f1:

”5|

3? u

‘1!

k,“ 2%;
~. '. 1:: .
‘ .v.
‘ _ A. 1...
43.33?

13 sf. 3
L' V4“-
(fr

|

'\

I.
q-‘z‘i V1.3-
I“ {35‘ £3.-
-E&'.4‘7,.“'~ A m

X
5‘3

: l
7.3 at“: .
fit‘a4fi-l..-.‘Lg.
‘-"‘a‘gr"“(‘ :.'.
‘r‘ :gh‘fc‘fi "a
H:- m~-.r‘rz-n‘-
:t “if!"
fin;

Ll

n ”13:51;
A .
*3 ‘73:.
~43"

3.
x

 

\

.
. .. aw:-
.. , “:3 -

. :51- '2

”~33 J1"...

"" .. ”.t:.‘.~ .

.. uwa'.-.“ '

In...“

fif‘fig
'7 . y"... .3" . 555%.:
n v H ~ . t - 1.4‘ -.
,‘ffiefi‘kflgfir‘é‘n .figéi‘.‘ WAT:
” 3p". m3£mzfifia§simxv 1:3? ‘
“35:" ":3: :53.“ :~;-:=-.\
mam Namath “v.1. .. :
‘fiSfi'?.‘€.;.’-" -._. -
7 “-0 "flaw—".1
’ k E?”- 1;?”
9.42;»:
.33 ‘ 3'35? ~25"
~ 1 $53» “3
. ' :15. .1 F ‘
1;: .2:
{RP-I ~1- 5:1&
“1

m"
u.
w

'7‘ 1’ \ “5.;‘F3 _
maxi. _
032333;. "953'

-' ‘5 ,gcqs.

.. .3 .
.. \. ”.1...
“13k 32%} 33“ ”:
”Mar, . 3".) _
“gap 1- .‘2. . V
x L. I a. - ‘ ‘7‘ l '
.. 333.. , . . ~.
~ ' .c‘t I ‘: L‘ z“ '1‘
v_4...;=:.:3¢;u ‘ '1: g; . ,3 .1: ”329'
333$?‘i gr ‘ r I
to“. 1. "’
1‘,

LG"

 

14¢“;Qfig: 3 352‘. 15': ’
”’9‘: . 3:91; 7,3353%.

$331333 33319333333
u‘ ..« \ ”t TEX-77%"

a
WMm-gv
L J .

ur- ‘ ‘ .
‘_‘ t
" ‘ .3 we ‘
-933», .9 {5% $135!?»
. ‘ a '-..
I}. 3“ .. s'&‘3
‘: £325“ ”tn-mamgflfi
. . ..,
i!»

.. 339’?ch
.35. 3.3;,

     

, 'L
3, ; 3.4%

xi?”

 

 

i
@.

 

 

 

168
MIIGCH

3 33333 33 333 3333333233333 33 3333333

3 12930

 

 

 

 

 

 

3

 

This is to certify that the

dissertation entitled

DISTRIBUTION AND CYCLING OF 15N-LABELED UREA
IN A KENTUCKY BLUEGRASS TURF

presented by

Eric David Miltner

has been accepted towards fulfillment
of the requirements for

 

Ph.D. degreein Crop and Soil Sciences

Major professor

Date May 12, 1994

MS U is an Affirmative Action/Equal Opportunity Institution O~ 12771

 

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

PLACE N RETURN BOX to man this checkout from your record.
TO AVOID FINES Mum on at Mon duo duo.

DATE DUE DATE DUE DATE DUE

 

351‘? 251933.

FEB Q 9 2000 i

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU loAnNflnnctNkocNEmd Oppontmltylmtmnion
mum

 

 

DISTRIBUTION AND CYCLING OF 15N-LABELED UREA
IN A KENTUCKY BLUEGRASS TURF

By

Eric David Miltner

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

1994

ABSTRACT

DISTRIBUTION AND CYCLING OF 15N-LABELED UREA
IN A KENTUCKY BLUEGRASS TURF

By

Eric David Miltner

The fate of 15N-labeled urea applied to Kentucky bluegrass (Boa
prams L.) turf was studied using intact monolith lysimeters. Soil type was a
Marlette fine sandy loam (fine-loamy, mixed mesic Glossoboric Hapludalfs).
Lysimeters were used only for the collection of leachate. PVC microplots were
installed for destructive soil sampling. Urea was applied at a rate of 196 kg N
ha‘1 yr'1 in five equal applications of 39.2 kg N ha'l, based on two application
schedules. The 'Spring' treatment was fertilized at approximately 38 day
intervals from late April through late September. The 'Fall' treatment was
fertilized from early June through early November. In 1991 only, the April
and November applications were made with 15N-labeled urea (25 atom %
excess). For the Spring treatment, 35% of the LFN was harvested in clippings
over two years. Approximately 30% of LFN was recovered from thatch 18
days after treatment (DAT). This value remained constant for the next year,
then declined. Only 8% of the LFN was recovered from soil 18 DAT. This
increased to 14% two years after application. Over the two year period, LFN
in leachate totaled 0.09 kg ha'l. For the Fall treatment, 38% of the LFN was
harvested in clippings over two years. Eighteen days after the November
application 62% of the LFN was recovered from thatch. This value declined to
35% by the following June. LFN in soil increased from 12% to 25% over two
years. Leachate LFN totaled 0.07 kg ha'1 over the two year period. Volatile

losses of N were suspected for both treatments, but a subsequent experiment to
measure this was inconclusive.

Two years following application of LFN, soils were incubated to
determine N mineralization rates. Over 50 days, 46.1 kg N ha'1 was
mineralized from the surface 10 cm of soil, of which 0.40 kg ha‘1 was LFN.
Addition of N as either NI-I4+ or grass clippings induced a priming effect on
the turnover of soil N.

The results of these experiments showed that the thatch layer and
microbial biomass played extremely important roles in immobilization of LFN.
Mineralization of organic N provides a significant source of N available for
plant uptake. Application of fertilizer N to turfgrass results in very little

leaching potential, even when applied in the late Fall.

To Laura, Ella, and Zelda

for all of your love, understanding, and friendship

iv

ACKNOWLEDGMENTS

I would like to express my gratitude to my advisor, Dr. Bruce Branham,
for his invaluable guidance and friendship during my studies. He provided me
with a great opportunity, offered strong support in the development of my
academic and research skills, encouraged me to develop my own special
research interests, and constructively challenged my assumptions and
hypotheses. Thanks also to the other members of my committee, Dr. Boyd
Ellis, Dr. Ted London, Dr. Eldor Paul, and Dr. Paul Rieke. Each truly made
significant contributions to my work. A special thank you is extended to Dr.
Paul Rieke for his friendship and interest in my studies. The assistance of Dr.
Dave Harris in mass spectrometry analysis is greatly appreciated. Thanks also
to Danielle McMahon for her technical assistance. I am indebted to Mark
Collins for his constant willingness to help when I needed it. I am grateful to
Duke and Terry Alden of Alden Enterprises, Barry Darling and members of the
University Farms crew, and Brian Graff for their assistance in lysimeter
construction and installation. I am also thankful for all of the assistance and
interaction from fellow technicians, graduate students, and student workers.
The cooperation and camaraderie of the entire turf group helps to make
Michigan State University the great place that it is. Further appreciation goes
to Walter Wilkie for his support and encouragement of my professional
development. Thanks also to the United States Golf Association for their

support of this research.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
INTRODUCTION

CHAPTER I: DISTRIBUTION OF SPRING AND FALL APPLIED
15N-LABELED UREA IN A KENTUCKY BLUEGRASS TURF

LITERATURE REVIEW
Plant Uptake
The Thatch Environment
Soil Residence and Transformation
Nitrogen Losses in Runoff
Losses Through Deep Percolation
Soil type effects
Nitrogen source effects
Establishment method effects
Nitrogen leaching from other grass crops
Gaseous Losses: Denitrification and Volatility
Late Fall Nitrogen Fertilization
MATERIALS AND METHODS

Lysimeter Construction
vi

ix

xii

12

14

14

16

18

19

21

23

28

28

Experimental Design
Sample Collection and Analysis
RESULTS AND DISCUSSION

Clipping Yield and Clipping and Verdure N
Clippings
Verdure

Nitrogen in Thatch

Soil Nitrogen
Biomass C and N

Relationships between inorganic and organic N
and N mobility

Leaching of Nitrogen
Mass Balance
CONCLUSIONS
CHAPTER II: IRRIGATION AND NITROGEN SOURCE EFFECTS
ON RECOVERY OF 15N-LABELED FERTILIZER APPLIED
TO TURFGRASS
LITERATURE REVIEW
Urease Activity
Effect of Nitrogen Rate
Effect of Nitrogen Source
Effect of Temperature

Effect of Moisture

Effect of Crop

vii

31

34

4O

40

40

55

55

62

63

7O

81

89

99

101

101

102

103

103

105

105

107

MATERIALS AND METHODS
RESULTS AND DISCUSSION

CHAPTER III: MINERALIZATION AND TURNOVER OF
SOIL NITROGEN IN A TURFGRASS COMMUNITY

LITERATURE REVIEW
Prediction of Mineralization Rates
Mineralization Characteristics of Some Soils
MATERIALS AND METHODS
Long-term Incubation (LTI) Experiment

Mineralization-1mmobilization-Turnover (MIT)
Experiment

RESULTS AND DISCUSSION
Long-term Incubation (LTI) Experiment

Mineralization-Immobilization-Turnover (MIT)
Experiment

Summary
APPENDIX

LIST OF REFERENCES

viii

109

112

117

117

117

120

124

124

125

127

127

131

133

137

145

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

LIST OF TABLES

Fertilization dates for Spring and Fall treatments
in 1991.

Schedule of pesticides applied to lysimeters 1 and 2
(Spring treatment) and 3 and 4 (Fall treatment).

Microplot sampling dates for Spring and
Fall treatments.

Analysis of Variance of Labeled Fertilizer Nitrogen
recovered for Spring treatment for a
split-plot design.

Analysis of Variance of Labeled Fertilizer Nitrogen
recovered for Fall treatment for a
split-plot design.

Total Nitrogen and Labeled Fertilizer Nitrogen content
of thatch, including organic and soil components,
for each sampling date for Spring treatment.

Total Nitrogen and Labeled Fertilizer Nitrogen content
of thatch, including organic and soil components,
for each sampling date for Fall treatment.

Summary of split-plot analysis of variance for concentration
of biomass carbon and total N and 15N (pg g'l)
in various soil pools for Spring treatment.

Summary of split-plot analysis of variance for concentration
of biomass carbon and total N and 15N (pg g'l)
in various soil pools for Fall treatment.

Concentrations of biomass C, N, and 15N (pg g'l)
by depth for eachsampling date for Spring treatment.

ix

33

35

36

41

42

59

61

64

65

66

Table 11.

Table 12.

Table 13.

Table 14.

Table 15 .

Table 16.

Table 17.

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

Concentrations of biomass C, N, and 15N (pg g’l) by
depth for eachsampling date for Fall treatment.

Inorganic, biomass, organic, and total N and 15N
concentrations by depth for the Spring treatment,
sampled on May 14, 1991.

Inorganic, biomass, organic, and total N and 15N
concentrations by depth for the Spring treatment,
sampled on May 26, 1992.

Inorganic, biomass, organic, and total N and 15N
concentrations by depth for the Spring treatment,
sampled on May 14, 1993.

Inorganic, biomass, organic, and total N and 15N
concentrations by depth for the Fall treatment,
sampled on November 26, 1991.

Inorganic, biomass, organic, and total N and 15N
concentrations by depth for the Fall treatment,
sampled on September 17, 1992.

Inorganic, biomass, organic, and total N and 15N
concentrations by depth for the Fall treatment,
sampled on November 30, 1993.

Recovery of Labeled Fertilizer Nitrogen in various
soil pools for each sampling date.

Flow-weighted means of nitrate and ammonium N
concentrations in leachate from Spring and
Fall treatments.

Water volume, inorganic N concentration, atom %
excess, and total amount of LFN leached for
selected samples.

Mass balance of total nitrogen for Spring and Fall
treatments at initial and final microplot sampling
Dates.

Recovery of Labeled Fertilizer Nitrogen form each
canopy increment and total percent recovery at each
sampling date for the Spring treatment.

X

68

71

72

73

74

75

76

8O

83

90

91

93

Table 23.

Table 24.

Table 25.

Table 26.

Table 27.

Table 28.

Recovery of Labeled Fertilizer Nitrogen form each
canopy increment and total percent recovery at each
sampling date for the Fall treatment.

Comparisons of Labeled Fertilizer Nitrogen recovery
in each canopy increment at 18 and 748 days after
treatment.

Summary of split plot analyses of variance at each
sampling time.

Percent recovery of LFN in each canopy level at each
sampling time.

Atom % excess 15N in each sampling increment at
selected sample dates for the Spring treatment.

Atom % excess 15N in each sampling increment at
selected sample dates for the Fall treatment.

xi

95

97

113

114

143

144

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

LIST OF FIGURES

Growing degree days and cumulative clipping yields for
Spring treatment during 1991 and 1992.

Growing degree days and cumulative clipping yields for
Fall treatment during 1992 and 1993.

Cumulative clipping yield and total nitrogen harvested in
clippings from microplots receiving the
Spring treatment.

Cumulative clipping yield and total nitrogen harvested in
clippings from microplots receiving the
Fall treatment.

Cumulative Labeled Fertilizer Nitrogen and Total Nitrogen
harvested in clippings from May 6, 1991 through

April 28, 1992 from microplots receiving the

Spring treatment.

Cumulative Labeled Fertilizer Nitrogen and Total Nitrogen
harvested in clippings from April 28, 1992 through

May 12, 1993 from microplots receiving the

Spring treatment.

Cumulative Labeled Fertilizer Nitrogen and Total Nitrogen
harvested in clippings from November 8, 1991 through
October 22, 1992 from microplots receiving the

Fall treatment.

Cumulative Labeled Fertilizer Nitrogen and Total Nitrogen
harvested in clippings from October 22, 1992 through
October 28, 1993 from microplots receiving the

Fall treatment.

xii

43

44

46

47

48

50

52

53

Figure 9.

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

Figure 18.

Figure 19.

Figure 20.

Labeled Fertilizer Nitrogen recovered in verdure,
cumulative clippings, and verdure plus cumulative
clippings at each sample date from microplots
receiving the Spring treatment.

Labeled Fertilizer Nitrogen recovered in verdure,
cumulative clippings, and verdure plus cumulative
clippings at each sample date from microplots
receiving the Fall treatment.

Cumulative drainage and precipitation plus irrigation
from lysimeters.

Nitrate-N concentration in leachate for each sample
collected for lysimeters receiving the
Spring treatment.

Nitrate-N concentration in leachate for each sample
collected for lysimeters receiving the
Fall treatment.

Cumulative Labeled Fertilizer Nitrogen and Total
Nitrogen in leachate from lysimeters receiving the
Spring treatment.

Cumulative Labeled Fertilizer Nitrogen and Total
Nitrogen in leachate from lysimeters receiving the
Fall treatment.

Nitrate-N and ammonium-N concentrations in soil
from the 0 - 5 cm depth during a 180 day incubation.

Nitrate-N and ammonium-N concentrations in soil

from the 5 - 10 cm depth during a 180 day incubation.

Total nitrate-N + ammonium-N accumulation in soils
during a 30 day incubation.

Total mineralization of 15N from unamended and
N-amended soils during a 30 day incubation.

Mineralization of 15N from NH4+-N and clipping-N

amended soils during a 30 day incubation, corrected
for mineralization from unamended check.

xiii

56

57

82

84

86

87

88

128

130

132

134

135

Figure 21.

Figure 22.

Figure 23.

Figure 24.

Figure 25.

Figure 26.

Diagram of lysimeter core (shaded portion) and
access area, side view.

Map of lysimeter plot area including locations of
lysimeters, microplots, and areas of disturbed soil.

Breakthrough curves for water and bromide applied
over surface of Lysimeters 1 and 2 under conditions of
saturated flow.

Breakthrough curves for water and chloride applied
around interior edges of Lysimeters 1 and 2 under
conditions of saturated flow.

Breakthrough curves for water and bromide applied
over surface of Lysimeters 3 and 4 under conditions of
saturated flow.

Breakthrough curves for water and chloride applied

around interior edges of Lysimeters 3 and 4 under
conditions of saturated flow.

xiv

137

138

139

140

141

142

INTRODUCTION

Turfgrass management is of major economic importance in the United
States. As of 1988 in Michigan alone, approximately 1.2 billion hectares were
occupied with turfgrass with an annual maintenance cost of $1 billion. Of this,
approximately $24 million is spent on fertilizers, not including the amount
spent by homeowners (Michigan Turfgrass Foundation, 1989). Nitrogen is the
fertilizer nutrient required in the greatest amount (Beard, 1973). The fate of
applied fertilizer nitrogen has both economic and environmental impacts. Based
on the figures given above, inefficient utilization of fertilizer N can obviously
result in excessive fertilizer expenditures. Perhaps a greater impact lies in the
ultimate destination of the nitrogen not utilized by plants. The obvious intent
of fertilizer application is for plant uptake and use in growth and development.
However, plants must compete with other organisms in the ecosystem.
Microorganisms in particular play an important role in immobilizing nitrogen
and making it unavailable for immediate plant uptake. The nitrogen is
eventually released, but residence time in the microbial biomass can keep it
unavailable for weeks, and it can be immobilized for even longer periods in
more complex forms of soil organic matter. Inorganic nitrogen can be held in
exchangeable forms on soil particles and eventually released into the soil
solution. From the soil solution, it may be lost to the atmosphere through
either volatilization or denitrification. It may be removed from the system by
surface runoff, or can be lost by deep percolation and eventual entry into
aquifers. This can result in a significant health hazard to animals and humans
if concentrations become high enough. The Environmental Protection Agency
has established a limit of 10 mg NO3’-N L'1 in drinking water supplies (U.S.

Environmental Protection Agency, 1976). Leaching of fertilizer nitrogen into

1

2

ground water supplies has become a topic of much attention in recent years,
and the turfgrass industry has received considerable scrutiny because of the
perception that nitrogen is over-applied to turfgrass, which is sometimes viewed
as an unnecessary crop with no real value to society. There are, however,
many benefits of turfgrass, including insulation from noise, protection form
glare, increased percolation of precipitation and less surface runoff of water,
softening of the ground surface for recreational use, cooling of the ground
surface, and aesthetic benefits, as well as many others. These many benefits
provided by turfgrasses justifies their continued cultivation and future research
that refines management practices and defines the environmental impact of these
practices. The objectives of this research were twofold:

1) to examine the fate of fertilizer nitrogen applied to turfgrass,
including the partitioning of this nitrogen into different pools in the
turfgrass system

2) to compare the leaching potential of fertilizer nitrogen applied in the

Spring (April) and the late Fall (November).
The answers to these questions will help us to understand both nitrogen use

efficiency and the environmental impacts of fertilizer nitrogen use in turf.

CHAPTER I

DISTRIBUTION OF SPRING AND FALL APPLIED 15N-LABELED
UREA IN A KENTUCKY BLUEGRASS TURF

LITERATURE REVIEW

A thorough review of the literature concerning the fate of nitrogen in
turfgrass was recently authored by Petrovic (1990). The review presented here
will not restate all of the literature included in the above published review, but
will concentrate largely on the use of 15N in monitoring fertilizer nitrogen fate.
The reader is referred to Petrovic (1990) for a broader discussion. The present
review will cover nitrogen dynamics throughout the turfgrass community
starting with the plant and moving downward through thatch and soil, and will
conclude with information on losses of nitrogen from the turfgrass community

through runoff, leaching, and gaseous transformations.

ElanLIlmak:

Studies of nitrogen uptake by turfgrasses are scarce. However, in
several experiments Bowman and colleagues have shown potential for very
rapid uptake and transport of N. Bowman et al. (1989b) foliarly applied 50 kg
N ha‘1 as 15N-labeled ammonium sulfate to field grown Kentucky bluegrass,
followed by 0.3 cm irrigation, and recovered 75% of the labeled fertilizer N in
plant shoots five days after application. The application was made in August
1985 and the turf had received no fertilizer nitrogen since the previous
November (50 kg ha‘l), and was in an N-deficient state as described by the

authors. Studying plants grown in solution culture, Bowman er al. (1989a)

3

4

found that perennial ryegrass (Lolium perenne L.) removed nitrate-N from
solution at a rate equivalent to 63 kg N ha"1 in 50 hours and 90 kg N ha'1 in
96 hours. Uptake of ammonium-N was even greater: 50 kg N ha"1 in 38 hours
and 128 kg N ha‘1 in 96 hours. These plants had been previously starved of N
for periods between one and four weeks. For turf that was maintained
continuously with nitrogen, uptake was approximately 50% of the N-deprived
turf for both NO3' and NH4+. Nitrogen uptake by Kentucky bluegrass (Poa
pratensis L.) was 50 and 77 kg N ha'1 from NO3' and NH4+ sources,
respectively, compared to 88 and 114 kg N ha'1 for perennial ryegrass over a
four day period. Increases in root dry weight were noted for plants grown in
the absence of N for one week or more, and the authors conclude that this was
partially responsible for the increased uptake in the deficient turf. They also
observed that NO3’-N uptake was reduced in perennial ryegrass grown in +N
solution culture by mowing between 30 min. and four days prior to application,
while uptake by plants receiving no N for 14 days prior to additional
fertilization was not affected by mowing.

Nitrogen uptake by a mixed stand of Kentucky bluegrass and red fescue
(Festuca rubra L.) grown on a sandy. loam soil under more typical management
practices was quite different (Starr and DeRoo, 1981). Over three years, total
nitrogen removed in clippings in plots where clippings were not returned
averaged 95 kg ha'l, an amount equivalent to 50% of the fertilizer N applied.
Where clippings were returned after mowing, harvested N averaged 137 kg ha‘1
(73%). Total nitrogen and fertilizer nitrogen uptake and dry matter
accumulation was rapid for a period of approximately three weeks following
application. Following this period, nitrogen uptake rates were relatively
constant at approximately 0.44 and 0.23 kg N ha‘1 day'1 where clippings were

returned and removed, respectively. Use of 15N-labeled ammonium sulfate

5

resulted in total fertilizer N uptake of approximately 30 kg ha'1 (one-third of
total application) during a 120 day period following a May application and 20
kg ha‘1 following a September application, regardless of clipping management.
Most of this uptake occurred during the first 30 days. Uptake rates of fertilizer
nitrogen during the steady state period (after 30 days) were very low, 0.05 and
0.04 kg N ha‘1 day'1 where clippings were returned and removed.

Studying perennial ryegrass under pasture conditions, Bristow er al.
(1987) applied 15N-ammonium nitrate at a rate of 60 kg N ha'l. They
observed recoveries in herbage of 33, 49, and 55% for the periods 28, 111,
and 370 days after application. Inclusion of stubble raised these recoveries to
53, 54, and 56%. Similar first-year recoveries in perennial ryegrass herbage of
46 to 53% of applied 15N-calcium nitrate were observed by Dowdell and
Webster (1980, 1984). Recoveries over a three to five year period were 52 to
60%.

Consistent levels of approximately one-third of applied fertilizer nitrogen
were recovered in soft chess (Bromus mollis L.) herbage over a three year
period when the application was made in October. This remained consistent
over different application schedules of 100 kg N ha”1 applied once or for three
consecutive years, or for a single application of 500 kg N ha'l. When the
fertilizer was applied in February at rates of 100 kg N ha‘1 either once or for
three consecutive years, recovery increased to approximately 59% (Jones et al. ,
1977). For subterranean clover (Trifolium subterranean: L.) recoveries from
October fertilizations of 100 kg N ha'1 one time and for three consecutive
years, and 500 kg N ha'1 one time were 11, 15, and 17%, respectively.
Herbage recovery of February applied fertilizer N approached 50%. For
crested wheatgrass [Agropyron desertorum (Fisch. ex Link) Schult.] fertilizer N

recovery in plant taps was approximately 30% in the first year following

6

application (Power an Legg, 1984). An additional 15% was recovered in plant
roots. Over a five year period approximately 40% was recovered in the
herbage. Seventy five percent of the total 15N recovered in the herbage was
recovered in the first year. Rekhi er al. (1982) applied 15N-labeled urea to
rice at a rate of 180 kg N ha“1 in either a single application or in three equal
application of 60 kg ha‘l. Recovery in grain and straw was 22 and 42% for the
single and split applications, respectively.

The literature illustrates wide variability in the potential for uptake and
transport to shoots of fertilizer nitrogen. Important variables include species,
nitrogen source, application rate, and management practices. Potential uptake

rates are very high, but in actual practice utilization is generally 60% or less.

Wm

Thatch has been defined as "a tightly intermingled layer of dead and
living stems and roots that develops between the zone of green vegetation and
the soil surface" (Beard, 1973). Mat is a mixture of soil and organic material,
generally occurring between the thatch layer and soil surface. Operationally, it
is difficult to distinguish discrete thatch and mat layers and the boundary of mat
and soil, generally leading to inclusion of mat with the thatch layer. This
region often serves as an important component of the turfgrass growth medium
(Hurto et al., 1980; Ledeboer and Skogley, 1967). Thatch possesses high
porosity, allowing for water and nutrient influx and air circulation, and does
posses some water holding capacity, although this is generally low due to the
high percentage of macropores (Hurto et al. , 1980). The fact that plant crowns
and roots can reside in the thatch layer has long been established (Ledeboer and
Skogley, 1967). The existence of enzymatic activity (Torello and Wehner,

1983) and active microbial populations (Mancino er al., 1993; Mancino er al.,

7

1988;) in the thatch layer has been documented. Because thatch does include
living plant tissue including crowns, rhizomes, stolons, roots, and probably soil
(depending on sampling technique), it is difficult to distinguish whether
nitrogen in the thatch layer is actually in living plant parts, plant material that
has senesced since actively taking up the N, microbial biomass components, or
non-living organic or inorganic components of the thatch. This can make
interpretation of this data difficult.

Thatch can act as a significant source and sink for fertilizer nitrogen.
Starr and DeRoo (1981) found a bluegrass/fescue thatch to have a total nitrogen
content equivalent to 280 to 510 kg N ha'1 where clippings were removed and
returned, respectivelyu Immobilization in the thatch of 38 and 42 kg ha"1 of
labeled fertilizer N (180 kg N ha'1 applied) was measured during one growing
season. In extracted Kentucky bluegrass cores in the laboratory, under suction
to simulate leaching, 46% of N applied as urea and 67% of N applied as IBDU
remained in the thatch after 15 days (Nelson et al. , 1980). Thirty minutes after
application of either calcium nitrate or ammonium sulfate to Kentucky
bluegraSs, Bowman et al. (1989b) recovered approximately 29% of the nitrate-
N and 46% of the ammonium-N in the thatch. Of the NH4+-N recovered,
53% was soluble and 47% was exchangeable, as determined by saturation
extraction with deionized water and KC] extraction respectively. The
application was made in 0.2 cm water and followed with 0.3 cm irrigation. An
additional 49% of NO3‘-N and 40% of NH4+-N was found in the underlying
soil. Total recovery was 82 and 89%, and declined rapidly by four hours after
application. By four days after application, almost no NO3'-N or NH4+-N
was recoverable by KC] extraction. The authors attribute the amount not

recovered to biological immobilization.

8

Data on the role of thatch in nitrogen fate is scarce. The ability of
thatch to immobilize large percentages of applied fertilizer nitrogen has been
demonstrated. However, the disposition (live plant material, microbial
biomass, non-living organic matter, soil) of that nitrogen in the thatch is
unclear. The developed thatch layer is unique to stands of perennial grasses,
and not all species of grass are prone to develop thatch layers. Generally only
those with secondary lateral shoots are affected (Turgeon, 1991). The
experiments reviewed on pasture grasses utilized species not prone to thatch

accumulation, and so cannot contribute data on this subject.

5 .1 E '1 II E .
Starr and DeRoo ( 1981) demonstrated that although uptake of fertilizer N

was rapid in the first 30 days following application, most of the nitrogen taken
up after that time was from other sources. Of the 180 kg N ha'1 applied as
15N, 26 and 37 kg ha'1 were immobilized in soil organic matter and 38 and 42
kg ha"1 were immobilized in thatch where clippings were removed and
returned, respectively. After the initial 30 day period, nitrogen uptake rates
ranged from 0.23 to 0.44 kg ha'1 day“1 where clippings were removed and
returned. Uptake of 15N-labeled fertilizer N was 0.04 and 0.05 kg ha’1 day’l.
They also determined a quasi-constant rate of uptake of 0.075 kg N ha‘1 day'1
from clippings returned during mowing. Clearly, the majority of the N uptake
was from N mineralized from soil or thatch. Over the entire growing season,
the sources of nitrogen recovered in clippings were as follows: approximately
one-third derived from fertilizer N, approximately one-third from soil N, and
approximately one-third from clippings returned. Of the N derived from
clippings, approximately one-third was from clippings returned in the current

year, and the remainder from clippings returned in the previous two years.

9

These previous two year's clippings constitute a source analogous to soil N:
that is, a long-term mineralizable source. The importance of clipping
management in nitrogen utilization by turf is obvious. Over four years of the
experiment, soil N in plots where clippings were returned increased 32% over
unfertilized controls and 45% over fertilized plots where clippings were
removed.

Increases in N03' concentrations in both a fine sandy loam and a sand
beneath a Kentucky bluegrass turf were observed during July and August by
Rieke and Ellis (1974). The increase was seen not only in plots fertilized in
May, but also where no fertilizer N had been applied. This was attributed to
mineralization of soil nitrogen. Geron er al. (1993) filled lysimeter containers
with a silt loam soil which had been excavated and mixed in equal amounts of
A, B, and C horizon material soil to simulate commercial construction
practices. Seven months later Kentucky bluegrass was established from both
sod and seed. During the first nine months following establishment, N03'-N
concentrations in percolate averaged approximately 20 mg L'l. During the
next year, percolate concentration averaged approximately 3 mg N L4. The
authors attribute the early high levels to a flush of mineralization of soil N
following disturbance. These studies indicate that immobilized soil nitrogen
can become an important source of N upon mineralization. This can have an
impact both on plant utilization and groundwater contamination.

Bristow et al. (1987) observed the dynamic nature of fertilizer nitrogen
transformations in the soil in a perennial ryegrass pasture. Two days after
application of 15N-labeled ammonium nitrate, 45 and 37% of the fertilizer N
was accounted for as soil mineral N and microbial biomass N, respectively.
After ten days, these levels had dropped to 20 and 13%. Soil organic matter

contained 4% of the fertilizer N at this time. After 28 days, soil mineral N,

10

microbial biomass, and soil organic matter contained 3, 3, and 8% of the
fertilizer N, respectively. During this time, cumulative amounts of fertilizer N
in above ground plant tissue were 11, 37, and 52%. A portion of the fertilizer
N was immobilized briefly by microbial biomass but was re-mineralized quickly
and taken up by the plants.

Three years after an October application of 100 kg N ha‘1 as 15N-
labeled ammonium sulfate to soft chess, 54 kg fertilizer N ha’1 remained in the
soil (Jones et al., 1977). Where the same rate was applied in three consecutive
years, 183 kg N ha’1 remained in the soil. When the fertilizer was applied in
February, soil immobilization accounted for 24 and 102 kg N ha'l. Greater
amounts were removed by the plants from the February applications.

Power and Legg (1984) found that between one and five years after
application of 15N labeled ammonium nitrate 25 to 40% of the fertilizer N
remained in the soil. There was a trend of decreasing levels of soil 15N over
time, indicating mineralization. They also observed that 15% of the applied N
was recovered in roots in the year of application. After three years, root N
accounted for 5% of the labeled fertilizer. They concluded application of N
fertilizer to a perennial grassland results in immobilization of significant
amounts of that nitrogen, primarily in plant roots and microbial biomass, and
that this immobilized N is then recycled in the system. Rekhi et al. (1982)
found approximately 40% of applied 15N urea in the soil following rice and
wheat crops. In the following season, 1.5 to 3% of the amount applied was
taken up as a result of mineralization.

Similar relative amounts of applied fertilizer N were recovered in soil
following application to Coastal bermudagrass (Cynodon dacrylon L.) but the
dynamics of the remaiping nitrogen was quite different than that observed by

Power and Legg. Of the 40% of the applied N (560 kg ha'l) remaining in the

11

soil, 12.5% was in plant roots, 6% in inorganic forms, 21% was immobilized
in organic forms, and 0.5% existed as exchangeable NH4+ (Kissel and Smith,
1978). In the following year, 17% of the applied N (almost one-half of what
remained in the soil) was recovered in the forage. Even if all root N was
transported into herbage, an additional 5% (28 kg ha'l) of the applied N was
mineralized and taken up by the plants. The actual amount is probably greater
than this.

Comparing the fate of 15N anhydrous ammonia in soils cropped with
perennial ryegrass or uncropped, Chalk and Keeney (1975) recovered
approximately 22%, 1%, and 0.5% of the applied N in organic, fixed NH4+,
and N03” forms in the cropped soil. In the fallow soil, approximately 9%,
2%, and 60-70% were recovered in these same forms. Plants play an important
role immobilizing nitrogen. The authors offered no explanation of the high
NO3‘ levels in the fallow soil, but it appears that nitrification occurred readily
but without plant uptake immobilization was limited. Unrecovered 15N in the
fallow soil was thought to be lost through biological and chemo—denitrification.

The above studies utilizing 15N labeled fertilizers illustrate that the
majority of fertilizer N uptake by plants occurs in the first growing season, and
in fact the first few weeks after application. Much of the remaining N in the
soil is present in organic forms, either as microbial biomass or plant material.
This immobilized nitrogen is mineralized and again made available to plants,
but at relatively low rates in succeeding years. Efficiency of fertilizer nitrogen
use is therefore low, but unused nitrogen is mostly immobilized and not

susceptible to off-site transport.

12
11° I . E [f

Morton er al. (1988) found very little potential for loss of fertilizer N in
runoff from a Kentucky bluegrass - red fescue lawn. The soil on the site was a
sandy loam with a high infiltration rate. They observed only two events over a
two year period when runoff occurred. The first was caused by rainfall on
snow-covered, frozen ground, and the other occurred following a 12.5 cm
rainfall when the soil was already very wet. Inorganic N concentrations during
these events were 1.1 to 4.2 mg L'l. Losses by runoff were less than 7% of
the total N lost in the aqueous phase.

On bermudagrass golf greens overseeded with perennial ryegrass,
appreciable concentrations of N in runoff were observed, varying with nitrogen
source , timing, and precipitation (Brown et al., 1977). When ammonium
nitrate was applied in February to turf receiving 1 cm irrigation on alternate
days, N concentrations in runoff were fairly constant at approximately 10 mg
L'l, except when approximately 30 mg L‘1 was detected following a 3 cm
rainfall event 25 days after application. When applied in September, a
detection of approximately 60 mg N L‘1 was made 10 days after application
during a daily irrigation cycle. Runoff was detected on limited occasions only
from organic sources of N (ureaformaldehyde, Milorganite) when applied in the
summer or fall, but none from ureaformaldehyde when applied in winter. In a
separate experiment, runoff losses were reported only from greens constructed
of sandy loam soil, and not from greens constructed of sand-based mixtures
(Brown et al., 1982). Losses were greater from inorganic sources (NH4NO3)
than from organic sources (Milorganite, IBDU).

Chichester (1977) compared nitrogen losses through runoff and leaching
under several different cropping systems in lysimeters. In one system,

Kentucky bluegrass maintained as a pasture was grown in lysimeters in a silt

13

loam soil on a 23% slope and received an average of 148 kg N ha’1 yr'l. In
another, corn was grown in a silt loam soil with 13% slope in a conventional
tillage system and received a yearly average of 219 kg N ha'l. A third set of
lysimeters containing silt loam on a 6% slope was cropped to minimum tillage
corn which received a yearly average of 255 kg N ha'l. The plots were
unirrigated but received very similar amounts of precipitation. Over four
years, the bluegrass plots averaged 0.5 cm runoff loss per year, the
conventional corn plots averaged 2.6 cm, and the minimum till corn averaged
0.9 cm. The bluegrass plots lost an average of 1.1 kg N ha‘1 yr‘1 as inorganic
N in runoff water. The conventional corn lost 3.5 kg N ha“1 yr‘l, and the
minimum till corn lost 1.9 kg N ha'1 yr'l. Loss of soil in runoff was 5200 kg
ha"1 yr“l carrying with it 10 kg N ha'1 yr'1 for conventional com, 35 kg ha'1
yr'1 with less than 0.1 kg N ha'1 yr'1 for minimum till corn, and less than 5
kg ha’1 yr'l transporting less than 0.1 kg ha’1 yr'1 for the bluegrass pasture.
Clearly the perennial pasture grass conserves not only nitrogen, but also water
and soil. The pasture system is more efficient than the conservation system for
corn, and losses from the grass are less even from a much more severe slope.
Although N losses from turf through runoff have been reported, they
generally occur only in isolated events triggered by high precipitation. The
turfgrass environment is generally regarded to be one of low runoff potential
due to the presence of thatch, which impedes lateral flow, and a dense root
zone near the soil surface which increases permeability. The perennial nature
of turfgrass stands and pasture grasses probably leads to development of similar
environments. This system allows for very little loss of water, soil, or nitrogen

through runoff.

14

Win

Percolation of fertilizer nitrogen through the soil and into aquifers is the
area of greatest environmental concern in the nitrogen fate puzzle. Nitrate-N is
considered a pollutant of drinking water, and the Environmental Protection
Agency has established a threshold limit of 10 mg L'1 NO3'-N (45 mg L'1
NO3') (USEPA, 1976). The turfgrass industry has received considerable
scrutiny in this area from the popular press. Turfgrass researchers have been
examining this problem for over 25 years. Soil texture, nitrogen source and
rate, and irrigation and precipitation levels all play important roles in

determining leachable nitrogen.

Soil type effects:

Following a single application of 290 kg N ha'1 as NH4NO3 on May 1
to Kentucky bluegrass growing on a fine sandy loam, Rieke and Ellis (1974)
observed a trend towards some downward movement of NO3‘ into late June.
On May 15 soil NO3‘-N concentrations were 44.2, 16.4, and 10.2 mg L‘1 at
depths of 0-15, 15-30, and 30-45 cm, respectively. On May 29 concentrations
at these depths were 41.1, 32.1, and 15.0 mg L'l. By June 24, concentrations
had decreased at each of these depths. From May 1 through June 24, NO3'-N
concentrations in the 45-60 cm depth increased from 5.8 to 10.9 mg L'l. On a
sand soil three weeks following application of 390 kg N ha‘l, concentrations of
N03‘-N in the 0-15, 15-30, and 30-45 cm depths were 46.0, 37.1, and 15.8 mg
L‘l. Two weeks later, concentrations were 8.4, 30.1, and 19.0 mg L'1 at each
depth, respectively. in the 45-60 cm depth, concentrations increased from 5.7
to 16.0 mg L‘1 between May 1 and June 19. In the second year of the study,

significant downward movement of NO3‘ was seen again on the sandy soil.

15

Nitrate concentrations in drainage water for sand-based greens were
observed to be higher than for sand-soil mixtures or sandy loam soil greens for
a variety of nitrogen sources (Brown, et al., 1982, 1977). They observed
leachate concentrations as high as 326 mg NO3' L'1 from sand greens, 314 mg
NO3‘ L'1 from mixed greens, and 180 mg NO3' L"1 from soil greens following
single applications of 163 kg N ha'1 as NH4NO3 (Brown et al., 1982).
Mitchell er al. (1978), studying NO3’ leaching from golf greens using eight
different soil mixtures, proposed that differences in leachate NO3'
concentration might be due more to the amount of total N in the soil than to
soil texture. Mean annual N03' concentrations in percolate from a sandy loam
soil in Connecticut supporting a bluegrass—fescue lawn were 4 mg N L'1 or less
under a variety of fertilization rates and irrigation schedules (Morton er al.,
1988). Over a two year period, Starr and DeRoo (1981) found no increase in
groundwater N03'-N concentrations (2.0 mg N L'l) beneath fertilized sandy
loam plots as compared to samples collected upstream (1.8 mg N L‘l). After
fertilizing the Kentucky bluegrass lawn with 15N labeled ammonium sulfate,
15N was detected in percolate only one time, but total concentration was near
background levels. Morton er al. (1988) and Starr and DeRoo (1981)
concluded that under management practices common to home lawns, the risk of
groundwater contamination from fertilizer nitrogen is extremely low. As a
basis for comparison, data of Wei] et al. (1990) from Maryland illustrates
groundwater nitrate levels under other planting systems. Sampling wells placed
in farmers' fields indicated mean NO3'-N levels in groundwater under manured
fields of corn and soybean of 27 mg N L'1 and 15 mg N L‘1 in unmanured
fields. In hardwood and pine forests, NO3'-N levels averaged 0.15 mg L'l.

16

Nitrogen source effects:

Rieke and Ellis (1974) observed downward movement of ammonium
nitrate through a fine sandy loam and a sand, but no downward movement of
Milorganite or ureaformaldehyde on these soils, or IBDU on the sand. On a
sand green, Brown et al. (1982) found that nitrate leaching was the greatest
from NH4NO3, intermediate from 12-12-12, and the least from Milorganite and
IBDU. Similar results were observed on sand/soil mixture and sandy loam
greens, except that the magnitude of N leached was smaller. They also saw a
small flush of NO3' in the leachate from Milorganite between 20 and 40 days
after application on the sand and sand/ soil mixture greens. Snyder et al. (1984)
found that more N was lost in leachate from ammonium nitrate than sulfur
coated urea when applied to bermudagrass (Cynodon dactylon X C.
transvaalensis) on a sandy soil with surface applications. It should be noted
that NH4NO3 was applied every two months and SCU was applied twice each
month, both at a rate of 50 kg N ha’1 month'l. However, injection of the
same monthly rate of N as NH4NO3 through the irrigation system in small
increments produced the smallest amount of N loss through leaching. Nelson et
al. (1980) found that 54% of applied urea and 15% of applied IBDU leached
from intact turf and soil cores in the laboratory following application of 295 kg
N ha'l. The amount of IBDU retained in the thatch layer was much greater
than the amount of urea, while similar amounts of each source were retained in
the soil. Over a 15 week period of laboratory incubation and leaching,
Bredakis and Steckel (1963) found that urea and ammonium sulfate yielded the
greatest amount of leachable nitrogen, followed by castor pumice and then

Milorganite, Uramite, and nitroform.

17

Irrigation and precipitation effects:

Studying the effects of irrigation and N rate on N03‘ leaching from a
Kentucky bluegrass/red fescue lawn on a sandy loam soil, Morton er al. (1988)
applied 97 or 244 kg N ha'1 as an equal mixture of urea and ureaformaldehyde
and irrigated at one of two rates: either to prevent moisture stress (low rate) or
at a rate of 3.75 cm wk'1 (overwatered). Where no fertilizer was applied, 1.9
and 2.8 kg N ha"1 was lost through leaching (mean concentrations 0.5 and 0.4
mg N L'l). At the low rate of fertility, losses were 3.0 and 13.7 kg N ha'1
(0.9 and 1.8 mg N L'l). At the high N rate, losses were 4.9 and 31.9 kg N ha‘
1 (1.2 and 4.0 mg N L'l). This data illustrates clearly that overwatering
produces unnecessary losses of nitrogen through leaching. However, even at
the high rate of fertilization, NO3’ concentration remained well below the EPA
threshold limit.

Snyder er al. (1984) compared nitrate leaching under daily irrigation to
replace evapotranspiration and an automated system which irrigated as
determined by soil moisture status. Fertilizer was applied as ammonium nitrate
(two month intervals) or sulfur coated urea (bimonthly) to the surface or as
ammonium nitrate through the irrigation system at the time of each irrigation
(fertigation). The rate was 50 kg N ha'1 month'1 for each treatment. The
automated irrigation system produced much lower nitrogen leaching losses
across all fertilization treatments. Daily irrigation resulted in overwatering and
higher nitrate leaching. Fertigation yielded less nitrate leaching than
conventional granular applications. Similar results were obtained when
comparing daily fertilization through an irrigation system with surface
fertilization every three weeks in conjunction with daily irrigation (Snyder er.
al., 1980). The authors collected leachate during three periods characterized

by different environmental conditions. When daily at approximately 2.5 times

18

ET, more N was leached from the conventional fertilization plots. During
another period when irrigation was less than ET but rainfall compensated for
the difference, conventional fertilization again produced more N leaching.
When irrigation and precipitation approximated ET, leaching losses between the
treatments did not differ. These studies indicate that metering fertilizer
nitrogen in small increments through the irrigation system as an effective way
of limiting losses by leaching. This holds true even when over irrigating,
presumably because the smaller amounts of N applied at one time are cycled
within the soil are rapidly immobilized.

Webster and Dowdell (1984) grew perennial ryegrass in lysimeters under
pasture conditions in clay loam and silt loam soils. Under conditions of natural
precipitation, 7.3 and 3.7% of the applied 15N-labeled calcium nitrate was
recovered in the leachate from the clay loam and silt loam soils, respectively.
Supplemental irrigation to provide 120% of mean annual precipitation yielded
leaching recoveries of 8.0 and 3.8%. In a third treatment, precipitation was
withheld for four weeks prior to fertilizer application. Following fertilization,
irrigation in an amount equivalent to the precipitation withheld from the plots
was applied over a two week period. Percolate recoveries increased to 8.2 and

17.9% , respectively.

Establishment method effects:

Method of establishment of turf can also have an impact on nitrate
leaching (Geron er al., 1993). Kentucky bluegrass established from seed on a
mixed silt loam soil yielded greater NO3’-N concentrations in leachate (15.9
mg L'l) than sodded bluegrass (9.5 mg L'l) during the first three months after
establishment. There was no significant difference during the next three

months, but for the subsequent 15 months the sodded plots yielded greater

19

concentrations than the seeded plots (7.9 and 2.9 mg N L‘l, respectively).
Greater root mass in seeded plots at that time was given as an explanation.
During the first nine months following establishment, average NO3'-N
concentration in the leachate was approximately 20 mg L'laveraged for both
establishment methods. For the next year, average concentrations were
approximately 3 mg L'l. The authors attribute this to mineralization following
soil disturbance. They caution that construction and establishment may be a

high risk time for nitrate leaching.

Nitrate leaching from other grass crops:

When ammonium sulfate was applied to soft chess at 100 kg N ha'1 in
October, 16 kg N ha'1 was recovered in leachate over three years (Jones et al.,
1977). When the same rate was applied for three consecutive years, the total
loss was 29 kg N ha‘l. A single application of 500 kg N ha“1 produced 68 kg
N ha’1 in the percolate over three years. For a both a single February
application of 100 kg N ha'1 and three annual applications of the same amount,
3 kg N ha'1 was detected in leachate. Single and repeated applications of 100
kg N ha'1 yr'1 and a single application of 500 kg N ha'1 yr"1 in October to
subterranean clover yielded leaching losses of 27, 68, and 106 kg N ha'l.
February rates of 100 kg N ha“1 in a single application or repeated produced
leaching of 9 and 17 kg N ha'l. Timing appears to be a very critical factor in
both of these cropping systems.

In bluegrass and corn systems, Chichester (1977) found mean inorganic
nitrogen concentrations in lysimeter drainage from bluegrass pasture over a four
year period to be 1.6 mg N L'l. Total loss of N was 5.8 kg ha'lyr’l. For
conventional and minimum tillage corn, concentrations were 50 and 39 mg N L'

1 and total losses through leaching were 167 and 149 kg N ha"1 yr'l. Low and

20

Armitage (1970) compared nitrogen leaching from meadow fescue (Fesruca
elatior), white clover (Tnfolium repens), and fallow soil in intact monolith
lysimeters (sandy loam soil). Over a three year period, total N recovered in
percolate was 3, 132, and 269 kg N ha"1 for the grass, clover, and fallow
lysimeters, respectively. Total N in rainfall during this time was 39 kg ha’l.
The perennial pasture grass was very effective in mining available nitrogen
from the soil. Dowdell and Webster (1980), growing forage perennial ryegrass
in lysimeters containing loamy sand soil, found that 2 - 5% of applied 15N-
calcium nitrate (400 kg N ha‘1 yr‘1 application rate) was recovered in leachate
in the first year. This was 60 - 70% of the total N in the leachate. Mean N03"
-N concentrations in three lysimeters were 7, 4, and 16 mg N L‘l. After the
first year, 0.1% or less of the applied 15N was recovered in leachate in each
successive year. Over three years, 15N labeled fertilizer comprised 45 - 60%

of the total N in the percolate.

The literature indicates that nitrate leaching can be widely variable
dependent upon a number of factors, including soil texture, nitrogen rate and
source, and irrigation and precipitation. Soil texture is an important factor
(Brown et al., 1982, 1977; Mitchell et al., 1978; Morton et al., 1988; Rieke
and Ellis, 1974; Starr and DeRoo, 1981), but is usually dictated either by
location or by use, as in the case of golf course putting greens. Good
management practices must be employed to limit nitrogen losses to
groundwater. Slow release sources provide less potential for loss than easily
soluble sources (Bredakis and Steckel, 1963; Brown, et al., 1982; Nelson et
al., 1980; Rieke and Ellis, 1974; Snyder er al., 1984;), but soluble sources can
be used if applied in reasonable quantities and watered judiciously. Frequent

application of small amounts is an effective method to limit leaching losses

21

(Snyder et al., 1984). Over irrigation can create a high potential for leaching,
especially in sandy soils (Morton et al., 1988; Snyder er al., 1984). Managers
must make the correct decisions in managing fertilizer applications in order to

conserve fertilizer nitrogen and protect the groundwater supply.

G I 'D"Ei' 11”.].

Denitrification is the stepwise reduction of N03‘ to N2, which can either
be mediated biologically or occur through a number of chemical reactions
(chemodenitrification). Several pathways are possible by which
chemodenitrification may occur. Examples include (Paul and Clark, 1989)

a) the reaction of a-amino groups with nitrite:

RNH2 + HNO2 —) ROH + H2O + N2
b) the reaction of nitrite with ammonium:

HN02 + NH4 —) N2 + H2O
c) and spontaneous decomposition of nitrous acid:

3HN02 -> HN03 + H2O + NO.

Biological denitrification is an enzymatic process mediated by soil
microorganisms. The general reaction sequence is:

N03‘ -9 N02' (-) NO) —> N20 —) N2.
There is some uncertainty as to whether or not NO is an intermediate in the
process. Both N20 and N2 can be terminal products and escape into the
atmosphere. The entire denitrification sequence requires more than one species
of organism to reach completion. The reduction of nitrate to nitrite is mediated
by many different bacteria. The subsequent steps are mediated by a much more
restricted group (Paul and Clark, 1989).

Presence of nitrate, limited oxygen availability, an available carbon

source, and presence of denitrifying bacteria are all required for denitrification

22

to occur. Conditions enhancing denitrification include high nitrate
availability, high organic matter content, high soil water content, near neutral
soil pH, and soil temperatures between 5 and 75°C (Paul and Clark, 1989).

Mancino et al. (1988) have been the only authors to publish data on
direct measurements of denitrification from turfgrass. They studied the effects
of soil texture, soil moisture content, and soil temperature on denitrification
from a Kentucky bluegrass sod using the acetylene reduction technique. For
silt and silt loam soil types, only 0.1 and 0.4% of N applied as KN03 (45 kg N
ha'l) was recovered as N20 when soil was at 80% saturation at 22°C for 10
days. Above 80% saturation, denitrification losses increased, with a maximum
of 5.4 and 2.2% loss from the silt and silt loam soils, respectively. For the silt
soil at a soil water content of 75% saturation, denitrification losses increased
linearly with temperature between 22 and 30°C from 0.02 to 0.11% of applied
N. Maximum losses from the silt and silt loam soil were 94 and 46% of
applied N over 10 days when the soils were at 100% saturation and the
temperature was 30°C.

Oxidation status of Kentucky bluegrass thatch and underlying silty clay
loam soil was measured under conditions of no irrigation, daily irrigation (0.6
cm day'l), and continuous irrigation (0.6 to 1.2 cm hr'l) on plots with either
natural drainage or drainage impaired by a plastic sheet at a 6 cm depth
(Thompson, et al., 1983). Over seven days, non-irrigated thatch and soil
remained well-aerated under both drainage regimes. Naturally drained plots
irrigated daily remained well aerated, but under continuous irrigation low
oxygen content after three days indicated reducing conditions. Impaired
drainage created low oxygen concentrations and reducing conditions in the
thatch after one and three days for the continuously and daily irrigated plots,

respectively. Reduced oxygen content of the underlying soils was measured at

23

one and seven days for each of these treatments. Temperatures in the thatch
were as high as 36°C at midday. This data indicates that conditions for
denitrification could exist where poorly drained soils are present and irrigation
is applied daily or more frequently. These types of conditions probably occur
with regularity on selected turfgrass sites.

Rekhi er al. (1982) made direct measurements NO, N20, and N2 for 70
days following application of 15N labeled urea to flooded rice. Loss as N20
or NO was 0.0001% of applied N. Loss as N2 was 7.2% and 4.8% of 180 kg
N ha'1 applied at one time or split in three equal applications.

The subject of volatility is addressed in detail in Chapter 2. Losses of N
from turf due to volatility have been measured directly by several researchers
(Bowman er al., 1987; Wesely et al., 1987; Torello et al., 1983; Nelson et
al., 1980; Volk, 1959). The activity of the enzyme urease in turf has also
been documented (Torello and Wehner, 1983).

Many researchers conducting mass balance experiments with 15N have
estimated total loss through gaseous evolution of N by attributing non-recovered
15N to denitrification and/or volatility. Starr and DeRoo (1981) attributed
losses from 24 to 36% of ammonium sulfate applied to Kentucky bluegrass to
denitrification and volatility. A loss of 47% of 500 kg N ha‘1 applied to clover

in the fall was attributed to these two mechanisms by Jones et al. (1977)

I Ellll' E 'l' .

Fertilization of cool season turfgrasses in the late fall (early November
in Michigan) has been a recommended practice for a number of years, primarily
due to early spring green up and positive growth response, without a large flush
of growth as is often seen with an early spring application. This practice has

received some attention and criticism from scientists and industry professionals.

24

It is sometimes perceived that there must be increased risk of groundwater
contamination by nitrates if N is being applied after the time that most plant
growth has slowed or stopped. One of the objectives of this experiment was to
assess this risk.

Wilkinson and Duff (1972) studied color and growth response, cold
tolerance, and chlorophyll content of Kentucky bluegrass to nitrogen
fertilization throughout the fall. Ammonium nitrate was applied to separate
plots at a rate of 100 kg N ha"1 on either October 1, October 15, November 1,
November 15, December 1, or December 15. Application on November 1 or
earlier resulted in enhanced green up and chlorophyll content throughout the
fall and winter. Fertilization on or after November 15 did not result in a
substantial fall color response, but these plants had higher chlorophyll contents
in mid-April during the spring green up period. Plants fertilized on October 1
or on or after December 1 had adequate cold tolerance, but fertilization
between October 15 and November 15 reduced cold tolerance. Incorporation of
N at this time apparently delayed the hardening process. Fertilization on
October 1 resulted in increased growth in the fall. Later applications did not
produce increased growth, but growth rates declined more slowly than the
unfertilized control. In the spring, plants which received fertilizer later in the
fall displayed greater growth rates. Based on these results, fertilization in early
November would appear to produce the greatest benefits: enhanced color but
no increased growth in the fall, and enhanced chlorophyll content and growth
response in the spring. Some cold tolerance may be sacrificed, however.
These responses are probably largely dependent on temperature, and so may
vary with location and year.

Response of 'Kentucky 31' tall fescue (Festuca arundinacea Schreb.) and

‘Cohansey' and 'Pencross' creeping bentgrasses (Agrosris palustris Huds.) to

25

nitrogen fertilization in the fall and winter were studied in Virginia (Powell et
al. , 1967). Ammonium nitrate was applied in single applications at a rate of 50
kg N ha'1 at the following times: October; October and December; October,
December, and February; and October through February, monthly. It was also
applied in January only and monthly from October through February at 100 kg
N ha‘1 month’l. Ureaformaldehyde was applied either in October only or in
October, December, and February at a total rate of 500 kg N ha'l. Both
species responded similarly in color and growth response. Each additional
increment of soluble N resulted in improved color between November and May.
For the single October application, however, turf was brown by January. A
single application of 100 kg N ha"1 in January produced a rapid color response
from brown to green. None of the treatments resulted in increased top growth
between December and March. Greater clipping yields resulted beginning in
March in response to increasing rates of soluble N. Yield from
ureaformaldehyde treated plots responded more slowly in March, but yield was
similar to the high rate soluble N plots by May. Carbohydrate concentrations
in stems were higher than in leaves from October through February. After
February, concentrations in leaves were as high or higher than in stems.
Carbohydrate concentration was generally higher in low N treatments.
Photosynthetic rates of the Cohansey bentgrass were measured on four dates
between January and March. Turf fertilized with higher rates of N showed
higher net photosynthesis.

Hanson and Juska (1961) measured root, rhizome, crown, and top
growth of 'Merion' Kentucky bluegrass between October and May in response
to nitrogen fertilization in the fall and winter. Root plus rhizome weights of
untreated controls increased 4.4 g plot'l between October and May. Plots

receiving 147 kg N ha'1 as NH4NO3 in September or September and October

26

showed additional increases of 0.8 and 2.2 g plot’1 respectively. These figures
include losses of 2.6 and 1.9 g plot'1 compared to the control during May.
Therefore, the balance at the end of April was additional growth of 3.4 and 4.1
g plot"l (increases of 57 and 68% as compared to the control). Increases in top
plus crown weights during this period were 52.9 g plot'1 for the unfertilized
check , and an additional 7.0 and 17.0 g plot'1 for September and September
and October fertilized turf. There were also decreases in May for top and
crown growth. The balance at the end of April was 14.9 and 21.2 g plot‘1
additional growth (89 and 126% increases).

These three experiments demonstrate that metabolic activity in cool
season turfgrass is maintained through the winter months. Although rates and
timing of fertilizer applications were not the same as current recommended
practices, the potential of late fall fertilization to stimulate root development
and spring recovery is clear. Concerns over the fate of unutilized nitrogen
applied at this time need to be addressed. Geron et al. (1993) examined this
issue in relation to establishment methods of Kentucky bluegrass and nitrogen
source and application timing effects. Between the summer of 1989 and the fall
of 1990 no significant differences were detected in N03’-N concentrations in
percolate from plots receiving 72.75 kg N ha'1 in November as compared to
plots receiving no November N (equivalent annual rates of 218 kg N ha’l).
However, during December of 1990 and January 1991 November fertilized
plots had a significantly higher NO3‘-N concentration in the percolate (3.37 vs.
2.39 mg N L‘l). The paper does not present data after this time. During the
winter and spring of 1990, concentrations were 14.07 and 1.10 mg N L‘1 for
plots not receiving no November N and 15.81 and 1.52 mg N L’1 for those
receiving N in November, but these differences were not significant. It should

be noted that these plots were established in May of 1989. High N03'-N

27

concentrations during the first winter were due largely to mineralization of
nitrogen following disturbance of the soil in the lysimeter filling process. In
addition, leaching characteristics of this system may be different than those of
established mature turf.

Chalk and Keeney (1975) applied 15N anhydrous ammonia to pots seeded
to ryegrass or left fallow under simulated fall and spring conditions. Recovery
in plants and soil were very similar for both treatments on two of three soils
tested. Total recovery averaged 96% of applied 15N for both spring and fall
treatments averaged over these two soils. For the third soil, recovery was 97%
for the spring treatment and 89% for the fall treatment. Losses from this
particular soil were 40 to 50% when fallow, and the authors attribute this to
denitrification. They explain that denitrification might be higher following
freeze-thaw cycles due to increased availability of organic matter serving as an

energy source for denitrifying bacteria.

MATERIALS AND METHODS

I . C .

Four intact monolith drainage lysimeters were constructed at the
Hancock Turfgrass Research Center at Michigan State University to study
nutrient and pesticide leaching. The soil type at the site is a Marlette fine
sandy loam (fine—loamy, mixed mesic Glossoboric Hapludalfs) and had been in
turfgrass for six years prior to construction of the lysimeters and was planted to
corn prior to that. Two of the lysimeters, constructed in 1989 and 1990, were
constructed with a bare soil surface following removal of the existing sod. The
other two, built in 1991, were constructed with newly established turfgrass
intact. The lysimeters are constructed of grade 304 stainless steel 0.5 cm thick
and are cylindrical in shape, 1.14 m in diameter (1m2 surface area) and 1.2 m
deep. The bottoms of the lysimeters have a 3% slope so that leachate collects
on one side where a drainage tube was later installed. To install each
lysimeter, a soil monolith was excavated by hand in depth increments of
approximately 20 cm, so that the diameter of the monolith was slightly greater
than the internal diameter of the container. After the first excavation, the
open-ended lysimeter container was placed over the monolith and downward
pressure was applied with a backhoe to slide the container over the monolith.
Excavation and enclosure of the soil core continued incrementally until the
container was filled and the soil surface was flush with the top edge of the
container. The soil core was excavated for an additional 15 cm so that it
extended below the bottom of the container. This exposed portion of the
column was then excavated by hand from two sides of the lysimeter, working
inward toward the center of the core. When approximately 20 to 25 cm of soil

was removed from each side, steel H-beams were placed under the cores,
28

29

oriented parallel to each other. A sheet of steel slightly wider than the
lysimeter was then placed on top of these beams and was pushed with a backhoe
so that it slid between the beams and the lysimeter core, severing the core from
the soil below. A round piece of 1.8 cm thick plywood, cut to fit the
lysimeter, was placed on top, the lysimeter was wrapped with chains, and the
entire core, including wood cover, steel sheet, and H—beams were lifted from
the excavation site using a small truck crane. The lysimeters were designed
with a pair of protruding axles on the outer surface on opposite sides at the
center of mass of the core. This allowed the lysimeters to be easily inverted
when lifted from the excavation pit. Once inverted, the core was unwrapped
and the steel beams and plate were removed, revealing the bottom soil surface
of the lysimeter core. Approximately 3 cm of soil was removed with a small
hand shovel and this void was then filled with 1 - 2 cm pea gravel that had been
rinsed in 1% sodium hypochlorite and water. The stainless steel bottom was
then welded in place, a hole was drilled, and a stainless steel drain tube
approximately 1.5 cm diameter and 10 cm long was welded into place. The
core was then reinverted.

Manhole structures had been constructed of 0.32 cm steel around a steel
beam framework. An access/work area measuring 1.14m wide x 1.35m long x
1.53m tall with a top entry hatch door was designed with a platform extending
from one side that the lysimeter core rested upon. The top surface of this
platform had a 3% slope to match the bottom of the lysimeter. When the core
was set on the platform, the top edge of the core was flushwith the soil surface
and the bottom of the core was approximately 0.6m above the floor of the
manhole allowing for placement of a sampling vessel. The manhole was set
into a different hole from which the cores were extracted. Cores were not

removed from the same site as their final destination because of a concern over

30

soil disturbance in the area of construction. The two locations were within 10m
of each other and were of the same soil type. After placing the cores on the
manhole structures, the two units were welded together along all external
contact areas to create a sealed manhole area with access to the lysimeter for
sampling. Excavation holes were then backfilled, matching soil types for each
distinct layer of the profile.

The lysimeters were installed in a 18m x 36m plot area reserved for all
future lysimeter studies. In May 1990 the entire area was sprayed with
glyphosate, prepared for seeding, and in June was seeded to 'Adelphi'
Kentucky bluegrass at a rate of 49 kg ha'l. Fertilizer (12-12-12) was applied
at a rate of 49 hg N ha'1 prior to seeding. Additional fertilizer was applied at
a rate of 24.5 kg N ha'1 on June 15 (as 12-12-12) and on July 11, July 24,
August 7, August 21, and August 27 as urea. Due to a high soil seed bank of
annual bluegrass (Poa annua L.) and establishment of the desired species at a
less than optimum time, the stand of turf that developed was of poor quality.
In September the area was again renovated with glyphosate and sodded with a
blend of Adelphi, 'Nassau', and 'Nugget' Kentucky bluegrasses (equal
proportions by weight at seeding). Urea was applied at a rate of 24.5 hg N ha'
1 on September 14 and 27 . Following establishment a thin piece of sheet metal
360 cm long and 5 cm wide was installed inside the interior edge of each
lysimeter to create a slight vertical extension of the lysimeter containers. This
piece extended approximately 1.25 cm above the soil surface. This prevented
the turf from both inside and outside the lysimeters from crossing the lysimeter
edges, and also prevented the possibility of surface runoff from the lysimeters.

In March 1991 microplots for soil sampling were installed adjacent to the
lysimeters. Use of confined microplots reduces the area to which 15N-labeled

fertilizer must be applied because lateral movement is not a concern. This was

31

necessary in order to conserve 15N. The microplots were constructed of 20 cm
diameter PVC, 60 cm in length. One edge of the microplots was beveled to
limit soil resistance at the leading edge during installation. They were installed
by pushing directly into the soil with a hydraulic cylinder unit constructed
especially for this purpose and mounted to a tractor. The thatch was first cut
with a knife around the inside edge of the microplot because it provided a great
deal of resistance against initial penetration of the PVC tube. The microplots
remained open at the bottom after installation. This method preserved the
native soil structure of the microplots and the surrounding area. A total of 64
microplots were installed in a randomized complete block layout with four
replicates.

In June of 1991 construction on the second pair of lysimeters started and
was identical procedurally to the first set, except that the established sod was
left in place. Complete enclosure of the cores and attachment of bottoms was
completed on August 1 and lysimeters were in place and the holes were
backfilled by September 12. These lysimeters were placed in the holes created
by excavation of the original set of cores near the center of the plot area.
These holes had been previously refilled but the same site was re-excavated so
that the area where soil was disturbed was minimized. This maximized the area
of the plot where native soil structure was preserved. These lysimeters were

completed in June.

E . l E .

The experiment was initiated in April 1991 with fertilizer application to
the two completed lysimeters. There were two treatments in the study, defined
by fertilizer application timing. Both treatments received a total of 196 kg N

ha'1 yr."1 as urea, applied in five equal applications of 39.2 kg N ha'l. The

32

'Spring' treatment received its first application in late April and succeeding
applications at approximately 38 day intervals, the final application being made
in late September. The schedule of application for the 'Fall' treatment was
similar except that there was no April fertilization and there was an additional
application made in early November. Exact dates for all fertilizations in 1991
are shown in Table 1. The April and November applications were made with
15N-labeled urea (24.9613 atom % excess in April, 25.2283 atom % excess in
November). Use of this stable isotope of nitrogen allowed for discrimination of
those particular fertilizer applications from all other N in the environment. For
1992 and 1993 fertilizer application dates were similar to those in 1991. The
15N-labeled material was applied in 1991 only.

The 15N-labeled fertilizer was applied by hand as a solution to each
lysimeter and microplot in 0.05 cm water. For the microplots this was
dispensed from a 125 ml polypropylene bottle which had five holes drilled in
the top. Lysimeter fertilization was performed with a plastic watering can.
Fertilization of each individual plot was immediately followed with irrigation
with an additional 0.45 cm water applied from the same container to ensure that
all of the fertilizer solution was rinsed from the container. After all lysimeters
and microplots were fertilized and irrigated, each plot was covered individually
with a piece of plywood and surrounding border areas were fertilized with non-
labeled fertilizer. This border fertilization and all other applications of non-
enriched fertilizer to microplots and lysimeters were made with a commercial
boom sprayer utilizing flood-jet nozzles at a rate of 561 L ha'1 (60 gal A'l).
This was immediately followed with 0.5 cm irrigation.

Soil samples were collected from the surrounding plot area annually,
submitted to the Michigan State University Soil Testing Laboratory for

analysis, and fertilized with potassium based on recommendations developed for

33

Table 1. Fertilization dates for Spring and Fall treatments in 1991.

 

 

__$P1i££__ Fall
April 26 ‘1’
June 4 June 4
July 12 July 12
August 19 August 19
September 27 September 27
November 8 ‘1'

 

'1’ These applications included l’N labelled urea.
All applications were at a rate of 39.2 kg N ha“.

34

turfgrass in Michigan. Annual rates of 49 and 98 kg K20 ha"l were applied in
1992 and 1993, respectively. Application of fertilizer phosphorus was not
necessary. Irrigation was applied as necessary to maintain a high quality turf.
A concurrent study investigating vertical mobility of common turfgrass
pesticides was conducted in the lysimeters. Several pesticide applications were
made and are recorded in Table 2. Pesticides applied to lysimeters were also
applied to corresponding microplots. Additional pesticides were applied

sparingly as needed to manage weeds and diseases.

5 l C 11 . l 3 l .

Leachate was collected from the lysimeters in 19 L glass jars. These jars
sat within 76 L plastic basins to catch overflow in the case of very large
drainage events. Jars were emptied when approximately 1/2 full, usually every
7 to 10 days, but sometimes more or less frequently depending on precipitation
patterns. Total mass of water was recorded and approximately 100 ml was
saved in a 125 ml polypropylene bottle and frozen until analysis.

Clippings were collected from all plots approximately weekly throughout
the growing season as dictated by growth rate. A pair of manual hand clippers
was used to cut the grass while holding a hand-held vacuum against the clippers
in order to collect clippings as quantitatively as possible. Clippings were dried
at 65°C for 72 hr then ground to pass an 80 mesh screen prior to analysis.

Microplots were excavated periodically for soil nitrogen analysis.
Sampling dates for each treatment are given in Table 3. This sample schedule
was designed so that for each treatment four sampling dates occurred in the first
year following 15N application and two sampling dates occurred in the second
year. For both treatments, the first sample was taken 18 days after application

and the final sample was removed approximately two years (748 days) after

35

Table 2. Schedule of pesticides applied to lysimeters l and 2 (Spring treatment)

 

and 3 and 4 (Fall treatment).
Application rate

_Date_ Pesticide kg a.i. had Lysimeters
8/12/91 isazofos 2.24 1 & 2
8/21/91 chlorothalonil 9.56 l & 2
9/17/91 dicamba 0.12 l & 2
9/17/91 2, 4-D 1.14 1&2

5/3/92 fenarimol 0.76 l & 2
6/ 18/92 propiconazole 0.84 1 & 2
7/21/92 triadimefon 1.53 1 & 2
7/21/92 metalaxyl 1.53 3 &4
7/21/92 chlorothalonil 9.56 3 & 4

8/5/92 metalaxyl 1.53 1 & 2

8/5/92 chlorothalonil 9.56 3 & 4
8/13/92 metalaxyl 1.53 3 & 4
8/20/92 chlorothalonil 9.56 3 & 4
9/4/92 chlorothalonil 9.56 3 & 4
9/4/92 metalaxyl 1.53 3 & 4

 

36

Table 3. Microplot sampling dates for Spring and Fall treatments.

 

 

 

Treatment
Date Sampled
May 14, 1991 Spring, Fall
June 21, 1991 Spring
October 14, 1991 Spring
November 26, 1991 Spring, Fall
May 26, 1992 Spring, Fall
June 29, 1992 Fall
September 17, 1992 Fall
November 30, 1992 Spring, Fall
May 14, 1993 Spring, Fall

November 30, 1993 Fall

 

37

application. The sample collected in May 1991 for the Fall treatment was used
for background N and 15N levels.

Samples were collected by first excavating the micrOplots intact. A 1.9
cm diameter soil probe was used tocollect a sample from the 60 - 100 cm depth
removing five core samples from directly below the microplot. After transport
to a laboratory, each PVC microplot was split open longitudinally to expose the
soil core within. The top segment of the core was separated from the rest of
the core just below the thatch layer using a kitchen knife. Verdure was defined
as all green plant material above the thatch layer, and was removed using
scissors. From the remainder of the excised section, soil was scraped away
until a layer of rhizomes was reached. These rhizomes defined the bottom of
the thatch layer. Soil removed during this operation was returned to the top of
the remainder of the core. Verdure and thatch samples were dried at 65°C for
72 hr. Because of different grinding methods, thatch samples were separated
into organic and soil components by hand massaging. Verdure and thatch
organic matter, as this material was labeled, were ground to pass an 80 mesh
screen with a Wiley mill. Thatch soil was prepared for total N and 15N
analysis by pulverizing with a rolling pin until it was a fine powder.

The remainder of the soil core was sectioned by depth into the following
increments: 0 - 5 cm; 5 - 10 cm; 10 - 20 cm; 20 - 40 cm; 40 - 60 cm.
Three subsamples were then removed from each increment. One was put into a
soil sample tin and dried at 65°C for 72 hr for soil moisture determination and
nitrogen analysis. Soils were not dried at 110°C because of the possibility of
organic matter decomposition at this temperature (Gardner, 1986). Another
subsample was stored at 5° for 2 - 5 days for microbial biomass determination.

The third subsample was frozen at -25°C.

38

Inorganic nitrogen (NO3', NH4+) was determined first by extracting
duplicate samples of dry soil with 1N KCl (5:1 v/w). Nitrate (Lachat
Instruments) and ammonium (Lachat Instruments) were determined by flow
injection analysis on a Lachat QuikChem autoanalyzer (Lachat Instruments,
Milwaukee, WI). Following this analysis, inorganic N was converted for 15N
analysis by the diffusion method of Brooks er al. (1989). A subsample of the
dried soil was prepared for total N and 15N analysis by pulverizing as
described above.

Microbial biomass was determined for each depth increment below the
thatch layer by the chloroform fumigation-incubation method (CFIM)
(Jenkinson and Powlson, 1976) using 25g dry weight equivalent of field moist
soil. Evolved CO2 was trapped in 2N NaOH and titrated with 1.5 N HCl.
Non-fumigated samples were also incubated but this control value was not used
in the calculations (Voroney and Paul, 1984). Samples were prepared for
biomass N and 15N analysis by extraction of fumigated samples with 1N KCl
followed by flow injection analysis and diffusion of NH4+ (Brooks er al.,
1989). Pre-fumigation NH4+ was accounted for by subtraction of NH4+
measured above as inorganic soil N.

Concentrations of NO3' and NH4+ in leachate were determined by flow
injection analysis. Samples were prepared for 15N determination by the
method of Brooks et al. (1989). Due to very low NO3' and NH4+
concentrations and inadequate volume for separate diffusions of NO3' and
NH4+, a single diffusion for both species was performed and 15N analysis of
leachate was for N03“ + NH4+.

Total nitrogen content of clippings, verdure, thatch organic matter,
thatch soil, and soil samples, and 15N enrichment of theses samples as well as

all diffusion samples was determined using a Europa Scientific Roboprep C-N

39

Biological Sample Converter and Tracermass mass spectrometer (Europa
Scientific USA, Cincinnati, OH).

The microplot data was analyzed statistically as a split-plot in time
design (Steel and Torrie, 1980), with time serving as main plots and depth as
subplots. In analysis of total N and labeled fertilizer N recovery, "depths”
included clippings, verdure, thatch, and soil (total of all soil depths). Clipping
data was totaled to include all clipping collection dates up to the time of the
respective soil sampling time. Concentration of N and 15N in the soil by depth
was analyzed in a similar manner in order to determine downward movement of
N through the soil profile. Single degree of freedom comparisons were made
contrasting 0 - 5 cm vs. 5 - 10 cm, (0 - 5) + (5 - 10) cm vs. (10 - 20) + (20 -
40) + (40 — 60) cm, and 10 - 20 cm vs. (20 - 40) + (40 - 60) cm depths.
Because of the time difference in application of 15N labeled fertilizer,
differences in distribution of the 15N, at least in the short-term, seemed
obvious. In addition, the scheduling of 15N applications and soil sampling
dates resulted in unequal time intervals between treatment application and soil
sampling for the two fertilization schedules. For these reasons, fertilizer
application timing was not included as a variable in the analysis. In this sense,
the two treatments were analyzed as separate experiments. The exception to
this is analyses of the first and last samplings following 15N applications,
which occurred 18 and 348 days following application for both treatments, in
which direct comparisons were made. Statistical analyses were conducted using

PC-SAS v. 6.04 (SAS Institute, Cary NC).

RESULTS AND DISCUSSION

The overall analyses of variance for LFN (labeled fertilizer nitrogen)
recovery by depth for each fertilization schedule in a split-plot design are
shown in Tables 4 and 5. Four ”depths” were included in this analysis:
clippings, verdure, thatch, and soil. Leachate N was not included because the
amounts were extremely low, LFN values being several orders of magnitude
less than in other components. Because of significant time x depth interactions
for both treatments, further analyses of N content at each depth over time were
conducted. Thatch and soil N were segmented into different components for
detailed analyses. The results will be presented in order of the depths
designated above: clippings, verdure, thatch, and soil. Treatment of leachate

data will follow.

$1.. 11'“ ”2].. IV! 11
Clippings:

Clipping yield closely followed growing degree day (GDD) accumulation
(10° C base, through November 30) for the first month of the growing season
in 1991 (Figure 1). For the remainder of the year however GDD accumulated
at a faster rate than did clipping yield. In 1992, which was a cooler year,
clipping yield outpaced GDD accumulation throughout the growing season for
the Spring treatment (Figure 1). Total GDD were 1617 in 1991 and 1165 in
1992, while clipping yields were 3668 and 4925 kg ha'l, respectively. The
turf exhibited greater growth under cooler temperature conditions. For the Fall
treatment in 1992 clipping yield accumulated rapidly as compared to GDD early
in the season, but then tapered off (Figure 2). In 1993 rates of accumulation of

GDD and clipping yield were similar throughout the year. For the Fall

40

41

.85 u a a “seesaw :

 

$.38 _: .38

m3. 842 S 3 Sum

2.8.: 3.: 3.83 M: 595 x 25
2.3% $58 23:. m 595
39m x 2:5

m3 m3: 2 3 8am

3.0 on a: m cassava
EN 3.: 8.3 8 ”Eu.
3M3 04.33932 gala We. 83.3w.

 

dummov “295% a Low 2.253.: wctam do» @2382 comouzz Doflztom 3.0%.:0 oo§t~> mo £3.22 .v 033.

42

.86 u m a anosammm 3..

 

:3? o: 33

23 «as? S :5 Sum

:Sé 5&2 328 2 .25 x 2::
2.3.8 8.2.“ 3.32 m 5&5
33— x 2:5

Rs 3.2: M: 3 arm

33 v3“ 8.2 m 8.3219;
:34. 8.9 M: .22 c 2::
lml gag. 3% Mn 3

 

.cwfiou 83-5% a .8 2258.: __am .8 3.2608 cowofiz “3:38”— vBoefi we 3:355 mo mam—22‘ .m 03am.

43

sfieq aexfiaq Buy/“019

.33 can 82
war—3. 8253: macaw H8 2:03 9:320 gun—:83 can 9:6 acumen 9:380 ._ 2sz

E258; Hot< $39
8m 8m 2:

 

 

 

Na QQG I ma 22> VA: 3 GOD ‘ a 20; ®

  

T

g

   

T

g

I

§

T

§

 

 

(Bu/3)!) 9191A Bmddno

é

44

.83 can ~3—
mcEE 2.258.. :8. .8 .22» 9.85:0 033388 E... 93.. 3.3.. 9.36.0 .N 0...»...

2258.8 .32 939
8» 8.. So can 8.. 8m 8m 8.

 

   
 

c _ h _ fin o

m can - Sou D
o I I.
3 Se 1 , an
a .. A
% ow 4 89. m
w 8o 1 .4. D.

.4 mu
0 an
r“ 4‘ 1 C80 U:
s 82 1 mx

8 new I- a 22> X. a mac 4. a 22> .mw

 

 

 

82 88

45

treatment 1993 was a warmer year than 1992 (1165 GDD in 1992, 1268 GDD
in 1993). In 1993 GDD was calculated only through September 30. GDD
through September 30 in 1992 was 1121. Clipping yields for the Fall treatment
were 4310 and 7542 kg ha'1 in 1992 and 1993, respectively. Clipping yields
followed similar patterns during the two years early in the season, but
accumulation rates slowed in 1992 and were maintained in 1993. Because GDD
was roughly the same in these two years it is difficult to attribute this
difference in clipping yield to temperature effects alone.

Cumulative clipping yield and cumulative nitrogen removed in microplot
clippings over the duration of the experiment are shown in Figure 3 for the
Spring treatment and Figure 4 for the Fall treatment. Clipping yield and
nitrogen content followed the same patterns within each treatment. Both
treatments showed similar trends in dry matter production and N content over
time. Flushes of growth and N content occurred for short periods during the
spring after which rates tapered off and were more gradual through most of the
growing season, with very little growth and uptake during winter months.
Total clipping yield over two years for the Spring treatment was 9220 kg ha'l,
containing 276 kg N ha'1 (mean of 3.0% N). The Fall treatment yielded
12,102 kg ha"1 in clipping weight containing 349 kg N ha‘1 (mean of 2.9% N).
Spring treatment clippings were collected from May to October in 1991 and
1992, while Fall treatment clippings were harvested during 1992 and 1993.
Because of these timing differences it is unclear whether yield differences were
due to treatment or environment.

Content of total nitrogen and labeled fertilizer nitrogen (LFN) in
clippings is broken down on an annual basis for both treatments in Figures 5
through 8. Following fertilization on April 26, 7.0 kg LFN ha'1 were removed
in clippings in the first 31 days (Figure 5). This was 18% of the LFN applied

46

(ml/33) sfimdduo U! N 19101.

cow

 

305.00.. 98.5 05 9.3000. 30380....
:8... 353:0 5 v0.83...— =0wo.._: .52 E... 20.» 9.3.50 03.22.50 .m 0...»...

2.0.5008. .08< 99.n—

con

09. SN

. L

 

Z .88. ® 20.» 9.3.50 IT

s s

(mt/Bx) may; Eutddqg

f
N
F!

 

 

econfi

47

(ml/3x) sfimdduo H! N 19101.

4:08.005 EB 05 9.3.000. $200.02..

:8... 353:0 5 00.0030: Suez: :32 can 203 330:0 03.0—08.6 .0 0.35

2.08.025 .0t< ESQ

 

 

 

 

 

O L . _ AU 0
1 80m 3
“H
8. -
m.
r 88 3
I A
com M
I 88 P
m
8? W
.. 8cm. hex
z .58. .®. 22. «5&5 .+..
09. 802

 

48

000.500.. $.95 05
m0m>m000. 8200.000 .00.. ~03 .wm =E< 095.5 82 6 >02 0.0.0 ”M03020
0_ 0030200 0090.32 300. 000 0090.32 .00=_..0m 00.0004 030030.00 .n 05me

0000500.; .204 £09

 

 

 

 

 

 

 

 

80 com com OS c

O _ _ . O
W
m
10¢ on 0 m.
m m.
U
N 2: - w H.
) NQwN? H.
Va; 9
an _ 1
m 9 IT. N
( OwfiI + ITI+I I Nm ml;
an
/
q
z 38. ® 2 .0N_=..0m IT Law

8m 2

49

and 50% of the total amount harvested in clippings for the Spring treatment.
Following this period of rapid uptake (average rate of 0.23 kg LFN ha'lday'l),
rates tapered off but remained steady throughout the summer and early fall. The
shape of the curve of LFN content between 31 and 158 DAT (days after
treatment) appears to be linear, but the curve was analyzed in two segments due
to the increased rate of total N uptake apparent following 116 DAT. This
change was associated with a fertilizer application made on August 19 (115
DAT). Mean daily uptake of LFN from 31 through 116 DAT was 0.036 kg ha'
1, and from 116 through 154 DAT was 0.043 kg ha‘l. Total N uptake rates for
these periods were 0.54 and 1.26 kg N ha'lday'l. Some of this large increase
in non-labeled N measured in the clippings was presumably fertilizer N from
the August application, but uptake of additional LFN was also influenced,
indicating a priming effect on mineralization of 15N incorporated into soil
organic matter. Applications made on June 4, July 12, and September 27 were
not followed by clear plant responses in N content. In fact, very little
additional N was harvested in clippings following the September fertilization.
In the first 31 days following fertilization, LFN accounted for 25% of the total
N recovered in the clippings. By November 8 (day 196) LFN harvested in
clippings totaled 12.1 kg ha'l, 31% of the total amount applied, and 9% of the
total N in clippings.

During the second growing season (1992), three different phases of
nitrogen accumulation, characterized by decreasing mean daily uptake rates,
were observed for Spring fertilized plots (Figure 6). Uptake was rapid from
367 DAT (April 28) through 384 DAT (May 15), with total N harvested in
clippings averaging 1.8 kg ha"1 day'l, and LFN averaging 0.035 kg ha“1 day‘
1. Following this period, uptake rates decreased, yielding two relatively steady

periods of uptake. Once again, increases in uptake following fertilization were

50

..00E.00.. w0..qm 0...
w0.>.000. 0.0.00.2... 0.0... mg. .2 >02 09.0.... «ma. .wu _..n.< 0.0.. 30.9.20
0. 00.0020: 0095.2 .08... 000 009....2 .00.....0n. 00.0004 03.0.0500 .c 0.09".

808.00.... .0...< 0.0Q
000 80 000 com 000 000
8~ _ _ .

 

 

I. - m.
m. 8N4

N

)

0d.

W 80-

Aux _ Se

in V
(Bu/3x) N 1921111195 P919091

30:.
z .50... ® 2 00.5.0... T

 

 

 

§

51

not generally observed, with the exception of the April application. Very little
additional uptake occurred through the fall and winter. The following spring, a
pulse of increased 15N and total N was observed as in previous springs, but
microplot clipping sampling for this treatment ended at this time due to
excavation of the last set of microplots. Total uptake of LFN over the two year
period was 13.9 kg ha"1 (35% of that applied), while a total of 276 kg total N
ha'1 was recovered.

Following application of LFN urea in November 1991, 1.8 kg ha“1 of
LFN was recovered in clippings on April 28, 1992 (171 DAT), the first
mowing date (Figure 7). A period of rapid uptake ensued through 188 DAT
(May 15), with mean uptake rates of 0.36 kg LFN ha'l day"1 and 1.92 kg total
N ha'lday‘l. Total uptake of LFN was 8.0 kg ha'1 by this time, 20% of the
applied amount and 53% of the total amount eventually harvested in clippings.
Nitrogen accumulation in clippings continued at decreasing rates throughout the
growing season. By 348 DAT (October 22) 12.4 kg labeled fertilizer N ha'1
had been harvested in clippings (32% of applied amount).

The next spring, rapid N accumulation occurred again in the clippings
until 555 DAT (May 19) (Figure 8). Fertilizer was applied on May 26 (562
DAT), and contrary to the more gradual uptake observed in the Spring
treatment following fertilization in May of 1992, uptake continued at a
relatively high rate throughout the growing season. Clipping harvests indicated
mean daily rates of 0.014 kg LFN ha'1 and 1.38 kg total N ha'1 for the Fall
treatment. Rates during a similar time period for the Spring treatment the
preceding year were 0.008 kg LFN ha‘l day'1 and 0.66 kg total N ha’1 day'l.
A direct comparison of total N uptake from lysimeters for both treatments from
June 9 to October 7, 1993 showed greater rates of uptake for the Fall treatment

than the Spring treatment (0.86 and 0.34 kg N ha'1 day‘l), indicating a

52

808.00.. :0... 0... w0.3000.
0.0.00.0.8 80.. .000. .2 .000.00 09.0.... .00. .w .0080>0Z 80... 000.00.?
0. 0.0.0020: 00w0...z .0.0... 000 00w0...z .03....0... 00.000. 03.0.0800 .0 0.09...

808.00.... .0..< 9.0m.
00.. 000 000 00.

 

0
V5
I

I0

§

 

(ml/Bx) N 1210.1.

N co
(ml/351) N 1921111193 991mm

9.
T
9I

Z .50... ® 2 .0N.....0n. IT

 

 

 

SN

0
'—¢

53

808.00.. :0... 00. 30.3000.
0.0.00.0.8 80.. .00. .3 .000.00 0300.0. N00. .00 .00200 80... 30.0000
0. 00.00200 030.32 .0.0... 000 030.37. .03....0". 02000.. 03.0.0800 .3 0.09".

808.00.... .0.u.< 30A.
000 000 000 00. 00.. 000

 

 

8.30. . 0.

§

(Bu/Bx) N ram.
8

In 0-
(mI/fix) N 1921111195 P919091

7. .50... ® 7. .03....0... IT

 

 

 

§

54

treatment effect. There may have been an environment effect as well, due to
different growing seasons.

Total nitrogen harvested in clippings during two years was much greater
in the present study (276 and 349 kg ha'1 for the Spring and Fall treatments,
respectively) than that observed by Starr and DeRoo (1981) (95 kg ha'l) over a
three year period. Expressed as percentage of fertilizer applied, these amounts
are 70% and 89% in the present study and 50% from Starr and DeRoo (1981).
Starr and DeRoo (1981) observed approximately 30% and 20% uptake of LFN
in the first 120 days following application of nitrogen in May and September,
respectively. In the present study, these values were 26% and 34% for April
and November applications. The measurement for this November application
occurred during the following spring. Calculated mean daily uptake rates of
fertilizer N as measured in clippings during a 'steady state‘ period following
initial surges in uptake were very similar in this study and in the work of Starr
and DeRoo (1981), 0.04 - 0.05 kg N ha‘1 day'l. Uptake rates of total
nitrogen, however, were at least twice as high in this experiment, resulting in
much greater total N uptake over the entire course of the study.

Measurement of uptake varies widely in the literature. Power and Legg
(1984) observed uptake of 30% of applied LFN in one growing season by
crested wheatgrass, which was similar to the findings of the present study.
However, Bowman et al. (1989b) observed much higher uptake rates by
nitrogen deficient Kentucky bluegrass in the field (12.5 to 19.25 kg N ha‘1
day”l vs. maximum rates of 1.92 kg total N ha'1 day'1 in the present study).
Bowman et al. (1989a) also observed uptake as high as 32 kg N ha’1 day‘1 by
perennial ryegrass in solution culture. Other researchers have shown that
perennial ryegrass pastures exhibited approximately 50% greater uptake of LFN

in herbage as compared to the turf in this study (Bristow et al., 1987; Dowdell

55

and Webster, 1980, 1984). This may be at least partially explained, however,
by greater uptake rates by perennial ryegrass as compared to Kentucky

bluegrass as measured by Bowman et al. (1989b) in solution culture.

Verdure:

The total amount of LFN harvested in the verdure decreased at each
sampling date for both the Spring (Figure 9) and Fall (Figure 10) treatments, as
would be expected as N was removed in clippings with each successive
mowing. The decreases in verdure LFN were associated with increases in
clipping LFN, indicating transport upward within the shoots to the leaf blades,
predominantly during the first growing season following application. However,
the amount of total shoot LFN (plotted as ”clippings + verdure") did not
change significantly throughout the study for the Spring treatment, and
differences that occurred for the Fall treatment showed no trend towards a
general increase or decrease. This indicates that after the first microplot
sampling (l8 DAT for both treatments) there was no significant transport of
additional LFN to above ground plant tissue.

Bristow et al. (1987) observed very similar results in a perennial
ryegrass pasture. Between 16 and 310 DAT of LFN as NH4NO3, total recovery
in herbage (cut above 5 cm) and stubble (soil surface to 5 cm) varied little

(49.9% to 55.8% of the applied LFN).

11° . I] I

Thatch was found to be a very important sink of fertilizer nitrogen and
also a large pool of total nitrogen. The amount of total N in the thatch was not
significantly changed over time with the regular addition of . fertilizer under

either fertilization schedule (Tables 6 and 7). There were fluctuations in the

56

808.00.. 38.0w 00.
30.3000. 0.0.00.0.8 80.. 0.00 0.080.. 0000 .0 30.0000 03.0.0800 00.0 0.00.0>

000 .30.00..0 03.0.0800 .0.00.0> 0. 00.0>000. 030.37. .03....00. 00.003 .0 0.09".

808.00.... .0..< 300.

com

L

coo

.

80

.

com

0

o

 

 

m...
H 030...

80> + .000 IT

000.800 4.

2:23, O

 

.0BN: .

<0

I I
O V)
.—4

(Bu/3x) N 1921111195 P919091

T
W
v-i

 

 

O
N

57

 

808.00.. :0”. 00.
30.3000. 0.0.00.0.8 80.. 0.00 0.080. 0000 .0 30.0000 03.0.0800 00.0 0.00.0>
000 .30.00..0 03.0.0800 .0.00.0> 0. 00.0>000. 030.32 .03....00 00.000... .0. 0.09...

808.00. .. 0...... 300.

com 80 80 com o
_ _ L . j C

md
II. 0.30".

  

T
W

800: ..
In.

C
(mt/3x) N 1921111195 99mm

 

 

 

 

I
V)
v—I

N.N
H 0.30".
.0.0> + .000 1T 30.0000 4 0.00.0 > .

 

 

O
N

58

amount of total N in the organic portions of thatch, but these fluctuations
showed no regular trend over time. A large proportion of the total N in thatch
was present in the soil component. Averaged for all samples, thatch was 89%
soil by dry weight.

Fertilizer N in the thatch layer was a much more labile pool as compared
to total N. Following April application, 31% of the LFN was found in thatch
18 DAT (Table 6). This amount remained fairly constant for the next year,
with the exception of the sample taken on October 14. The 32% of LFN found
in the thatch in November 1991 was considerably higher than the 19% reported
by Starr and DeRoo (1981). Between May and November of 1992 LFN in
thatch decreased significantly, presumably due to mineralization and subsequent
transport deeper into the soil, or loss through volatility. Plant uptake was
shown previously not to be a factor later than 18 DAT. Examination of the
data over the two year period shows an apparent trend in decrease of LFN in
the organic component and an associated increase in the soil component (with
the exception of the final sample with respect to soil LFN). Thatch is extremely
heterogeneous, and this N could be present in living plant material (crowns,
roots, rhizomes), incorporated into plant material that senesced since the time
of uptake, adsorbed to previously senesced plant material, incorporated into
microbial biomass in either the organic or soil components of the sample, or
could be present in the soil component as organic, exchangeable, fixed, or
soluble inorganic N. Because the technique used for thatch separation
segregated components based on particle density and size, the shift in LFN
from organic to soil components indicates breakdown of the organic material
into smaller components which migrated downward into the soil portion of the
thatch layer or were small enough to be included in the soil portion in the

screening process. This breakdown could have included complete

59

 

 

o... 4.0 00
0.. 3. 0.0
3 3 E
02 v... 3
m0. 0.. 0.0
I. 3. 2.
0.0. 0.. Z
.0. Eu 0..
_. 2 0.
4% I=|ow| gal

030.07. 0000.0... 00.0004

 

m7. m7. m0
.2. 3H 00.
.00 80 0..
0.0 0.0 .om
00.. 30 o...
2.0 m0.” .0
90 03 N:
0.... Sn 0..
483% .I00Iml '03qu
0030.07. .08...

.38 00.8.
80:.
83 S .
800.0
800: _
5:5.
800.0
30.0.

I|I||0.0Q

 

808.00.. w0..0m .0. 0.00 3000800
0000 .0. 0.0000088 :00 000 0.0090 00.00.00. 00.00. .0 80.000 030.02 0000.0... 00.0004 000 030.07. .30... .0 0.00...

60

mineralization of thatch organic material or a more partial breakdown of large
pieces of organic material into smaller ones, and was probably a combination of
the two.

Recovery of 62% of LFN from the thatch layer occurred 18 DAT
following application in November (Table 7). By the following May this had
not changed significantly. By June 29 there was a significant decrease in both
total thatch LFN and LFN in the organic component compared to the May
sample, and soil LFN increased compared to the November 1991 sample.
These trends were maintained throughout the study. Breakdown of LFN in the
organic component and transfer to the soil component occurred in a similar
manner as for the Spring treatment. This breakdown cycle occurred in 1992 for
both treatments, even though April LFN was applied in 1991. It is unclear why
thatch LFN remained stable throughout the summer following application for
this Spring treatment.

The total amount of LFN in thatch shortly after application of the Fall
treatment was striking. This difference in early recovery from the thatch
accounted for most of the difference in total recovery between the two
treatments. As thatch LFN decreased in the Fall treatment over time, so did
total recovery. Most of the LFN in thatch from the Fall treatment in the
November 1991 sampling was in the organic component. Large amounts of
LFN were probably not in live plant material in the thatch layer because
recoveries in clippings and verdure were similar for the two treatments during
the first spring following application. If November-fertilized plants were
storing this additional LFN, it would be expected to result in increased LFN
content in clippings of these plants as compared to those fertilized in the
spring. Root tissue could account for some of the additional thatch LFN in the

Fall treatment. Some other organic pool was also probably acting as a short-

 

61

Z 00 on 02 m7. 8. .80. 00....

 

 

0.0 0.0 0.. .00 000 mm. 0000...
0.0 0.0 0.m 00m SN mum «0.02m
0.0 0.0 «N man 000 00 N030...
0.0 0.0 0N 00.0 .00 mm «00.0
0.0. 0.0 0.0 00m 000 00 N0\0N\0
0. .N 0.0 0.0. 0N0 .00 0 . N ~0\0N\m
m.0~ 0... 0.0m 000 m0v mm. .0BN...
_- 00 0..
I03... 40.3 04.33 .03.: I_._olmI .0030. .003.
030.07. .000...0... 00.0004 030.07. .0.0...

 

808.00.. :0... .0. 0.00 0000800.
0000 .0. 080000800 000 000 0.0090 00.00.00. 00.00. .0 80.000 0000.02 000.800 00.0004 000 0000.07. .30... .0 0.00...

62

term storage reservoir over the winter for LFN. Significant losses of thatch
LFN occurred with the onset of warmer temperatures. Root turnover would
probably not be rapid enough to account for all of the loss. The differences in
early recovery from the thatch between Spring and Fall treatments, significant
rapid loss of LFN from the thatch for the Fall treatment, and total recovery
data (presented later) indicate the possibility of gaseous losses of N from the

thatch layer.

5 'l 11'

Soil N and 15N were partitioned into inorganic, biomass, organic, and
total N pools in each of the sampling depths of the soil profile. Total,
inorganic, and biomass N were directly measured and organic N was calculated
by subtracting inorganic from total N. Biomass carbon was also measured.
Samples from the top five depths (to 60 cm total depth) only were used in
statistical analysis. These samples were contained within the microplots at the
time of excavation. The sixth sample (60 - 100 cm) was collected from below
the microplots with a soil probe, and it was found upon analysis that this
method resulted in contamination of the sample. This data, therefore, was not
used. Complications were encountered in measuring biomass N. In the
biomass determination, NH4+ from a non-fumigated, non-incubated control
sample is subtracted from the amount of NH4+ detected in a chloroform-
fumigated, incubated sample, resulting in NH4+ released due to fumigation.
In the soil from this experiment NH4+ levels in non-fumigated samples
remained relatively high at depths below 10 cm. The combination of
immobilization of liberated NH4+ in fumigated samples and high control
NH4+ resulted in negative numbers when this calculation was performed for a

large number of the samples below the 10 cm depth. The resulted was negative

63

values for biomass N. Data was examined for approximately 50 samples from
the O - 5 and 5 - 10 cm increments where this problem did not occur and a
biomass C:N ratio of 5 .7 :l was calculated. This ratio was used for all samples
to estimate biomass N from biomass C. For soil sample 8, removed in
November 1992, a large number of base traps used to trap C02 in the
incubation process became unusable due to evaporation. Reliable numbers
could not be obtained, so there is no data for biomass C or N from this sample.
Tables 8 and 9 present summaries of the statistical analyses of biomass C and
soil N and 15N pools for the Spring and Fall treatments, respectively. Time by
Depth interactions were seen for all parameters except biomass C and N in the
Spring treatment, where depth only was a significant factor. Pursuant to these
interactions biomass C, N, and 15N and the relationships between inorganic and

organic N and 15N were examined in greater detail.

Biomass C and N:

In Tables 10 and 11 concentrations of biomass C, N, and 15N at each
depth for each sampling date are given. In this analysis means for each
parameter at each depth were compared over time to distinguish long-term
trends in biomass activity. Values for C and N follow identical trends due to
the constant ratio used to estimate biomass N. For this reason, explicit
discussion of biomass N will be avoided, but can be implied from biomass C
trends. For the April fertilized treatment there were no significant differences
for any parameter at any depth (Table 10). There appeared to be a trend
towards increasing biomass C over time, especially during the first year. This
held true at all depths throughout the profile. Almost all of the 15N in biomass
was confined to the upper 10 cm of the profile, although there were trace

amounts in the 10 to 40 cm depth. No 15N was detected in biomass below 40

64

Table 8. Summary of split-plot analysis of variance for concentration of
biomass carbon and total N and 1’N (ug g“) in various soil pools for

 

 

 

 

Spring treatment.

Source of Variation
Sgil Pogl _Lme_ Replication Depth Time‘Depth
Biomass C * *
Biomass N * *
Inorganic N * * “ * * *
Organic N It a: at t t It 1!
TOtal N t i t t I! t *
Biomass lsN * * * * *
Inorganic lsN * * * * * *
Organic lsN * .. a: g. ...
Total ”N * t t 0 g:

 

‘2" Significant at P = 0.05 and 0.01, respectively.

65

Table 9. Summary of split-plot analysis of variance for concentration of
biomass carbon and total N and 1’N (pg g") in various soil pools for
Fall treatment.

 

 

Source of Variation

 

 

Spil Pppl Time Replication Depth Time‘Depth
Biomass C * "' * * "' *
Biomass N ** ** **
Inorganic N " * "' * "' *
Organic N * * "' * * *
TOtal N t * i t t * *
Biomass lsN * * * * * *
Inorganic l’N * * * * *
Organic lsN * * * "' * *
Total lsN * * * "' * *

 

‘3" Significant at P = 0.05 and 0.01, respectively.

Table 10. Concentrations of biomass C, N, and 15N (pg g") by depth for each

66

sampling date for Spring treatment.

 

Depth
L00
0-5 C
N
ISN
5-10 C
N
UN
10-20 C
N
ISN
20-40 C
N
lSN
40-60 C
N
[SN

1991

 

(It
\
p...
A

424
74
0.64

370
65
0.08

245
43
0.03

90
16
0.01

87
15

6/1

367
64
0.40

241
42
0.08

217
38
0.01

213

37

207
36

 

10/14 11/26
737 609
127 109
0.68 0.53
275 374
47 67
0.10 0.14
135 327
23 58
0.03 0.04
150 193
26 34
0.01 0.02
O 210
O 38
0 0

312
56

460
29
0.03

211
38
0.02

63
11

1993

M
\
#
A

717
126
0.38

609
107
0.14

220
39
0.02

96

17

105
18

FPLSD
(0.05)

NS
NS
NS

NS
NS
NS

NS
NS
NS

NS
NS
NS

NS
NS
NS

 

NS = not significant

67

cm at any time. In the 0 to 5 cm and 5 to 10 depths, the data shows that the
biomass maintained an active role in the cycling of the applied 15N over the
entire two years of the study. Between May and November of 1991 the
percentage of 15N in biomass as compared to total biomass N in the upper 0 to
5 cm decreased from 0.86 to 0.48%, while it increased from 0.12 to 0.21% in
the 5 to 10 cm depth. This indicates downward movement of the 15N over this
time. This movement could occur either as inorganic N or in senesced roots
followed by incorporation into microbial biomass. After November of 1991
these percentages decreased at both depths.

For fall fertilized turf, microbial biomass populations appeared to
decrease over the first year following application (Table 11). This was not
significant in the 0 to 5 cm depth, but was significant at 5 to 10 cm. This is in
contrast to the trend observed for the Spring treatment. In May of 1993
biomass levels increased and were significantly higher than in the previous June
and September for the two upper depths. Measured biomass C was very low
for the final sampling date throughout the profile, although this was significant
only at the two upper depths. Because these values were so much lower than
any others measured, including those from a similar time of year in 1991, it is
suspected that the 1993 values are not a real effect but are due to some
procedural error. However, a systemic error should have affected all samples
equally, but values for lower depth samples are not entirely inconsistent with
data from earlier dates. The reason for these inconsistencies is not clear.

Biomass 15N decreased significantly in the top soil depth between
November 1991 (shortly after application) and the following May. If these
values are converted to kg LFN ha’l, this equates to a decrease from 2.15 to
1.07 kg LFN ha'l, a release of approximately 1 kg LFN. Differences in

biomass 15N during 1992 were not significant, but a significant decrease

68

Table 11. Concentrations of biomass C, N, and 1’N (ug g") by depth for each
sampling date for Fall treatment.

 

Depth 1991 ........... 1992 ................ 1993 ......
Lem) FPLSD
11/26 5/2 6/29 /17 5/1 11/30 (0.05)
0-5 (2 703 614 426 486 929 60 364
N 126 110 75 as 163 10 60

15N 0.90 0.45 0.28 0.36 0.50 0.03 0.33

5-10 C 468 342 153 171 302 58 114
N 83 61 27 30 53 10 19
15N 0.15 0.11 0.05 0.04 0.10 0.02 NS

10-20 C 273 212 100 207 258 76 NS
N 49 38 18 36 45 13 NS
15N 0 0.03 0.01 0.03 .03 0.01 NS

20-40 C 250 107 179 72 322 81 NS
N 45 19 31 13 56 14 NS
lsN 0 0 0 0 0.01 0 NS

40-60 C 148 145 112 117 71 98 NS
N 26 26 32 21 12 17 NS
15N 0 0 0 0 0 0 NS

 

NS = not significant

69

occurred between May and November of 1993. This decrease was tied to the
low biomass C measurements in the November sample. Similarly to the Spring
treatment, percent of 15N in biomass as compared to total biomass N decreased
during the first year after application in the 0 to 5 cm depth, from 0.71% in
November 1991 to 0.42% in September of 1992, and continued to decrease
through 1993. In contrast to the Spring treatment, an increasing trend did not
occur in the second depth (0.18%, 0.18%, 0.13%, and 0.19% in 11/91, 6/92,
9/92, and 5/93, respectively).

Calculations of the percentage of 15N as compared to total N in biomass
raise an interesting point. Using the O to 5 cm depth from the Spring treatment
as an example, 0.86% of biomass N was 15N 18 days after 15N application.
Thirty eight days later, and 17 days after an application of an equivalent
amount of non-labeled N, the value decreased to 0.63%. In October, following
two additional non-labeled N applications, the value was 0.54%. If the
biomass was utilizing the applied N similarly from all applications, dilutions in
the percentage of 15N in the biomass would be expected with each subsequent
application. This was not seen however. It appears that either the microbial
biomass is not as active in the cycling of fertilizer N applied during the summer
as compared to that applied in the spring, or there is preferential cycling of
15N over 14N. This interpretation should be viewed with caution, however.
Biomass 15N values are estimates, calculated based on constant biomass C:N
ratios as described. In addition, total amounts of 15N were extremely small.
Considering the importance of the microbial biomass in nitrogen cycling,

further investigation of this effect is warranted.

70

Relationships between inorganic and organic N and N mobility:

Inorganic N made up approximately 2% of the total soil N in the profile
in the spring treated plots in May 1991 (Table 12). Of the 98% of the N in
organic forms, approximately 7% was present as microbial biomass. From
LFN, 14% and 17 % of the 15N found in the upper two depths was in inorganic
forms. Below 10 cm all 15N detected was in organic forms. Of the organic
15N, 75% or more was present as microbial biomass. One year later,
distribution of total N was similar (Table 13). Of the LFN applied one year
previously, between 0 and 11% of that recovered in various soil depths was in
inorganic forms. Of the organic 15N measured (89 to 100% of total), 24 to
38% was in microbial biomass. Two years after initiation of the study, 4% of
the 15N applied was in inorganic forms, and 17 to 24% of organic 15N was in
microbial biomass (Table 14).

For the Fall treatment, trends in total N were similar to the Spring
treatment. Inorganic N averaged 2% of total N in the first fall (Table 15) and
1% (Tables 16 and 17) thereafter. Biomass N was 9% and 8% of organic N in
1991 and 1992, respectively (Tables 15 and 16). Because of problems with
biomass determination in the November 1993 sample as discussed previously,
data for this pool for 1993 is not reliable.

Eighteen DAT of LFN in November, an average of 11% of the 15N was
present in inorganic forms. Of the organic 15N present in the top two depths,
80 to 100% was in microbial biomass (Table 15). One year following the late
fall application, 0 to 7% of the 15N was in inorganic forms, and 14 to 43% of
organic 15N was microbial biomass (Table 16). Two years after application,
98 to 100% of 15N present in soil was in organic forms (Table 17).

For both fertilization timings, the importance of microbial biomass in

immobilizing LFN that reaches the soil is clear. The biomass comprised 6 to

71

{i

ii

no.0
vod
N_.o
mad

 

.569

&*

**

 

o o
56 Ed
vod mod
2 .o mod
mud mod

g 4.36
Avw wnv Z:

**

*§

§*

**

o
o
0
mod
2.0
2134.4.

 

mm—
5N
was.
omw
owN_

 

Bop

50.58%“: .86 23 mod n m 8 835.3% 3......

 

62 2
m2 2
E 2.
m4» m6
:2 i

g g

a we 2 38

ooéN
.m> cauc—

ooé—
.m> o — -o

on
.m> mlo

“33:80
v 093..
9 Stem
m _ 0N6—
: on
wN muo

643mg AEJIQ
£65

 

 

.86. .3 a: :6 Baa:
.8253: macaw of com 53c .3 82355280 2: new 2 .83 can .oEamS £22.85 diam—05 .N_ 2nd.

72

§§

**

N00
50.0
00.0
3.0
E..—

 

3
O
E-

ii

i}

 

**

**

*§

**

~00
0
_0.0
00.0
m0.0
€643

 

_0.0 0
50.0 N00
00.0 8.0
000 2.0
E..— 34.0
salsa .44..de
”70 05 Z:

00—
034
005.
20—
0mm—

 

38.

22686.2 .85 23 mod n m a 885:5 4......

0:
094
>00
«52
0 _ 2

6464a

 

9.0 03 Z .88.

00-0N
.m> 0N0—

00-2
.m> 0 _ -0

0 0 -m
.m> m0
385.80
_ _ m 00-0w
N 9.00
on 0 0N0—
0m N.

mm

07m
00 Z 0-0

136805 1.01106“ .5:— ~83

59.5

 

 

.82 .8 32 .8 8383

.2853: 05.00 05 .8 5000 .3 8333:0860 Z: 05 Z 33 05. .2508 £3803 6:80.55 .2 035.

73

50300008.. .30 0:0 m0.0 H m 8 E85005 2......

 

 

 

 

 

 

 

8-8
3.. .1. a... a... .m> 0N0—
8-2
3.. a... a... a... .m> 070
2a
3.. a... .1. .1. .m> m0
“wants—GU
o o o o 5 62 2 N 8.3
o o o o SN EN 2 m 3-8
.3 2 .o 8.0 5o 36 ca. 3 a 8-2
«2 40¢ 2 .o :3 86 MS... 8. a 2A..
m2 «3 ”no 86 22 £2 62 a We
.38. 64.3.3.0 4% 6.24.33 Bop madam 4.4% 86994 4%
p.» we 2.. :0 we 2 .33 5 a

 

.82 .2 .32 8 Bass.

82:89: 0580 2: 50 5000 .3 828.8083 2: .8“ Z .83 8a .8508 .3853 .2508:— .v_ 035.

74

§*

*§

**

m0.0
00.0
00.0
0.0
mm.

 

.50...

*§

§§

*§

 

**

**

ifi

**

**

.7!

0
0
.00
V00
N..0
3.1.4.34.

 

3.0 0
00.0 0
00.0 .00
m..0 m..0
.... 0.0
354.6 .340
A70 03 Z:

00.
.00
000.
0.0.
N00.

 

.50...

0.3.82.3. ..0.0 .80 no.0 u m 8 830.005 8......

.40.

.0N
«No.
8..
.00.

634.6

 

9-0 010 Z .50...

ON
0.4
0.4
.40
ON.

1.28.0.0

0
0.
h.
b.
.m

0.50.100...

 

00-0w

.m> 00-0.

00-0.
.m> 0.-0
07m
.m> m0

:3.800

000.4

0v-0N

0N-0.
0 . -m
0.0

190.1
5.60

 

.00. .3 308262 :0 00.083

.8053: :0... o... .0.. 53.. .3 80.8.8288 Z: 03 Z .50. .80 6.30.0 6380... 6.50.0... .m. 0.00...

75

0.2.08.8. .80 .05 8.0 u m 0. 0.80.005 t...

 

**

ii

***

00.0
00.0

 

 

*i

**

0
0
0
00.0
50.0
00.0%

 

ii

** {.0

** ii

0 0

0 0
00.0 00.0
00.0 .00
00.. 00.0

0430.10 $80.00.
9.0 03 Z:

 

 

 

00-00
.m> 00-0.
00-0.
.m> 0.0
070
.m> 0-0
”0.0.800
00. 00. .N . 00-0..
000 000 0. 0 3.00
0:. m... 00 0 00-0.
30 000 00 0 0.0
000. 000. N0 0. 0-0
.30... 04.00 .0 1080.0. 0.80% .80.
530
9.0 03 Z .80...

 

.000. .0. 800.8000 00 3.908

82080.. :0". o... .0.. 080.. .3 80.8.8300... Z: .80 Z .80. .80 6.0.0.0 .8200... 6.50.00. .0. 0.00...

76

0.03.0030. ..0.0 000 00.0 n a. .0 8000.005 2......

 

§§

if

00.0
0.0
00.0
00.0

 

if

**

 

*§

**

i!

o
o
o
S...
8.0
0.30.34.

 

0 0
00.0 0
0.0 .0.0
00.0 00.0
00.0 00.0

Olga 0030.0.
9.0 03 7...

.0.

00.

0.0
00..
0000

 

.03...

 

 

00-00
.0> 000.
00-0.
.m> 0.0
0.0
.m> 00
000.800
00. n. v 00.0.~
v0. 0. 0 0.100
. .0 0. 0 00-0.
00.. 0. 0 070
000. 0. 0. 0-0
00.00wa 0008. 0. ..m.... wawflodl. .80.
.0000
9.0 01.2.0.0...

 

.000. .00 .0080>0Z 00 00.080...

808.00.. :0... 0... .0.. 0800 0.. 000.008.0008 7.: 000 Z .30. 000 6800.0 .8080... 6800.00. .0. 0.00...

77

9% of the immobilized total N, but shortly following fertilizer application the
biomass immobilized at least 75% of the LFN that reached the soil. Although a
relatively high percentage of inorganic LFN was present as compared to total N
in inorganic forms, most of the LFN had been protected from leaching through
rapid incorporation into organic forms. Total organic N was present in more
stable forms, but LFN was also converted to less labile forms over the two
years of the experiment. Because fertilizer N was present in inorganic forms to
a larger extent than total N, fertilization could lead to increased leaching
potential. However, the total amount of N introduced to the soil from a single
fertilizer application in this study was only about 0.07% of total N. The impact
of a single fertilizer application on environmental quality would seem to be
insignificant. The use of 15N, however allows for accurate tracing of this
small addition through the soil profile.

Tables 12 through 17 also show single degree of freedom orthogonal
contrasts for comparison of mobility potentials of 15N pools over time. As an
example, if a difference exists between two adjacent soil depths at Time 1 but
no longer exists at Time 2, the data would probably indicate downward
movement from the upper to the lower layer over this time period. If a lack of
difference later becomes a significant difference, it could indicate that N has
moved out of the lower layer of the pair of interest into the depth below.
Analysis of 15N concentrations for the Spring treatment indicates potential
vertical movement of LFN. In 1991, there was no inorganic 15N below 10 cm
(Table 12). In May of 1992, inorganic 15N was present in the 10 to 20 cm
depth (Table 13). The statistics indicate a significant increase in this depth as
compared to 20 to 60 cm. In May of 1993 this relationship was unchanged, and
it is encouraging that there was no 15N below 20 cm (Table 14). However,

because inorganic N is very labile in soils this may not be the best indicator of

78

potential leaching. Examination of organic 15N data may provide additional
information on the potential for long-term vertical mobility. Between May of
1991 (Table 12) and May of 1992 (Table 13) total 15N and organic 15N levels
increased throughout the profile, presumably due to movement of LFN
downward from the thatch layer. Biomass levels remained consistent at these
depths. The increase in 15N at lower depths may have been due to root
production. Statistically no change in the relationships between depths
occurred because increases occurred at all depths. In May of 1993 total and
organic 15N were no longer detected below 20 cm. There was therefore a
significant difference between the 10 to 20 and 20 to 60 cm depths for biomass,
organic, and total 15N. Since the previous year, 15N had either been
transported upward in plants or mineralized and moved downward out of the
sampling area.

For the Fall treatment, inorganic 15N was found in the 10 to 20 cm
depth in November 1991 (Table 15) but not below 10 cm after that time. In
1991, organic and total 15N were greater in the 20 to 60 cm depth than the 10
to 20 cm depth. In September 1992 15N was not detected below 20 cm (Table
16). As for the Spring treatment this 15N was either transported upward in
plants or mineralized and moved through the soil. In November 1993 organic
15N was again detected in the 20 to 40 cm increment. This caused non-
significance in the contrast of 10 to 20 cm vs. 20 to 60 cm, indicating
significant downward movement in the organic form since September 1992.

Although vertical mobility of LFN is indicated by this data, involvement
of organic material indicates cycling of fertilizer N through plant material
before potentially being lost to deep soil depths. This is in sharp contrast to the
perceived notion of fertilizer rapidly flushing through a system, presumably in

soluble forms. Consider data from May 1992 for the Spring treatment (Table

79

13) and November 1993 for the Fall (Table 17), which represent the greatest
concentrations of LFN below the 10 cm depth for each treatment. If all of the
LFN were mineralized and leached, this would total 2.10 and 1.68 kg LFN ha’1
for the Spring and Fall treatments, respectively. These values represent 5%
and 4% of the amounts applied. This is a very liberal estimate considering the
nature of the nitrogen cycle and the fact that significant rooting exists to depths
of 20 cm or more. The extremely small amounts of 15N involved illustrate the
power of 15N in nitrogen cycling research.

Table 18 presents data on LFN in inorganic, biomass, and organic pools
as well as total soil LFN on a mass balance basis for each sample date. The
total amount of LFN in the soil shortly after application was only 8% and 12%
of the total applied for April and November applications, respectively. Very
little was present in mobile inorganic forms. Microbial biomass played an
important role in immobilization. Over time the amount of LFN in the soil
increased for both treatments, but remained predominantly in organic forms.
The proportion of organic 15N present as microbial biomass decreased with
time, indicating incorporation into more stable forms. This immobilization
pattern is positive from an environmental standpoint, in that organic forms of
N, even in the relatively labile microbial biomass pool, are relatively immobile.
From a fertilizer utilization point of view, it appears that fertilizer N that enters
the soil is readily immobilized and is therefore unavailable for immediate
uptake. Much of the non-biomass organic 15N however is probably plant root
material, and represents utilized LFN. As roots senesce they enter the soil
organic matter cycle, which then constitutes a significant source of the applied
fertilizer N. Investigation of the dynamics of this soil organic nitrogen pool
and its mineralization would provide greater understanding of fertilizer nitrogen

utilization in turf.

80

Table 18. Recovery of Labeled Fertilizer Nitrogen in various soil pools for each

 

 

 

sampling date.
Imam Date W1 r ni Blooms 91211122 Total
kg LFN ha'l
Spring 5/14/91 0.44 2.22 2.73 3.16
6/21/91 0.12 1.48 4.19 4.31
10/14/91 0.22 2.16 5.94 6.16
11/26/91 0.23 2.00 6.51 6.74
5/26/92 0.31 1.92 7.64 7.95
11/30/92 0.25 ----- 6.31 6.56
5/14/93 0.09 0.99 5.26 5.34
Fall 11/26/91 0.47 2.65 4.31 4.77
5/26/92 0.28 1.78 3.48 3.76
6/29/92 0.12 0.99 2.64 2.76
9/17/92 0.28 1.28 6.02 6.31
11/30/92 0.33 ----- 5.76 6.10
5/14/93 0.16 2.02 8.63 8.79
11/30/93 0.21 0.22 9.76 9.96

 

81

1.2211111211332932

The cumulative water balance for the entire time of the study is shown in
Figure 11. Between April 26, 1991 and December 23, 1993 a total of 277 cm
of precipitation and irrigation fell on the lysimeters. Precipitation accounted
for 221 cm of this, and irrigation totaled 56 cm. A total of 137 cm of water
was collected as leachate from the Spring treatment, and 108 cm was collected
from the Fall treatment. Collection of leachate from the Fall treatment did not
begin until November 8, 1991. Subtracting out the period from April 26 to
November 8 for the Spring treatment to equalize collection intervals, the total
amount of leachate was 125 cm. Total leachate amounts were not significantly
different between the two treatments regardless of which Spring value was
used. Table 19 shows flow weighted means of N as NO3' and NH4+,
collected in the leachate. There were no significant differences between
treatments for any of these measurements. Brown et al. (1982) also reported
NH4+ present in leachate from golf greens. Soil mixtures with greater
proportions of "soil" as compared to sand produced higher proportions of
NH4+. Flow-weighted mean NO3'-N concentrations were well below the EPA
threshold limit of 10 mg L‘l. These findings agree with the conclusions of
Starr and DeRoo (1981) and Morton et al. (1988) who stated that the potential
for groundwater contamination resulting from fertilization of lawn turf is
negligible.

All four lysimeters showed elevated levels of NO3"-N at the initiation of
the study. For Lysimeters l and 2 (Spring treatment), concentrations above 1
mg L'1 occurred until September 3 and July 22, respectively (Figure 12). July
22 was the only day during this period for both lysimeters when 15N
enrichments in the leachate were above background (0.00065 and 0.00015 atom

96 excess for lysimeters 1 and 2, respectively). For the majority of the study

82

. 222:3»— Eoc :23me 23 5283895 can owaEEc 0:33:50 .2 0.53m

E08389 House. 93n—

 

 

 

 

82 com cow 2:. com
_ 1......4
a e:
Nara A 24%
_ V1.4 41.“ ES
_ 85: 4‘41.
82% .
4 Q...
332 441.4
.444
14.4.24 .35 + .985 ..4 BE 63589 IT mctqm .0?qu :9

 

o
o
low
B
.2: m.
J
M
-oflmwul
9
-oowxafl
m
fiomm

 

 

§

83

Table 19. Flow-weighted means of nitrate and ammonium N concentrations in
leachate from Spring and Fall treatments.

 

 

 

Treatment Nitrate-N Ammonium-N
mg L“
Spring 0.3 1 0.12

Fall 0.63 0.14

 

84

82

3.” \0A‘ufl’ 4.‘u
n.. 9w
C

 

#58305 93.25 2: magnum:
28252: 8.. 382.8 2953 some .8.“ 8228. E cocabcoocoo 2-38:2 .S 2:me

E258; cots... $39
84

I
|
I
9!
u
'

3R3

N 3853.3 10' fl 33.593 ll

 

_ QNNP

Q.

0)

 

 

N

('I/3W) N—eme

fl'

85

NO3'-N concentrations were well below 1 mg L’l. In the fall of 1993
concentrations began to increase slightly to approach 1 mg L'l.

For lysimeters receiving the Fall treatment concentration above 1 mg L‘1
occurred during the first month of the study, but declined throughout the winter
and spring (Figure 13). During this first month, there was only one occasion
when 15N enrichment was above background (0.00320 atom % excess,
Lysimeter 4, 12/3/91). From June 8 to July 14 1992 concentrations near or
above 1 mg L'1 were detected in Lysimeter 4. NO3'-N concentrations were in
general slightly higher in leachate from the Fall treatment than the Spring
treatment during much of the study. Beginning in the fall of 1993 N03’-N
levels were elevated in leachate from the Fall treatment as well. The elevated
concentrations of NO3'-N which occurred for all lysimeters in the early
samples were probably due to enhanced mineralization caused by soil
disturbance during construction of the lysimeters. This effect was also noted by
Dowdell and Webster (1980) and Geron et al. (1983).

The total amount of nitrogen collected in leachate from lysimeters
receiving the Spring treatment was 5.77 kg N ha'1 (Figure 14). Total LFN in
leachate was 0.09 kg ha‘l, which was 0.23% of the amount applied and 1.6%
of the total N in leachate. For the Fall treatment total N in leachate was 8.30
kg ha"1 (Figure 15). LFN contributed 0.07 kg N ha‘l, or 0.18% of the amount
applied and 0.8% of the total N leached. There were no significant differences
in total N or LFN in leachate between treatments. The vast majority of the
LFN leached for both treatments occurred in a few isolated incidents, as is
evident in Figures 12 and 13. Details on sample composition for these events
appears in Table 20. Note that the data in Table 20 is for individual

lysimeters, and that in Figures 12 and 13 are for means of two lysimeters. The

86

.2258: :5— 2: $5338
3225?: 3.. 380:8 038.8 some 8.. 3282 5 5388350 2.2922 .2 0.53m

Bongo; “3:. $95

 

 

cow coo ace com o
\7— 9 _ _ o
\. / a me: l...‘ .obmktmtom...
.. ‘ P... ##1ti

”70.3. .0 00.“ t .u w
x... .s \ N
I.
3:: n
B
N m...
.. N
w

meE r m
90
/
n1
(

r v.

v uses—E»; Iol m 5856.3 IT

 

 

 

87

.3253: 9:25 2: 31:88 2208?:

:8: 8223— .: cowobmz :38. 28 5952 525.8"— 3353 03336.20 .3 053m

(Bu/3x) N mm.

2.258; 8&4. when

 

 

 

 

 

 

 

o8“ cow coo co»4 com o
ad _ .1... 9...... V .......). 0X) («A Sod
V «M043. 0 «MA» 0 0x .mmav..,mwfio«owv /4\.,....,... 7w .8 ...mmflvmm... l4.
\ .....
ed - S w - o8... m
5:“ m.
382 So a 3%: d
. .... fl .

o 4. Va... 85: - 96 o 9
m.
H
TILE. 8:3 a
6.0 1 1 80.0 J
N
O O m
cm. onelT 25.x. - 832%
g B
(

CA; o2 .o

88

45:53.: :3 2: 2:332 2038?:

:5: 2282 E 5352 .88. 2E cowozmz Suzuki @233 352255 .2 23E

E258; “2.2 mama

 

 

 

cow cow cow 8N o
c6 _ 0 0. .m...ou/)vm<w| “mm" 0 0 on.” .w A Good
i...) . )
ed. V . Sod
.Iw —m\w\:
0
m. . NQO—R .
o .1 - o3 c
N ,
\vv
Op O01 VWQNMA _ NQQ: 1 O00.0
Ur nab—R
max
9w- I owod
_ Z 30H IT 2%..— X
QOQQ
99 oofio

 

 

(Bu/3x) N 1921111195 99mm

89

magnitude of these leaching events as represented in the figures is, therefore,
one-half of that as presented in Table 20.

Precision of 15N content in leachate should be noted. Background 15N
enrichment of drainage waters were calculated from a mean of 22 samples
collected prior to application of LFN. The mean and standard deviation for
atom % 15N were 0.37121 and 0.00439, respectively. Atom % excess of each
sample was calculated based on this background measurement. In most
samples, 15N enrichment was near background levels. In some cases,
conversion to atom % excess resulted in negative values. For the purpose of
mass balance calculations, these were set to zero, because it is impossible to
have ”negative leaching.” Therefore, total values of LFN in leachate were
probably inflated. Consider the early 15N detection from the lysimeters
receiving the Spring treatment. The value of 0.00320 atom % excess in this
sample was actually less than the SD (0.00439) of the atom % of the
background samples. This raises the question of whether or not this ”detection”

of 15N was real.

MasLBalancc

Mass balances for total N for both treatments on the first and last
microplot sampling dates are given in Table 21. The data column for the first
date includes N from fertilizer and an estimate of N from atmospheric
deposition over the two year period covered by the samplings. Totals of initial
N amounts plus added N compared to final N amounts indicate whether N is
being gained or lost by the system. For the Spring treatment two data sets are
included for initial N. The first data is total N measurements from microplots
sampled for the Spring treatment in May 1991 following addition of the first

fertilizer application. Comparison of this data with data from May 1993

90

 

«mod

wvod

Sod
1-..“: mg---
342.34...

353.6 mm; mm; 8?}: m =fim

:33. 82 82 333 _ 2:8

88....» X: 82 833 N 9.2%
2....-.— wE--- ..... Eu .....

3% zlgaaols 354% Slum g 3%

 

.8228 3228
2a venomo— qumo 2:025 .39 EB .3028 ..\o 202 222.2223 2 2.232: .0222, .023 .om 22¢.

91

 

 

 

 

 

 

8% ~30 vmom 33 3.3 .50...

22:83.0

8 8 8 2.23852

can m 2. com 2 202.520...

m o m.m 0o md 02:20..

w . av 2% mm»... .35. Son ..om

.8 m3. .3. N9. 2... 202:...

2. 8 .v w m 8 02.55

a: o 02 3“ 3. 25.3.6

..2 z 2.

lelomlm... 221% a +3 a 02.4%.

205% 20232.23

 

.32.. 2.228 22822 .25 .22 .52. 2 3202205 =2... .22 2.55 2.. comes... .55 .2 022.2. 0.22 ..N 0.2...

92

indicates an increase in total nitrogen of 324 kg ha'1 in addition to the amount
added from fertilizer and atmospheric deposition. Total soil N in this sample
was much lower than total soil N on all other sampling dates for both
treatments. Because of this, the second set of data from May 1991, derived
from non-fertilized microplots harvested at the same time, was included in
Table 21. Comparing this data with that from 1993, decreases in total N in
thatch and soil were partially offset by the increase of N in clippings and a
small amount recovered in the leachate, but an overall decrease of 862 kg N ha'
1 resulted. This was equivalent to 15% of the initial plus added N and was
approximately double the amount of N added as fertilizer. Previous data has
shown that approximately 20% of the fertilizer N may have been lost through
volatility, but this would account for only about 87 kg N ha‘1 of the loss
calculated here. The remainder of the N was probably lost through
denitrification.

Results for the Fall treatment were similar. Losses of N occurred from
the soil and thatch while increases were seen in clipping N and leachate. An
overall loss of 707 kg N ha'l, or 11% of the total amount, was observed. A
loss of 20% of fertilizer N would account for 79 kg ha'1 of the total loss
calculated here, but most of the loss is probably attributable to denitrification.

The overall analyses of variance for total LFN recovery were presented
previously (Tables 4 and 5). Total recoveries of LFN from each canopy
increment at each sampling appear in Tables 22 and 23 for the Spring and Fall
treatments, respectively. Cumulative recovery in clippings at each sample date
increased and recovery in verdure decreased correspondingly for both
treatments. Thatch was an important sink for LFN. Over 30% of Spring
applied LFN was recovered from thatch 18 DAT (Table 22). A relatively

stable amount was retained in thatch throughout the first year for the Spring

93

 

we
on
no
mm
2.
mm
on

 

museum .x.

 

 

m2 m2 5N
3 .mN wood :6
modm wood 36
noon wood was
omém ooo.o who
vwdm ooo.o 2.0
ovdm ooo.o .né
om.om ooo.o 26
:38. alga—4&4 mad.

..2 2.5 3

 

3n Se 3d
an 8.0 3.2
an.» 85 2.:
Ne: NS 8.2
3.2 8.». 8.2
m: can 3.:
was 8.» a:
2.2 2.: :3
find alga» Edna

Qmamm
mQER
NQoQ:
NQon
KEN:—
~338—
Kinky
_o\v:m

clam

 

4:253: macaw of So 23 @2953

some 3 E382 2.023 :33 ES “5820:1353 some .58 5on Z ..onEtom “co—2.3 ho bo>ooom .NN 03a“.

94

treatment. During the second year thatch LFN levels decreased significantly,
coinciding with a trend of decreasing total recovery. Recovery from soil
increased significantly between May and October 1991, and remained
significantly higher through November 1992. Total recovery did not differ
significantly over time, ranging from 64 to 92%. This was similar to the range
in recovery reported by Starr and DeRoo (1981) of 64 to 76%.

Over 62% of the Fall applied LFN was recovered from thatch at 18 DAT
(Table 23). Recovery from thatch was similar in the May 1992 then decreased
sharply and significantly by the end of June 1992. For the remainder of the
study there was a trend towards decreasing thatch LFN, and in November of
1993 the amount was significantly lower than in June of 1992. Soil LFN
tended to increase while thatch LFN was decreasing. The largest decrease in
thatch LFN was not associated with an increase in the soil, but rather coincided
with a significant decrease in total recovery. Total recovery from the
November 1991 and May 1992 samples was significantly greater than from all
other dates, but the latter dates were not different from one another.

Decreases in thatch LFN from both treatments were usually associated
with increases in soil LFN, but not in equivalent amounts. The loss from
thatch was also related to decreasing total recovery. It was suspected that loss
of LFN in gaseous forms was occurring. This was first suspected following
harvest of the first sample for the Spring treatment, which resulted in total
recovery of only 78%. The apparent loss from the Fall treatment over six
months after application was surprising. It is unknown whether this was due to
volatility or denitrification, but the rapid rate of loss would favor volatility.

As described previously, the timing of LFN applications and soil
sampling generally precluded direct comparisons between treatments.

However, there were two sample dates which coincided in DAT for the two

95

 

 

 

 

 

m3 m2 8a 8s 82 ”no 8......
a a: Sod 3... $8 2.0 8.: 85m:
8 a: 8o... 2.» an” 2.. 3.2 8.2%
t .3 ~86 3 92 E. and 832:
3 cm .86 .3 .2 2a 8.2 Sta
3 SA .8... o: 3.: a: 2: 8.23
8. 2.. ooo.o o: 8.: a.» a; 33%
2: :2. ooo.o E... ”a: 5.: o 58:.
..2 25 w.
@383. x .38. 3% no: 3 8.84% fig male

 

.258th ..a... 2.. .0.. 83 wit—cam
some 3 .0968. .582. .82 .38 22:26:. .8058 ..ouo 8.5.. comes... Snatch. 3.3%....0 boson... .mm 0.9....

96

LPN applications, and direct comparisons were made for these times. For the
Spring treatment, the May 1991 and May 1993 samples were 18 and 748 DAT,
respectively. For the Fall treatment, the November 1991 and November 1993
samples coincided. This provided an opportunity to examine the disposition of
LFN after equal time intervals for the two treatments. Comparisons of
differential responses of treatments over time were not of primary interest, and
so time was not included as a factor. Data for each sample time was analyzed
separately, and effects of treatment and depth were examined. As before, the
depth factor represented clippings, verdure, thatch, soil, and leachate. For
both sample times, treatment x depth interactions were significant (P = 0.01
and P = 0.05 for times 18 and 748, respectively). Comparisons were then
made between treatments at each depth (Table 24). Totals of all depths within
each treatment were analyzed separately and are included in Table 24.

The significant difference in clipping harvest at 18 DAT is meaningless
because clippings were not harvested during this period for the Fall treatment.
Content of LFN in verdure was not different. This was a surprising result
because transport of this much N into shoots of fall fertilized turf was not
expected. The difference in fertilizer N in thatch was great and coincided with
a difference in total recovery. These were encouraging results after the
suspected volatile loss from the April application. However, this difference in
thatch LFN was apparently equalized the following spring when significant
LFN loss from thatch occurred for the Fall treatment. Over the two year
period, more LFN was recovered from clippings of late fall fertilized turf.
More LFN was also present in the soil for the Fall treatment, and total recovery
was significantly higher. The higher level in the soil could raise concern about
potential groundwater contamination, but soil 15N was present predominantly

in organic forms which are relatively immobile. This higher soil LFN level

97

Table 24. Comparisons of Labeled Fertilizer Nitrogen recovery in each canopy
increment at 18 and 748 days after treatment.

 

 

 

Days After Canopy Treatment
1121mm mm mm Ea_ll
kg LFN ha'1
18 Clippings 0.94 0“
Verdure 14.25 14.01
Thatch 12.15 2428*
Soil 3. 16 4. 77
Leachate 0.00 0.00
Total 30.50 43.06"
748 Clippings 13.89 15 .02“
Verdure 0.68 0.27
Thatch 5.23 6.69
Soil 5.34 996*
Leachate 0.01 0.07M
Total 25.15 32.01"

 

*,*"' Means within a row significantly different at P = 0.05 and 0.01,
respectively.

98

may actually indicate greater fertilizer use efficiency by the late fall fertilized
turf. Leachate LFN was also greater for the Fall treatment than the Spring
treatment at 748 DAT. As was shown in Figure 14 however, leachate LFN
increased in the Spring lysimeters during the fall of 1993 (after 800 DAT), and
was not significantly different from the Fall treatment in December when data
compilation was completed. This parameter was more dependent on season
than DAT. The comparisons at 748 DAT should be viewed with some caution.
Between November 1992 and May 1993 total recovery for the Spring treatment
decreased by 12% (Table 22). The reason for this is not known, but this
probably had an effect on the statistical comparison. Recovery in November
1992 was 76% for the Spring treatment, which is comparable to the 81%
recovery for the Fall treatment in November of 1993. The differences detected

between treatments at 748 DAT may not be true treatment differences.

CONCLUSIONS

Under both fertilization schedules more than 35% of applied LFN was
removed in clippings over a two year period, the majority of that occuring in
the first growing season after application. Application of nitrogen in the late
fall yielded greater uptake of LFN as well as greater total N uptake and higher
clipping yield. Thatch played an important role in storage of fertilizer nitrogen
on both short-term and long-term scales, but may also be an environment that is
conducive to gaseous losses of nitrogen. Initially following application, only 8
and 12% of the LFN applied in April and November, respectively, was found
in the soil. These numbers increased to 14 and 25% over two years. Soil
microbial biomass played an important role in early immobilization of this
applied nitrogen. Only small proportions of this N were present as inorganic N
and so potential mobility of the soil LFN pool was low. Mean flow-weighted
NO3' concentrations in leachate were 0.31 and 0.63 mg N L‘1 for the Spring
and Fall treatment, respectively. Total LFN recovered in leachate was
approximately 0.2% of the amount applied over two years. Leachate will
continue to be monitored due to the detections of higher N concentrations in the
Fall of 1993.

This study indicates that potential for groundwater contamination due to
nitrogenous fertilizer application to turfgrass is minimal. Application of N in
early November, an agronomically favorable practice, does not pose an
increased risk as compared to an April application. In fact, utilization of late
fall applied N may be more efficient. ‘

Questions posed by the results of this research include the potential for
losses through volatility and denitrification, the role of the thatch layer in

storage and cycling of nitrogen, and immobilization and mineralization
99

100

potentials of turfgrass soils. Research in these areas would more fully describe

the nitrogen cycle in the turfgrass ecosystem.

CHAPTER II

IRRIGATION AND NITROGEN SOURCE EFFECTS
ON RECOVERY OF 15N-LABELED FERTILIZER
APPLIED TO TURFGRASS

LITERATURE REVIEW

Ammonia volatilization is the process by which NH4+-N is converted to
gaseous NH3 and lost to the atmosphere. The general reaction is:

NH4+ + H20 + OH' -+ NH3 + 21120
and requires an aqueous alkaline environment as the reaction indicates (Tisdale
and Nelson, 1975). In order for loss of nitrogen to occur by volatilization, it
first must be available in the NH4+ form. Ammonium can be supplied either
from mineralization of N from soil organic sources, application of NH4+
containing fertilizers, or application of fertilizers containing other N sources
which are subsequently converted to NH4+. One of the most common forms
of nitrogenous fertilizer is urea [CO(NH2)2] which is hydrolyzed in the soil by
the following reaction:

CO(NH2)2 + 2H20 —) (NH4)2CO3.
This reaction is mediated by the urease enzyme. Ammonium carbonate is not
stable and dissociates yielding NH4+ and C032', which consumes H+ from
soil solution, driving pH up. Therefore, even in neutral or slightly acidic soils,
microenvironments surrounding urea particles or droplets can become alkaline,

resulting in good conditions for volatility.

101

102

II E . .

Overrein and Moe (1967) incubated two soils (Chalmers silt loam, pH
6.5, O.M. 4.5%; Plainfield sand, pH 5.7, O.M. 1.2%) with urea at rates of
224, 448, and 896 kg N ha'1 and studied the effects of several different
environmental parameters on ammonia volatilization. They found the rate of
urea hydrolysis in the Chalmers silt loam at 24% moisture to increase linearly
with application rate at 28°C. At 4°C the response was linear for the first two
rates, but hydrolysis dropped off for the high rate of urea, presumably due to
more limited activity of urease at the lower temperature and an effective
saturation of the enzyme. At 28°C rates of hydrolysis were approximately 110,
220, and 450 kg N ha’1 day-1 from the 224, 448, and 896 kg N ha-1
application rates. Volk (1966) observed almost no hydrolysis of urea when
applied to air dry soils incubated in the laboratory. In the field, hydrolysis and
volatile loss was greatly reduced on dry soils as compared to moist soils.

Urease activity in a Kentucky bluegrass thatch was approximately 25
times higher than in underlying soil (Torello and Wehner, 1983). Activity in
thatch of dormant turf was approximately 40% of that of actively growing turf.
Ten days of greenhouse incubation increased the activity of the dormant thatch
urease by about 40%. Activity was almost twice as great in the upper 1 cm of
a 2 cm thatch layer as compared to the lower 1 cm. There was also great
variability between two different sampling locations. Bluegrass leaves also had
high urease activity, approximately 18 times greater than the soil. Bowman er
al. (1987) found similar trends in urease activity in Kentucky bluegrass shoots,
thatch, and soil, although the magnitude of activity was not as great (15 times
greater in thatch compared to soil). Shoot and thatch tissue had comparable
levels of activity on a dry weight basis, but activity in the thatch was double

that in plant shoots when expressed on an area basis. Most of the urease

103

activity (96%) in plant shoots was in sheath tissue (including dead and
senescing tissue) as opposed to leaf blades. The top 1 cm of soil had five times

the urease activity of the 1 - 3 cm layer.

Wm

Ammonia volatilization rates from a Chalmers silt loam following
application of 224, 448, and 896 kg N ha'1 as urea were measured each day for
five days, and were found to increase curvilinearly (Overrein and Moe, 1967).
A greater proportion of NH4+ was bound to soil at low application rates than
high rates. The highest rates of volatilization, occurring on day two, were
approximately 0.5, 2, and 9 kg N ha’1 clay'l for the three fertilizer rates.
Total volatile losses over five days were 0.7, 1.0, and 1.8% of the applied N.
Volk (1959) noted a similar effect when pelletized urea was applied to four
different warm season turfgrasses. Percent lost through volatilization averaged
13.6, 18.0, and 20.6 for application rates of 28, 56, and 112 kg N ha'l,
respectively. Simpson and Melsted (1962) observed that approximately 30% of
N applied as urea was lost by volatility from a bluegrass forage sod over 10
days when applied at 56, 112, and 168 kg N ha'l. Total loss increased with
rate of application for a number of grass crops tested. Wesely et al. (1987)
also found that total N loss increased with increasing application rates, but the
percent N lost remained relatively constant. For application rates of 17 and 34

kg N ha’l, total losses over four days were 35 and 31 % of the applied N.

Volatile losses of N from ammonium sulfate, urea, and ammonium

nitrate applied to Meloland clay loam (calcareous, pH 8.0, moisture content

75% of field capacity at time of application) were 25, 16, and 11% of applied

104

N over 70 days (Martin and Chapman, 1951). Loss from NH4NO3 was low
because only one—half of the N was in the NH4+ form at the time of
application. Applied to four different sandy loam soils of pH 7.7, 7.5, 7.1,
and 6.7, loss of N from ammonium sulfate was 14, 19, 2, and 4%. Losses of
urea-N were 14, 16, 18, and 36%. Losses of NH4NO3-N were 7, 7, 1, and
1%. Volk (1959) measured volatile losses from pelletized urea averaging 30%
from four turfgrasses in the field, as compared to 0.3% loss from NH4NO3.
Soil pH was not given for this particular experiment, but other soils in this
paper ranged in pH from 4.4 to 6.7. The only significant loss of ammonium
sulfate-N occurred from a Perrine marl with a pH of 7.8 In the above studies,
application of ammonium salts to alkaline soils resulted in volatile loss, but
application to neutral or acidic soils resulted in very little loss. Urea applied to
acidic, neutral, or alkaline soils resulted in volatile loss.

Torello et al. (1983) studied the effects of different urea sources and
application techniques on volatilization from a Kentucky bluegrass turf grown
on a silt loam soil with pH 6.4. Applied at 293 kg N ha‘l, significantly more
urea-N volatilized (10.3%) than N from two sulfur coated urea sources (2.3 and
1.1%). Sulfur coated urea with a dissolution rate of 37.5% produced greater
volatile loss than formulations with 27.2 and 18.0% solubility (0.7, 0.2, and
0.2% lost, respectively; application rate of 98 kg N ha'l). Comparing two
ureaformaldehyde products formulated for application through a spraying
system, they found that the formulation with 50% free urea yielded greater
volatile loss than the product with 35% free urea (4.5 vs. 3.2%, application
rate of 49 kg N ha'l). In a fourth experiment, greater loss was measured from
urea applied as a solution (4.6%) than when applied in a prilled form. This
was explained by the fact that the solution application remained primarily on

leaf blades, where it was subject to high urease activity (Torello and Wehner,

105

1983) and the lack of exchange sites provided no protection against volatile loss
as would occur in the soil (Volk, 1959). Interestingly, Titko et al. (1987)
observed the opposite effect. Greater loss occurred from granular urea than
spray applied urea. Temperature and spray solution volumes were essentially
the same in both experiments [24°C, 1629 L ha"l (Torello et al., 1983); 22.2°
C, 1630 L ha‘1 (Titko er al., 1987)], but N rates were different (49 hg ha'1
and 73 kg ha'l, respectively). Torello et al. (1983) noted that roughly
equivalent amounts were lost from spray applied urea and ureaformaldehyde
combination products. The UF components were not hydrolyzed on leaf tissue,

and they concluded that volatile loss was from the urea components only.

W

Increases in volatile loss of N from ammonium sulfate, urea, and
ammonium nitrate with increasing temperature were observed by Martin and
Chapman (1951). They attributed this to greater soil drying. Volk (1959) also
noted this temperature affect. Titko et al. (1987) observed significantly higher
loss from both granular and dissolved urea from Kentucky bluegrass at 22.2°C

as compared to 10°C, but no further significant increase at 32.2°C.

Wm

Martin and Chapman (1951) concluded that initial soil moisture content
had no effect on total volatile losses, although they stated the reason for the
temperature effect they observed was a greater degree of soil drying. This
reasoning seems to be contradictory. Titko et al. (1987) subjected turf to
alternate wetting and drying cycles every 24 hours, and maintaining constant
relative humidity of air passing through the system (68 :1: 4%), found greater

loss from both dissolved and granular urea as compared to plots not subject to

106

the wetting/drying cycles. Martin and Chapman (1951) also saw additional loss
with each increment of a wetting/drying cycle. Wesely er al. (1987) observed
that the greatest rates of volatility occurred at times of drying of moist leaf
tissue. Titko et al. (1987) observed lower volatile losses from turf maintained
at 68% relative humidity compared to 31 % relative humidity.

Irrigation of turf with 1 cm of water during the first hour following
application of either granular or solution urea reduced volatile loss
significantly. Bowman et al. (1987) applied 50 kg N ha'1 as urea in 0.2 cm of
water to Kentucky bluegrass followed by irrigation amounts of 0, 0.5, 1.0, 2.0,
and 4.0 cm. With no irrigation, 25% of the applied N was lost in the first 24
hours. Increasing depths of irrigation decreased losses over this period to
approximately 7, 4, 2, and 2%. When urea was applied in 0.05 cm of water
with no supplemental irrigation, loss was lower than for the 0.2 cm depth of
application (18% vs. 25 %). If no supplemental irrigation was applied
following application in 0.2 cm of water, 21, 47, and 8% of the applied N was
recovered from shoots, thatch, and soil (0 - 3 cm depth). When 0.5 and 1 cm
of irrigation was applied, 5, 26, and 26% and 1, 25, and 27% were recovered
from these environments, respectively. From application in 0.05 cm, 75, 24,
and 1% was recovered from shoots, thatch, and soil. The vast majority of the
N recovered in the soil in all cases was in the upper 1 cm. Nitrogen not
recovered was attributed to biological immobilization. Although apparent
distribution in the thatch layer for the 0.5 and 1.0 cm treatments was the same,
the lower loss rate from the 1 cm treatment was probably attributed to deeper
incorporation within the thatch layer where urease activity is lower (Torello and

Wehner, 1983).

107

Mm

Simpson and Melsted ( 1962) found that volatility losses from a bluegrass
sod were considerably higher than from other grass crops. Hydrolysis began
immediately and rates were high for the first two to three days for all crops.
Volatility was measured in bluegrass on the first day, but for all other crops
there was a delay of 2 - 3 days before the onset of N evolution. Bluegrass
samples were taken from the field, whereas all other crops were planted from
seed in the greenhouse. The authors attribute the difference in results to the

organic litter in the bluegrass sod.

Volatile losses of fertilizer N from turfgrasses have been well-
documented. Bowman et al. (1987) and Torello and Wehner (1983) have
shown that urease activity in turfgrass thatch is much higher than the soil
below. Elevated temperatures have been shown to contribute to volatility
(Martin and Chapman, 1951; Titko et al., 1987; Volk, 1959), as well as
drying conditions (Martin and Chapman, 1951; Titko et al., 1987; Wesely et
al., 1987). These three factors combine to make the thatch layer an area of
high volatility potential. Some researchers have shown that higher percentages
of N are lost through volatility as application rate increases (Overrein and Moe,
1967; Volk, 1959). Others have observed increased volatile loss with increased
rate, but no marginal increases in loss (Simpson and Melsted, 1962; Wesely et
al., 1987). Urea exhibits higher volatility rates as compared to modified urea
or nitrate sources (Martin and Chapman, 1951; Torello er al., 1983; Volk,
1959). Irrigation immediately following application also reduces losses
(Bowman et al. , 1987). Although ammonia volatility can be a source of loss of

fertilizer N applied to turf, choice of the correct fertilizer source, application

108

during cooler and/or drier times of the day, and proper irrigation following

application should limit such losses.

MATERIALS AND METHODS

Recoveries in the range of 75% to 90% of spring applied urea and
recoveries in excess of 100% for fall applied urea in the experiment in Chapter
I indicated that substantial losses of urea-N may have occurred through
ammonia volatility under spring conditions. This experiment was designed to
account for possible volatile losses under these conditions and determine the
effect of irrigation on volatile losses of ammonia in the cool humid climate of
Michigan.

In May of 1993 60 microplots 30 cm in depth were constructed of 20 cm
diameter PVC pipe. These microplots were beveled on one end and installed in
the same manner as those in the previous experiment. Soil and turf conditions
were as described in Chapter 1. Three fertilizer treatments were applied to
these microplots in a randomized complete block layout with four replicates.
Two treatments consisted of fertilization with 15N-labeled urea (19.1991 atom
% excess) at a rate of 38.0 kg N ha'l. One of the urea treatments received a
total of 0.5 cm irrigation with the application (UL) and the other received 2.0
cm (UH). Bowman et al. (1987) observed a reduction of volatile loss form 7%
to 2% using these irrigation practices. The third treatment was fertilized with
15N-labeled potassium nitrate (KNO3) (19.4546 atom % excess) at a rate of
37.2 kg N ha'1 watered-in with 0.5 cm water. Presumably, there should be no
volatile loss of N from KNO3, at least in the short term. Comparison of KNO3
and UL treatments should illustrate the potential loss due to use of urea, and
comparison of UL with UH should illustrate the effect of irrigation. All
treatments were applied in 0.05 cm water from 125 m1 polypropylene bottles

and immediately watered-in with the appropriate amount of water as described

109

110

in Chapter I. Treatments were applied on June 11, 1993. Air temperature was
24°C, wind was from the east at 13 km h'l, and relative humidity was 53%.

Five designated sampling times for the experiment were 1 hr, 4 days, 20
days, 40 days, and 80 days after treatment. Bowman et al. (1987) found that
volatile losses of N from urea under similar conditions occurred primarily in
the first 12 hours after application, with slight additional losses occurring up
until 72 hours. Volk (1966) also observed that the majority of total loss due to
volatility occurred within 72 hours, and Wesely et al. (1987) found that
volatilization rates peaked within 48 hours after application. It was
hypothesized that the largest loss in this experiment would occur between 1
hour and 3 days after application. At each of these times one microplot for
each sample in each replication was excavated and processed as described in
Chapter I. Exceptions were that soil depth increments extended to 20 cm only,
and only one soil subsample was saved from each depth, to be dried and ground
for analysis. All plant tissue and soil samples were ground as before and
analyzed for total N and 15N enrichment on a Europa Scientific Roboprep C-N
Biological Sample Converter and Tracermass Mass Spectrometer. Data from
the experiment in Chapter I indicated very little downward fertilizer N
movement. It was assumed that 15N not recovered in plant and soil samples
from within the microplots was lost to the atmosphere. Therefore, volatile
losses of 15N were implied by difference.

Microplots designated for excavation at 1 hr after treatment were
partially excavated before hand because of time constraints. Soil was removed
from around each microplot, leaving the enclosed core standing freely but still
attached to the soil below. After application of 2 cm irrigation to designated
treatments, water was seen draining out of the bottoms of three of the four

exposed microplots receiving this treatment. This drainage appeared to be from

111

the edges of the cores, and movement may have been primarily by preferential
flow down the tube walls.

Data was converted to percent recovery since the application rates were
slightly different for the two sources, and analyzed as a split-plot with
treatment as main plot and ”depth" as sub-plot for each sample time. Once
again, depths were comprised of clippings, verdure, thatch, and soil (total of
all depths). Analysis was conducted with PC-SAS v. 6.04 (SAS Institute,
Cary, NC).

RESULTS AND DISCUSSION

There were significant treatment by depth interaction effects on percent
LFN recovery at 1 hr and 20 and 40 days after treatment (Table 25). Data was
further analyzed to separate treatment differences at each canopy level on each
sampling date (Table 26). One hour after application there was significantly
more LFN recovered from soil from the urea high irrigation treatment (UH)
than from the urea low irrigation (UL) or potassium nitrate (K) treatments,
indicating deeper placement as a result of increased irrigation. Flow of
irrigation water from the bottoms of the partially excavated microplots
receiving the UH treatment did not result in significant additional losses of
LFN. Less than quantitative recovery of LFN is indicative of procedural
errors, but the lack of significant differences between treatments indicate that
these occurred uniformly across treatments. Lack of differences in total
recoveries between all treatments and the similarity in location of LFN between
UL and K treatments shows that there was no significant volatile loss of N from
urea as compared to KNO3. Bowman et al. (1987) observed less than 1%
volatile loss of N 1 hr following application of urea followed with 0.5 cm
irrigation. Although total recovery was much less in their study (57%),
Bowman et al. (1987) also found that 56% of the amount of LFN recovered 1
hr after application was in the thatch layer. This was very similar to the 53%
of the total recovered in the 0.5 cm irrigation treatments in the present study.

Samples collected four days after treatment yielded recoveries well in
excess of 100% for the UL and K treatments (Table 26). Recovery for the UH
treatment (85.3%) was significantly lower than UL and K, but similar to total

recovery for UH in the 1 hr sample Total recoveries for UL and K treatments

112

113

Table 25. Summary of split plot analyses of variance at each sampling time.

 

Source of Sample time
Variation
u 4 days 20 days 40 days
Treatment **
Replication
Depth * * ‘ * ‘ * * *

Treatment x
Depth " § §
§, " Significant at P = 0.10 and 0.01, respectively.

114

Table 26. Percent recovery of LFN in each canopy level at each sampling time.

 

 

lime Treatment Clippings Mr; M §_9_i_l Total

1 hr UL 0 29.1 46.4 11.5 87.0
K 0 29.7 46.5 12.4 87.6
UH 0 16.9 38.4 30.0 85.3
FPLSD NS NS NS 13.1 NS
(0.05)

4 days UL 0 51.8 62.7 11.8 128.5
K 0 46.9 57.7 15.4 119.9
UH 0 28.4 37.7 19.2 85.3
FPLSD NS 12.7 NS NS 22.6
(0.05)

20 days UL 16.7 24.9 42.6 5.4 83.3
K 25.4 24.1 36.8 4.4 85.9
UH 13.7 15.1 31.2 13.2 73.1
FPLSD 6.8 NS NS 2.8 NS
(0.05)

40 days UL 24.7 22.4 33.2 8.7 88.9
K 35.1 20.2 27.2 6.2 88.8
UH 20.6 15.1 24.7 12.2 72.7
FPLSD 6.6 4.4 NS NS NS

(0.05)

 

115

were between 83 and 89% for all sample times except 4 days. Large increases
in LFN were seen in verdure and to a lesser extent thatch, but the reason for
the overall discrepancy is unclear.

At 20 and 40 DAT, total recoveries were not significantly different
between treatments. Recovery of LFN in clippings was greater for K than UL
and UH at both times. LFN in verdure was significantly higher for UL and K
than UH at 40 DAT and also appeared to be higher at 20 DAT, but this
difference was not significant. Recovery from soil was higher for UH than UL
and K at 20 and 40 DAT. This data indicates that deeper placement of
fertilizer N due to irrigation moved the fertilizer past the zone of greatest plant
uptake.

The percentage of LFN in the thatch remained high throughout the study,
but did appear to decrease over time within each treatment (disregarding the 4
day sample because of the high recovery). The decreases in thatch LFN were
accompanied by increases in the amount recovered in clippings + verdure,
indicating either uptake of fertilizer N from the thatch or transport of plant
LFN within the thatch layer upward in the plant.

The potential for volatile losses of N from turfgrass has been well
documented (Volk, 1959; Nelson et al., 1980; Torello and Wehner, 1983;
Nelson et al., 1980; Torello et al., 1983; Bowman et al., 1987; Titko et al.,
1987; Wesely et al., 1987). In addition, the thatch layer has been described as
an environment conducive to volatilization. Although data in this experiment
indicates that a large portion of applied fertilizer nitrogen is resident in thatch
immediately after application, decreases in LFN from thatch could be accounted
for through LFN harvested in plant tissue above the thatch layer. Under the
conditions of the present experiment, volatile losses of N were not a significant

factor.

116

Based on the results of this experiment, no conclusions can be made
concerning non—recovered LFN in the experiment in Chapter 1. Although
volatile losses were suspected in that experiment, the present data shows that
non-quantitative recovery can occur even with a fertilizer source not prone to
volatility. However, volatility was not measured directly in the present
experiment, which may be a contributing factor in these results. Further
investigation into the nature of thatch and its ability to immobilize N, especially
following fall applications, and subsequent losses of N from thatch during

warming periods, are warranted.

CHAPTER III

MINERALIZATION AND TURNOVER OF SOIL NITROGEN
IN A TURFGRASS COMMUNITY

LITERATURE REVIEW

WWW

Soil organic matter constitutes a major reserve of nitrogen that can
contribute to agricultural productivity. Estimating mineralization rates and
availability of this nitrogen can lead to a better understanding of the plant-soil
system and more efficient use of inputs. Stanford and Smith (1972) conducted
fundamental research in use of long-term incubation of soils to construct
estimates of N mineralization potential. Studying 39 different soils incubated
and leached intermittently to remove mineralized N over a period of 30 weeks,
they formulated the following model for estimating N mineralization potential:

Nm = No * (l-e'kt)
where

Nm = total mineralized N in time t

N0 = potential N mineralization

k = mineralization rate constant.
They found this model to have the greatest utility when applied to data for
incubations from weeks 2 through 30. When the higher mineralization rates
observed in the first two weeks were included, the fit of the model was not as
accurate. For 29 soils included in the analysis, the average rate of

mineralization was 0.054 wk'l (5.4% of mineralizable N released per week).

117

118

They concluded that total mineralization potential could be most accurately
predicted from the sum of mineralization in the first two weeks plus the model
prediction for the long-term release of N.

Molina er al. (1980) reexamined the data of Stanford and Smith (1972)
and found that the following two pool model for nitrogen mineralization
provided a more accurate estimate:

Y = s * (1 - e-ht) + (1-5) * (1 - e-kt)
where

Y = cumulative mineralization at time t

S = a fraction of potentially mineralizable N

h = mineralization rate for fraction S

k = mineralization rate for fraction (1 - S)

t = time.

This model implies the existence of two different pools of mineralizable soil N
which are released at different rates. Based on the data, they calculated a labile
pool (S = 15.8% of total N) which decomposed at rate h = 1.131 wk'1 and a
more resistant pool (1 - S = 84.2%) with a decomposition rate of k = 0.0424
wk'l. By non-linear least squares (NLLS) analysis it was determined that this
model provided a more accurate explanation of the data. Smith et al. (1980)
also determined that NLSS was a more precise test of fit as compared to least
squares analysis of log-transformed data for the single exponential model.

Deans et al. (1986) also employed the two pool model of Molina er al.
(1980), and further concluded that the single model underestimated No (total
mineralization) and overestimated the rate of decomposition k. A simplified
version of the two pool model was constructed by Bonde et al. ( 1988):

Nm =Na*(1-e-ht) + Ct

where

119

Nm = total N mineralized in time t

Na = total labile mineralized N (S “ No)

Ct = linear expression representing slope of mineralization curve of

resistant pool.
For incubations of fallow soils and soils cropped to cereals with either no added
fertilizer, added fertilizer, or added manure, mineralization rates were best
described by the simplified model. The authors proposed that a constant
amount of mineralization could occur when microbial populations reach some
maximum level and equilibrium is established, and substrate is in abundant
supply. Where the cropped soil was amended with straw, the full two pool
model provided the best fit.

Boyle and Paul (1989) also employed the two pool model to describe N
mineralization from sewage sludge amended plots. They found the period of
mineralization of the labile fraction to be 11 weeks, with mineralization of the
recalcitrant fraction predominating after that time.

A number of possible limitations to the use of the incubation method
have been suggested. Bremner (1965) questioned the practices of pre-leaching
the soil, using soil amendments such as vermiculite or sand to enhance leaching
characteristics, the use of nutrient solutions to extract mineralized N, and using
suction to equilibrate soil moisture content prior to incubation. Each of these
methods was used by Stanford and Smith (1972) in their original work. Smith
et al., (1985) noted that organic N was removed with inorganic N when soils
were leached with CaClz, but changing the volume or concentration of the
extractant did not significantly affect this. Failure to account for partially
mineralized N or removal of substrate could lead to errors in estimating
mineralization potential. Juma and Paul ( 1984) explored a variety of extraction

procedures, including autoclaving and different chemical extractants. They

120

concluded that any one extractant could not account for all sources of
mineralizable N, and that further work in this area could lead to better

discrimination between various mineralizable N pools.

ll' 1' . Cl . . f5 5 .1

Mineralization potential and rate constants vary with methodology and
soils. Stanford and Smith ( 1972) reported mineralization potentials for various
soils ranging from 18 to 358 mg kg'l. Based on percent of total soil N, the
range was 4.6 to 40.6%. Mineralization rate constants ranged from 0.035 to
0.095 wk‘l, with a weighted mean of 0.054 wk'1 for 29 of the soils. Smith et
al. (1980), using the single pool model, reported No values between 65 and 92
mg kg'1 and rate constants between 0.124 and 0.240 wk"1 for three different
soils. Fitting the data of Stanford and Smith (1972) to the two pool model,
Molina et al., (1980) determined rate constants of 1.131 for the labile
component (15.8% of N0) and 0.0424 for the more resistant component (84.2%
of No). Deans et al. (1986) fit data from several researchers for soils from a
number of states to the single and two pool models. Based on the double
model, mineralization potentials ranged from 76 to 881 mg kg'l. Labile
fractions comprised between 1 and 25% of total mineralizable N. Rate
constants for the labile components ranged from 0.29 to 3.50 wk'l, and from
0.004 to 0.069 wk"l for resistant fractions.

Nordmeyer and Richter (1985) found that addition of organic
amendments to incubated soils altered the N mineralization dynamics. Addition
of 6000 and 12000 Mg leaves ha'1 (190 and 380 kg N ha'l) from sugar beet
(Beta vulgaris) prompted turnover and increased mineralization in an amount
approximately equivalent to the amount of added N. Approximately 2/3 of the

added material was quickly decomposable, and the remaining 1/3 was more

121

stable. Rates of decomposition of the labile soil N were reduced while rates for
more stable N were increased in response to the addition of leaf material.
Following addition of wheat straw (Triticum aestivum), immobilization rates
increased for 30 days compared to the control. Mineralization rates were
greater than immobilization, resulting in slight net mineralization for amended
soils during this period, but net mineralization was greater in untreated soils for
100 days as compared to soils receiving 900 Mg straw ha‘l. For soils amended
with twice the amount of straw, net mineralized N remained lower than the
control for the entire 150 day incubation.

Boyle and Paul (1989) found total mineralization and mineralization rates
increased in response to addition of sewage sludge to soil. Rates were 0.010,
0.013, and 0.020 wk’1 for soil receiving 0, 45, and 180 Mg sludge ha'lyr'l
for 8 years. Sludge was not applied for three years prior to the incubation
experiment, and barley (Hordeum vulgare L.) was grown in the field each year.
These rates do not reference either labile or stable fractions, but apply to the
entire soil N pool. Additional quantities of 266 and 458 mg N kg’1 were
mineralized in treated soil as compared to the control. They also observed a
decrease in microbial biomass of over 50% in the first 20 weeks of the
incubation followed by stabilization. If a C:N ratio of 7:1 is assumed, N
mineralized from microbial biomass represented approximately 8% of total
mineralized N averaged for all treatments.

Bonde et al. (1988) also saw increased mineralization due to added N
sources. Mineralizable N over a 40 week incubation was 93 ug g'1 for fallow
soil. Soil cropped to cereals with no additional N mineralized 107 ug g'l.
Additions of calcium nitrate (80 kg N ha‘1 yr'l), calcium nitrate plus straw (80
kg N ha‘l, 1800 kg C ha'l), and calcium nitrate plus manure (80 kg N ha'l,

1800 kg C ha'l) to cropped soils resulted in mineralization of 114, 167, and

122

165 ug g‘l. Between 6.2 and 8.8% of the total N pool was mineralized.
Reductions in the size of the microbial biomass pool were also reported. At 4,
9, and 47 weeks of the incubation, the biomass pool was 36%, 23%, and 8% of
its initial size. Biomass N was reduced to 9% of the initial amount at 47
weeks. A model, analogous in form to the model for N mineralization, was
developed to describe the decline in biomass:

Bt = Ba * or“) + Br * (ex-1“)
where

Bt = microbial biomass at time t

Ba = labile biomass

Br = resistant biomass

h = rate constant, labile biomass

k = rate constant, resistant biomass.

Biomass N in the soil ranged from 3.9 to 6.8% of the total N. Of the N
mineralized, 55 to 89% was from biomass.

In contrast, Juma and Paul (1984) determined that 15 to 25% of N
mineralized in 12 weeks was from microbial biomass. Due to computational
differences, this number can be recalculated as 6 to 10% in order to be
compared to data of Bonde, et al. (1988). Muramoto et al. (1982) measured
mineralization flushes from oven dried (70°C) and air dried soil and found that
77% and 55% of the mineralized N was biomass N. They observed a strong
linear relationship between decline in biomass C and increase in mineralized N,
in a ratio of approximately 10:1.

The literature indicates that mineralization potentials and rates vary
widely among soils. The nature of labile and resistant components of organic N
has not been completely defined, and values for mineralizable biomass N,

considered generally to be labile, are vastly different. Aside from these

123

difficulties and inconsistencies, mineralizable N remains a very important
subject. Its estimation and/or measurement can provide very useful information
in the understanding of the nitrogen cycle.

Determinations of mineralizable nitrogen from turfgrass soils as
measured by laboratory incubation procedures were not found in the literature.
Starr and DeRoo (1981) observed increases in soil nitrogen due to fertilization
and clipping management practices, and predicted that as much as 50% of this
added N would be mineralizable. By their estimates, an additional 40 kg ha'1
of N could be available for plant growth. Several authors have recognized the
importance of mineralization and its potential impact on turf growth and
environmental quality (Starr and DeRoo, 1981; Mitchell et al., 1978; Geron er
al. (1993). The objectives of this experiment were to measure mineralizable N
from soil with a turfgrass cover. The effects of added fertilizer and leaf

clippings on turnover of soil N were also studied.

MATERIALS AND METHODS

Two laboratory incubation experiments were conducted to study the
dynamics of nitrogen cycling in a soil beneath a Kentucky bluegrass turf. The
first experiment was designed to assess the potential for mineralization of N
from soil over 180 days. The second experiment was designed to study the

effects of added N sources on short-term cycling of N.

- I ° x r'

A variation of the long term incubation experiment of Stanford and
Smith (1972) was initiated in August of 1993. The soil used was from the 0-5
cm and 5-10 cm depths of microplots harvested in May 1993 from the
experiment described in Chapter I. Only plots which had received 15N-labeled
urea (39.2 kg N ha'l, 24.9613 atom % excess 15N) in April of 1991 were
used. Samples to be analyzed on twelve dates were weighed out with four
replicates for each date and each depth. For each sample, approximately 12g of
field moist soil was weighed into a 50 ml glass beaker. The beaker was
covered with Parafilm to prevent moisture loss and placed into a 120 ml plastic
specimen cup (Premium Plastics, Inc., Chicago, IL). Three holes were
punched in the lid of the cup using a dissecting needle. A thirteenth set of
samples, consisting of approximately 30g of soil, was also prepared. This set
was placed into 1 L glass canning jars for trapping of C02 for measurement of
microbial biomass. All samples were placed in a controlled environment
chamber which remained at 24°C, and the samples were not exposed to light.
Twelve sample times of 0, 1, 3, 10, 20, 30, 50, 70, 90, 120, 150, and 180

days after treatment (DAT) were designated, and at each of these times four
124

125

soil samples for each depth were removed for analysis. In addition, at each
sample time, base traps of 2 ml of 2N NaOH, contained in glass scintillation
vials, were placed into the glass jars. At each subsequent sample date the traps
in the jars were removed and capped and new traps replaced them.

The beakers containing the soil samples were removed from specimen
cups and placed in an oven at 65°C for 48 hr to dry the soil. Soil was then
extracted with 1N KCl (5:1 v/w) and analyzed for NO3'-N and NH4+-N by
flow injection analysis with a Lachat QuikChem autoanalyzer (Lachat
Instruments, Milwaukee, WI). This was followed by diffusion and trapping of
N (Brooks, er al., 19 ) for measurement of 15N enrichment of each N species.
Separate diffusions were conducted for NH4+ and N03” due to the different
volumes needed to yield appropriate amounts of total N for analysis.
Enrichment was measured with a Europa Scientific TracerMass mass
spectrometer (Europa Scientific USA, Cincinnati, OH). Base traps were
titrated with 1 N HCl. All 15N data is reported after correction for background

15N in untreated soil.

i r i i n- m 'liz ion-T rn v r MIT Ex rim

This 30 day incubation experiment utilized the 0-5 cm increment of the
soil described above. Samples were prepared by the same procedure and
sample dates were identical. In this study, three treatments based on addition
of supplemental N prior to incubation, were used. The first treatment (check)
received no supplemental N. The same experimental units were utilized for this
treatment and the corresponding samples in the LTI study. The second
treatment (NH4+-N) consisted of addition of 18.5 pg/g of supplemental N as
(NH4)ZSO4. This rate was equivalent to a field application rate of 40 kg N

ha'l. The third treatment (Clipping N) received an equivalent amount of

126

supplemental N applied as Kentucky bluegrass grass clippings. Samples were

processed and analyzed as described for the LTI experiment.

RESULTS AND DISCUSSION

I _ 1 l . :1 III E r’

Samples collected at day 50 during the incubation had greatly elevated
levels of NH4+, approximately 12 and 30 times greater than surrounding
samples in the 0-5 and 5-10 cm depths, respectively. Concentrations of 15N in
these samples were not elevated. Contamination of these samples in the lab
following incubation was suspected, and they were not included in the analysis
for NH4+. Nitrate-N levels appeared to be consistent with surrounding
samples and were used in the analysis.

Cumulative concentrations of total and 15N-labeled NH4+ and NO3’ for
the 0-5 cm depth are shown in Figure 16. For the first 10 days there were few
changes in N levels in the soil. Rapid mineralization occurred during the 10 to
50 day period, as evidenced by increasing NO3' concentrations. After 50 days,
there was a slight decrease in both total and labeled NO3' concentrations.
Because these values are cumulative, this indicates that either previously
mineralized N was becoming immobilized, or that denitrification was occurring.
Use of parafilm on the beakers to retain moisture probably also precluded Oz
diffusion, resulting in anaerobiosis and subsequent denitrification after 50 days.
From 0 to 50 days, the net NO3‘ and 15N03' accumulations (final
concentration - initial concentration) were 33.5 and 0.128 pg g'l, respectively.
By 180 days these values had decreased to 29.3 and 0.113 pg g'l. On a field
area basis, this equates to mineralization of 23.5 kg N ha'1 during the first 50
days. Mineralized nitrogen from the LFN applied two years previously totaled
0.36 kg N ha’l. Concentrations of NH4+-N remained relatively stable for the
first thirty days, then declined for the remainder of the incubation. Over the

course of the experiment, NH4+-N decreased from 11.6 to 3.2 pg g'l, and
127

128

(3/3n)vHN91 pu13 £ONSI

.5333... >3 ow. a «5:6
:53 Eu m - o 2.. 8o... ..8 c. 32.8.5280 2-83.888.“ 23 2.3.3.2 .o. 2:2".

5333:. ..o 3.5

 

 

 

 

 

 

ooN om. oo. om o
o _
«mm
N
m
..o. W
p
N
H
.7
N o 1 m
crux
r on
«em—Am. VA #32 :1 mOZm. 0 mOZ .
m.o oo

 

129

15NH4+ decreased from 0.026 to 0.004 pg g'l. This coincided with the
decrease in N03‘ concentrations and indicates nitrification followed by
denitrification.

Figure 17 shows that mineralization rates had peaked by 30 days in the
5-10 cm depth, as compared to a 50 day mineralization flush in the upper
depth. Decreases in both NO3’ and 15N03' between 30 and 50 days again
indicates probable denitrification. Additional accumulations of NO3' continued
from days 50 through 180. Net NO3’ accumulations were 10.0 and 18.4 pg g'1
at 30 and 180 days, respectively. Associated 15N03' levels were 0.019 and
0.036 pg g'l, respectively. On a field area basis this equates to 8.1 and 14.8
kg N ha-1 over 30 and 180 days. or the labeled fertilizer N applied, 0.060 and
0.116 kg N ha"1 were mineralized. The pattern of nitrification of NH4+ was
similar to that in the upper depth. Concentrations of total NH4+ decreased
from 5.3 pg g"1 to 2.4 pg g'1 over the entire incubation, while 15NH4+
decreased from 0.004 to 0.001 pg g'l.

In the upper depth, total NO3‘ increased by a factor of 6 over 180 days,
while 15N03' increased by a factor of 9. In the lower depth, these ratios were
3 and 4 respectively. The 15N was in a more labile fraction as compared to
total N. This is in agreement with data from Chapter I (Table 13), which
shows 6% of total N and 23% of 15N present in microbial biomass in the 0-5
cm depth. Long-term incubation of soil from the 5-10 cm depth yielded nearly
equal amounts of net NO3' and 15NO3' accumulation during the 0 to 30 day
and 30 to 180 day periods. This is in contrast to data from the 0-5 cm depth,
in which net immobilization occurred after 50 days. The recalcitrant fraction
of organic matter in the lower depth is apparently more available than the
recalcitrant fraction in the upper depth. This data is inconclusive, however,

because denitrification probably masked the true mineralization rates.

130

(3/311) VHNSI pun £01091

dong—=05 an“. c3 a war—.6

:32. :8 o— 1 m 2: 59a :8 5 3335588 2183.568.“ can 2-28:2 .5 053m

553:2: we when
2:

 

mmod 1

mod 1

mvcd

 

 

 

 

E22 X

 

viz 1‘1 mOZE o.

mOZ .

(3/3“)17HN DUB EON

 

 

O
v

131

Net mineralization of soil organic material in this soil resulted in the
release of 41.6 kg N ha'1 in the first 50 days of incubation. During this period
0.40 kg N ha"1 of the LFN applied two years previously was mineralized. This
was equivalent to 1.7% and 5.6% of the total N and fertilizer N, respectively,
present in the soil. The fertilizer N mineralized also represents approximately
1% of the amount of LFN applied. Typical fertilizer N applications to turf are
in the range of 24.5 to 49 kg N ha'1 month‘l, and so a contribution of 25 kg

N ha'1 month'1 from mineralization, as seen in this experiment, would be a

significant source of N relative to fertilizer N additions.

 

Total concentrations of (NO3' + NH4+) extracted from soil for each
treatment at each time are shown in Figure 18. The curves were corrected for
the amount of inorganic N present at time 0 (after addition of supplemental N).
The resultant curves show the total inorganic N released through mineralization
during the incubation. In the check treatment there was a lag in mineralization
for the first 10 days. Between 10 and 30 DAT 13.9 pg g'1 of total N was
mineralized, resulting in net mineralization of 11.2 pg g'1 over 30 days. This
was equivalent to 8.1 kg N ha'1 on a field area basis. Addition of NH4+-N
stimulated N turnover almost immediately. During the 30 day incubation 41.6
pg g'1 of inorganic N accumulated, equivalent to 30.1 kg N ha'l. Response to
addition of grass clippings was intermediate to the other two treatments. A
total of 23.1 pg g'1 or 16.7 kg ha'1 of inorganic N was released during the
incubation. Addition of supplemental N stimulated mineralization of soil N as
compared to the unamended check. The response to NH4+-N was greater than

the response to organic N due to the high availability of the NH4+-N source.

132

cm

£22335
>2. cm a 9:26 38 E cog—assoc“ 2153:6885 + 2-28:: .58. .3 2:32

 

 

 

 

cm 2 S m o
_ _ _ _ OM-
1 .0: N
..... .. .. 1:12‘ 0
‘131 A: Co
1“! +
- 9
a m
.V
- em 4%
Go
/
arex
2.336 1T zémz mw 3.8.6 4

 

 

cm

133

Total concentrations of 15N03'-N accumulated during the 30 day
incubations are shown in Figure 19. With no supplemental N added, a net of
0.051 pg g’1 of 15NO3'-N was released by mineralization. This represents
2.3% of the total organic 15N in the soil at the beginning of the incubation and
0.14 kg ha'1 of the LFN applied two years previously. Figure 20 presents the
l5N03"-N released through mineralization from the amended soils corrected for
the amount mineralized from the check treatment at each time, resulting in net
15N mineralization due to addition of supplemental N. More 15NO3'
accumulated in N-amended soils than in the check treatment over the 30 day
incubation. Following supplemental NH4+-N an additional 0.065 pg g'1 of
15N03'-N was released through mineralization compared to the check.
Following addition of grass clippings the additional l5N03’-N mineralized was
0.023 pg g'l. This additional mineralization of labeled soil N in response to N
addition compared to the check treatment indicated a priming effect on soil
microbial activity. The priming effects of NH4+-N and clipping-N resulted in
additional mineralization of 2.9% and 1.0% of the soil 15N , respectively. On a
field area basis total mineralization of the LFN applied two years previously
was 0.32 kg ha'1 and 0.23 kg ha'1 in the NH4+-N and clipping-N amended

soils, respectively.

Summary

These experiments indicate that soil organic matter can be a substantial
source of N in soils with a turf cover. Inorganic N is made available through
mineralization processes, and this effect is stimulated by addition of
supplemental N through either synthetic fertilizers or plant materials. Nitrogen
from a single fertilizer application two years prior to the incubation

experiments was found to be mineralized at faster rates as compared to the bulk

134

£03235 .33

cm a 93:6 “:8 33:25-2 35 33:25:: :83 22 he 53332058 38H. .2 9:33

:o333E 30 $35

 

Om mm ON 3 S w o
— — r _ _ c
- ........ \‘
...... ‘ 1 no.0 m
Cu
\h \W
.3
1 2.0 an
(

 

2-2-:2 .mw 2-336 1T 3.85 4

 

 

Ed

135

.285 33:25:: :83 5333805.: 3.3 32823 £63335 .33 on
a $353 335 33:25 2-3.3830 35w 21+v22 :83 22 .8 33333052 .ou 2222

3633205 we when

 

 

on R 8 2 2 n o
7 _ 2 _ _ “Ce?
I
S
N
m
3.0 o W
- .- m
.- m.
44\ - 8... m-
m m
aw
2.336 1T 2.22 mw- mm

 

 

 

2.3

136

soil N, indicating its presence in the more labile fraction of soil organic matter.
Nitrogen was mineralized most rapidly during the first 30 to 50 days of these
experiments, but continued throughout the 180 day incubations. Anaerobic
conditions and denitrification my have occurred after 50 days in the long term
incubation experiment.

These experiments indicate that the benefits from fertilizer application
are in excess to the amount of N applied. Although applied N may be quickly
immobilized, this stimulates release of organic N in excess of the amount
applied. This is an especially important consideration in clipping management.
Return of clippings to turf not only alleviates the disposal problem, but

enhances microbial activity and associated nitrogen and carbon cycling.

APPENDIX

APPENDIX

.30? 0.5 .83 338 can Eomtoa 33:3 28 338mm»— mo ESwSQ AN 0.5me

 

 

TIE mall

 

 

1|lll‘

'IIIIL

‘llll‘

rIIIIL

 

 

 

 

 

 

 

8%

 

 

muons

 

E WC

83222 _HIIIIU_

 

_
_ =8

 

 

 

Ea

$80853 .

 

z|v

137

138

.zom cosh—3:. .3 32a

can £83835 £8259: me 20:82 wig—2: no.3 83 e825». .3 a“: .nu 23E

E mm;

EON

 

 

 

J_

 

 

. I... .
«an.

 

ENm;

 

 

 

 

100

139

 

Q Water, Lys 1 -+- Water, Lys 2

—l— Br, Lys 2

9 Br, Lys 1

 

 

 

 

l l I

§ § §

(8111) .13 eApelnumg

6

4
Hours After Treatment

Figure 23. Breakthrough curves for water and bromide applied over surface of Lysimeters 1

and 2 under conditions of saturated flow.

140

.Bec c2822 mo $832.8 .825 N can _ E20533
we mow—co 8:85 2:63 tom—mam 028:3 can .833 he 82:0 smack—.335 .vm 9:.me

E2585. cot< @503

 

 

o m w m N _ o
O p _ L p _ o
w
2 L 8m
3
n m
m. cm, - 82 w
e m.
U. B
A n...
a on. - 82 M.
m
.. m
m 91 ,. coca )
1 m
mu. mm
II\
on- 1 83
N .3 as; ...+.... :3 as; o N ”3 .6 ll :3 .6 l...

 

 

 

8
§

141

.30: @2833 we 82:23 825 v 23

m 320534 no 38:5 B>o com—qua 02595 can 833 3.. noise smack—33cm .mm 235

('1) .1919 AA eApelnumg

E253; by? @502

 

w o v N o
_ L _ “n x t
o \‘fl‘fl‘u G H. «T L1
\s\\\ .0.0 Q +
\ ..O. .+ .1
DNA . ....0 4-4 I
\ .nU
, .0. .....+
\ o + -
we .0. +
$1 \ .0.
H0 ..+
.. .. 0...... ......
8 1 O... .. O u ..+._+
\\ ...IT... 1
8L .
v 93 £855 ......+ m 23 .333 .....o. v r3 ...m ll m 93 ...m |¢|
o2

 

 

 

8 8 3 $3:
(8) .13 oApelnumg

§

GQ

142

.320 @2838 he 2532.8 .525 v can m 2208?»;
ac woman 020an 2:5; @233 02320 was .833 .80 «2:3 £98233on .3 83E

ufiOEuflQhH. h0¢< mhflcm

 

 

 

 

w e .v N o
O _ _ 0 4m. .... O
0...”. +...-1:
...-0... .+.
0-0.0 ...-......
G .+..+ c I 8%
3 Cal 0. ._+ o 0
m .. .. .. 82 n
n _ o. .. w
I O. + n
e e. - . .. I
m. 0.0 i. 0 1 8W— m.
m . a
0.... .. . y T
Du 81 O ....+ + 0 §N D
m ...T.- -....T-.. ....+.. .. \ s; \w}
) .+. o x I 8WN 8
n1 - u {x
I..\ Ow 1 a u -
r econ
V mhw— Juan; 3+ ..... M mhxu £033 ...0.-.. V 9A1— .—U |..T| m 951— .—U l0|
02 8mm

143

 

ooood
wooed
@306
$36
«a: .o

www.md
:hmd

mead
moo—.0
va—R

ooood
:36
aged
:mcd
3:6

Nammd
momvd

mvvmd
hmmmd

NQoQ:

88.0
020.0
momod
owned
w: _ .o

mmwmd
owood

ohwmd
momvd
NQon

Naod
mzod
wooed
wmmod
wowed

35nd
aoamd

hvowd
oahmd

33m:—

good
@806
mmood
aged
Nmbod

vwomd
Shag

awvh
>056
_o\v_\m

Euo©-ov
Eoov-om
Eoom-o_
:50— -m

EU m I O
58

=8

.8qu Bags
£85.;
oSEo>

mwcfific

loohzom

 

.2553: macaw 2: ho.“ 83.. 29:8 3828 3 E2565 mEEEam some 5 Z: «888 a\. :82 Km 03¢

144

 

ooocd
wvmcd
mm-Od
hovod
m: _ .o

mwomd
voamd

cm: .0
36— .o

onQ Z

0256
2590
filed
wowed
N02 .0

etymd
hobvd

omhmd
ammvd

va—R

ooood
wooed
030.0
886
2 _ _ .o

3.906
336

omomd
wmcmd

onQ:

>086
good
good
Smod
ammod

mmnmd
wa—od

vwomd
waved

NQBNR

386
5290
«wood
38.0
wmood

gum—d
3»va

Mamba

SEQ:

Eoocuov
Eoovuou
Eoom-o_

Eco—-m

Eom-o
U=cm

:8

.33.: 35MB
£89;

oSEo>
35%—=0

lulluuoSom

 

45830: :5 2: 8m 82% 29:3 3828 3 308205 mEEEam some 5 Z: 9,888 o\.. ES< .3 03¢.

LIST OF REFERENCES

LIST OF REFERENCES

Beard, J. B. 1973. Turfgrass: Science and culture. Prentice-Hall,
Englewood Cliffs, NJ.

Bonde, T.A., J. Schnfirer, and T. Rosswall. 1988. Microbial biomass as a
fraction of potentially mineralizable nitrogen in soils from long-term
field experiments. Soil Biol. Biochem. 20:447-452.

Bowman, D.C., J .L. Paul, and W.B. Davis. 1989a. Nitrate and ammonium
uptake by nitrogen-deficient perennial ryegrass and kentucky bluegrass
turf. J. Amer. Soc. Hort. Sci. 114:421-426.

Bowman, D.C., J.L. Paul, W.B. Davis, and S.H. Nelson. 1987. Reducing
ammonia volatilization from kentucky bluegrass turf by irrigation.
HortScience 22:84-87.

Bowman, D.C., J.L. Paul, W.B. Davis, and S.H. Nelson. 1989b. Rapid
depletion of nitrogen applied to kentucky bluegrass turf. J. Amer. Soc.
Hort. Sci. 114:229-233.

Boyle, M., and E.A. Paul. 1989. Carbon and nitrogen mineralization kinetics
in soil previously amended with sewage sludge. Soil Sci. Soc. Am. J.
53:99-103.

Bredakis, E.J., and J .E. Steckel. 1963. Leachable nitrogen from soils
incubated with turfgrass fertilizers. Agron. J. 55:145-147.

Bremner, J .M. 1965. Nitrogen availability indeces. In C.A. Black et al. (ed.)
Methods of soil analysis, part 2. Agronomy 9:1324-1345.

Bristow, A.W., J .C. R den, and DC. Whitehead. 1987. The fate at several
time intervals of 1 N-labelled ammonium nitrate applied to an established
grass sward. J. of Soil Sci. 38:245-254.

Brown, K.W., R.L. Duble, and J .C. Thomas. 1977. Influence of management
and season on fate of N applied to golf greens. Agron. J. 69:667-671.

145

146

Brown, K.W., J .C. Thomas, and R.L. Duble. 1982. Nitrogen source effect on
nitrate and ammonium leaching and runoff losses from greens. Agron. J.
74:947-950.

Chalk, P.M., and D.R. Keeney. 1975. Fate of 15N-labeled anhydrous
ammonia under simulated fall and spring conditions. Agron. J. 67:41-45.

Chichester, F.W. 1977. Effects of increased fertilizer rates on nitrogen
content of runoff and percolate from monolith lysimeters. J. Environ.
Qual. 6:211-217.

Deans, J .R., J .A.E. Molina, and C.E. Clapp. 1986. Models for predicting
potentially mineralizable nitrogen and descomposition rate constants.
Soil Sci. Soc. Am. J. 50:323-326.

Dowdell, R.J., and C.P. Webster. 1980. A lysimeter study using nitrogen-15
on the uptake of fertilizer nitrogen by perennial ryegrass swards and losses
by leaching. J. of Soil Sci. 31:65-75.

Dowdell, R.J., and C.P. Webster. 1984. Effect of drought and irrigation on
the fate of nitrogen applied to cut permanent grass swards in lysimeters:
experimental design and crop uptake. J. Sci. Food Agric. 35 :1092-1 104.

Geron, C.A., T.K. Daanneberger, S.J Traina, T.J. Logan, and J.R.Street.
1993. The effects of establishment methods and fertilization practices on
nitrate leaching from turfgrass. J. Environ. Qual. 22:119-125.

Hanson, A.A., and F.V. Juska. 1961. Winter root activity in kentucky
bluegrass (Poa pratensis L.). Agron. J. 53:372-374.

Hurto, K.A., A.J. Turgeon, and LA. Spomer. 1980. Physical characteristics
of thatch as a turfgrass growing medium. -Agron. J. 72:165-167.

Jones, M.B., C.C. Delwiche, and W.A. Williams. 1977. Uptake and losses of
15N appplied to annual grass and clover in lysimeters. Agron. J. 69:1019-
1023.

Juma, N .G., and EA. Paul. 1984. Mineralizable soil nitrogen: amounts and
extractability ratios. Soil Sci. Soc. Am. J. 48:76-80.

Kissel, DE, and SJ. Smith. 1978. Fate of fertilizer nitrate applied to coastal
bermudagrass on a swelling clay soil. Soil Sci. Soc. Am. J. 42:77-80.

Ledeboer, F.B., and C.R. Skogley. 1967. Investigations into the nature of
thatch and methods for its decomposition. Agron. J. 59:320-323.

147

Low, A.J., and E.R. Armitage. 1970. The composition of the leachate
through cropped and uncropped soils in lysimeters compared with that of
the rain. Plant Soil 33:393-411.

Mancino, Charles F., M. Barakat, and A. Maricic. 1993. Soil and thatch
microbial populations in an 80% sand:20% peat creeping bentgrass putting
green. HortScience 28: 189-191.

Mancino, C.P., W.A. Torello, and DJ. Wehner. 1988. Denitrification losses
from kentucky bluegrass sod. Agron. J. 80:148-153.

Martin, J .P., and H.D. Chapman. 1951. Volatilization from ammonia from
surface-fertilized soils. Soil Sci. 71:25-34.

Marumoto, T., J .P.E. Anderson, and K.H. Domsch. 1982. Mineralization of
nutrients from soil microbial biomass. Soil Biol. Biochem. 14:469-475.

Mitchell, W.B., A.L. Morehart, L.J. Cotnoir, B.B. Hesseltine, and D.N.
Langston, III. 1978. Effect of soil mixtures and irrigation methods on
leaching of N in golf greens. Agron. J. 70:29-35.

Michigan Turfgrass Foundation. 1989. 1988 Michigan turfgrass industry
report. Michigan State University Cooperative Extension Service, East
Lansing, MI.

Molina, J .A.E., C.E. Clapp, and W.B. Larson. 1980. Potentially
mineralizable nitrogen in soil: the simple exponential model does not

apply for the first 12 weeks of incubation. Soil Sci. Soc. Am. J.
44:442-443.

Morton, T.G., A.J. Gold, and W.M. Sullivan. 1988. Influence of
overwatering on nitrogen losses from home lawns. J. Environ. Qual.
17:124-130.

Nelson, K.E., A.J. Turgeon, and J .R. Street. 1980. Thatch influence on
mobility and transformation of nitrogen carriers applied to turf. Agron. J.
72:487-492.

Nordmeyer, N., and J. Richter. 1985. Incubation experiments on nitrogen
mineralization in loess and sandy soils. Plant Soil 83:433-445.

Overrein, L.N., and P.G. Moe. 1967. Factors affecting urea hydrolysis and
ammonia volatilization in soil. Soil Sci. Soc. Amer. Proc. 31:57-60.

148

Paul, E.A., and F.E. Clark. 1989. Soil Microbiology and Biochemistry.
Academic Press, San Diego, CA.

Petrovic, A.M. 1990. The fate of nitrogenous fertilizers applied to turfgrass.
J. Environ. Qual. 19:1-14.

Powell, A.J., R.E. Blaser, and R.E Schmidt. 1967. Physiological and color
aspects of turfgrasses with fall and winter nitrogen. Agron. J.
59:303-307.

Power, J .F., and J .0. Legg. 1984. Nitrogen-15 recovery for five years after
application of ammonnium nitrate to crested wheatgrass. Soil Sci. Soc.
Am. J. 48:322-326.

Rekhi, R.S., O.P. Meelu, and R.K Gupta. 1982. Lysimeter studies on
recovery of 15N-labeled urea in wetland rice. Plant Soil. 66:57-67.

Rieke, P.E., and B.G. Ellis. 1974. Effects of nitrogen fertilization on nitrate
movements under turfgrass. p. 120-130. In E.C. Roberts (ed.) Proc. 2nd
Int. Turfgrass Res. Conf. ASA, Madison, WI. 19 - 21 June 1972.
Blacksburg, VA.

Simpson, D.M.H., and S.W. Melsted. 1962. Gaseous ammonia losses from
urea solutions applied as a foliar spray to various grass sods. Soil Sci.
Soc. Amer. Proc. 26:186-190.

Smith, J.L., R.R. Schnabel, B.L. McLean, and G.S. Campbell. 1980.
Potential errors in the first-order model for estimating soil nitrogen
mineralization potentials. Soil Sci. Soc. Am. J. 44:996-1000.

Snyder, G.H., B.J. Augustin, and J .M. Davidson. 1984. Moisture sensor-
controlled irrigation for reducing N leaching in bermudagrass turf. Agron.

J. 76:964-969.

Snyder, G.H., E.O. Burt, and J.M. Davidson. 1980. Nitrogen leaching in
bermudagrass turf; daily fertigation vs. tri-weekly conventional
fertilization. p. 185-193. In Proc. 3rd Int. Turfgrass Res. Conf. ASA,
Madison, WI.

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

Starr, LL, and H.C. DeRoo. 1981. The fate of nitrogen fertilizer applied to
turfgrass. Crop Science. 21:531-536.

149

Thompson, D.C., R.W. Smiley, and M. C. Fowler. 1983. Oxidation status
and gas composition of wet turfgrass thatch and soil. Agron. J.
75:603-609.

Tisdale, S. L., and W. L. Nelson. 1975. Soil fertility and fertilizers. 3rd ed.
Macmillan Publishing Co., Inc., New York, NY.

Titko, S. III, J .R. Street, and T.J. Logan. 1987. Volatilization of ammonia
from granular and dissolved urea applied to turfgrass. Agron. J.

79:535-540.

Torello, W.A., and DJ. Wehner. 1983. Urease activity in a kentucky
bluegrass turf. Agron. J. 75:654-656.

Torello, W.A., D.J. Wehner, and A.J. Turgeon. 1983. Ammonia
volatilization from fertilized turfgrass stands. Agron. J. 75:454-456.

Turgeon, A.J. 1991. Turfgrass Management. 3rd. ed. Prentice-Hall,
Englewood Cliffs, NJ.

U.S. Environmental Protection Agency. 1976. Quality criteria for water. U.S.
Gov. Print. Office, Washington, DC.

Volk, G.M. 1959. Volatile loss of ammonia following surface application of
urea to turf or bare soils. Agron. J. 51:746-749.

Volk,. G.M. 1966. Efficiency of fertilizer urea as affected by method of
application, soil moisture, and lime. Agron. J. 58:249-252.

Webster, C. P., and R.J. Dowdell. 1984. Effect of drought and irrigation on
the fate of nitrogen applied to cut permanent grass swards in lysimeters:
leaching losses. J. Sci. Food Agric. 35:1105-1111.

Weil, R.R., R.A.Weismiller, and R.S. Turner. 1990. Nitrate contamination

of groundwater under irrigated coastal plain soils. J. Environ. Qual.
19:441-448.

Wesely, R.W., R.C. Shearman, B.J. Kinbacher, and S.R. Lowry. 1987.
Ammonia volatilization from foliar-applied urea on field-grown kentucky
bluegrass. HortScience 22: 1278-1280.

Wilkinson, J. F., and D. T. Duff. 1972. Effects of fall fertilization on cold
resistance, color, and growth of kentucky bluegrass. Agron. J.
64:345-348.