EPA-600/3 76-042
April 1976
Ecological Research Series
NITRATE REMOVAL FROM WATER AT THE
WATER-MUD INTERFACE IN WETLANDS
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal
species, and materials. Problems are assessed for their long- and short-term
influences. Investigations include formation, transport, and pathway studies to
determine the fate of pollutants and their effects. This work provides the technical
basis for setting standards to minimize undesirable changes in living organisms
in the aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-76-042
April 1976
NITRATE REMOVAL FROM WATER
AT THE WATER-MUD INTERFACE IN WETLANDS
by
W.H. Patrick, Jr., R.D. Delaune, R.M. Engler and S. Gotoh
Louisiana Agricultural Experiment Station
Louisiana State University
Baton Rouge, Louisiana 70803
Grant No. 800428
Project Officer
William Sanville
Marine and Freshwater Ecology Branch
Corvallis Environmental Research Laboratory
Corvallis, Oregon 97330
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
CORVALLIS ENVIRONMENTAL RESEARCH LABORATORY
CORVALLIS, OREGON 97330
-------
DISCLAIMER
This report has been reviewed by the U.S. Environmental Protection
Agency, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names of commercial products
constitute endorsement or recommendation for use.
ii
-------
ABSTRACT
I. The oxidized and reduced layers in flooded soil were character-
ized by vertical distribution of the oxidation-reduction (redox) po-
tential and concentrations of manganous manganese, ferrous iron, sulfide,
nitrate, and ammonium. Redox potential was measured with a special
motor-driven assembly which advanced a platinum electrode at a rate
of 2 mm/hour through the flooded soil profile. Vertical distribution
of reduced forms of manganese, iron, and sulfur and of nitrate and am-
monium was determined by freezing and slicing the flooded soil into
segments 1 or 2 mm thick. The apparent thickness of the oxidized layer
was different when evaluated by the distribution of the various com-
ponents in the profile, with the sulfide profile indicating the thickest
oxidized zone, the manganese profile indicating the thinnest oxidized
zone, and the iron profile showing an intermediate thickness. The thick-
ness of the oxidized layer increased with duration of flooding.
II. The floodwater NO^ removal rate of intermittently-flooded fresh
water swamp soils and continuously-flooded saline marsh soils of south-
ern Louisiana was quantitatively characterized in a laboratory study.
Of the two areas studied, the marsh area was the more effective sink
for N(>3 contaminated waters with an average initial removal rate of 9.15
ppm N/day. After correcting for the rate of NO^ diffusion, the micro-
bial NC>3 removal rate was calculated to be 7.64 ppm N/day. The swamp
soil had a removal rate of 4.38 ppm N/day. The tnicrobial NO^ removal
rate for this area, after correcting for diffusion, was 2.50 ppm N/day.
Studies on samples of floodwater separated from the soil showed the ac-
tive site of microbial NOg reduction to be the soil-water interface or
within the soil, but not in the floodwater. Additions of organic matter
to a mineral soil flooded for rice (Oryza sativa L.) culture decreased
the thickness of the aerobic-anaerobic zone at the soil-water interface
and increased the rate of NOg reduction.
III. The concentration of atmospheric 0? over a flooded soil is a factor
in determining the amount of N lost by denitrification. Large increases
in N loss occurred from the first few increments of oxygen with little
further loss occurring above 20 percent 02« The thickness of the aerobic
layer was also governed by the amount of $2 in the air. Nitrogen loss
was generally related to the thickness of the aerobic layer, even though
appreciable loss occurred at 5 and 10 percent 0_ where the aerobic layer
was relatively thin.
IV. Ammonium nitrogen in a flopded soil or sediment exposed to oxygen
from the water column undergoes sequential nitrification and denitrifi-
cation. Ammonium in the aerobic surface layer of soil or sediment is
iii
-------
nitrified and the resulting concentration gradient between this layer
and the underlying anaerobic layer causes ammonium to diffuse upward
into the aerobic layer where it also undergoes nitrification. Nitrate
produced in the aerobic layer then diffuses down into the anaerobic
layer where it is denitrified to No and ^0. Approximately one-half
of the nitrogen involved in the nitrification-denitrification process
is ammonium originally present in the surface aerobic soil or sediment
layer with the remainder diffusing up from the underlying anaerobic
layer. If oxygen is absent or limiting, nitrification will either not
occur or will occur at a lower rate, resulting in a reduced amount of
nitrate available for the denitrification process.
V. The ©2 reduction rates, NO^ reduction rates, and the effects of
added 0« on NO, reduction and redox potential in four flooded or inter-
mittently flooded soils from the swamp and coastal marshes of Louisiana
were quantitatively characterized in a laboratory study. The NO^ added
either to the shallow floodwater or mixed with the soil in a suspension
rapidly disappeared. Eighty to ninety ppm NO^ was reduced and disap-
peared from the soil suspensions in 1 to 4 days and from the floodwater
over a soil in 10 to 20 days. No NO^ was lost from floodwater removed
from the soils. Oxygen depletion in the soil suspensions occurred in
15 minutes to 4 hours. Redox potential curves exhibited a characteris-
tic inflection after 02 disappearance in all soils studied. Nitrate
reduction did not appear to be inhibited by as much as 16 ppm 0~ dis-
solved in the soil suspensions.
iv
-------
CONTENTS
Page
Abstract ill
List of Figures vi
List of Tables viii
Sections
I Introduction 1
II Characterization of the Oxidized and Reduced Zones
in Flooded Soil 6
III Nitrate Removal from Floodwater Overlying Flooded
Soils and Sediments 23
IV The Role of Oxygen in Nitrogen Loss from Flooded
Soils 37
V Nitrification-Denitrification Reactions in Flooded
Soils and Sediments: Dependence of Oxygen Supply
and Ammonium Diffusion . . 44
VI Effect of Dissolved Oxygen on Redox Potential and
Nitrate Reduction in Flooded Swamp and Marsh Soils 54
VII References . . . 78
-------
FIGURES
No. Page
1 Oxidized and reduced layers of a flooded soil 2
2 Apparatus used for measuring redox potential
profiles in flooded soils 9
3 Oxidation-reduction (redox) potential profiles in
flooded Crowley silt loam at various times after
flooding 10
2+
4 Vertical distribution of reduced manganese (Mn )
in a flooded Crowley silt loam at various times
after flooding 14
5 Vertical distribution of reduced iron (Fe2+) in a
flooded Crowley silt loam at various times after
flooding 15
6 Vertical distribution of reduced sulfide (S2~) in
a flooded Crowley silt loam at various times after
flooding 16
7 Vertical distribution of ammonium in a flooded
Crowley silt loam at various times after flooding 18
8 Vertical distribution of nitrate in a flooded
Crowley silt loam at various times after flooding 19
9 Thickness of the surface oxidized layer as measured
by the manganese, iron, sulfur, and redox potential
profiles after 13 weeks flooding 21
10 Floodwater NOZ removal in relatively undisturbed
cores of a fresh water swamp soil as affected by
diffusion and reduction 28
11 Floodwater N0~ removal in relatively undisturbed
cores of a salt water marsh soil as affected by
diffusion and reduction 28
12 The effects of additions of rice straw on the de-
velopment of redox potential profiles in a flooded
riceland soil 33
vi
-------
Figures - continued.
No. Page
13 The effects of additions of rice straw on the re-
moval of floodwater NO" by diffusion and reduction 33
14 Loss of total N (labeled + unlabeled) from flooded
soils as affected by 0 content of air 39
15 Loss of labeled N from flooded soil as affected by
02 content of air 40
16 Effect of 02 content of air on thickness of aerobic
soil layer after 90 days' incubation 40
17 Processes involved in the sequential conversion of
organic nitrogen to elemental nitrogen in flooded
soils and sediments 46
18a The conversion of N-labeled ammonium to elemental
nitrogen in a flooded soil. The distribution of la-
beled ammonium among various nitrogen fractions fol-
lowing incubation under 30% oxygen 49
18b The conversion of -^N-labeled ammonium to elemental
nitrogen in a flooded soil. The effect of oxygen
content on conversion of labeled ammonium to various
nitrogen fractions 50
19 Total oxygen consumption, oxygen converted to ni-
trate, and oxygen converted sequentially to nitrate
and then to carbon dioxide as affected by time of
incubation and oxygen concentration 53
20 Apparatus used to study nitrate reduction, oxygen
consumption and redox potential in stirred suspen-
sions of sediment 59
21 Changes in oxygen content and redox potential with
time in stirred suspensions 62
22 Changes in nitrate content with time in stirred sus-
pensions 64
23 Nitrate removal following repeated additions of ni-
trate to the floodwater overlying flooded soil
66
24 Nitrate removal from floodwater in contact with
soil surface and removal from floodwater separated
from soil 68-69
25 Effect of added oxygen on nitrate reduction and redox
potential in stirred suspensions 71-75
vii
-------
TABLES
No. Page
1 Quadratic regression equations and correlation coef-
ficients representing the rates of nitrate removal from
the floodwater of fresh water swamp soils and salt water
marsh soils from the Bayou Sorrel area of the Mississippi
River alluvium and Barataria Bay area of Louisiana 29-30
2 Quadratic regression equations and correlation coef-
ficients representing the rates of nitrate removal from
the floodwater of flooded Crowley silt loam soils 35
3 Nitrogen content of flooded soil in ppm after 120 days'
incubation with and without addition of 0.5 percent
rice straw (four replications) 42
4 Description of soils used in study 57
5 Certain chemical properties of soils used in this study 58
viii
-------
SECTION I
INTRODUCTION
Flooding causes marked changes in some of the chemical and biologi-
cal processes taking place in the soil. These changes are caused by the
curtailment of atmospheric oxygen entering the soil. In drained soils
there are enough air-filled pores to permit adequate diffusion of atmos-
pheric oxygen into the soil. When these soil pores are filled with water
and the soil surface is covered with several inches of water, however, the
amount of oxygen reaching the soil surface is reduced by a factor of over
10,000. Even though the supply of oxygen reaching the soil is greatly
curtailed by flooding, microbial respiration is not diminished. The im-
balance between the oxygen requirement of the soil and the oxygen supply
reaching the soil surface results in a rapid depletion of soil oxygen with-
in hours after flooding. The only part of a flooded soil that is oxygen-
ated is a thin layer at the-soil surface since oxygen diffusing through
the flood water penetrates only a short distance into the soil before it
is consumed (see Figure 1). An important property of a flooded soil is the
thickness of this surface oxygenated layer. In soils that have a low mi-
crobial requirement for oxygen this layer is relatively thick, 2 to 5
centimeters, while in soils that have high oxygen requirement the layer
may be no thicker than 1 to 2 millimeters. The oxygen requirement is usu-
ally governed by the supply of decomposable organic matter.
In the surface oxygenated or aerobic soil layer, microbial and chemical
conditions are very much like those in the drained soil. In the underlying
oxygen-free soil layer pronounced biological and chemical changes are set
-------
Air
Water
Soil
Air
Water
Oxidized
Soil Layer
Reduced
Soil
Figure 1. (Top) A diagram showing the relative rate of oxygen movement through
the flood water (small arrow) and potential consumption rate of oxygen in the
soil (large arrow). Reduced conditions develop in the soil when the oxygen
supply through the flood water is not sufficient to meet the requirements for
oxygen in the flooded soil.
(Bottom) Differentiation of a waterlogged soil into a surface oxidized or
aerobic layer and an underlying reduced or anaerobic layer as a result of a
limited oxygen supply reaching the soil surface.
-------
in motion when oxygen disappears. As long as oxygen is present oxidized
components of the soil such as nitrate, manganic compounds and ferric com-
pounds are not biologically or chemically reduced. After oxygen disappears
following flooding, many soil microorganisms are able to substitute one
or more of these oxidized chemical components for the oxygen required in
respiration. Nitrate, the higher oxides of manganese, and hydrated ferric
oxide are reduced if oxygen becomes absent or limiting and if an energy
source (organic matter) is available to the microorganisms. Nitrate and
manganic compounds are readily reduced since the energy required for their
reduction is low and a number of species of microorganisms are capable of
carrying out this process. Ferric compounds are more difficult to reduce
but the large amount of reducible ferric iron in most soils makes ferric
compounds an important oxidation-reduction component of the soil. These
reactions are carried out by facultative anaerobes, microorganisms which
can substitute other reducible compounds for respiratory oxygen. Sulfate
can also be reduced to sulfide by anaerobic microorganisms but this reaction
is carried out only under strictly anaerobic conditions by a few species of
microo rganisms.
The difference in ease of reduction of the inorganic oxidation-reduction
systems in the soil results in a more or less sequential reduction of the
various components following flooding. Free oxygen in the soil is reduced
first and is at least partially depleted before nitrate and manganese com-
pounds begin to be reduced. Ferric compounds are reduced after nitrate and
manganic manganese and, if no oxygen enters the soil, sulfate will be re-
duced to sulfide. Almost always, however, all oxygen and nitrate have dis-
appeared from the soil before iron reduction begins. The availability of
-------
an energy source and the absence of oxygen are the only necessary require-
ments for these reduction reactions to occur, outside of favorable environ-
mental conditions, since the microorganisms that carry out these reactions
occur in almost all soils, even those where flooding has not been known to
take place.
Nitrogen reactions in flooded soils differ markedly from those of
well drained soils. In well drained soils decomposition of organic matter
releases ammonium nitrogen which is then oxidized biologically to nitrate.
In flooded soils ammonium nitrogen accumulates since the conversion of
ammonium to nitrate cannot take place in the absence of oxygen.
Not only is nitrate not produced in any part of a flooded soil except
the thin surface aerobic layer, nitrate added to a flooded soil is subject
to denitrification which results in loss of nitrogen from the soil as N«
or N.O.
Denitrification losses of nitrogen are undesirable in lowland rice
agriculture since nitrogen that might be taken up by the crop is lost from
the soil, however such losses from the soil-water system may be beneficial
in removing toxic levels of nitrate from water or soil. This process is
also extremely important in maintaining the nitrogen balance of the earth,
since denitrification in lakes, oceans, swamps, marshes, and both flooded
and nonflooded soils constitutes the major mechanism by which elemental ni-
trogen is returned to the atmosphere.
The objectives of the research carried out in this project were: (1)
to find out how rapidly, how completely, and by what mechanism nitrate is
removed by biological reduction from shallow surface water in swamps, marsh-
es and flooded soils, and (2) to determine the oxidation-reduction properties
-------
of the water-mud interface that control or influence the reduction of
nitrate to nitrogen gas.
-------
SECTION II
CHARACTERIZATION OF THE OXIDIZED AND
REDUCED ZONES IN FLOODED SOIL
The presence of a surface oxidized layer and an underlying reduced
layer at the surface of a flooded soil has important effects on biological
and chemical transformations in the soil. Microbiological processes tak-
ing place in the oxidized zone are similar to those carried out in drained,
aerated soil by aerobic microorganisms while the processes taking place
in the underlying oxygen-free reduced layer involve facultative and true
anaerobes. Most of the inorganic redox systems in soils occur in the oxi-
dized form in the surface oxygenated zone (ferric iron, manganic manganese,
nitrate nitrogen, sulfate sulfur) and in the reduced form in the underlying
oxygen-free zone (ferrous iron, manganous manganese, ammonium nitrogen, sul-
fide sulfur). For the purposes of this study the oxidized layer was con-
sidered to be the surface layer of soil containing little or no reduced
iron, manganese, and sulfur and the reduced layer was considered to be the
layer under the oxidized layer in which ferrous iron, manganous manganese,
and sulfide exist. Inorganic nitrogen is present as nitrate in the oxi-
dized layer and as ammonium in the reduced layer (Pearsall and Mortimer,
1939).
The differentiation of a flooded soil or sediment into two distinct
zones as a result of limited oxygen penetration into the soil material was
first described by Pearsall and Mortimer (1939) and first investigated
thoroughly by Mortimer (1941, 1942). The supply of oxygen at the soil sur-
face and the consumption rate of oxygen in the soil were recognized as the
-------
major factors determining the thickness of the oxidized surface layer.
The two layers existing at the surface of flooded soil can also be
characterized by differences in the oxidation-reduction or redox poten-
tial. The various inorganic and organic redox systems in the soil con-
tribute to this potential. Mortimer (1941, 1942) used a special assembly
of platinum electrodes to obtain a profile of redox potential in lake
muds. He found that high oxygen consumption in the mud was associated
with a thin oxidized layer and that low oxygen consumption was associated
with a thick oxidized layer. Gee (1950) and Alberda (1953) demonstrated
the presence of a thin oxidized layer underlain by a reduced layer in a
flooded soil profile.
Although the oxidized and reduced layers can be characterized by re-
dox potential profiles, the distribution of oxidized and reduced chemi-
cal components in these layers would provide better information on the
exact nature of the processes going on in the two layers of a flooded soil.
Pearsall and Martimer (1939) measured various redox components in the water
,*
above a lake bottom, in the oxidized surface mud layer, and in the under-
lying oxygen-free mud and found that the surface layer of mud contained
the oxidized forms of iron, inorganic nitrogen and sulfur while the under-
lying mud contained reduced forms of these elements. Howeler and Bouldin
(1971) measured the distribution of various forms of iron in a flooded pro-
file and showed that ferric iron was present in the oxidized layer and ab-
sent in the reduced layer while ferrous iron was absent in the oxidized
layer and present in the reduced layer.
The study reported here was designed (1) to develop a method for
measuring the oxidation-reduction or redox potential vertically through
-------
the soil profile in order to determine the thickness of the oxidized layer
and (2) to determine the distribution of the reduced form of reducible
soil components such as inorganic manganese, iron, sulfur, and the dis-
tribution of nitrate and ammonium nitrogen within the oxidized and reduced
layers.
MEASUREMENT OF REDOX POTENTIAL PROFILES IN FLOODED SOIL
A technique was developed by which a small platinum electrode (ap-
proximately 1 mm long) was driven downward through an undisturbed flooded
soil at a rate of 2 mm/hour. The electrode was constructed by sealing 18
gauge platinum wire into 5 mm diameter pyrex tubing with 1 mm of the wire
to complete the cell. The platinum and calomel electrodes were connected
to a vacuum tube voltmeter (Beckman Zeromatic pH meter) and the output from
the meter recorded on a Sargent MR recorder with a chart speed of 10.16
cm/hour (4 inches/hour). With this arrangement 2.54 cm (1 in) of chart
was equivalent to 0.5 cm depth. A sketch of the system is shown in Figure
2. The platinum electrode was positioned 5 ± 0.5 mm above the surface of
the soil at the beginning of each profile measurement.
By use of this system it was found in a laboratory experiment that
differentiation in the redox potential of a Crowley silt loam surface soil
developed rapidly after submergence, Figure 3. The redox profile measured
after only one day of flooding showed little change from the soil surface
downward because reducing conditions had not had time to set in, but a
week of submergence caused a marked change with depth. If it is assumed
that redox potential values greater than + 200 mV (at pH 7) denote oxi-
dized conditions and values less than this value denote reduced conditions
-------
Electric motor with _
low speed gear drive
!
Recorder
r
Millivolt
meter
Calomel
electrode j^j fez &•
Flooded
soil
pt
electrode
Figure 2. Apparatus used for measuring redox potential
profiles in flooded soils.
-------
i 10
i
««-«
Q.
5 .
"5
30
Floodwater
-100 0 100 200 300 400 500
Oxidation — reduction potential (mV)
Figure 3. Oxidation-reduction (redox) potential
profiles in flooded Crowley silt loam at
various times after flooding. Incubation
times shown are 1 day to 13 weeks.
10
-------
(Pearsall, 1938; Patrick and Mahapatra, 1968), it can be readily seen
that the depth of the surface oxidized layer increased with time of in-
cubation. For the sample shown the thickness of the oxidized layer in-
creased from approximately 6 mm at 2 weeks to about 12 mm after 13 weeks
incubation. This increase with time was apparently due to the decline
in the rate of organic matter decomposition which permitted oxygen from
the overlying flood water to penetrate deeper into the soil before being
reduced. All depths of the flooded soil had a pH value of approximately
6.8.
The method developed appears to provide a reliable way to character-
ize the oxidized and reduced layers of flooded soil in terms of the thick-
ness of the oxidized layer and the relative redox potential values in both
layers. The sharpness of the boundary can also be ascertained from the
redox potential curve. Most of the change in redox potential occurred with-
in a 8 mm zone. Even though the slow rate of advancement of the electrode
should have provided a reasonable time for the electrode to reach near-
,1
equilibrium conditions at any given depth, the slow negative drift of a
platinum electrode when placed in a biological reducing medium caused the
low redox potential values (those below about + 100 mV) to be slightly
higher than they would have been if the electrode had been left at a given
depth for several hours.
After obtaining redox potential-depth curves such as those shown in
Figure 3, the dividing line between the aerobic surface layer and the under-
lying anaerobic layer was estimated. The midpoint between the two points
of maximum change in slope was chosen. It is interesting to note that this
point occurred in all cases at approximately + 200 mV, a redox potential
11
-------
(at pH 7) usually considered to be in the transition zone between aerobic
and anaerobic soil conditions.
DISTRIBUTION OF INORGANIC REDUCIBLE SUBSTANCES IN FLOODED SOILS
The oxidized-reduced double layer was characterized by measuring the
vertical distribution of the reduced form of several inorganic components,
manganous manganese, ferrous iron, and sulfide. Reduced forms of manga-
nese and iron are relatively soluble and can be rather easily extracted
from the soil, while the oxidized forms of these elements are insoluble
and difficult to extract. Sulfide can also be readily extracted and ana-
lyzed.
Samples of Crowley silt loam surface soil were placed in 113.4 g (4
oz)—cylindrical polyethylene containers (which were also used for the
redox potential measurements) and submerged under approximately 2.54 cm
q r
(1 in) of water. For the sulfide experiment S-labeled sulfate was uni-
formly mixed with the soil before flooding. No additional sources of iron
and manganese were added. Samples were incubated for various periods
ranging from 1 day to 13 weeks, rapidly frozen, and then sectioned with
a large microtome into horizontal slices 1 or 2 mm thick. Freezing did
not disrupt the soil as much in the plastic containers as in glass contain-
ers. The soil samples were immediately extracted for the ions to be meas-
2+
ured. Manganese and Fe were extracted with -normal sodium acetate of pH
2.8, followed by appropriate chemical tests for the two elements (manga-
nese by atomic absorption spectrometry and iron by the colorimetric di-
pyridyl method). Separate samples were incubated for the sulfide study.
All of the sulfide in the thin sections was converted to hydrogen sulfide
12
-------
and collected in dilute sodium hydroxide solution following acidification
of the soil with sulfuric acid by use of a modified Conway diffusion cell.
Radioactive sulfide was determined by a flowing gas proportional G. M.
Counter. It was found convenient to use this tracer technique to follow
sulfide-distribution because of the difficulty of analyzing the small a-
mount of total sulfide in the approximately 1 g soil sections,
Manganous manganese distribution in a flooded soil profile is shown
in Figure 4. As early as 1 day after submergence there was a slight in-
crease in reduced manganese below the 5-cm depth. The amount of reduced
manganese reached a maximum after 1 week. As the incubation period in-
creased the thickness of the oxidized layer (or the layer in which man-
ganous ion was virtually absent) also increased. For long periods of wa-
terlogging it was apparent that oxygen was penetrating deeper into the soil
and increasing the thickness of the manganous-free zone.
The profiles of ferrous iron (Figure 5) differed from the manganese
profiles in two important respects. First, the thickness of the oxidized
layer relatively free of the reduced form of iron was much greater than
for manganese, and second, no iron reduction was evident 1 day after flood-
ing and very little reduction had occurred after one week. Maximum re-
duction was reached 8 weeks after flooding. As with the manganese profiles,
the longer the period of incubation, the deeper the oxidized or ferrous-
free layer.
Sulfide distribution following flooding is shown in Figure 6. Sulfate
reduction, similar to iron reduction, was relatively slow to develop, with
maximum sulfide in the anaerobic layer occurring after 8 weeks of incuba-
tion. The thickness of the oxidized (sulfide-free) layer was greater than
for iron.
13
-------
E
E
I
150 300 450
Manganese — ppm
600
Figure 4. Vertical distribution of reduced
manganese (Mn ) in a flooded Crowley
silt loam at various times after flooding.
Incubation times shown are 1 day to 13
weeks.
14
-------
E
E
I
Q.
0)
o
1
500
1000 1500
Iron-ppm
2000 2500
Figure 5. Vertical distribution of reduced
iron (Fe ) in a flooded Crowley silt
loam at various times after flooding.
Incubation times shown are 1 day to
13 weeks.
15
-------
f 10
I
S 15
20
25
500 1000
Sulf ide - cpm
1500
Figure 6. Vertical distribution of reduced sulfide
(S^~) in a flooded Crowley silt loam at various
times after flooding. Incubation times shown
are 2 weeks to 13 weeks.
16
-------
The surface oxidized layer probably serves as an effective sink for
the reduced forms of manganese, iron, and sulfur since manganous, ferrous,
and sulfide ions diffusing upward to this zone would be rapidly oxidized
to insoluble manganic and ferric compounds and to elemental sulfur.
DISTRIBUTION OF NITRATE AND AMMONIUM NITROGEN IN FLOODED SOIL
Nitrate and ammonium distribution in the flooded profile was deter-
mined using an experimental procedure similar to that used for manganese,
iron, and sulfide. Samples of the same Growley silt loam, into which 200
ppm N as ammonium sulfate was well mixed, were incubated under flooded
conditions in polyethylene containers, frozen, and sectioned as described
above. A water extract of the 2-mm sections was analyzed for nitrate by
the phenoldisulphonic acid method. Ammonium in the soil was extracted in
a modified Conway cell with dilute sodium hydroxide and the ammonia ab-
sorbed in dilute acid and determined by nesslerization.
The distribution of ammonium and nitrate nitrogen in the flooded
soil profile is shown in Figures 7 and 8. Ammonium nitrogen became depleted
in the surface layer, decreasing from 200 ppm to 50 ppm in 3 weeks in the
surface 8 mm. There was a gradual increase in ammonium nitrogen with depth
to about 15- to 20-mm depth. Apparently ammonium nitrogen was depleted
from the upper part of the reduced soil zone by diffusion to the oxidized
layer where nitrification could occur.
Although nitrate accumulated in the surface oxidized layer the maximum
concentration was less than 10 ppm. There appeared to be a deepening of
the nitrate-containing layer with time down to the 15- to 20-mm depth. The
reason for the low concentration of nitrate in the oxidized layer was very
17
-------
E
I
|
o
o
10
15
20
25
10w
day
50 100 150 200
Ammonium N—ppm
250
Figure 7. Vertical distribution of ammonium in a
flooded Crowley silt loam at various times after
flooding. Incubation times shown are 1 day to
14 weeks.
18
-------
I io
t
0)
o
= 15
20
25
Ida*
23456789
Nitrate N—ppm
10
Figure 8. Vertical distribution of nitrate in a
flooded Crowley silt loam at various times
after flooding. Incubation times shown are
1 day to 14 weeks.
19
-------
likely due to nitrate being lost through diffusion and denitrification at
the same time that it was being produced from ammonium oxidation. Be-
cause of its mobility nitrate formed in the oxidized layer readily dif-
fuses down into the anaerobic layer where denitrification will occur. The
process: ammonium diffusion from the reduced layer to the oxidized layer •*
ammonium oxidation to nitrate (nitrification) -»• nitrate diffusion from the
oxidized layer to the reduced layer -»• denitrification, can explain the de-
crease in ammonium in the oxidized layer and the small amount of nitrate
present in the aerobic layer at any one time. Such a process will also
explain the large loss of nitrogen from flooded soils which receive addi-
tions of ammonium nitrogen (Tusneem and Patrick 1971; Broadbent and Tusneem
1971).
CONCLUSIONS
The thickness of the oxidized surface layer of a flooded Crowley silt
loam soil was least for manganese, intermediate for iron, and greatest
for sulfide. The differences in thickness of the oxidized layer could be
predicted on the basis of the ease of reduction of the compounds of these
elements since manganese was easiest, iron intermediate, and sulfate most
difficult to reduce. Sulfate required a lower redox potential before re-
duction could occur and hence was reduced at a lower depth in the flooded
soil. This progressive increase in thickness of the surface oxidized layer
can be best seen in Figure 9. It should be noted that the redox potential
profile corresponded most closely to the iron profile in terms of the thick-
ness of the oxidized layer.
The oxidized layer was depleted of ammonium nitrogen and accumulated
20
-------
I ->
E
E
l
£ 10
Q.
0)
Q
O
"> 20
0 500 0
Manganese Iron
(ppm) (ppm)
2500 0 1500*500 -100
Sulfide Red ox potential
(cpm/g) fmV)
Figure 9. Thickness of the surface oxidized layer as measured by
the manganese, iron, sulfur, and redox potential profiles after
13 weeks flooding.
-------
nitrate, although the amount of nitrate did not account for the large de-
crease in amnonium. Nitrification and denitrification were probably both
proceeding at the same time in the oxidized and reduced zones.
22
-------
SECTION III
NITRATE REMOVAL FROM FLOODWATER OVERLYING
FLOODED SOILS AND SEDIMENTS
Quantitative studies of the rates of floodwater nitrate (N0~) re-
moval by flooded soils and sediments are essential to an understanding
of the capacity of wetland areas to act as a sink for N0~ polluted waters.
Oxygen-deficient conditions, existing in continuously-flooded and sea-
sonally flooded areas, help to govern the reactions, availability, or
removal of several important sediment nutrients or pollutants. This is
especially true for the nitrogen system as NO" is readily reduced under
oxygen-deficient conditions.
Microbial NO^ reduction, or denitrification, in a submerged sediment
or soil, as a significant mechanism for removal of large amounts of soil
nitrogen, is a well-known phenomenon. Losses of nitrogen by denitrifica-
tion have been shown to range from 15 to 1,000 ppm for time intervals vary-
*
ing from 1 to 4 days. Patrick (1960) recorded a N0~ reduction rate of 15
ppm/day in a reduced soil and except where rates in excess of 1,000 ppm of
NO- were added, the NO" reduction rate was a zero-order reaction with re-
spect to NO" concentrations.
The rate of denitrification in flooded soils or sediments may be af-
fected by several mechanisms, some of which are: depth of overlying water,
organic matter content of the soil, development of the aerobic-anaerobic
zone at the mud-water interface, and diffusion to the zone of active deni-
trification. The profile differentiation is characterized by two distinct
layers of soil or sediment: (1) a surface-oxidized layer of varying depth,
23
-------
and (2) an underlying reduced layer (Pearsall 1950). The oxidation-
reduction (redox) potential of an oxidized zone is generally higher than
300 mV and NO^ is usually considered to be relatively stable. However,
in the anaerobic layer which is less than 300 mV, active and complete
NO- reduction will take place (Patrick and DeLaune 1972). Diffusion of
N0~ to the active sites of denitrification plays an active role in con-
ij
trolling the rate of floodwater NOo removal. The concentration gradient
decreases in the direction of active denitrification sites and convection
currents and perturbation by wind tends to keep the floodwater of these
relatively shallow areas well-mixed. Soil organic matter content, as a
microbial energy source, has a significant effect on the rate of N0~ re-
duction, as shown by the work of Bremner and Shaw (1958) in which 1,000
ppm NO was lost in 4 days from a submerged soil to which an energy source
had been added.
The objectives of this study were to quantitatively characterize the
floodwater NO^ removal capacity of two flooded areas and to differentiate
between N0~ removal from the floodwater by diffusion alone, and NOl removal
by a combination of diffusion and NOT reduction. The areas selected for
this evaluation were the intermittently-flooded, fresh water swamp soils
of the lower Mississippi River alluvial areas of Louisiana and the continu-
ously-flooded salt water marsh soils of the Barataria Bay area of Louisiana.
In addition, the effect of additions of an energy source on the thickness
of the aerobic portion of the aerobic-anaerobic zone and its relationship
to NOn removal in Louisiana riceland soil was studied.
24
-------
MATERIALS AND METHODS
Floodwater NO- Removal by Two Flooded Soils
Three core samples each of a fresh water swamp soil (Bayou Sorrel,
Louisiana) and salt water marsh soil (Barataria Bay, Louisiana) were
obtained by driving thin wall aluminum pipes, 15.2 cm in diameter, into
the soil to a depth of about 55 cm and removing them intact to the lab-
oratory. The core samples were taken at sites about 150 m apart. The
fresh water swamp soil was a mineral soil containing about 4% organic
carbon and the salt water marsh soil was an organic soil containing about
12 to 14% organic carbon. The marsh soil is located in the inter tidal
zone of the Gulf of Mexico. The floodwater remained over the cores and
additional floodwater was obtained for each core. Each core was sealed
at the bottom and covered to prevent leakage and evaporation. The cores
were then incubated in the laboratory at 30C while the additional water
samples were stored at 4C. The floodwater of each core was then regulated
to a depth of 7.6 cm and received 25 ppm NO^-N (KNO_) which was evenly
distributed throughout the floodwater. The N0~ removal capacity of these
soils was then evaluated by a routine analysis of the floodwater for NOI
over a 10-day incubation period. This was accomplished by lowering a
pipette exactly 2.5 cm from the water surface and withdrawing a 2-ml sam-
ple. The samples were frozen for later analysis by the phenoldisulfonic
(PDS) acid method (Jackson 1958). To further explore the active site of
N0« removal, 1 liter of floodwater was removed from about 1 cm above the
surface of each soil core, mixed with KNO,. and sampled periodically for
N0~. To differentiate between N0~ diffusion from floodwater to soil alone
25
-------
and by diffusion plus denitrification, the soil cores were sterilized
with formaldehyde to give a 1.0% concentration throughout the core.
The soil cores were then incubated for 7 days and the sterility checked.
Potassium nitrate was then added to the cores as before and the flood-
water sampled as previously described. Prior to NOo analysis with PSD
the formaldehyde was destroyed with 30% 1^02.
Floodwater NOg Removal and Thickness of the Aerobic Layer as Affected
by Addition of Organic Material
Crowley silt loam (a Louisiana riceland soil containing about 0.7%
organic carbon) was mixed with varying amounts of finely-ground rice
(Oryza sativa L.) straw to give quantities of soil containing 0.0, 0.1,
0.5, and 2.0% added organic matter (OM). Duplicate 400-g portions of the
treated soils were weighed into 9 by 18 cm glass jars and flooded with
570 cc H20 to a depth of 7.6 cm. The flooded soils were preincubated for
21 days to ensure reducing conditions and establishment of aerobic-
anaerobic zones at the soil-water interface (Mortimore 1941). After pre-
incubation, NOo as KNOo was added to the floodwater as previously described.
To characterize the aerobic-anaerobic layers of the treated soils, a redox
profile of the soil-water interface to a depth of 20 mm was made as de-
scribed by Patrick and DeLaune (1972). To differentiate N03 diffusion
from microbial NOo reduction the flooded soils were sterilized and treated
as previously described.
To facilitate interpretation of the changes in floodwater NO^-N con-
centration, a quadratic regression analysis of each group of data from
the various treatments was made. The data were calculated as the curve
26
-------
most closely approximating a set of data points, £ (time) and £ (N07-N
concentration) , expressed by the equation
y = at + bt2 + c
with a^ as the initial rate of NO^ removal and £ as the initial NO, con-
centration. To differentiate diffusion of NO~-N from microbial NO^-N
removal, the equation depicting N07 loss in the sterile series was alge-
braically subtracted from the equation for NO" loss in the nonsterile
series. The resulting equation describes the rate of N0~ removal by mi-
crobial action only.
RESULTS AND DISCUSSION
Floodwater NOI Removal by Two Flooded Soils
Floodwater NO^ removal rates by three undisturbed soil cores of the
fresh water swamp of the Bayou Sorrel area of the lower Mississippi River
alluvium are presented as rate curves in Figure 10. Equations repre-
senting the best fit curves are shown in Table 1. Total N0~ removal
characterized by diffusion of NO, and microbial N0~ reduction was rela-
tively rapid with all N0~ disappearance in 10 days or less. The initial
rates of total N07 removal as underscored in Table 1 were 4.12, 4.03, and
4.98 ppm NO«-N for cores A, B, and C of the mineral soil. Decreases in
N0~ concentration due to diffusion only are shown in Figure 10 for the
sterile series and initial rates of diffusion are underscored in Table 1.
Little variation was noted in the initial diffusion rates of cores A and
B (1.61 and 1.62 ppm NO^-N), but the diffusion rate in core C was slight-
ly higher (2.38 ppm NO-j-N) . The rate curves, representative of microbial
NO, removal only, are presented in the lower portion of Figure 10 and are
27
-------
Core A
Diffusion
(sterile)
CoreB
CoreC
Reduction
2468 100 2 46 k »0 2 4
Time in itoys
Figure 10. Floodwater NO^ removal in rela-
tively undisturbed cores of a fresh water
swamp soil as affected by diffusion and
reduction.
Corel
Core 2
Core 3
Time in days
Figure 11. Floodwater N0~ removal in relatively
undisturbed cores of a salt water marsh soil
as affected by diffusion and reduction.
28
-------
Table 1. QUADRATIC REGRESSION EQUATIONS AND CORRELATION COEFFICIENTS REPRESENTING THE RATES OF
NITRATE REMOVAL FROM THE FLOODWATER OF FRESH WATER SWAMP SOILS AND SALT WATER MARSH SOILS FROM
THE BAYOU SORREL AREA OF THE MISSISSIPPI RIVER ALLUVIUM AND BARATARIA BAY AREA OF LOUISlANAt
NOo-N removal by
diffusion and reduction
(nonsterile)
NO-j-N removal
by diffusion
(sterile)
NOZ-N removal
by microbial
reduction
ppm/N
Bayou Sorrel, 7.6 cm H20 Depth
NJ
0.20t2 + 21.43
r - 0.996**
y = -4.03t + 0.18t2 + 20.05
r = 0.993**
-4.98t
0.987**
21.60
Core A
y = -1.6lt + 0.05t2 + 25.21
r = 0.980**
Core B
y = -1^62t_+ O.OSt + 22.33
r = 0.984**
Core C
y = -2.38t + 0.13t2 + 26.34
0.962**
2.51t - 0.15t" + 3.78
y = 2.41t - O.lOt + 2.28
y = 2.60t - 0.16tz +4.74
Table 1 continued.
-------
Table 1 (continued): QUADRATIC REGRESSION EQUATIONS AND CORRELATION COEFFICIENTS REPRESENTING THE
RATES OF NITRATE REMOVAL FROM THE FLOODWATER OF FRESH WATER SWAMP SOILS AND SALT WATER MARSH
SOILS FROM THE BAYOU SORREL AREA OF THE MISSISSIPPI RIVER ALLUVIUM AND BARATARIA BAY AREA OF
LOUISIANAt
NOq-N removal by
diffusion and reduction
(nonsterile)
NOg-N removal
by diffusion
(sterile)
NO^-N removal
by microbial
reduction
u>
o
10.60t + 1.46t + 17.4
0.969**
•8.92t + 0.75t^ + 26.6
0.969**
y - -7.92t + 0.61tz + 25.0
0.948**
ppm/N
Barataria Bay 7.6 cm H20 Depth
Core 1
y - -1.24t + O.OSt2 + 15.39
r - 0.737**
Core 2
y = -1.34t + 0.06t2 + 25.7
r - 0.858**
Core 3
7 " -1.72t + O.OSt2 +2.09
r - 0.949**
y - 9.36t - 1.38tT - 2.01
y = 7.58t - 0.69tz - 0.9
y = 6.20t - 0.53tz - 4.1
** Significant at the 1.0% level of statistical probability.
t The curve most closely approximating a set of data points where _t is time in days and %_ is ppm
NOo-N and is expressed by the equation £ = a_t + bt2 + £ where r_ is the correlation coefficient.
-------
the algebraic differences of the two upper curves. The regression equa-
tions for these curves are shown in Table 1. The initial rates of micro-
bial N0~ removal were similar for the three cores (2.51, 2.41, and 2.60
ppm N07-N). These data demonstrate a rapid movement of NO" out of the
relatively shallow floodwater of these soils with most NO" disappearance
attributed to microbial removal. However, when the floodwater was incu-
bated out of contact of the soil there was no decrease in the NO^ con-
centration of the water showing that there was no active microbial NO^
removal in the water.
Total N0~ removal from the saline floodwater overlying the organic
3
marsh soils of Barataria Bay was very rapid, as shown by the rate curves
in Figure 11, with complete disappearance of N0~ after 6 days. Rates of
initial NO, removal were calculated and are underscored in Table 1. These
rates were 10.60, 8.92, and 7.92 ppm NO~-N for cores 1, 2, and 3. As
contrasted to the swamp soil, the rate of total NO" removal was much high-
er; however, a greater variation in rate was noted among the marsh cores.
The rate curves for NO" diffusion are also presented in Figure 11 with
the initial rate of NO" diffusion underscored in Table 1. These rates
were 1.24, 1.34, and 1.72 ppm NO^-N for cores 1, 2, and 3 and were similar
to diffusion rates calculated for the fresh water cores. Consequently,
the calculated rates of microbial NO" reduction were appreciably faster
than those for the fresh water soil and were 9.36, 7.36, and 6.20 ppm
NO~-N for cores 1, 2, and 3, respectively. These data showed a very ac-
tive microbial NO" removal demonstrating that these marsh soils can act
as a sink for significant amounts of NO". As with the fresh water soil,
no NO" was reduced in floodwater separated from the marsh soil cores.
31
-------
Floodwater NO, Removal and Thickness of the Aerobic Layer as Affected
by Addition of Organic Material
The effects of additions of various amounts of organic matter as
ground rice straw on the thickness of the aerobic part of the aerobic-
anaerobic zone and on NO, removal rate from the overlying floodwater were
studied on Crowley silt loam, a riceland soil relatively low in native
organic matter. The increasing additions of organic material greatly in-
creased microbial activity, consequently resulting in thinner aerobic
layers at the soil-water interface. The redox profile curves used to
characterize the aerobic-anaerobic double layer to a depth of 20 mm are
shown in Figure 12. The depth of the aerobic zone (redox potential great-
er than 300 mV) extended 14 mm from the soil surface for the untreated
soil; a sharp decrease in depth occurred with increased additions of or-
ganic matter. The thickness of the aerobic layers of soil were 11 and
2.5 mm for the 0.1 and 0.5 additions of rice straw. For the 2.0% addi-
tion, reducing conditions extended to the soil surface with little de-
tectable aerobic zone. Rice straw additions had two effects on NO
reduction; the thinner aerobic zone decreased the distance over which
NO" diffused to the site of denitrification, the anaerobic soil layer,
and the added energy source increased the total capacity for NOT reduc-
tion. It appears from the redox profile of the soil receiving the 2.0%
addition of rice straw that NO" would be actively denitrified at or very
near the soil-water interface. Consequently, there would be no dilution
of NOT by diffusion into the soil. The increased microbial activity
placed a greater consumptive demand on free oxygen diffusing through
the floodwater to the soil-water interface and in the case of the 2.0%
32
-------
soil-water Interface
2.0% R£J
20
•400 •300 -200
(oxidized) Redox Potential- mv
100
(reduced)
Figure 12. The effects of additions of
rice straw on the development of redox
potential profiles in a flooded rice-
land soil. R.S. s rice straw.
4 6 $10 °0 24 6 8 fe
Time in days Time in days
Figure 13. The effects of additions of rice
straw on the removal of floodwater NO^ by
diffusion and reduction. R.S. = rice
straw.
33
-------
level of rice straw the free oxygen appeared to be consumed at the soil-
water interface and did not move into the soil. The effects of the addi-
tions of rice straw on NO" removal rates are presented in Figure 13 and
calculated rates are shown and underscored in Table 2. The initial
rates of total NO" removal were drastically affected by the addition
of the energy source and were 1.47, 2.60, 2.67, and 4.79 ppm NOT-N per
day for the 0, 0.1, 0.5, and 2.0 additions. Nitrate removal due to dif-
fusion is shown by the curve for the sterile series. Diffusion of NO-
was similar for all rice straw treatments and is presented as an average
for all treatments. Microbial NO" reduction appeared to be the dominant
3
removal mechanism, especially after the first few days.
The rates of microbial NO" reduction as the difference of the ster-
ile and nonsterile curves are also presented in Figure 13. These curves
depict a sharp increase in NO" reduction with increasing organic matter.
The initial rates of NO" reduction, as underscored in Table 2, were -0.17,
0.96, 1.03, and 3.15 ppm NO~-N per day for the 0, 0.1, 0.5, and 2.0% ad-
ditions of rice straw, respectively. The negative value was due to the
fact that microbial reduction plus diffusion was not appreciably greater
than diffusion alone for several days in the untreated soil.
CONCLUSIONS
This study has shown that both fresh water swamp and salt water
marsh soils of Louisiana can remove N0~ rapidly from the overlying flood-
water. Of the two kinds of soil studied, the salt water marsh was the
most effective sink for NO" with an average removal rate of 9.15 ppm N/day.
Converting these data to a hectare basis would result in a N0~ removal
34
-------
00
Ul
Table 2. QUADRATIC REGRESSION EQUATIONS AND CORRELATION COEFFICIENTS REPRESENTING THE RATES OF
NITRATE REMOVAL FROM THE FLOODWATER OF FLOODED CROWLEY SILT LOAM SOILSt
Organic
matter
addition
0.0%
0.1%
0.5%
2.0%
NO^-N removal by NOl-N removal
diffusion and reduction by diffusion
y =
r =
y =
r =
y =
r =
y -
r =
(nonsterile) (sterile)
-1.47t + O.OSt2 + 28.1 y = -1.64t + 0.12t2 + 28.1
0.963** r = 0.934**
-2.60t + O.llt2 + 28.5
0.988**
-2.67t + 0.07t2 + 27.6
0.983**
-4,79t + 0.21t2 + 26.5
0.994**
N07-N removal
by microbial
reduction
y = -0.17t + 0.07t2 + 0
y = 0.96t + O.Olt2 - 0.4
y = 1.03t + 0.05t2 + 0.5
y = 3.15t 4- 0.09t2 + 1.16
** Significant at the 1.0% level of statistical probability.
t The curve most closely approximating a set of data points where t. is time in days and y_ is ppm
NO~-N and is expressed by the equation y = at + bt2 + C where £ is the correlation coefficient.
J - ~— "™" *"""
-------
rate of about 7.4 kg N/ha per day. The fresh water swamp, also an
effective sink for NQ,, had an average removal rate of 3.5 kg N/ha per
day, about one-half that of the salt water marsh. Further studies
showed that additions of organic matter as ground rice straw regulated
the thickness of the aerobic-anaerobic double layer of a flooded soil
and drastically affected the rate of floodwater NOT removal. The soil
receiving the largest addition of organic matter had the highest rate
of NO" removal. This effect was due to increased microbial activity which
caused a thinner aerobic-anaerobic zone and in turn decreased the dis-
tance over which NOT could diffuse to the underlying anaerobic zone and
increased the NO. reduction capacity of the anaerobic zone.
36
-------
SECTION IV
THE ROLE OF OXYGEN IN NITROGEN LOSS
FROM FLOODED SOILS
Recent experiments utilizing 15N as a tracer have shown that N loss
from flooded soils occurs only when the flooded soil is exposed to 0?
(Tusneem and Patrick 1971; Patrick and Tusneem 1972; Broadbent and Tusneem
1971). The development of a thin oxygenated surface layer of soil permits
the biological oxidation of ammonium N to nitrate which in turn diffuses
into the underlying anaerobic layer where denitrification occurs. Although
the experiments cited above showed the effect of exposing the flooded soil
to air containing 21 percent 02, no information is available on the effect
of different concentrations of 02 on the magnitude of N loss. This study
deals with the loss of labeled N from flooded soil incubated under ^-0™
mixtures containing various concentrations of 0~. The effect of 0~ content
on the thickness of the surface aerobic soil layer where nitrification oc-
curs was also measured.
MATERIALS AND METHODS
Crowley silt loam soil with an initial total N content of 0.087 per-
cent was sampled from the plow layer of a field used for lowland rice, dried,
and passed through a 40-mesh sieve. The dry soil was tumble-mixed for sev-
eral hours with 230 ppm N as (NH(4)2SO^ which contained 10.228 atom percent
l^N excess. Duplicate 100-g samples of the soil were weighed into 8-oz.
wide-mouth glass bottles and flooded with an equal weight of water. The
depth of soil in the bottle was 4 cm with 2.7 cm of overlying flood water.
37
-------
Incubation vas carried out in the dark at 30°C for periods up to 120 days
under O^-No m^xtures containing 0, 5, 10, 20, 40, and 80 percent 02. The
bottles were incubated in the dark in large sealed containers through
which flowed a slow stream of moist air of the desired 0- content. Sam-
ples were analyzed in quadruplicate periodically for total N (labeled
plus unlabeled) in both the inorganic (nitrate + nitrite + ammonium) frac-
tion and the organic fraction (Bremner 1965).
The thickness of the aerobic surface layer of soil was determined by
a method described earlier (Patrick and DeLaune 1972) in which the redox
potential profile was measured by advancing a slow-moving platinum elec-
trode downward through the flooded soil.
RESULTS AND DISCUSSION
The magnitude of N loss was governed by the amount of 02 in the atmos-
phere over the flooded soil. Losses of both total N (Figure 14) and la-
beled N (Figure 15) increased as the 02 content increased. Maximum losses
of approximately 183 ppm total N (17 percent of total) and 155 ppm labeled
N (67 percent of that added) at 120 days were recorded. Most of the N loss
occurred at the expense of the inorganic fraction (which was present almost
entirely as added labeled N). A small amount of labeled N entered and re-
mained in the organic fraction. Little loss of N occurred at zero 02. The
small amount lost at zero 02 was probably the nitrate present at the begin-
ning of incubation. Nitrogen loss increased greatly with small increases
in 02 content. An increase in 0? content above 20 percent had little addi-
tional effect on N loss. Apparently the 0 content of the earth's atmos-
phere is sufficient to cause near-maximum denitrification loss of N in flood-
ed soil.
38
-------
No Oxygen
5 percent Oxygen
iivu
WOO
900
800
01
noo
woo
900
800
0
1100
1000
900
800
t
^Unaccounted for N
Inorganic N
Organic N
10 percent Oxygen
""^.Unaccounted forN
Inorganic N ""'~~-~^^
Organic N
r
40 percent Oxygen
^.^Unaccounted for N
Inorganic N Vv--— ££.,
Organic N
) 20 40 *0 80 100 1:
i
i
•0 (
""* — v Unaccounted for N
Inorganic N
\.___
Organic N
. .....
20 percent Oxygen
~~y Unaccounted for N
Inorganic N ^~^----~__;
Organic N
.
80 percent Oxygen
V Unaccounted for N
\-^*-*-^.^
Inorganic N N.
Organic N
I 20 40 60 80 100 12
Time-Days
Figure 14. Loss of total N (labeled + unlabeled)
from flooded soils as affected by C>2 content
of air.
39
-------
No Oxygen
5 percent Oxygen
1
200
150
100
50
0
200
150
100
50
0
200
150
100
SO
Q
^Unaccounted for N
. Inorganic N
'~"~ Organic N
10 percent Oxygen
-'•^.^Unaccounted for N
V
X
. Inorganic N
"~"T Organic N
40 percent Oxygen
"~ "v Unaccounted for N
' X
^'-
"^S.-i;
Inorganic N
Organic N
0 20 40 60 80 100 120
'-S.. - Unaccounted for H
^" "x
NE
. Inorganic N
""* , Organic N
20 percent Oxygen
' — \ Unaccounted for N
V :?•:
V
Inorganic N
""""^. Organic N
80 percent Oxygen
"" • ' ";
%^|jlnaceountexl ifofS
il-;.^? ':i
'
°0 20 40 60
Oxygen - percent
80
Figure 16. Effect of 02 content of
air on thickness of aerobic soil
layer after 90 days' incubation.
40
-------
Because of the role of the aerobic soil layer in supporting nitri-
fication, a study was made of the effect of 02 content on the thickness
of this oxygenated layer, since a previous experiment (Patrick and Tusneem
1972) indicated that a thick aerobic layer would result in higher N loss
than a thin layer. The thickness of the aerobic layer at the various 02
levels after 90 days of incubation is shown in Figure 16. The first in-
crements of 02 (up to 10 percent) created only a thin aerobic layer while
02 contents of 20 percent and higher resulted in a layer 2 to 3 cm thick.
The extent of N loss did not correspond precisely to the thickness of the
aerobic layer since appreciable loss occurred at 5 and 10 percent 02 where
only a thin aerobic layer existed. Although the thickness of the aerobic
layer is important in helping to determine the amount of nitrate formed
from ammonium, it is not the only factor governing this process. The
rate of ammonium diffusion from the underlying reduced soil to the aerobic
surface layer, regardless of how thin this layer is, contributes to the
amount of nitrate formed.
Because of the known effect of organic matter in immobilizing N in an
organic form not subject to denitrification, an additional treatment was
included in which the soil received 0.5 percent ground rice straw and was
incubated for 120 days under the various 0_ concentrations. The results
obtained (Table 3) show an appreciable conservation of N as a result of
added organic matter, with N loss being decreased 30 to 40 ppm labeled N
and slightly more total N. Analyses of inorganic and organic N fractions
showed that the rice straw immobilized an appreciable amount of the inor-
ganic N, thereby protecting it from nitrification and subsequent denitri-
fication.
41
-------
Table 3. NITROGEN CONTENT OF FLOODED SOIL IN PPM AFTER 120
DAYS' INCUBATION WITH AND WITHOUT ADDITION OF 0.5 PERCENT
RICE STRAW (FOUR REPLICATIONS)
Oxygen
%
0
5
10
20
40
80
Total
N
Rice No rice
Straw straw
1081
1049
1018
1016
1056
957
1077
994
944
947
951
920
Labelled
N
Rice No rice
straw straw
196
159
138
140
153
104
207
127
107
92
117
76
42
-------
The results of this study show that the amount of 0- over a flooded
soil is a factor in determining the amount of N lost by denitrification.
Large increases in N loss occurred from the first few increments of oxy-
gen with little further loss occurring above 20 percent 02. The thickness
of the aerobic soil layer was also governed by the amount of 02 in the air.
Nitrogen loss was generally related to the thickness of the aerobic layer,
even though appreciable loss occurred at 5 and 10 percent Oj where the
aerobic layer was relatively thin. Ammonium diffusion from the underlying
anaerobic zone to the aerobic surface layer is probably a significant fac-
tor in determining the rate of N loss in flooded soils. Adding organic
matter to the soil significantly reduced N loss.
43
-------
SECTION V
NITRIFICATION-DENITRIFICATION REACTIONS IN FLOODED SOILS AND SEDIMENTS:
DEPENDENCE OF OXYGEN SUPPLY AND AMMONIUM DIFFUSION
Denltrification in soils, swamps, marshes and water bodies is the
major process by which elemental nitrogen is returned to the atmosphere.
Denitrification takes place when a deficiency of oxygen causes certain
facultative anaerobic bacteria to use nitrate in place of oxygen as an
electron acceptor for respiration. Although denitrification in soils as
a result of temporary anaerobic conditions caused by variations in soil
moisture has been recognized for years (Russell, 1961), the extent of
and mechanisms involved in nitrogen loss in continuously flooded systems
are not as well documented.
Ammonium is the predominant inorganic form of nitrogen in oxygen-
deficient flooded systems and denitrification cannot take place unless
conditions exist where nitrification of the ammonium can first occur.
The development of a thin oxygenated surface layer as a result of dis-
solved oxygen penetrating a short distance into the flooded soil or sedi-
ment before being consumed allows the oxidation of ammonium to nitrate
to take place in this oxygenated or aerobic surface layer. Nitrate pro-
duced from this reaction diffuses down into the underlying oxygen-free
layer where it is denitrified.
Recent experiments have shown that appreciable denitrification will
occur in flooded soils if both oxygen from the atmosphere and ammonium
from the flooded soil are available (Tusneem and Patrick, 1971; Broadbent
and Tusneem, 1971). Although the nitrogen converted from nitrate to gaseous
44
-------
forms in flooded systems is derived from ammonium oxidized to nitrate
in the aerobic layer, the amount of nitrogen gas usually greatly exceeds
the amounts of ammonium and nitrate present in the aerobic layer at any
one time. Ammonium movement from the underlying anaerobic layer to the
surface aerobic layer is apparently necessary to account for the large
nitrogen losses that occur in flooded systems. Removal of ammonium by
nitrification in the aerobic layer creates a concentration gradient which
causes ammonium in the underlying anaerobic layer to diffuse upward to
the aerobic layer where it undergoes nitrification. Nitrate formed by
this process readily diffuses down toward the nitrate-free anaerobic layer
where it is denitrified to nitrogen gas by serving as an electron acceptor
in the oxidation of organic matter. These processes are illustrated in
Figure 17. As indicated in the equation, nitrate is an intermediate in
the overall pathway between ammonium and elemental nitrogen.
In the experiments described in this report, the processes discussed
above for governing nitrogen loss in flooded soils and sediments are ex-
amined. Particular attention is given to the roles of oxygen supply to
the aerobic layer and ammonium diffusion from the anaerobic layer to the
aerobic layer in determining nitrogen loss through the nitrification-
denitrification process. Nitrogen-15 labeled ammonium was used to trace
nitrogen through the various nitrogen forms.
MATERIALS AND METHODS
The experiment was carried out using large (45.5 cm length x 4.5 cm
diameter) pyrex tubes fitted at the top with a stopcock and a ground glass
junction for entry to a mass spectrometer. All connections in the apparatus
45
-------
02
Diffusion
Nti
Diffusion
I
i «,-
Organic N
NO
N2O
Flood Water
Aerobic Layer
Anaerobic Layer
Nitrification - Denitrif ication Reaction :
24NHJ + 48 O2
+ 24H2O
5C6H12O6 +24H"*" — * 12N2+30CO2 + 42 H2O
24NH4+5C6H12O6+48O2 »• 12 N2 + 30CO2 * 66H2O +24H+
Figure 17. Processes involved in the sequential conversion of organic nitro-
gen to elemental nitrogen in flooded soils and sediments.
-------
were ground glass and tested for leaks. Fifty grams of soil (Crowley
silt loam) and enough water (50 ml) to provide a 1.5 cm water layer over
the soil were added to each flask. Ammonium sulfate with an N enrich-
ment of 29.944 atom-% was thoroughly mixed with the soil at the rate of
250 yg N per g soil. The effect of oxygen in the atmosphere over the
flooded soil on the conversion of ammonium nitrogen to nitrogen gas was
studied in two experiments. In one experiment, an atmosphere of 30% oxy-
gen and 70% argon was maintained over the flooded soil. Duplicate flasks
were removed after various periods of incubation for analyses of five la-
beled nitrogen fractions: ammonium, organic (protein) nitrogen, nitrate,
nitrous oxide, and elemental nitrogen. The amount of labeled nitrogen
was determined in each of these fractions by appropriate procedures em-
ploying a mass spectrometer (Bremner, 1965). In the other experiment,
conditions were the same as described above except that variable oxygen
contents ranging from 0 to 50% were maintained over the flooded soil for
100 days, after which the distribution of labeled nitrogen was determined.
Side tubes containing ascarite granules were placed in each flask to absorb
excess carbon dioxide. The amount of oxygen consumed was also determined
by mass spectrometer analysis. The size of the flask and the amount of
soil were chosen so that large changes in the oxygen content did not occur
during incubation.
Separate experiments were carried out to measure the magnitude of
ammonium diffusion from the anaerobic layer to the aerobic surface layer.
One method consisted of incubating columns of flooded soil containing 200
yg ammonium N per g soil and determining ammonium distribution by slicing
the soil columns horizontally into 2 mm sections. The distribution curves
47
-------
were used to calculate the ammonium diffusion from the anaerobic layer
to the aerobic layer by measuring the decrease in ammonium in the anaer-
obic layer with time. Another approach to the estimation of ammonium
diffusion from the anaerobic layer into the aerobic layer was to measure
the total amount of ammonium converted to nitrogen gas and nitrate and to
subtract from this value the amount of ammonium originally present in the
aerobic layer. Diffusion was not involved in the nitrification of am-
monium originally present in the aerobic layer. In order to use this
method it was necessary to allow the reaction to proceed long enough so
that all of the ammonium originally present in the aerobic layer had been
converted to nitrate. For both the above methods, it was also necessary
to accurately determine the thickness of the reddish-brown oxidized fer-
ric oxide layer overlying the grayish-brown ferrous oxide layer. The
thickness of the aerobic layer determined by this method was found to
correspond closely with the thickness of the nitrate-containing layer.
RESULTS AND DISCUSSION
Results of the experiments dealing with the effect of oxygen on am-
monium conversion to nitrate and then to nitrogen gas are shown in Figure
18. The ammonium content of the system decreased with time with a buildup
of nitrogen gas evident after two weeks. No nitrous oxide (N»0) was pres-
ent until the last sampling at 100 days. Nitrate was present after 30
days, but did not accumulate. The most striking result of this experiment
was the rapid conversion of ammonium to nitrogen gas after 30 days. This
length of time was approximately the same as that required for the devel-
opment of a pronounced aerobic surface layer in the flooded soil. A small
48
-------
vO
300
Figure 18a. The conversion of 15N-labeled ammonium to elemental nitrogen in a flooded
soil. The distribution of labeled ammonium among various nitrogen fractions following
incubation under 30% oxygen.
-------
300
-V,
Figure 18b. The conversion of 15N-labeled ammonium to elemental nitrogen in a flooded
soil. The effect of oxygen content on conversion of labeled ammonium to various
nitrogen fractions.
-------
amount of labeled ammonium nitrogen was incorporated into the organic
fraction at the beginning of incubation and did not change appreciably
in concentration during the 100-day period.
The oxygen content of the atmosphere over the flooded soil during
the 100-day incubation period had a marked effect on the stability of am-
monium. Where no oxygen was present almost all of the added labeled am-
monium remained in the ammonium form except for the small amount incor-
porated into the organic fraction. Where oxygen was present, however, at
least part of the ammonium was converted to nitrogen gas, with the amount
increasing as the oxygen content increased up to 26%. A higher oxygen
content resulted in a thicker aerobic layer and slightly more nitrate, but
no additional nitrogen gas.
These experiments utilizing tracer nitrogen demonstrate that ammonium
in a flooded soil is converted to nitrate and then to elemental nitrogen.
Measurements of the thickness of the aerobic surface layer and of the a-
mount of labeled nitrogen gas produced showed that approximately twice as
j
much ammonium was converted to nitrate and elemental nitrogen as was orig-
inally present in the aerobic layer. Calculation of ammonium diffusion
from the anaerobic layer to the aerobic layer based on direct measurement
of ammonium distribution with depth throughout the aerobic-anaerobic lay-
ers also showed that approximately half of the ammonium nitrogen involved
(720 pg cm~^ or 72 kg ha"-*-) was derived from ammonium diffusing from the
anaerobic layer to the aerobic layer. The measured diffusion coefficient
for ammonium was 2.5 x 10~ cm2 see" as determined by the method of
Phillips and Brown (1964). Similarly, the diffusion coefficient for nitrate
diffusing from the aerobic layer to the anaerobic layer was^l.5 x 10
51
-------
2 —1
cm sec . These results are applicable for flooded soil or sediment
systems in which no downward percolation occurs that could prevent upward
diffusion of ammonium. In most flooded soils and shallow water bodies,
percolation of water is extremely slow.
Oxygen consumption in the flooded soil is shown in Figure 19. Most
of the oxygen consumed initially was probably used for oxidation of or-
ganic carbon by heterotrophs, but after several days a considerable por-
tion was also being used for nitrification. Some of this oxygen was
present as nitrate, while some had been converted from nitrate to carbon
dioxide in the denitrification reaction (see equation in Figure 17). In-
creasing the concentration of oxygen over the flooded soil increased the
thickness of the aerobic layer and increased the amount of oxygen con-
sumed. Nitrification and organic carbon oxidation accounted for approxi-
mately equal amounts of oxygen.
CONCLUSIONS
The results of this study show that ammonium nitrogen in an anaerobic
soil or sediment exposed to atmospheric oxygen undergoes sequential nitri-
fication and denitrification. The source of the ammonium reacting with
atmospheric oxygen in the aerobic layer consists of the ammonium initially
present in the aerobic surface layer of soil or sediment plus an approxi-
mately equal amount diffusing to the aerobic layer from the underlying an-
aerobic layer. It is likely that a significant part of the ammonium pres-
ent in flooded soils and shallow water sediments follows this pathway and
makes a major contribution to the elemental nitrogen being returned to
the atmosphere.
52
-------
o
CO
O)
E
I
Q.
3
(0
§ 90
O
c
0)
s
6 60
O2 consumed
in N2 production
O2 present in NO3
30
40 60
Time — Days
100 days incubation
Total O2
consumption
0\
O2 consumed in N2 production
20 30
Percent Oxygen
Figure 19. Total oxygen consumption, oxygen converted to nitrate, and
oxygen converted sequentially to nitrate and then to carbon dioxide
(see equation in Figure 17) as affected by time of incubation and
oxygen concentration.
53
-------
SECTION VI
EFFECT OF DISSOLVED OXYGEN ON REDOX POTENTIAL AND
NITRATE REDUCTION IN FLOODED SWAMP AND MARSH SOILS
The NO~-N content of some natural bodies of surface water has in-
creased slightly in recent years. This increase has been attributed to
a number of sources, including the use of N- fertilizers for field crops
and the surface disposal of N-containing wastes. The increase in NO_-N
content of surface water supplies, part of the pollution load, has made
it necessary to learn more about the mechanisms by which N is returned
from the water to the atmosphere. In low lying swamp and marsh areas
that receive local drainage from agricultural areas where large amounts
of N- fertilizers are used, NO^-N is rapidly removed from the water
through denitrification after the NO^-N has diffused through the inter-
face separating the floodwater from the anaerobic mud layer (Engler and
Patrick, 1974).
An anaerobic zone exists in a flooded soil because 0? moving from the
atmosphere through the floodwater can penetrate only a short distance in
the soil before being utilized by aerobic and facultative anaerobic micro-
organisms. The layer below the zone of 0? penetration becomes anaerobic
rapidly after flooding (Patrick and DeLaune, 1972). In the anaerobic lay-
er, many facultative anaerobic bacteria use NOo-N as a substitute for ©2,
reducing the NOl-N to N» gas or N_0 which returns to the atmosphere
(Pearsall, 1950; Mitsui, 1954). The reduction process is accentuated
where there is an ample supply of organic detritus or energy source in
the mud for bacterial metabolism. Although this mechanism of nitrogen
54
-------
loss has been known to exist, little information is available on the ef-
fect of dissolved ()„ in the overlying and pore water on the process.
At the surface oxygenated sediment-water or soil-water interface,
microbial and chemical conditions are very much like those in drained or
oxidized soil. In the underlying 0 -free soil or sediment layer anaer-
obic biological and chemical transformations take place in the absence
of 0_. In the presence of 0? the oxidized constituents of the soil such
as NO^-N, Mn compounds, and Fe3+ compounds are not biologically or
chemically reduced (Mortimer, 1941). After 0- disappears following
flooding of an oxygenated soil or sediment many soil micro-organisms
are able to substitute one or more of these oxidized chemical compo-
nents for theO- required in respiration. Nitrate, the higher oxides of
Mn and hydrated ferric oxide are reduced more or less sequentially if
Oj becomes absent or limiting and if an energy source (organic matter)
is available to the micro-organisms (Turner and Patrick, 1968) . Nitrate
and Mn^+ compounds are readily reduced since the energy required for
their reduction is low and a number of species of micro-organisms are
capable of carrying out this process.
The objectives of this study were to measure Q^ reduction rates,
N0~ reduction rates, and the effects of added 0 on NO^ reduction and
redox potential in four swamp and marsh soils from Louisiana. In addi-
tion, laboratory experiments were set up to determine the site of N03
reduction in these systems and to evaluate the role of NO^ diffusion
from the overlying water into the reduced soil in the overall N03 re-
duction process.
55
-------
MATERIALS AND METHODS
Four soils taken from Mississippi River floodplain swamps and the
Coastal Marsh of Louisiana were selected for this investigation. Both
fresh and brackish water areas were represented. Bulk samples of soil
were placed in several layers of large polyethylene bags, sealed to pre-
vent moisture loss and minimize oxygen contact, and iced to reduce bio-
logical activity. Floodwater overlying the soils was collected in sep-
arate containers and both soils and floodwater were stored at 2 to 4°C
until used in the experiment. A brief description of the four soils is
given in Table 4. All soils were from areas subject to flooding all or
part of the year. Selected chemical properties of these soils are pre-
sented in Table 5. Experiments were conducted at 30°C.
0. Removal from Soil Suspensions
Suspensions (soil:water = 1:2) of each of the four soils were
placed in the flask shown in Figure 20 and stirred continuously. The
container was completely filled so that no air space existed. Oxygen
gas was bubbled through the suspension until the dissolved 0_ content
as measured with a membrane-covered polarographic 0, electrode increased
to 16 ppm (on a water basis). The electrode was standardized with air
saturated distilled HO at 30°C. The flask was sealed and both dissolved
0 and redox potential were monitored continuously for several hours.
Redox potential measurements were made with a bright platinum electrode
and calomel half cell connected to a vacuum tube voltmeter. The filled
flasks contained 1850 ml of suspension.
56
-------
Table 4
DESCRIPTION OF SOILS USED IN STUDY
Soil No. Description
1 This is an unclassified clay from a
marsh that is intermediate between
brackish and salt water. This soil
is flooded throughout the year.
2 This soil is a muck taken from a
fresh water swamp and consists of
an organic surface horizon 4 to
15 inches thick underlain by a soft
clayey horizon. This soil belongs
to the Barberry series.
3 This soil is a muck from a slightly
saline marsh with surface layers of
well decomposed plant materials
underlain by four feet of soft clay.
The sample is of the Lafitte series.
4 This is a clay that is subject to
flooding by salt water but which
was unflooded at the time of sam-
pling and belongs to the Lafitte
series.
Control This is a Crowley silt loam, a
Louisiana Coastal Prairie soil used
extensively for rice cultivation and
containing about 0.7% organic carbon.
57
-------
Table 5
CERTAIN CHEMICAL PROPERTIES OF SOILS USED IN THIS STUDY
Extractable
t
Soil
No. pH
** Organic
Moisture Carbon
Total
Sulfide
P
percent
1 7.6
2 5.7
3 6.3
4 5.7
236.1
219.7
257.7
114.2
4.9
8.4
35.9
4.3
965.3
0.0
0.0
0.0
267
145
139
164
K
Ca
Mg
parts per million
616
368
624
616
4400
4400
7200
640
1946
1424
6280
1070
** Grams of water per 100 grams of soil.
P extracted with 0.1N HC1 + 0.03N NH.F; other elements extracted with
0.1NHC1. *
58
-------
Calomel.
hall cell
Reaction
chamber
Magnetic stirrer
Figure 20. Apparatus used to study nitraLe reduction, oxygen
consumption and redox potential in stirred suspensions of
sediment.
59
-------
No" Reduction by Anaerobic Soil Suspensions and from Floodwater Dredging
Soil
Suspensions (soil:water = 1:2) of each of the four soils were pre-
pared by mixing the soil in a Waring blender with enough N0~ as KNO_
added to the suspension to provide 80 to 100 ppm N (on dry weight basis)
to each suspension. For each soil, ten-30g suspensions were placed in
test tubes and the tubes purged with Ar. The soils were incubated an-
aerobically and samples removed periodically for N0_ analyses. The soil
solution was separated by centrifugation and NO -N measured with an Orion
specific ion electrode. Nitrate removal was followed until complete.
Samples of each of the four soils were also placed in one quart wide
mouth jars and flooded with approximately 6.5 cm of floodwater that
had been sampled with the soil. Nitrate-N was added three times at rates
of 50, 90, and 50 ppm to the overlying floodwater and the NO content of
the water routinely determined with an Orion specific ion electrode.
When the N0_ content decreased to a low concentration the next NO, addi-
tion was made and the rate of reduction again observed.
NO- Reduction in Floodwater Separated from Soil
The experimental procedure was similar to the previous experiment ex-
cept that the water collected from the field sites was placed in the jars
exclusive of any soil and NO -N was added at a rate of 100 ppm-N. As a
control of floodwater was placed over each corresponding soil and incu-
bated with 100 ppm NO~-N added to the water.
Effect of Added Q^ on NO Reduction
The apparatus used for this experiment is shown in Figure 20. Enough
60
-------
N0~ as KNO was mixed into the 1:2 soil:water suspensions to give ap-
proximately 40 ppm N and CL was bubbled into the suspension until the
dissolved 0- concentration of the water phase reached 16 ppm. Redox po-
tential and concentrations of 0^ and NO -N were monitored as before. In
a parallel series the same amount of N0_ was added but with no 0^ addi-
tion. The small amount of 0 detected in these suspensions was due to
air contamination during mixing. The containers were completely filled
with suspension so that no air space existed. As a comparison, an air
dry Crowley silt loam (a Louisiana riceland soil) was carried through
the identical procedure. The Crowley soil had been stored air dry six
months prior to this study.
It should be noted here that in some of the NO- disappearance stud-
ies the apparent NO concentration did not quite reach zero. This is
an artifact characteristic of Cl interference with the NO specific ion
electrode. Zero NO concentration is indicated by a low constant NO^
reading.
RESULTS AND DISCUSSION
0~ Removal from Soil Suspensions
Reduction of G£ for the four suspensions was quite rapid with loss
of 16 ppm dissolved 0 occurring over a range of time from a few minutes
to four hours (Figure 21). Soil 1 contained considerable S~ and this
highly active reduced material aided in depleting the added 0- in 15
minutes. The redox potential of this suspension reached a minimum value
in about 1 hour. Oxygen addition and subsequent removal had a notice-
able effect on the redox potential curves. The point at which 0
61
-------
Soil 1
Soil 2
16
12
8
4
Eh
L°2,
0 1 2 3 4 5
01 2345
0 5
4-1
0)
012345
012345
Figure 21. Changes in oxygen content and redox potential
with time in stirred suspensions.
62
-------
disappeared was clearly marked by an inflection in the redox potential
curve except for soil 1 which consumed all of the 0 in a few minutes.
The point at which 0 was depleted was marked by a rapid drop in redox
potential after which the potential leveled off probably as a result of
being poised by the other soil redox systems. This inflection was par-
ticularly apparent in soil 4 which exhibited the slowest 0. consumption.
Since there are a number of oxidation-reduction systems controlling the;
redox potential in a flooded soil, the disappearance of the oxidized
component of one system, such as dissolved oxygen, will result in the
potential decreasing rapidly to a new level at which other oxidation-
reduption systems, such as thpse associated with Mn and N0~, will again
cause it to be stabilized at the lower potential (Turner and Patrick,
1968). Dissolved 02 in flooded soil is more difficult to measure in
the field than is redox potential and following redox potential changes
may be of value in estimating the point at which oxygen is depleted in
flooded soils.
N0« Reduction by Anaerobic Soil Suspensions and from Floodwater Over-
lying Soil
A series of experiments were conducted in which the rates of N0~
reduction by the four soils and from the floodwater were measured.
These two experiments were designed to determine both the potential rate
of NOl reduction and the actual rate that might be encountered under
field conditions.
As shown in Figure 22, N0~ reduction was rapid in all of the soils
with total removal in 1.5 days for soil 1 and 3 and in about 4 days for
63
-------
Soil 1
Soil 2
120
100
80
0123
Soil 4
Figure 22. Changes in nitrate content with time in
stirred suspensions.
64
-------
soil 2. Except for soil 4, N0~ reduction was essentially linear with
time, indicating a zero order reaction. For soil 4 however, NOl re-
duction was initially rapid but then slowed after one day.
Although the previous experiment showed that NOl mixed with the
reduced soil samples was rapidly reduced, no information was provided
on removal of N07 from the overlying floodwater. Under field conditions,
N0~ containing waters are rarely, if ever, well mixed with the soil.
Since most NO. in runoff water draining into back swamp areas would ac-
cumulate in the floodwater, it is important to determine how rapidly
floodwater N07 overlying anaerobic swamp sediments is subject to re-
duction. Because of the presence of dissolved 0« in the floodwater,
it is possible that NO, reduction in the floodwater would be inhibited
by the presence of the more easily reduced O™. As described in earlier
research (Engler and Patrick, 1974), Figure 23 shows that NO, in the
floodwater overlying a marsh or swamp soil was lost at a significant
rate, although not nearly as rapidly as when mixed with the anaerobic
soil.
Rapid NO,, removal was noted from the water overlying soil 1 while,
as was the case for oxygen removal, soil 2 showed the slowest removal.
Probably the most significant finding of this experiment was the ability
of the soil-water system to continue to reduce NO- following successive
applications of N07 to the floodwater. On a hectare basis, the total
reduction of N0~ for the entire three-month experimental period amounted
to approximately 70kg/ha. This amount of NO^-N is much greater than
would be expected to drain into a back swamp from adjacent agricultural
areas and indicates that these flooded soils are capable of rapid
65
-------
100
75
50
0)
ro 25
T3
O
O
a100
2 75
50
25
20
SoiM
1
40
Soil 2
60
80
100
Surface—open to air
i
Flood wofer
A (6 cm)
Soil (6 cm)
^
V
20
40 60
Time—Days
80
100
Figure .23. Nitrate removal following repeated additions of
nitrate to the floodwater overlying flooded soil.
66
-------
100
Soil 3
20
40 60
Time—Days
100
80
100
Figure 23 (continued): Nitrate removal following repeated
additions of nitrate to the floodwater overlying flooded
soil.
67
-------
Soil 1
Soil 2
100
I
Soil 3
100
75
50
25
100
75
50
25
12345
Soil 4
12345 V0 1 2 3 4 5
Time—Days
Figure 24. Uitrate removal from floodwater in contact with
soil surface and removal from floodwater separated from
soil.
68
-------
removal of N(£ draining into the area from adjacent areas.
NOo Reduction in Floodwater Separated from Soil
The results of the previous experiment showed that a rapid deple-
tion of NCL occurred in the floodwater overlying soil. It would be of
value to determine if any of this reduction takes place in the flood-
water itself or if it is essential for the N0~ to diffuse into the re-
duced soil layer before reduction occurs. An experiment was therefore
designed to measure NO- reduction in the floodwater overlying the re-
duced soil and to also measure N0~ reduction in another sample of the
same floodwater that had been removed to a separate jar containing no
soil. Results of the comparison are shown in Figure 24 with the upper
curve of each graph representing NOZ content of the floodwater separated
from the soil. No NOl disappeared from the separated floodwater. How-
ever, NO, disappearance was rapid when the floodwater remained in con-
tact with the soil.
It is of interest to consider the possible mechanisms by which NO"
added to a shallow, oxygenated water layer overlying an anaerobic soil
is reduced. Little reduction is likely to take place in the oxygenated
layer because of the presence of the more easily reduced 0-. Conse-
quently, for N0_ reduction to occur, NO ions must diffuse or leach into
the anaerobic soil zone underlying the surface oxidized or aerobic soil
layer. There is no significant NO" adsorption onto the mineral soil or
organic matter surfaces and the N0~ ion is free to respond to the sharp
concentration gradient that exists between the floodwater and the re-
duced soil layer. Though the diffusion of NO into the reduced layer is
69
-------
relatively slow, it is continuous since the concentration gradient per-
sists as long as NO is present in the floodwater. The anaerobic soil
layer thus serves as an effective sink for NO^.
The removal of NO" from floodwater does not necessarily mean that
all potentially toxic forms of inorganic nitrogen have heen removed from
the soil-water systems. The NO ion represents the first product of the
denitrification process and an accumulation of NOl could be more toxic
than the equivalent concentration of N0_. All reported studies have
shown, however, that under conditions comparable to those in the field,
only trace amounts of NO (2 to 5 ppm) accumulate in flooded soils.
Apparently, the reduction of N0~ proceeds at a potentially higher rate
than does the reduction of N0~, since NO reduction is seldom a rate
limiting step in denitrification.
Effect of Added 02 on NOj Reduction
In general, 16 ppm added 0_ did not appear to significantly hinder
N0~ reduction in the soil suspensions. As shown in Figure 25, NO, re-
•J J
duction began immediately after NO addition in both oxygenated and
J
nonoxygenated samples and most of the added NO had disappeared by the
end of 24 hours. The added 0 was depleted in 0.5 to 5 hours after
addition with the S~-containing soil (soil 1) consuming 02 most rapidly
and soil 4 requiring the longest time for 0,. removal for the swamp soils.
The agricultural mineral soil required 6 hours to consume all the 0..
The redox potential of the suspension again responded markedly to the
disappearance of oxygen. At the time of 0_ depletion, the redox poten-
tial decreased sharply. This inflection in redox potential was discussed
70
-------
Soil 1
50
£40
Q
Q
I
i
0)
30
eo 20
i_
•M
Z 10
01—
Nitrate
Time—Hours
> 500BT "1
| 400
to
~ 300
c
0>
^^J
o 200
Q.
g 100
73
0)
cr 0
L\ Redox
I V potential
\ \
\ \ 02 added
"\ V /
V'V-/.*— ___^
^w
^•> O— 0— 0 n ^
•^^M ^^^H
20 r— —
E 16 I Oxygen
* I
I 12 t
c
fl>
on 8
X
0 4
n
I
1 ^-O2 added
^X^ ^
•till iiil iiil iiiliiiliii
24
Figure 25. Effect of added oxygen on nitrate reduction and
redox potential in stirred suspensions.
71
-------
Soil 2
50
a
a
I
£40
30
•
a
20
O2 added
,O2 added
-•-
-Ou
X>2 added
I
. . i
Nitrate
Redox
potential
Oxygen
-cH
. I ... I . . .
8 12 16
Time—Hours
20
24
Figure 25 (continued): Effect of added oxygen on nitrate
reduction and redox potential in stirred suspensions.
72
-------
Soil 3
50
1.40
a
I
30
a>
+••
(0
20
10
Nitrate
O2 added
o \
\ O2 added
4i V
~ v. .%»... i .
°0 4 8
Redox
potential
Oxygen
12 16
Time—Hours
. I
20
24
Figure 25 (continued): Effect of added oxygen on nitrate
reduction and redox potential in stirred suspensions.
73
-------
Soil 4
50
1.40
o.
z 30
o>
z 10
0
500
f 400
"w
'.g 300
o>
| 200
g 100
0
Q.
I
20
16
0) 8
X
5 *
added
added
I
1
Nitrate
Redox
Potential
Oxyqen
. -1
• 1
8 12 16
Time —Hours
20
— n
24
Figure 25 (continued): Effect of added oxygen on nitrate
reduction and redox potential in stirred suspensions.
74
-------
Crowley Silt Loam
Redox
potential
200 ^
8 12 16
Time—Hours
20
20 i
<
El6
Q.
o.
1 12
c
01 8 '
>
X
0 4
«
, O2 added
^*v /^ Oxygen
X
\
\
\
^
\\
i i i I i *\ if i i i I i i i 1 i i i 1 i i i
24
Figure 25 (continued): Effect of added oxygen1on nitrate
reduction and redox potential in stirred suspensions.
75
-------
in a previous section, however, no change in NO, removal rate could be
detected at this point. The Crowley soil which was thoroughly oxidized
at the beginning of incubation showed a pattern similar to that of the
anaerobic soils.
Soil 1 which, had previously shown the most rapid reduction of 02 and
NO., also showed the same trend in this experiment. Demand for 0- was so
rapid that an inflection in the redox potential curve could not be de-
tected. There was, however, about a 75 mV higher potential noted in the
oxygenated series. Soils 2, 3, and 4 also showed no distinct effect of
added 0 on N0~ removal from the suspension, however the redox potential
curves showed the characteristic inflection points as 0» disappeared.
Soils 3 and 4 exhibited highly poised redox potentials (about 200 mV
higher where 0_ was added) while soil 2 showed a difference in Eh of
about 75 mV between the 0 treated and untreated systems. The addition
of 16 ppm of 0 (twice that of average atmospheric 0_ saturated suspen-
sions) did not appear to inhibit NO, disappearance in any of the soil
suspensions, but did result in a characteristic inflection in the redox
potential curves. It should also be mentioned that this was a very small
amount of p? in terms of the reducing capacity of these anaerobic soils.
The agricultural or comparison soil (Crowley soil) had a much higher
initial redox potential which was characteristic of a very oxidized soil
system and contained significant atmospheric 0. at the beginning of in-
cubation as shown in the curve where no 0 was added. The same charac-
teristic inflection of redox potential at the time of 0_ depletion was
again demonstrated. As in the swamp soils, NO reduction was not sig-
nificantly affected by 0 addition.
76
-------
CONCLUSIONS
The experiments carried out in this study demonstrate that the
flooded swamp and marsh soils in the interior swamp and coastal marsh
areas of Louisiana have a high capacity for reduction of NO . Nitrate
added either to the shallow floodwater or mixed in the soil rapidly dis-
appeared under laboratory conditions. Nitrate reduction does not take
place in the floodwater, but is dependent on the NO" moving downward in-
to the anaerobic soil layer. The N0~ reduction capacity is much greater
than is required to handle any amount of NO~-N that is likely to be
present in these areas, either from oxidation of ammonium that has been
released from organic matter or from NO derived from local runoff from
adjacent agricultural areas. Nitrate reduction does not appear to be
inhibited by the presence of a small amount of molecular 02 in flooded
soils.
77
-------
SECTION VII
REFERENCES
Alberda, Th. 1953. Growth and root development of lowland rice and its
relation to oxygen supply. Plant Soil 5: 1-28.
Bremner, J. M. 1965. Isotope-ratio analysis of nitrogen-15 tracer
investigations. In: Methods of Soil Analysis, 1965. C. A. Black
(ed.). Amer. Soc. Agron., Madison, Wis. pp. 1256-1286.
Bremner, J. M., and K. Shaw. 1958. Denitrification in soil. I.
Methods of investigation. J. Agr. Sci. 51: 22-39.
Broadbent, F. E., and M. E. Tusneem. 1971. Losses of nitrogen from
some flooded soils in tracer experiments. Soil Sci. Soc. Am. Proc.
35: 922.
Engler, R. M., and W. H. Patrick, Jr. 1974. Nitrate removal from
floodwater overlying flooded soils and sediments. J. Envir. Qual.
3: 409-413.
Gee, J. C. de. 1950. Preliminary oxidation potential determination
in a "sawah" profile near Bogar (Java). Contrib. Gen. Agr. Research
Bogar 106 (Cited by Alberda, 1953).
Howeler, R. H., and D. R. Bouldin. 1971. The diffusion and consump-
tion of oxygen in submerged soils. Soil Sci. Amer. Proc. 35: 202-208.
Jackson, M. L. 1958. Soil Chemical Analysis. Prentice-Hall, Inc.,
Englewood Cliffs, N. J. p. 498.
Mitsui, S. 1954. Inorganic Nutrition, Fertilization, and Amelioration
for Lowland Rice. Yodendo Lt., Tokyo.
Mortimer, C. H. 1941. The exchange of dissolved substances between
mud and water in lakes. J. Ecol. 29: 280-329.
Mortimer, C. H. 1942. The exchange of dissolved substances between
mud and water in lakes. J. Ecol. 30: 147-201.
Patrick, W. H., Jr. 1960. Nitrate reduction rates in a submerged soil
as affected by redox potential. Trans. Int. Congr. Soil Sci. 7th,
Madison, Wis. II: 494-500.
78
-------
Patrick, W. H., Jr., and R. D. DeLaune. 1972. Characterization of the
oxidized and reduced zones in flooded soil. Soil Sci. Soc. Amer. Proc.
36: 573-576.
Patrick, W. H., Jr., and I. C. Mahapatra. 1968. Transformation and
availability to rice of nitrogen and phosphorus in waterlogged soils.
Advan. Agron. 20: 323-359.
Patrick, W. H., Jr., and M. E. Tusneem. 1972. Nitrogen loss from
flooded soil. Ecology 53: 736.
Pearsail, W. H. 1938. The soil complex in relation to plant communi-
ties. I. Oxidation-reduction potentials in soils. J. Ecol. 26: 180-
193.
Pearsail, W. H. 1950. The investigation of wet soils and its agricul-
tural implications. Emp. J. Exp. Agr. 18: 289-298.
Pearsall, W. H., and C. H. Mortimer. 1939. Oxidation-reduction poten-
tials in waterlogged soils, natural waters and muds. J. Ecol. 27:
483-501.
Phillips, R. E., and D. A. Brown. 1964. Ion exchange diffusion II.
Calculation and comparison of self- and counter-diffusion coefficients.
Soil Sci. Soc. Amer. Proc. 28: 758-63.
Russell, E. W. 1961. Soil Conditions and Plant Growth. Ed. 9. John
Wiley and Sons, Inc., New York, N.Y.
Turner, F. T., and W. H. Patrick, Jr. 1968. Chemical changes in water-
logged soils as a result of oxygen depletion. Trans. 9th Intern. Congr.
Soil Sci. (Adelaide, Australia) IV: 53-65.
Tusneem, M. E., and W. H. Patrick, Jr. 1971. Nitrogen transformations
in waterlogged soil. La. Agr. Expt. Sta., Louisiana State University,
Bull. 657, pp. 735-737.
79
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/3-76-042
2.
3. RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
Nitrate Removal from Water at the Water-Mud
Interface in Wetlands
5. REPORT DATE
April 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
W. H. Patrick, Jr., R. D. Delaune, R. M. Engler and
S. Gotoh
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Louisiana Agricultural Experiment Station
Louisiana State University
Baton Rouge, Louisiana 70803
10. PROGRAM ELEMENT NO.
XH1627
11. CONTRACT/GRANT NO.
R-800428
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Corvallis Environmental Research Laboratory
200 SW 35th Street
Corvallis, Oregon 97330
13. TYPE OF REPORT AND PERIOD COVERED
Final 10/70 - 3/75
14. SPONSORING AGENCY CODE
EPA/ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The oxidized and reduced layers in flooded soil were characterized by vertical dis-
tribution of the oxidation-reduction (redox) potential and concentrations of manganom
manganese, ferrous iron, sulfide, nitrate and ammonium. The apparent thickness of
the oxidized layer was different when evaluated by the distribution of these various
components in the profile. Flood water NOj removal rates of intermittently-flooded
freshwater swamp soils and continuously-flooded saline marsh soils indicated that the
area of NC^reduction was in the soil, added organic matter increased the rate of NOo
reduction and the reduction rate was approximately twice as fast in the marsh soil as
in the swamp soil. Atmospheric 02 over a flooded soil increased denitrification up
to a concentration of approximately 20%. The N loss appeared to be related to the
thickness of the sediments aerobic layer. Ammonium nitrogen in a flooded soil
exposed to 02 from the water column undergoes sequential nitrification and denitrifi-
cation. Ammonium nitrogen is nitrified in the aerobic zone, diffuses to the anaerobic
zone where it is denitrified to ^ and ^0 and then diffuses from the system. Oxygen
reduction rates, NO^ reduction rates and the effects of added 02 on N03 reduction and
redox potential in flooded soils were investigated and indicated that N03 added to
floodwater rapidly disappeared, ©2 loss from sediments occurred rapidly and NOg
reduction was not inhibited by up to 16 ppm 02 dissolved in soil suspensions.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Nitrate*
Ammonium*
Oxidation-Reduction*
Nitrification*
Denitrification*
Soils*
Marsh
Swamp
Microbial
Diffusion
Organic Matter
Oxygen
02D
06ACF
07B
08H
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
88
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
80
ft U. S. GOVERNMENT PRINTING OFFICE: 1976-697.054184 REGION 10
------- |