000 co_~m>m_0 E ONm NM a 79. D F o :38
O\O '72 ox no
858. 8.8:m> EmE.8.h 8.2 2 82.8“. 2802 ..om 5
.838: :o. 8.2:... 88.8858 E9. 2.x. .0. 83.9 9388:... mm 8.8:

113

 

 

 

 

Bow .8 .de:onEm8>> 858.08 8850 E ad ms 8>o.0 8m

~ .3 m:
9.8 .. .3 8 m: I:

N .3 9.

mam: .8 8 26...QO 9.82:8“. .P .3 mm m... I: 83.0 :8”.
N .3 5-..:

:82 ..: .m 888.2 8:8. =9 5.; 8...: P .3 8.2 N: I: .o 08.: 8.6.0 :8
I. .3 :3:

mm? .8 8 8.88... w .3 mo :80. Em 836 By.
89.988: Va .3 8.2

59 .EPEmI :8 858: :8. 5.2. 85.2 F .3 on E80. ...m 85.0 8m

8:. .8988 9.: 5:55”. 88...... e 8.2%: 2.. :8: ES. .8 .96.: :8
ox.

850w .cmEEoo .8E88F 8.2 .x. 8.8.5:. :8 no.0

83.: .8. 8. 8.2.x.%88...JamoE8 E9. 2.x. 8. 838., 8888:... cm 8.8..

114

REFERENCES

F ahmam, DE, and J .R. George. 1993. Dinitrogen fixation and nitrogen transfer
among red clover cultivars. Can. J. Plant Sci. 73: 1047-1054.

George, J. and P.W. Singleton. 1992. Nitrogen assimilation traits and dinitrogen
fixation in soybean and common bean. Agron. J. 84: 1020-1028.

Hardarson, G. 1994. International F AO/ IAEA program on biological nitrogen
fixation. p.189-202 In: P.H. Graham, M.J. Sadowsky, and GP. Vance (eds.)
Symbiotic Nitrogen Fixation. Kluwer Academic Press, Netherlands.

Hardarson, G., F. Zapata, and S.K.A. Danso. 1984. Effect of plant genotype and
nitrogen fertilization on symbiotic nitrogen fixation by soybean cultivars. Plant
Soil 82:397-405.

Heichel, G.H. and KL Henjum. I991. Dinitrogen fixation, nitrogen transfer and
productivity of forage legume-grass communities. Crop Sci. 31:202-208.

Heichel, G.H., C. P. Vance, D.K. Barnes, and K. I. Henjum. 1985. Dinitrogen
fixation and N and dry matter distribution during four year stands of birdsfoot
trefoil and red clover. Crop Sci. 25:101-105.

Herridge, D.F., F.J. Bergersen, and MB. Peoples. 1990. Measurement of nirtogen
fixation in soybean in the field using the ureide and natural 15N abundance

methods. Plant Physiol.

Kohl, D.H., G.S. Shearer, and J .E. Harper. 1980. Estimates of N2 fixation based on
differences in the natural abundance of 15N in nodulating and non-nodulating
isolines of soybean. Plant Physiol. 66:61-65.

Mallarino, A.P., W.P. Wedin, R.S. Goyenola, C.H. Perdomo, and GP. West. 1990.
Legume species and pr0portion effects on symbiotic dinitrogen fixation in
legume-grass mixtures.

Neuhausen, S.L., P.H. Graham, and J .H. Orf. 1988. Genetic variation for
dinitrogen fixation in soybean of maturity group 00 and 0. Crop Sci. 28:769-722.

Ofsu-budu, K.G., K. Noumura, and K. Fujita. 1995. N2 fixation, N transfer and
biomass production of soybean cv. Bragg and its supemodulating mutant
ntleO7 in sorghum mixed cropping at two rates of N fertilizer. Soil Biol. Bioch.

27:311-317.

Sparrow, S.D., V.L. Cochran, and EB. Sparrow. 1995. Dinitrogen fixation by seven
legume crops in Alaska. Agron J. 87:34-41.

Vasilas, B.L., R.L. Nelson, J .J . Fuhrmann, and T.A. Evans. 1995. Relationship of
nitrogen utilization patterns with soybean yield and seed-fill period. Crop Sci.
35:809-813.

Vasilas, B.L., R.L. Nelson, and RM. Vanden-Hervel. 1990. Effect of method and
plant sample on nitrogen fixation estimates in soybean. Agron. J. 82:673-676.

115

 

Warembourg, F .R., F Lafont, M.P. Fernandez. 1997. Economy of Symbiotically
fixed nitrogen in red clover (Tn’fofium pratense L.)

116

 

 

"‘8888