Denitrification in wetlands as a means of water quality improvement

MISSING IMAGE

Material Information

Title:
Denitrification in wetlands as a means of water quality improvement
Series Title:
Florida Water Resources Research Center Publication Number 48
Physical Description:
Book
Creator:
Graetz, D. A.
Krottje, P. A.
Erickson, N. L.
Fiskell, J. G. A.
Rothwell, D. F.
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Notes

Abstract:
In recent years, much attention has been focused on the use of wetlands for nutrient removal from polluted waters such as secondarily treated sewage effluent and agricultural drainage waters. In the case of non-point sources such as agricultural runoff, wetland treatment may be the only feasible means of pollution abatement. Nitrification and subsequent denitrification can be an important route for the removal of nitrogen from water flowing through a wetland ecosystem. The objectives of this research were (1) to determine the extent of nitrification and denitrification occurring in a marsh ecosystem receiving secondarily treated sewage effluent, and (2) to measure denitrification and nitrous oxide evolution rates in a variety of flooded wetland soils. Nitrification and denitrification rates were measured in simulated marsh ecosystems, i.e., soil :water columns containing approximately 30 cm of marsh soil and either 15 or 30 cm of overlying water (secondarily treated sewage effluent). Nitrification was apparently rate limiting when nitrogen was supplied in the form of ammonium. Therefore, efficiency of nitrogen removal would be increased by allowing the nitrogen in wastewater to nitrify prior to entering the marsh. First-order denitrification was observed in soil :water columns. Average nitrate removal rates, assuming a floodwater nitrate concentration of 10 mg N liter-1, were 1.2 kg N ha-1day-1 without plants and 2.0 kg N ha-1day-1 with plants. Denitrification potentials of Florida wetland soils were determined using 15 soils with a wide range of texture, organic matter content, and soil reaction (pH). Denitrification followed apparent first-order kinetics, with rate constants varying from 0.040 day-1 to 0.192 day-1. Nitrate removal rates, assuming a 10 mg N liter-1 floodwater nitrate concentration, ranged from 0.6 to 2.9 kg N ha-1day-1. Denitrification rate was significantly correlated with soil organic carbon content and with soiI pH. Nitrous oxide accounted for from less than 0.2 to 6.5% of nitrate consumed. The experimental data obtained in this work indicate that many Florida wetlands could function as efficient treatment systems for nitrate-bearing wastewater and drainage water. Soils high in organic matter with a pH range of 5 to 7 would provide the most rapid removal rates.

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the source institution and holding location.
System ID:
AA00001522:00001


This item is only available as the following downloads:


Full Text







ji


Publication No. 48


DENITRIFICATION IN WETLANDS AS A

MEANS OF WATER QUALITY IMPROVEMENT


By


D. A. Graetz
(Principal Investigator)


Soil Science Department, IFAS
University of Florida
Gainesville


i 'Fi ,W t' ." s'k ..* -' '" -' S


r
















Denitrification in Wetlands


as a Means of Water Quality Improvement



By



D. A. Graetz, P. A. Krottje, N. L. Erickson,

J. G. A. Fiskell and D. F. Rothwell



PUBLICATION NO. 48

FLORIDA WATER RESOURCES RESEARCH CENTER

RESEARCH PROJECT TECHNICAL COMPLETION REPORT

OWRT Project Number B-035-FLA



Matching Grant Agreement Number
14-34-0001-8073


Report Submitted:


May, 1980


The work upon which this report is based was supported in part
by funds provided by the United States Department of the
Interior, Office of Water Research and Technology
as Authorized under the Water Resources
Research Act of 1964 as amended.













ACKNOWLEDGEMENTS


The authors wish to extend appreciation to the Office of Water Re-
search and Technology, United States Department of the Interior, for
financial support of this work. The administrative assistance of Dr.
W. H. Morgan (former Director) and Dr. James Heaney (present Director)
of the Florida Water Resources Research Center is greatly appreciated.
Thanks are also due Ms. Mary Robinson for accounting assistance. Sup-
plemental financial support of the Center of Environmental Programs,
Institute of Food and Agricultural Sciences, University of Florida is
gratefully acknowledged.

Thanks are also due Drs. S. E. Bayley and J. Zoltek, Jr. for allow-
ing us to cooperate on their study of the Clermont marsh, and to Dr.
K. R. Reddy for his constructive advice and assistance. The assistance
of U.S.D.A. Soil Conservation Service personnel in locating sampling
sites was extremely helpful. Special thanks go to Candace Cantlin and
Diane Sompongse for their assistance in the field and in the laboratory.











TABLE OF CONTENTS


ACKNOWLEDGMENTS . . . . .

LIST OF TABLES . . . . .

LIST OF FIGURES . . . . .

ABSTRACT . . . . .

INTRODUCTION . . . . .

CHAPTER I LITERATURE REVIEW . . .

Wetlands as Wastewater Treatment Systems .

Nitrification . . . .

Denitrification . . . .

Denitrification Kinetics: Real and Apparent

Additional Mechanisms for Nitrate Loss .

Nitrous Oxide Production . . .

Summary . . . . .

CHAPTER II NITRIFICATION AND DENITRIFICATION IN
RECEIVING SECONDARILY TREATED SEWAGE

METHODS AND MATERIALS . . .

Soil and Water Sampling . . .

Description of Experiments . . .

Analytical Procedures . . .

RESULTS AND DISCUSSION . . .

Nitrification . . . .

Denitrification . . . .

SUMMARY AND CONCLUSIONS . . .


CHAPTER III DENITRIFICATION

METHODS AND MATERIALS .


. .


. .

. .



. .

. .

. .



. .


. .








A MARSH SO
EFFLUENT







. .


. .


Page

i

v




vii




2

2

2

5

7

9

11

12


IN FIFTEEN FLORIDA WETLAND SOILS

. . . . .


37

37








Page

Soils and Soil Sampling . . . . 37

Descriptions of Experiments . . . .. 42

Method of Measuring N20 Evolution . . . 43

Analytical Methods . . . . . 43

RESULTS AND DISCUSSION . . . . 46

Comparative Denitrification Rates . . . 46

Nitrous Oxide Evolution . . . . 56

Long Term Denitrification and Nitrous Oxide Evolution 61

SUMMARY AND CONCLUSIONS . . . . 66

LITERATURE CITED . . . . . . 68










LIST OF TABLES


Page

1. Effect of charcoal filtration and inoculation with soil 19
water extract on nitrification in marsh and lake waters
without soil.

2. Ammonium removal rates from overlying water of soil:water 24
columns containing ammonium amended sewage effluent.

3. Nitrate removal as affected by water source, associated 28
cation and presence of soil.

4. Nitrate removal rates from overlying water of soil:water 29
columns containing nitrate amended sewage effluent.

5. Comparative denitrification rates of 14 representative 53
Florida wetland soils.

6. Selected chemical and physical characteristics of surface 54
horizons (top 10 cm) of the soils studied.

7. Denitrification rate constants as predicted by Equation 5. 57

8. Nitrous oxide evolution rates from 14 representative 60
Florida wetland soils.

9. Long term (8 week) denitrification rates in Samsula and 63
Pelham soil columns.










LIST OF FIGURES


Page

1. Nitrate diffusion fronts in flooded soil systems (a) with 8
no aerobic zone, and (b) with aerobic zone.

2. Experimental soil:water columns for nitrification and 16
denitrification studies.

3. Nitrification in soil:water columns flooded with 15 cm of 21
ammonium amended sewage effluent, with [o] and without [*]
plants.

4. Nitrification in soil:water columns flooded with 30 cm of 22
ammonium amended sewage effluent, with [o] and without [*]
plants.

5. Nitrification in limed soil:water columns flooded with 23
15 cm of ammonium amended sewage effluent, with [o] and
without [*] plants.

6. Ammonium and nitrate concentrations in ammonium amended 26
marsh soil during in situ incubation in polyethylene bags.

7. Nitrate removal from overlying water of soil:water columns 30
(with [o] and without [*] plants) containing nitrate amended
sewage effluent to a depth of 15 cm. Nitrate was added on
day 0 and day 18.

8. Nitrate removal from overlying water of soil:water columns 31
(with [o] and without [*] plants) containing nitrate amended
sewage effluent to a depth of 30 cm. Nitrate was added on
day 0 and day 18.

9. Nitrate removal from overlying water of limed soil:water 32
columns (with [o] and without [*] plants) containing
nitrate amended sewage effluent to a depth of 15 cm.
Nitrate was added on day 0 and day 18.

10. Nitrate and ammonium concentrations in nitrate amended 34
marsh soil during in situ incubation in polyethylene bags.

11. Location of soil sampling sites. 38

12. Apparatus for obtaining soil cores with minimum soil 41
compaction.

13. System for determining nitrous oxide evolution from flooded 44
soil cores.









LIST OF FIGURES
(continued)
Page

14. Changes in floodwater nitrate concentrations with time 47
in soil:water columns of 14 soil types.

15. Changes in floodwater nitrate concentrations with time 51
in two soil:water columns of Surrency soil.

16. Nitrous oxide evolution from flooded soil cores in 59
relation to floodwater nitrate concentrations.

17. Relationship between reciprocal of denitrification rate 62
constant and percent nitrous oxide evolution for 14
soil types.

18. Long term (10 week) nitrous oxide evolution from flooded 65
Samsula and Pelham soil columns.









ABSTRACT


In recent years, much attention has been focused on the use of wet-
lands for nutrient removal from polluted waters such as secondarily
treated sewage effluent and agricultural drainage waters. In the case
of non-point sources such as agricultural runoff, wetland treatment may
be the only feasible means of pollution abatement. Nitrification and
subsequent denitrification can be an important route for the removal of
nitrogen from water flowing through a wetland ecosystem. The objectives
of this research were (1) to determine the extent of nitrification and
denitrification occurring in a marsh ecosystem receiving secondarily
treated sewage effluent, and (2) to measure denitrification and nitrous
oxide evolution rates in a variety of flooded wetland soils.

Nitrification and denitrification rates were measured in simulated
marsh ecosystems, i.e., soil:water columns containing approximately 30
cm of marsh soil and either 15 or 30 cm of overlying water (secondarily
treated sewage effluent). Nitrification was apparently rate limiting
when nitrogen was supplied in the form of ammonium. Therefore, efficiency
of nitrogen removal would be increased by allowing the nitrogen in waste-
water to nitrify prior to entering the marsh. First-order denitrification
was observed in soil:water columns. Average nitrate removal rates, assum-
ing a floodwater nitrate concentration of 10 mg N liter-1, were 1.2 kg N
ha-lday-I without plants and 2.0 kg N ha-lday- with plants.

Denitrification potentials of Florida wetland soils were determined
using 15 soils with a wide range of texture, organic matter content, and
soil reaction (pH). Denitrification followed apparent first-order kinetics,
with rate constants varying from 0.040 day-1 to 0.192 day-1. Nitrate re-
nmoval rates, assuming a 10 mg N liter-1 floodwater nitrate concentration,
ranged from 0.6 to 2.9 kg N ha-lday-1. Denitrification rate was signifi-
cantly correlated with soil organic carbon content and with soil pH.
Nitrous oxide accounted for from less than 0.2 to 6.5% of nitrate consumed.

The experimental data obtained in this work indicate that many
Florida wetlands could function as efficient treatment systems for nitrate-
bearing wastewater and drainage water. Soils high in organic matter with
a pH range of 5 to 7 would provide the most rapid removal rates.









INTRODUCTION


Considerable concern exists over detrimental effects of human
activities on natural waters. One major problem is excessive nitrogen
input from such sources as domestic sewage, urban and agricultural run-
off, and certain industrial wastes. In waters where nitrogen is the
limiting nutrient, excessive input can accelerate eutrophication,
thereby detracting from the aesthetic, recreational, and commercial
value of the waters. Nitrogen pollution may also be a health hazard, as
nitrate-N levels above 10 ppm in water for human or animal consumption
are known to result in methemoglobinemia, a potentially fatal disorder.

A variety of methods exist for removal of nitrogen from wastewater
with tertiary treatment, including ammonia stripping and induced nitri-
fication-denitrification. These generally require substantial capital
investment and energy expenditure. Diversion of wastewater into wet-
lands has recently attracted interest as a low-cost, effective alter-
native for removal of nitrogen and other undesirable materials. In the
case of non-point sources of pollution such as agricultural runoff,
wetland treatment may be the only feasible means of pollution abatement.

Nitrate is an important form of nitrogen in wastewater. Denitri-
fication, the microbial reduction of nitrate to gaseous products (pri-
marily denitrogen and nitrous oxide), is among the mechanisms for
removal of nitrate by wetlands. An undesirable aspect of denitrifi-
cation is release of nitrous oxide into the atmosphere, as this compound
reacts with and destroys ozone in the stratosphere. Ozone is important
in shielding the earth from solar ultraviolet radiation. Depletion of
the stratospheric "ozone layer" and consequent high levels of ultra-
violet radiation would be likely to have a number of negative effects on
terrestrial lifeforms, including increased incidence of skin cancer in
humans.

Knowledge of wetland denitrification rates and soil characteristics
affecting these rates is incomplete, and knowledge of wetland nitrous
oxide production is almost nonexistent. It is hoped that the results
presented here may be useful in the evaluation of wetland sites for
wastewater treatment, and perhaps contribute to the development of
regional or global nitrogen budgets. Specific objectives include:

(1) To determine nitrification and denitrification rates in a
marsh soil receiving secondarily treated sewage effluent.

(2) To measure denitrification rates in a variety of flooded
wetland soils and to determine quantitative relationships
between denitrification rate and selected soil characteris-
tics in order to develop a predictive equation for denitri-
fication rates in specific soils.

(3) To investigate nitrous oxide evolution from wetland soils in
terms of amounts evolved and factors influencing evolution
rates.







CHAPTER I


LITERATURE REVIEW

Wetlands as Wastewater Treatment Systems

A variety of wetland ecosystems, including freshwater marshes
(Fetter et al., 1978; Zoltek et al., 1979), hardwood swamps (Boyt et
al., 1977), and cypress domes (Odum and Ewel, 1977) have been investi-
gated as wastewater treatment systems. Studies have included systems
with primarily downward percolation (Odum and Ewel, 1977), primarily
lateral flow (Boyt et al., 1977), and intermediate situations (Zoltek et
al., 1979). In general, wetlands have been quite efficient in removal
of nitrogen, phosphorus, B.O.D., and coliform bacteria from both second-
arily treated sewage and agricultural and urban runoff. Removal of 90%
or more of incoming nitrogen and phosphorus is not uncommon, although
much lower efficiencies have been reported (Fetter et al., 1978).

Denitrification, plant uptake, and microbial assimilation (immobil-
ization) are possible nitrogen removal mechanisms. The last two pro-
cesses can remove both ammonia and nitrate, the primary inorganic
nitrogen forms in wastewater. By definition, denitrification involves
only nitrate. Ammonia entering a wetland must therefore be nitrified
(biologically oxidized to nitrate) before it can be denitrified. Nitri-
fication can occur in the aerobic floodwater or surface soil zone and
the nitrate produced is subject to denitrification after diffusion into
the deeper, anaerobic soil (Patrick and Reddy, 1976). The kinetics of
this situation are quite complex, and either nitrification or denitri-
fication may be rate limiting (Reddy and Graetz, 1980).

A significant portion of the nitrogen entering a wetland may
already be in the nitrate form (Fetter et al., 1978; Boyt et al.,
1977). In one study, an artificial wetland treatment system was speci-
fically designed to maximize nitrification during flow of runoff through
pasture prior to entering the actual wetland (Small, 1978). Thus, the
nitrogen burden on the wetland was mostly nitrate, with denitrification
the primary removal mechanism.


Nitrification

Nitrification is a biologically mediated two-step reaction in which
ammonium is first oxidized to nitrite and then to nitrate (Alexander,
1977). The organisms responsible for these reactions are chemoauto-
trophs and therefore are not dependent upon organic carbon, but rather
use carbon dioxide as the carbon source. Nitrifier numbers have been
found to vary from zero to a million organisms per gram of soil, depend-
ing on conditions. Their numbers are generally less in water samples,
with ammonium oxidizers ranging from zero to a few thousand and nitrite
oxidizers from zero to a few hundred (Matulewich et al., 1975). Nitri-
fiers are reported to decrease to very low numbers in climax ecosystems
(Rice and Pancholy, 1973). These low numbers have been partially
explained by increasing concentrations of tannins (Rice and Pancholy,
1978) and other carbon containing compounds (EPA, 1973; EPA, 1975) that
have been reported to have an inhibitory effect, although this has









been debated. Alexander (1977) attributed this "inhibition" to the
depletion of inorganic nitrogen (i.e., ammonium and nitrate) which is
assimilated by the proliferating heterotrophic microorganisms decom-
posing the carbonaceous compounds. Free ammonia at high pH values has
been found inhibitory (Alexander, 1977) as well as accumulation of
certain trace metals (Liang and Tabatabai, 1978) and certain pesticides
(Alexander, 1965).

Although autotrophic microorganisms are most often responsible for
nitrification, heterotrophic nitrification has recently been observed.
Results indicate that under appropriate conditions of pH, sufficient
carbon and ammoniacal nitrogen (and low competition), heterotrophic
bacteria (and actinomycetes) can produce small quantities of nitrite as
well as other more oxidized inorganic and organic nitrogen compounds,
i.e., hydroxylamine, 1-nitrosoethanol (Alexander, 1977; Verstraete and
Alexander, 1973). Heterotrophic nitrate formation is usually minimal but
it has been shown that certain fungi (in culture) are capable of oxidiz-
ing nitrite to nitrate. Recently, work done by Tate (1977) on an
organic soil in South Florida with a pH of 7.1, showed that the popula-
tions of Nitrosomonas and Nitrobacter autotrophss) were only sufficient
to account for 0.1% of the nitrate present. With selective culturing
techniques, large populations of Arthrobacter (heterotrophs) were found.
When these bacteria were inoculated into sterile soil, nitrite was
produced. Thus, it appears that in some cases, heterotrophic nitrifi-
cation can be an important source of nitrate, but the general signi-
ficance of this process has not yet been determined.

Nitrification rates are very sensitive to environmental conditions.
The most obvious consideration is that the nitrifiers exist under
aerobic conditions with ammonium present. In flooded soils, oxygen
supply is a potential limiting factor for nitrification since few
nitrifiers are present in the overlying aerobic waters and oxygen
diffusion into the soil is slow. White et al. (1977) suggested the
coupling of photosynthetic oxygen production in surface water with
nitrification since nitrification rates were higher at midday than
morning and higher in July and August than any other months. Other work
by White et al. (1977) showed daytime nitrate production three times
that at night. Consequently, development of an oxidized surface soil
layer is crucial to the nitrification process which requires 4.3 mg of
0 to oxidize one milligram of ammonium nitrogen to nitrate (White et
a., 1977). Reddy et al. (1976) described the development of an oxi-
dized layer in initially anaerobic soil:water columns amended with 200
mg NH4-N per gram of soil. In the first week a thin aerobic soil layer
developed on top of the anaerobic soil, which increased to 0.5, 1.25,
1.5, 1.6, and 2.0 cm after 7, 15, 30, 60, and 120 days respectively, at
28 C. Patrick and Delaune (1972) also found the aerobic layer to in-
crease to 1.5 and 2.0 cm with time. The rate of development and thick-
ness of this layer is determined by the oxygen supply and amount of
readily decomposable organic matter, being slower and thinner with more
organic matter due to increased microbial activity and higher oxygen
demand (Engler and Patrick, 1974; Patrick and Reddy, 1976). The


-3-








small size of this aerobic layer has been implicated as the limiting
factor in nitrification (and consequently denitrification) in many
flooded soils (Patrick and Tusneem, 1972). Ammonium is supplied to the
oxidized soil layer by diffusion from eutrophic surface waters or from
the anaerobic subsurface soil by diffusion and physical disruption (by
benthic fauna and gas ebullition) (Chen et al., 1972). In experiments
using flooded columns packed with ammonium amended soils, it was re-
ported (Patrick and Reddy, 1976) that approximately half of the ammonium
nitrified was initially present in the aerobic surface layer of soil and
the remaining half diffused to the aerobic layer from the anaerobic
subsurface soil. The diffusion coefficients calculated in the above
experiments for ammonium moving in a saturated Crowley silt loam were
0.216 cm2/day (Patrick and Reddy, 1976; Reddy et al., 1976).

Reddy et al. (1976) reported that nitrification followed zero-order
kinetics (based on ammonium loss). Other work (based on nitrite plus
nitrate production) indicated a sigmoid relationship where the peak
production varied with the initial concentration of ammonium (Alexander,
1977).

Alexander (1975, 1977) reported a significant relationship between
nitrate production and pH. In neutral to slightly alkaline soils,
nitrifiers are present in the largest numbers. The optimum range for
nitrate production has been reported to vary from 6.6 to 9.0 (Alexander,
1977; Keeney and Gardner, 1970; EPA, 1975). At high pH values, ammonium
changes to ammonia which is detrimental to Nitrobacter populations, and
as a result, oxidation of nitrite to nitrate is inhibited. Lime has
been reported to stimulate nitrification (Alexander, 1977; Tilsdale and
Nelson, 1975) by serving as a pH buffer and a carbon source. On the
other end of the pH scale, nitrification has been reported below pH 4
but generally does not occur at a significant rate until a pH of 5.0 to
5.5 is reached (Alexander, 1977; Tilsdale and Nelson, 1975). The
nitrification process also lowers pH due to the release of hydrogen ions
as ammonium is converted stepwise to nitrate (Broadbent, 1973). The
extent of the lowering depends on the amount of ammonium nitrified and
the buffering capacity of the system.

Temperature has also been shown to have a marked effect on nitri-
fication with an optimum range between 30 and 350 C (Alexander, 1977;
Broadbent, 1973). Above 40 C and near freezing, the amount of nitrate
or nitrite produced becomes negligible (Tilsdale and Nelson, 1975).
White et al. (1977) reported nitrification (in stream bed sediments) to
cease when the water temperatures dropped below 17 C.

Since the nitrifiers are obligate aerobes, they are more sensitive
to excess moisture than dry conditions, although nitrate is not produced
at very low moisture levels. Chemical breakdown of nitrite (and release
as gaseous nitrous oxide) has also been implicated under well aerated,








low moisture conditions. Lance and Whisler (1972) showed the effect of
varying cycles of flooding and drying on nitrification of ammonium in
wastewater and found essentially all the ammonium was nitrified with
short frequent cycles: 2 days flooded followed by 5 days of drying.
Other factors such as season or depth in the profile are important only
as they are related to factors such as temperature, pH, oxygen content,
and nutrient availability.

Denitrification

Biological denitrification is generally regarded as the use of
nitrate and other oxidized forms of nitrogen as alternatives to oxygen
in the respiratory process. A large number of heterotrophic bacteria,
and a smaller number of autotrophs are capable of this process (Payne,
1973). Virtually all denitrifiers can respire aerobically, and some can
obtain energy through fermentation. Because of the diversity of deni-
trifying organisms, denitrifier populations are relatively high in most
soil habitats. With a few possible exceptions (Delwiche and Bryan,
1976), denitrification occurs only under at least partially anaerobic
conditions, since oxygen is used preferentially to nitrate when both are
present. In flooded systems, anaerobic conditions exist below a surface
layer of aerobic soil, the thickness of which is controlled by oxygen
concentration in the overlying water and oxygen demand of the soil
(Bouldin, 1968).

As with most biochemical processes, denitrification rate tends to
increase with rising temperature until heat induced enzyme deactivation
begins to occur (Focht, 1974). Slow denitrification has been reported
at 30C (Nommik, 1956), with rates increasing rapidly up to about 350 C
(Bremner and Shaw, 1958; Bailey and Beauchamp, 1973). Rates increase
somewhat more slowly from 35 to about 60-65 C, dropping off rapidly at
higher temperatures (Nommik, 1956; Bremner and Shaw, 1958).

Amount and source of available carbon is perhaps the most important
soil characteristic determining denitrification rate. Many reports link
rapid denitrification with high levels of organic matter, in both
flooded and unflooded systems (McGarity and Meyers, 1968; Stefanson,
1972; Engler and Patrick, 1974). Added organic material tends to
stimulate the process more than endogenous organic matter because the
latter is generally humified and thus resistant to degradation (Bremner
and Shaw, 1958; Reddy et al., 1978). Alternation of aerobic and anaero-
bic conditions has been found to increase denitrification by increasing
carbon decomposition (Lance and Whisler, 1972; Reddy and Patrick, 1975).

At least three studies have attempted to establish a quantitative
relationship between organic carbon and denitrification. Stanford et
al. (1975a) related hot water extractable carbohydrate to denitrifi-
cation rate for 30 widely varying soils. An r2 value of 0.82 was
obtained for carbohydrate vs. apparent first-order denitrification rate,
while total organic carbon vs. apparent first-order denitrification rate
gave an r2 of 0.69. Burford and Bremner (1975) studied 17 soils and
obtained a 0.99 r2 value for zero-order denitrification rate vs. "miner-
alizable" carbon as determined by CO2 evolution. Total organic carbon








yielded only a 0.59 r2 value. In both of these studies, denitrification
was assessed in batch studies under total anaerobiosis. The relevance
to in situ denitrification rates may therefore be limited because oxygen
is usually present to some degree in natural systems. Andersen (1976)
studied nitrate loss from aerobic floodwater overlying five lake sedi-
ment cores. Correlating "available carbon" (determined by 02 consump-
tion) with denitrification, he obtained r2 values of 0.92 and 0.70 for
low (2-3 ppm) and high (10 ppm) nitrate-N concentrations, respectively.
These lower correlations than those obtained by Burford and Bremner
(1975) may reflect such complicating factors as nitrate transport and
the presence of oxygen.

Plant roots have been widely reported to enhance denitrification,
particularly in soils with low endogenous carbon (Stefanson, 1972;
Bailey, 1976). This has been attributed to increased organic matter or
lower oxygen levels in the root zone, or both (Volz et al., 1976;
Alexander, 1977). Sherr and Payne (1978) conducted an interesting study
on the effect of Spartina alterniflora roots on denitrifying activity.
Above ground parts of the plants were cut and roots pruned in order to
terminate the below ground effects of living plants. Denitrifying
activity was found to be unaffected five months after perturbation, but
significantly lowered after 18 months. This was taken to indicate that
root exudates were relatively unimportant and that the main effect of
roots was due to long term maintenance of underground biomass. Another
important implication of this study is that root pruning, such as would
occur in the taking of soil cores, may have little short term effect on
denitrification.

There is a definite relationship between soil pH and denitrifi-
cation, but the strength of this relationship is subject to question.
A number of reports indicate that nitrate loss rates are highest between
pH 7 and 8, and drop off considerably below pH 6 (Wijler and Delwiche,
1954; Nommik, 1956; Bremner and Shaw, 1958). Some of these reports
(Nommik, 1956; Bremner and Shaw, 1958) are suspect because pH variation
was accomplished by amendment with acid or base, resulting in non-
adapted microbial populations. Brezonik (1977) feels that the variety
and adaptability of denitrifying organisms reduces the importance of pH
in determining nitrate consumption rates. There are several reports of
relatively rapid denitrification at pH values approaching 5.0 (Van
Cleemput et al., 1975) and even below 5.0 (Ekpete and Cornfield, 1965;
Erickson, 1978; Gilliam and Gambrell, 1978).

A factor which has not been widely investigated is duration of
exposure of the soil to nitrate. As with any microbial substrate,
addition of nitrate to soil results in certain adaptive and selective
changes in the soil population (Nash and Bollag, 1974; Doner and McLaren,
1978). Insufficient nitrate adaptation time has been invoked as the
cause of irregular nitrate consumption rates in short term studies
(Starr and Parlange, 1975). In one of the very few truly long term
denitrification experiments, a soil column was leached with a solution
of 100 ppm nitrate-N for over 1,000 days (Day et al., 1978). Nitrate
removal rate in the top 12.5 cm was found to decrease several fold over
the first 30 days, and more slowly thereafter. This effect was









attributed to depletion of available organic matter. Such a dramatic
effect might not occur under other experimental conditions. Leaching
tends to maximize organic matter loss through removal of the soluble
fraction (Doner et al., 1974) and would be expected to maximize oxida-
tion by continually and rapidly resupplying nitrate. Reddy found
relatively little change in nitrate removal rate from floodwater over-
lying an organic soil during a 25 day period (K. R. Reddy, personal
communication).

Denitrification Kinetics: Real and Apparent

The literature regarding denitrification kinetics is somewhat
confusing, with zero-order, first-order, and Michaelis-Menton kinetics
reported. Zero-order kinetics are reported most frequently (Nommik,
1956; Broadbent and Clark, 1965; Cooper and Smith, 1965; Van Cleemput et
al., 1975). Some authors (Focht, 1974; Doner et al., 1974) regard these
reports to be indicative of Michaelis-Menten kinetics with very low Km
values, rather than true zero-order kinetics. In sewage treatment
systems, Michaelis-Menten kinetics with Km values of about 0.05 ppm
nitrate-N have been reported for denitrification (Requa and Schroeder,
1973; Stensel et al., 1973). Doner et al. (1974) reported "very low" Km
values for soil column studies but did not specify numbers. A much
higher Km value of 125 ppm nitrate-N has been reported for a continuous
flow soil column (Ardakani et al., 1975). This unlikely value may be
due to failure to achieve a true steady state or failure to account for
diffusion of nitrate into microsites (Galsworthy et al., 1978).

First-order denitrification kinetics have also been widely re-
ported, primarily with experimental designs involving floodwater over-
lying soil (Bowman and Focht, 1974; Stanford et al., 1975). The rela-
tionship between presence of floodwater and first-order kinetics prompted
a thorough study by Phillips et al. (1978) and Reddy et al. (1978).
They determined both experimentally and theoretically that zero-order
nitrate depletion in the soil together with continuous resupply by
diffusion from floodwater results in what appears to be a first-order
nitrate loss rate. The initial boundary conditions for the model of
Phillips et al. (1975) involve an even distribution of nitrate in
floodwater and soil solution. In such a situation, the nitrate diffusion
front does not reach an equilibrium, and a quasi-steady state is not
attained until after the majority of nitrate in the system has been
denitrified. For this reason the model is applicable to batch type
studies where the above condition is met, but not to column studies
where initial soil nitrate concentration is zero.

Bouldin and coworkers addressed the conditions which would occur in
soil columns more directly. First, a model was developed for oxygen
consumption by sediments (Bouldin, 1968). Applying Fick's diffusion
laws and assuming zero-order oxygen consumption within the sediment, it
was determined that the second derivative of oxygen concentration with
respect to depth is a constant. This results in a parabolic shaped,
steady state diffusion front as shown in Figure 1(a). Since oxygen
demand is constant in the volume of sediment exposed to oxygen, total














Nitrate concentration -;


water


(a)


soil


Nitrate concentration-+


( b)


Figure 1. Nitrate diffusion fronts in flooded soil systems (a) with
no aerobic zone, and (b) with aerobic zone.









consumption is proportional to depth of penetration. Both are pro-
portional to the square root of oxygen concentration in the floodwater
and also to the square root of volumetric sediment oxygen demand, as in
the following equation:

dQ = (2CAD)2 [1]
dt
where Q is quantity of oxygen (mg) crossing the soil:water interface; C
is floodwater oxygen concentration, mg liter-1; A is rate of oxygen
consumption, mg cm-3day- ; D is oxygen diffusion coefficient, cm day-1;
and t is time, in days. These relationships were later verified experi-
mentally (Howeler and Bouldin, 1972).

The above model would be expected to apply to denitrification if it
were assumed that nitrate concentration in the floodwater decreased
fairly slowly (as would be the case with all but very shallow water
depths) and that nitrate consumption is uniform from the soil:water
interface downward. The second assumption is not valid when oxygen is
present in the water, since denitrification is theoretically zero in the
aerobic soil layer. Bouldin et al. (1974) addressed this situation by
developing a model which assumes no nitrate consumption in the aerobic
zone and maintenance of nitrate concentration at zero at the aerobic:
anaerobic interface. In this case, the second derivation of concen-
tration with respect to depth is zero, and the diffusion front is as
shown in Figure 1(b). The total consumption is now directly propor-
tional to floodwater concentration (i.e., apparent first-order) and
directly proportional to the thickness of the aerobic zone. The assump-
tion that nitrate reduction is rapid enough to maintain zero concen-
tration at the aerobic:anaerobic interface is obviously and admittedly
(Bouldin et al., 1974) erroneous, and is adopted for the sake of simpli-
fying the model. As a result, there is no allowance in the model for
volumetric nitrate consumption rate. This problem may be insignificant
since oxygen consumption rate and denitrification rate are closely
related. A truly accurate model would involve a combination of Figures
1(a) and 1(b) and would be somewhat complex, mathematically. A number of
investigators (Van Kessel, 1976; Reddy et al., 1979) have obtained high
r2 values when applying apparent first-order denitrification kinetics to
nitrate consumption data under boundary conditions similar to those
assumed by the model of Bouldin et al. (1974), indicating that this
model may be adequate. Apparently no attempts to correlate nitrate loss
to square root of floodwater concentration (as in the 1968 Bouldin
model) have been published.

Additional Mechanisms for Nitrate Loss

Although reduction of nitrate to nitrite in nature is thought to be
a solely biological process, under certain conditions nitrite can be
chemically decomposed into gaseous products. Four major types of
chemodenitrification are generally recognized: (1) self-decomposition
of nitrite into NO; (2) reaction of nitrite with ammonia, urea or amino
acids yielding N2 (Van Slyke reaction); (3) reaction with organic matter
yielding mostly N2; and (4) reaction with heavy metals resulting









in NO, N20 or N2 (Broadbent and Clark, 1965). The first three require
a pH in the vicinity of 5.0 or lower. The last can occur at a pH as
high as 8.0 (Moraghan and Buresh, 1977). There is evidence for the
occurrence of all of these reactions in sterile, nitrite amended soils
(Van Cleemput et al., 1976; Nelson and Bremner, 1970). Assessment of
the natural significance of chemodenitrification is difficult due to
inability to inhibit biological nitrite decomposition without inhibiting
biological nitrite production (Van Cleemput et al., 1975). However, it
is generally felt that chemodenitrification is insignificant in nature
(Burford and Stefanson, 1973; Focht, 1974).

Dissimilatory reduction of nitrate to ammonia is analogous to
denitrification, but the end product is ammonia rather than gaseous N
compounds. As in denitrification, the reduction is for the purpose of
obtaining respiratory energy rather than obtaining cell nitrogen. This
process was first described over 20 years ago (Verhoeven, 1956) but has
only recently received much attention. Although populations capable of
this process may be large (Tiedje et al., 1979), the process tends to
occur only under intensely reduced conditions (Buresh and Patrick,
1978). Since nitrate is produced under oxidizing conditions, the
majority of nitrate would likely be denitrified before reaching condi-
tions sufficiently reducing for ammonia production. For this reason,
Buresh and Patrick (1978) feel that dissimilatory reduction to ammonia
is insignificant in flooded soils. Sorensen (1978) feels that this
process may be of equal significance to denitrification in marine
sediments.

A significant portion of the nitrate entering soils can be assimi-
lated by microorganisms and converted to organic nitrogen. Assimilation
of added nitrate is usually low when soil C:N ratio is narrow, since
nitrogen assimilation needs would be met by mineralized N. For this
reason, assimilation is lower with a well humified carbon source than
with fresher material (Alexander, 1977).

Reports of assimilation under saturated conditions range from about
10% to 37% of applied nitrate (Andersen, 1976; Chen et al., 1972; Chen
and Keeney, 1974). There is some indication that percent assimilation
is related to concentration of added nitrate, 37%, 14-26%, and 17%
having been reported for concentrations of 2 ppm-N (Chen et al., 1972),
10 ppm-N (Chen and Keeney, 1974), and 100 ppm-N (Stanford et al., 1975b),
respectively. This would be expected in view of the relatively low
assimilation needs and high respiration needs of anoxic communities
(Tusneem and Patrick, 1971). Doner et al. (1974) found that net assimi-
lation was very low in relatively long term column studies with contin-
uous nitrate supply. It was felt that once rapid increase in microbial
biomass ceased and stable population developed, assimilation became
nearly zero. An increase in assimilation relative to denitrification at
low temperatures has been reported (Craswell, 1978), possibly indicating
that denitrification is more affected by cold than is assimilation.









Nitrous Oxide Production


The amount of nitrous oxide (N20) evolved during denitrification
varies greatly. Ratios of N2:N20 can range from greater than 100:1
(Denmead, 1979) to less than 1:10 (Stefanson, 1972). The majority of
reports are in the vicinity of 10:1, and the Council for Agricultural
Science Technology (C.A.S.T., 1975) has estimated an average worldwide
ratio of 16:1 for agricultural soils.

Such factors as high organic matter content, high moisure content,
fine soil texture, and low oxygen concentration have all been shown to
decrease N20 emission (Nommik, 1956; Stefanson, 1972; Focht and Ver-
straete, 1977), indicating that conditions of high electron acceptor
demand result in more complete reduction of nitrate to N2. Guthrie and
Duxbury (1978) reported an exception to this trend with lower N2:N20
ratios occurring in flooded histosols than in the same soil when drained.
Ratios of N2:N20 have been reported to decrease at low temperatures
(Bailey and Beauchamp, 1973).

Nitrous oxide production tends to increase with decreasing pH
(Wijler and Delwiche, 1954; Nommik, 1956; Burford and Bremner, 1978).
Blackmer and Bremner (1978) have shown that this may be due to an
indirect effect of pH. These investigators reported that high nitrate
concentrations inhibited N20 reduction, thereby increasing emission of
this gas. The effect was greater at low pH than at high pH. In the
absence of nitrate, N20 was reduced as rapidly at low pH as at high pH.
It was concluded that the effect of pH was to control the degree of
nitrate inhibition of N20 reduction.

Prior exposure to denitrifying conditions (i.e., anoxia and nitrate)
can also influence N2:N20 ratios. Blackmer and Bremner (1979) observed
that nitrate application resulted in an increase in N2:N20 ratio over a
48 hour period, presumedly due to induction of N20 reduction activity in
the denitrifier population. Guthrie and Duxbury (1978) noted indica-
tions of a more long term increase in N2:N20 ratio due to population
selection in nitrate-amended soil.

Nitrous oxide can be produced from processes other than denitri-
fication. As mentioned above, chemodenitrification can potentially
result in N20 production, but the extent of this in nature is unknown.
Yoshida and Alexander (1970) reported that axenic cultures of Nitro-
somonas can produce N20 from nitrite during nitrification of ammonia.
This had been thought to be insignificant in nature (Focht, 1974) until
recently when it was found that considerable N20 production can occur
during nitrification in aerobic soils (Bremner and Blackmer, 1979;
Denmead et al., 1979b). Dissimilatory reduction of nitrate to ammonia
can also result in N20 production in pure culture (Tiedje et al., 1979),
but this has not been demonstrated in situ.


-11-








Summary

Pertinent literature on nitrification, denitrification and N20
production in soils can be summarized as follows:

(1) Many of the factors affecting nitrification and denitrifi-
cation in soils are known in a general, qualitative way.
Nitrification is favored by high oxygen availability and
near neutral pH, and may be inhibited by certain organic
substances. Poor aeration, high organic carbon availability
and neutral pH tend to accelerate denitrification. In
flooded systems, transport of nitrate into the anerobic
zone is an important rate-determining factor which many
studies have ignored.

(2) In wetland systems, the potential exists for rapid nitri-
fication of incoming ammonia in the upper, aerobic soil zone,
and possibly in the water column.

(3) It is likely that denitrification accounts for the majority
of nitrate consumption in wetland soils, particularly under
conditions of heavy nitrate enrichment.

(4) Relatively little is known about rates of N20 production
during denitrification. Average reported N2:N20 ratios
are in the vicinity of 10:1. Low electron acceptor demand
and low pH tend to favor N20 production. Soil processes
other than denitrification, including nitrification and
dissimilatory reduction of nitrate to ammonia, may be
important N20 sources, at least under certain conditions.

(5) The data base for nitrification, denitrification, and
N20 production in wetlands is at present quite narrow.


-12-









CHAPTER II


NITRIFICATION AND DENITRIFICATION IN A MARSH SOIL
RECEIVING SECONDARILY TREATED SEWAGE EFFLUENT


The objective of this study was to determine the feasibility of
using natural wetlands to remove nitrogen via denitrification from
secondarily treated municipal sewage effluent. Since nitrogen in the
form of nitrate is a prerequisite for denitrification, nitrification was
included as part of the study. Nitrogen removal was investigated in
laboratory and in situ experiments on a freshwater marsh in Central
Florida.

METHODS AND MATERIALS

Soil and Water Sampling


Soil samples were obtained from a 32 hectare freshwater marsh
located in the town of Clermont, Florida, about 100 miles (160 kilo-
meters) south of Gainesville on Route 27. The soil was classified as a
variant of Brighton peat typicc medifibrist) characterized by 1.5 meters
of relatively undecomposed, acidic, organic soil overlying 1.5 meters of
sand over clay of an undetermined depth.

Soil used in "batch" type studies was collected in bulk from a zone
between 8 and 45 cm deep. Soil was not taken from the top 8 cm because
of the dense network of roots near the surface. Bulk samples were
stored in a polyethylene pail (with lid) until used. Soil used in the
column nitrification and denitrification studies was collected as intact
columns, some with plants and some without. This was done by selecting
an appropriate site (i.e., a young, healthy plant, or no plant), using
a shovel to make the initial circular cut through the peat, roots,
and/or rhizome, then placing polyvinyl chloride (PVC) pipe (70 x 10 cm,
length x diameter) over the plant and pushing it into the soil as far as
the cut was made with the shovel. This was followed by hammering the
column about 45 to 50 cm into the soil. Compaction was less than 10%.
The next step involved rocking the column back and forth in place to
break the bottom loose and then pulling the column up with a hand under
the bottom to prevent the soil from falling out. A "knock-out test cap"
was then pushed into the bottom of the column. Silicone rubber sealant
was used around the cap to stop leaks. In the lab, the columns were
allowed to stabilize for two to four weeks (with those containing plants
placed under a window) while receiving only enough deionized water to
keep the soil saturated.

Water used in these investigations was obtained from the following
sources: marsh surface water, oxidation pond water and secondarily
treated sewage effluent from the Clermont sewage treatment plant, and
Lake Alice water (University of Florida campus). All water samples were
collected and stored in polyethylene containers. Lake and marsh water


-13-








was used within 24 hours. Effluent was stored at 40C for a maximum of
three days. Water samples containing particulate matter were allowed to
settle and the water used was siphoned off the top.

Plant species were in the process of change during the study period,
due to changing weather patterns and lower water levels. Sagittaria
lancifolia (arrowhead) and Pontederia cordata var. Lancifolia (pickerel
weed) were the original dominant species which were invaded by Panicum
sp. (maidencane) and Hibiscus sp. (marsh hibiscus).

Description of Experiments

Preliminary Nitrification Study

An initial study was designed to determine the effects of dissolved
organic compounds (tannins) and bacterial inoculation on nitrification
in marsh water. Yellow colored, dissolved organic compounds were removed
from the marsh water by adding powdered activated charcoal, mixing for
one hour, and filtering (#42 Whatman filter paper). A suspension of
nitrifiers was obtained by shaking a mixture containing 100 grams of
soil known to contain nitrifiers with 150 ml of deionized water for one
hour and allowing the soil to settle out. Ten milliliter aliquots were
used as inoculum for the appropriate treatment. All samples were amended
with ammonium (as ammonium sulfate) to a concentration of about 50 ppm
NH4-N, placed in polyethylene containers (in triplicate) and incubated
at 250C for 30 days. After 25 days, it was noted that the pH had dropped
considerably so it was adjusted up to 7.2 (0.1) with 0.5N NaOH. Twenty
milliliter samples were treated with phenyl mercuric acetate (PMA) and
stored at 40C until pH, ammonium and nitrate were determined, usually
within 24 hours.

Preliminary Denitrification Study

An initial laboratory study using test tubes of nitrate-amended
solutions and marsh soil was designed to ascertain if denitrification
would occur. Solutions of potassium nitrate in deionized water, in
marsh water, and in oxidation pond water and of calcium nitrate in
deionized water were prepared to a concentration of about 45 ppm N03-N.
A 10 ml aliquot of each solution was added to a test tube (eight repli-
cates) containing 6.0 (0.2) grams of marsh soil and incubated at 250C
for 11 days. The same design without the soil was employed to ascertain
if denitrification would occur in the water alone. Sampling involved
pouring off the solution and analyzing it for ammonium and nitrate.

Nitrification and Denitrification Studies Using Soil:Water Columns

In order to better simulate actual marsh conditions, additional
experiments for both nitrification and denitrification were conducted
using intact soil:water columns, half with and half without Sagittaria
lancifolia (arrowhead). Each column consisted of 45 cm of soil and
either 15 or 30 cm of overlying effluent. Treatments in both studies


-14-









were the same (Figure 2). Four replications, two with and two without
plants, of each of the following treatments were used:

1. Control effluent to 15-cm depth.
2. N amended effluent to a 15-cm depth.
3. N amended effluent to a 30-cm depth.
4. N amended effluent to a 15-cm depth with 10 grams CaCO3 added
to adjust pH.

The amount of calcium carbonate was calculated using Yuan's (1974)
double buffer method to bring the top 15 cm of soil to a pH of 7.0. It
was applied on the surface of the soil, to the appropriate columns,
several days before a study began. A second and third application of
the appropriate form of N was pipetted into the individual soil:water
columns as a more concentrated solution.

In all cases, evapotranspiration was accounted for (two to four
times weekly) by adding deionized water to maintain the appropriate
depth. Columns in both studies were aerated to maintain dissolved
oxygen near saturation values. The overlying water in the columns was
sampled (about 15 ml) at about half the water depth, i.e., 7.5 or 15
cm. Phenyl mercuric acetate was added and the sample was stored at 4 C
until ammonium and nitrate were determined, usually within 24 hours.

In Situ Nitrification and Denitrification Studies

For the purpose of obtaining natural nitrification and denitri-
fication rates, an in situ study was conducted. Eno (1960) and more
recently Struble (1977 -found polyethylene bags acceptable for this type
of study due to their permeability to gases (specifically oxygen and
carbon dioxide) and impermeability to ions such as nitrate and ammonium.
Polyethylene bags were filled with 100 grams marsh soil and subjected to
the following treatments:

1. Control, shallow soil in bags placed at 8 cm.
2. Control, deep, lime soil plus lime in bags buried at 30 cm.
3. Nitrification, shallow soil amended with ammonium in bags
buried at 8 cm.
4. Nitrification, deep soil amended with ammonium in bags
buried at 30 cm.
5. Denitrification, shallow soil amended with nitrate in
bags buried at 8 cm.
6. Denitrification, deep soil amended with nitrate in bags
buried at 30 cm.
7. Denitrification, deep, lime soil amended with nitrate
plus lime in bags buried at 30 cm.

Bags were flattened to maximize surface area and heat sealed. For
the nitrification study, soil was amended with ammonium (as ammonium
chloride) to bring the final concentration to about 35 ppm NH4-N.


-15-















FIER


'AIR STONE








-ROOTS


Figure 2. Experimental soil:water columns for nitrification and
denitrification studies.


-16-










For the denitrification study, soil was similarly amended with nitrate
(as potassium nitrate). Control soils received no nitrogen. Lime used
in all cases was 0.85 gram of finely powdered calcium carbonate per bag
which was the amount extrapolated from an earlier application of Yuan's
double buffer method and was intended to bring the pH of the soil to
7.0.

Bags were retrieved at selected intervals. Temperature, pH, and
dissolved oxygen readings were taken at each site. After bags were
removed, sample pH was determined, and the contents of each bag were
extracted with 2N KCI. The extractant was analyzed for ammonium and
nitrate.

Analytical Procedures

Extraction Procedure

Ammonium, nitrite, and nitrate were extracted from the soil by
transferring the 100 grams of wet organic soil into a polyethylene
sample bottle, adding 100 ml of 2N KCI solution, shaking briefly and
then allowing the suspension to stand for two hours. The mixture was
shaken again before filtration (#40 Whatman filter paper) and finally
stored, with PMA, at 4C until inorganic nitrogen was determined
(Bremner, 1965).

Determination of Ammonium and Nitrate

Ammonium and nitrite plus nitrate values were obtained through
steam distillation of water samples or potassium chloride extractants
(Bremner, 1965).

Determination of pH

An Orion 701 or 401 meter was used in conjunction with a combin-
ation pH electrode.

Dissolved Oxygen Determination

A YSI Model 54 oxygen meter with a YSI Model 5419 probe was used to
determine concentrations of dissolved oxygen.


-17-







RESULTS AND DISCUSSION

Nitrification


Laboratory Investigations

Since nitrification has been reported to be predominantly an
autotrophic process (Alexander, 1977) and therefore not in need of
organic carbon compounds, initial studies were designed to determine if
nitrification would occur in the marsh water alone, i.e., without any
soil. Preliminary results indicated that nitrification did not occur in
the marsh water even though dissolved oxygen and pH were at acceptable
levels. Thus, further studies were conducted to identify the reasons
for the lack of nitrification. Since nitrification is reportedly in-
hibited by certain types of dissolved organic compounds (Rice and
Pancholy, 1973; EPA, 1973; EPA, 1975), marsh water was filtered through
charcoal to remove dissolved organic. Since it was also possible that
nitrifying organisms simply were not present in the marsh water, an
inoculant containing nitrifiers was added as a second treatment. Water
from a eutrophic lake on the University of Florida campus (Lake Alice)
was included for comparison purposes. All samples were aerated, amended
with ammonium, and incubated in polyethylene containers for 30 days.

Results of charcoal filtration and nitrifier inoculation are given
in Table 1. Evaporation caused an increase in ammonium concentration so
the effects of the two treatments can best be compared using the nitrate
data. Charcoal filtration had little effect on nitrification rate,
however, inoculation increased the rate in both the marsh and lake
water. These results indicate that nitrification will occur in over-
lying marsh water if nitrifiers are present, but at least in this marsh,
nitrification is minimal due to a lack of nitrifiers. The nitrification
that occurred in the uninoculated lake water indicated that some nitri-
fiers were already present which is a likely consequence of Lake Alice
receiving sewage effluent. Having this nitrifying population already
present would also explain why nitrification started sooner in the lake
water than in the marsh water.

Since the nitrification process releases hydrogen ions and water
has a low buffering capacity, the pH dropped substantially during the
course of the experiment. After 25 days, the pH had dropped approxi-
mately two units and was below that normally acceptable for nitrification
which necessitated artificially adjusting it upward. Five days after
adjustment, it had again dropped two to three pH units in actively
nitrifying samples. In a natural aquatic system, this could rapidly
limit nitrification unless the system was adequately buffered to main-
tain an acceptable pH range. Dissolved oxygen was not limiting through-
out the study period.

The above experiment suggested that nitrification did not occur in
the overlying natural water of this marsh, but that nitrification would
occur in the water if sewage effluent were added. Published work has
indicated that nitrification will also occur in the thin, aerobic sur-
face zone of a submerged soil. Thus, the next phase of the present
study was designed using soil:water columns to simulate marsh conditions.


-18-










Table 1. Effect of charcoal filtration and inoculation with soil water extract
on nitrification in marsh and lake waters without soil.

Treatments Days

Water Char. .


source filt. I


noc. 0 3 7 11
------------------NH4-N, mg

yes 52.4 49.0 51.5 51.3
yes 50.1 46.5 49.6 47.1
no 52.4 51.9 52.1 54.1
no 50.1 51.1 48.7 52.2
yes 52.3 45.9 38.6 33.0
no 52.3 46.5 40.6 35.5

------------------NO3-N, mg

yes 0 0.58 1.0 3.6
yes 0 0.70 1.0 2.7
no 0 0.10 1.0 0.18
no 0 0 1.0 0.50
yes 0.35 0 6.2 12.3
no 0.35 1.1 3.4 9.5


Marsh
Marsh
Marsh
Marsh
Lake
Lake


Marsh
Marsh
Marsh
Marsh
Lake
Lake


Marsh
Marsh
Marsh
Marsh
Lake
Lake


Marsh
Marsh
Marsh
Marsh
Lake
Lake


oxygen
8.1
8.2
8.2
8.0
8.0
8.1


18 25 25' 1 30
liter1-------------------
52.2 59.7 --- 54.7
50.8 50.1 --- 52.9
57.6 61.6 --- 65.6
54.0 60.1 --- 63.8
32.3 33.0 --- 27.0
36.8 39.1 --- 40.8

liter------------------
4.9 7.9 --- 17.0
4.9 11.9 --- 13.4
0.46 0.10 --- 0.8
0.28 1.2 --- 0.8
12.8 -- -- 23.2
10.3 11.5 --- 11.5


, mg
8.2
8.0
8.2
8.3
8.0
8.1


1 iter-1
7.4
7.3
7.2
7.5
7.2
7.2


--- 7.3
--- 7.2
--- 7.2
--- 7.3
--- 7.2
--- 7.0


-------------------------pH---------------------------


6.6
6.5
6.6
6.5
7.9
7.9


6.6
6.6
6.7
6.5
7.9
8.0


6.2
6.1
6.6
6.5
7.4
7.7


5.1
4.9
6.5
6.1
5.0
5.4


4.1
3.9
5.6
4.5
4.3
5.3


4.4
4.2
6.0
5.5
4.2
5.2


t pH adjusted to 7.2


(0.1) with 0.5 N NaOH after 25 day samples taken.


-19-


yes
no
yes
no
no
no


yes
no
yes
no
no
no


-------------Dissolved
yes yes 8.0 8.0 8.1
no yes 7.8 8.1 8.1
yes no 8.0 8.2 8.2
no no 7.8 7.8 8.1
no yes 8.3 7.7 7.3
no no 8.3 7.9 7.4


yes
no
yes
no
no
no


yes
yes
no
no
yes
no








Plants play an important role in controlling nitrogen forms and con-
centrations in marsh systems through plant uptake and decomposition as
well as the introduction of root exudates and oxygen into the rhizo-
sphere soil. For this reason, they were included in this study with
water depth and pH as variables. Soil:water columns (10 cm I.D.)
containing 45 cm of soil and either 15 or 30 cm of overlying secondarily
treated sewage effluent were used. Treatments consisted of a control
with sewage effluent to a depth of 15 cm, ammonium-amended effluent to
a depth of 15 cm, ammonium-amended effluent to a depth of 30 cm, and
ammonium-amended effluent to a depth of 15 cm and pH adjusted with
CaCO3. Four replications, two with and two without plants, of the above
treatments were used.

Concentrations of ammonium and nitrate observed over a 52 day
period in the overlying water are depicted graphically in Figures 3
through 5. Overlying water in the columns was not aerated initially and
oxygen diffusion was not sufficient to provide adequate aeration. Con-
sequently oxygen became depleted within the first few days of the study,
curtailing nitrification. This resulted in the leveling off or increase
of ammonium concentrations in columns without plants. The increase in
ammonium concentration was probably due to ammonification, which does
not require oxygen. This increase was especially evident in the limed
columns as would be expected since ammonification is favored at neutral
pH levels (Hubbell, 1971). The effect was not observed in columns with
plants because the plants continued to assimilate ammonium. Subsequently,
(day 9), all columns were artificially aerated.

Nitrification rate appears to be related to ammonium concentration;
however, due to the limited amount of data, first-order kinetics cannot
be assumed. No attempt will be made to quantitatively evaluate the
kinetics of nitrification in flooded systems. It is likely that nitri-
fication kinetics under the present experimental conditions are quite
complex, since a portion of the nitrification is occurring in the water
column or at the soil:water interface, while a portion occurs at some
small depth within the soil. The former would be expected to be a zero-
order process, while the latter is likely non-zero-order because of
diffusion dependence (as discussed for denitrification in Chapter 1).
Ammonium removal rates given in Table 2 are derived from the slopes of
removal curves at approximately 15 to 25 mg liter-1 ammonium-N concentration.

In the 15 cm water depth columns with plants, all added ammonium
was depleted in approximately 15 days despite the above mentioned
aeration problem (Figure 3). At 18 days, additional ammonium was added
to both the treatments with and without plants. Ammonium removal con-
tinued to be rapid with plants and, in contrast to the earlier days of
the study, ammonium removal was also relatively rapid in the columns
without plants. Only small amounts of nitrate were detected in either
column (Figure 3). This can be explained partly by the fact that some
nitrate was removed by the plants and, more importantly, by nitrate
diffusion into the anaerobic sediment and subsequent denitrification.
Ammonium removal rate (Table 2) was greater with plants (2.9 kg N ha-I
day-1) than without plants (1.5 kg N ha-lday-1).


-20-










Water depth 15cm


Ammonium N


0 10 20 30 40 50
days


Nitrate N


0 10 20 30 40 50
days


Nitrification in soil:water columns flooded with 15 cm of
ammonium amended sewage effluent, with [o] and without [*]
plants.


-21-


Figure 3.











Water depth


Ammonium N


0 10 20 30 40 50
days


Nitrate -N









I


days


Nitrification in soil:water columns flooded with 30 cm of
ammonium amended sewage effluent, with [o] and without [*]
plants.


-22-


20


10


Figure 4.


30 cm












pH Adjusted, water depth 15cm


Ammonium -N


0 10 20


days


Nitrate -N


0 10 20 30 40 50
days

Nitrification in limed soil:water columns flooded with
15 cm of ammonium amended sewage effluent, with [o] and
without [*] plants.


-23-


Figure 5.














Table 2. Ammonium removal rates from overlying water of
soil:water columns containing ammonium amended
sewage effluent.


Treatment


15 cm water depth

30 cm water depth

15 cm water depth, pH adj.


Without Plant With Plant

---------kg N ha-lday1---------

1.5 2.9

1.2 5.3

11.9 11.2


Rates assume an ammonium-N concentration of 20 mg liter-1


-24-









Increasing the water depth to 30 cm did not greatly affect the
ammonium removal rate (Table 2) but because the total amount of added
ammonium was greater, the removal time was somewhat longer (Figure 4).
As with the 15 cm water depths the ammonium removal rate was greater
with plants (5.3 kg N ha-Iday-i) than without plants (1.2 kg N ha-1
day-1). Nitrate concentrations (Figure 4) were low throughout the
study period for the same reasons noted for the 15 cm water depth columns.

Increasing the pH significantly increased ammonium removal rate
(Figure 5 and Table 2). Ammonium concentrations decreased from about 40
ppm N to less than 5 ppm within 3 days. The fact that nitrification was
responsible for much of this decrease is indicated by the high (20 ppm
N) nitrate concentrations observed in the overlying water. Nitrifying
bacteria require a pH near neutrality to thrive and are considerably
inhibited by pH values below 5. There was little difference in ammonium
removal rates in pH adjusted treatments with or without plants possibly
because the bacteria were able to effectively compete with the plants
for the ammonium. The effect of the plants is, however, very evident
from the nitrate concentrations. Nitrate concentrations were consistently
lower in the presence of plants.

Control columns consisted of secondarily treated effluent to a 15-
cm depth overlying marsh soil with no treatment except aeration.
Concentrations of ammonium and nitrate in the controls with and without
plants remained low throughout the study period.

When columns were dismantled at the end of the 52 days, pH in the
top 15 cm of soil for all except the limed columns averaged around 5.5,
dropping down to about 4.5 over the next 30 cm. Surface pH of the limed
columns approached neutrality but decreased with depth (within 10 cm) to
values similar to the other treatments since the lime had not been
incorporated into the soil.

In Situ Investigations

Since nitrification did not appear to occur naturally in the water
alone but did occur when in contact with soil in the column experiments,
an in situ study was performed to determine location of and rates at
which this oxidative process would occur in the soil profile. Poly-
ethylene bags containing marsh soil amended with ammonium were placed at
shallow (8 cm) and deep (30 cm) positions in the marsh soil.

Ammonium and nitrate concentrations in the sample bags are shown in
Figure 6. Ammonium values declined with time at the 8 cm depth but
remained relatively constant at the 30 cm depth. This indicates that
nitrification was occurring at the 8 cm level but not at 30 cm. During
the time of this in situ study, the water table flucuated at and slightly
below the soil surface. Under these conditions, the 8 cm depth may have
been under aerated conditions periodically due to the depth of the water
table. However, oxygen has been reported to diffuse from the roots of a
number of marsh plants into the surrounding waterlogged soil (Armstong,
1967) and this mechanism is likely responsible for providing much of the
oxygen required by nitrification. Roots were largely absent at the


-25-













Ammonium -N


30 1


0--0


shallow
deep
control,shallow
control ,deep


012 I i I I
0 5 10 15 20 25
days



Nitrate -N


days


Figure 6.


Ammonium and nitrate concentrations in ammonium amended
marsh soil during in situ incubation in polyethylene bags.


-26-


20r


E


101


E








30 cm depth of the deep samples. Small increases in nitrate concen-
tration were observed at both depths. At the 8 cm depth the increase in
nitrate concentration did not correspond in quantity to the decrease in
ammonium concentration, indicating that some denitrification and/or
plant assimilation was also occurring at this depth. The presence of
nitrate at the 30 cm depth is likely an artifact of the sample extraction
technique after removed from the soil.


Denitrification

Laboratory Investigations

Preliminary laboratory denitrification studies were conducted to
obtain approximate denitrification rates and ascertain how these rates
would be affected by the presence or absence of soil, associating cat-
ion, and water source. This entailed a "batch" type study which ex-
cluded plants. Water samples from several sources were amended with
potassium or calcium nitrate salts and incubated with and without soil.
The data (Table 3) shows a steady decline in nitrate concentrations with
time, for all soil-water treatments. No clear effect of water source
nor corresponding cation on nitrate removal rate was observed. The most
likely nitrate removal mechanisms are denitrification and microbial
immobilization, with the former being the dominant reaction. Solutions
not in contact with soil showed no appreciable change (slight increase
due to evaporation) which indicates that denitrification did not occur
in these waters in the absence of soil. This is in agreement with
similar work done by Engler et al. (1976) and likely was due to insuf-
ficient readily available organic carbon, even in oxidation pond water,
to support denitrification.

After finding that water source and associating cation had little
effect on the denitrification rate, an investigation was initiated using
soil:water columns to simulate the marsh environment. Potassium nitrate
and secondary sewage effluent were used as the nitrate and water source,
respectively. Water depth, pH, and plants were used as variables in a
manner similar to that used in the nitrification study.

Nitrate concentrations in the overlying water of each treatment are
presented graphically in Figures 7 through 9. Nitrate removal appeared
to follow first-order kinetics. (Denitrification kinetics, including
methods for determining reaction order, are discussed in Chapter 3.)
Both first-order rate constants (day-l) and nitrate removal rates (kg N
ha-lday-1, assuming a floodwater nitrate concentration of 10 mg N
liter-l) are given in Table 4. Although useful for theoretical and data
manipulation purposes, first-order rate constants can be misleading
because they express nitrate removal as a function of the amount of
nitrate present. That is, if floodwater depth (and therefore the amount
of nitrate in the system) is doubled while removal rate remains the
same, the first-order constant will be reduced by one half. For this
reason, when floodwater depth is a variable, nitrate removal rates are
of more value for comparison purposes than are rate constants.


-27-











Table 3. Nitrate removal as affected by water source, associated
cation and presence of soil.

Days

Treatment 0 2 5 11
------ NO3-N, mg liter- ------

Ca(NO 3)2 in deionized water (no soil) 40.5 41.8 41.3 42.3
KNO3 in deionized water (no soil) 45.0 45.2 46.6 47.3
KNO3 in oxid pond water (no soil) 51.6 53.0 54.7 58.0
KNO3 in marsh water (no soil) 43.2 44.9 44.8 49.2

Ca(N03)2 in deionized water (over soil) 40.5 30.9 25.5 12.2
KN03 in deionized water (over soil) 45.0 33.9 26.3 15.0
KNO3 in oxid pond water (over soil) 51.6 43.0 30.0 21.5
KNO3 in marsh water (over soil) 43.2 35.0 26.5 14.3

------ % N03-N remaining------
Ca(N03)2 in deionized water (over soil) 100 76.3 63.0 30.1
KNO3 in deionized water (over soil) 100 75.2 58.4 33.4
KNO3 in oxid pond water (over soil) 100 83.3 58.2 41.6
KNO3 in marsh water (over soil) 100 81.0 61.3 34.3


-28-

















Table 4. Nitrate removal rates from overlying water of soil:water columns
containing nitrate amended sewage effluent.

Treatment Nitrate removal rate Denitrification rate constant

---kg N ha"'day"---- ------------day ------------

15 cm water depth
without plant 1.9 0.127
with plant 1.5 0.102


30 cm water depth
without plant 0.8 0.026
with plant 2.9 0.098


15 cm water depth, pH adj.
without plant 1.0 0.067
with plant 1.6 0.105

Assuming floodwater nitrate-N concentration of 10 mg liter-].


-29-




















Water depth 15cm


40 -



30-



20-



I0



0 --

0








Figure 7.


Nitrate N


10 20 30 40 50
days


Nitrate removal from overlying water of soil:water columns
(with [o] and without [*] plants) containing nitrate amended
sewage effluent to a depth of 15 cm. Nitrate was added on
day 0 and day 18.


-30-





















Water depth


40 Nitrate N



S 30


E
20



I0




0 1.0 20 30 40 50
days






Figure 8. Nitrate removal from overlying water of soil:water columns
(with [o] and without [*] plants) containing nitrate amended
sewage effluent to a depth of 30 cm. Nitrate was added on
day 0 and day 18.


-31-


30cm





















pH Adjusted, water depth 15cm


Nitrate -N


E 20




10"



0 .i










Figure 9.


) I0 20 30 40 50


days


Nitrate removal from overlying water of limed soil:water
columns (with to] and without [*] plants) containing
nitrate amended sewage effluent to a depth of 15 cm.
Nitrate was added on day 0 and day 18.


-32-











Removal rates for the treatments without plants ranged from 0.8 to
1.9 kg N ha-lday-1. Results of the three treatments are difficult to
compare because of the low number of replications and the high spacial
variation found in wetland soils. (Chapter 3 contains further discussion
of spacial variation found in a number of wetland soils.) The two water
levels were originally utilized because it was felt that water depth
would influence the aeration status of the system, with possibly less
oxygen reaching the soil with the deeper water depth, thereby increasing
the rate of denitrification. However, the need to artificially aerate
the water column negated the water depth effect. It was also antici-
pated that pH adjustment would increase denitrification rate. Since the
pH adjusted treatment had the lowest rate, it was assumed that this
value reflected natural variation rather than a true pH effect. (Lack
of pH effect is probably the result of lime having been applied to the
soil surface while denitrification is a subsurface process. Incorpora-
tion of lime into the soil would be impractical in the field due to
plant destruction, etc.) Thus, the overall denitrification rate in
these systems was taken as the average of the three treatments. There-
fore, a removal rate of 1.2 kg N ha-'day-1 (at 10 mg nitrate-N liter"1)
would be expected for this soil. Removal rates at other nitrate con-
centrations would vary in direct proportion to the concentration, since
the reaction is first-order.

Inclusion of plants in the system substantially increased nitrate
removal rates. A number of factors may have been responsible for this.
Plant roots may have provided more favorable conditions for denitrifi-
cation via root exudates or other influences. Downward movement of
floodwater into the soil due to transpirational pumping may have in-
creased denitrification rates by accelerating nitrate transport to
denitrification sites. Plants may also have assimilated nitrate,
augmenting denitrification as a removal mechanism. Because plants in
different columns varied both in size and vigor, as well as for reasons
noted for the columns without plants, removal rates for columns with
plants, averaged over the three treatments, were used as an indication
of typical removal rates. Values (again assu ing 10 mg nitrate-N
liter-1) ranged from 1.5 to 2.9 kg N ha-lday" with an average removal
rate of 2.0 kg N ha-lday-1.

Ammonium levels in the overlying water of all treatments were less
than 1 mg N liter-] at all except the initial sampling date. The con-
trol treatment also had low nitrate values. This indicates that either
very little ammonium was being released from the soil, or that it was
being nitrified and then rapidly denitrified.

In Situ Investigations

Nitrate removal as a function of depth and incorporated lime was
investigated by in situ incubation of nitrate and lime amended marsh
soil in polyethylene bags. Soils amended with nitrate showed rapid
nitrate removal, with 25 mg liter-1 nitrate removed in 2 to 8 days,
depending on treatment (Figure 10). The effect of depth on nitrate
removal is shown by comparing data for unlimed, shallow site (3.1 _g N
per gram wet soil per day) with unlimed, deep samples (5.1 pg N g-lday-).


-33-

















Nitrate N


E
10




O L
0




20



0





0
0



Figure 10.


-- shallow
*- deep
o--o deep, lime
a-a control, shallow
&-a control, deep


days


Ammonium N


5



Nitrate and ammonium
marsh soil during in


10 15 20
days


concentrations in nitrate amended
situ incubation in polyethylene bags.


-34-










It appears that at the shallow depth there are both aerobic (likely
in close proximity to plant roots) and anaerobic sites. Thus, nitri-
fication and denitrification were occurring simultaneously at this
depth. This is also indicated by the small amount of ammonium accumu-
lation and slower nitrate removal compared to the 30 cm depth.

Limed soils amended with nitrate and placed deep in the profile
removed 12.8 pg N g 'day-l. When this is compared to a similarly
placed, unlimed sample in which 5.1 pg N g-lday-l was removed, it is
obvious that liming enhanced the removal rate. Apparently, denitri-
fication is limited by the natural soil pH of 4.7. The immediate
stimulation of denitrification by raising the pH is indirect evidence
that most nitrate removal occurs biologically rather than chemically,
since the latter is inhibited at higher pH values. Another interesting
aspect of liming is the increased ammonification shown by limed treat-
ments (Figure 10). This ammonium accumulation did not occur in shallow
samples, probably due to the presence of small amounts of oxygen, but
occurred in the deep, more anaerobic samples.









SUMMARY AND CONCLUSIONS


Nitrification and denitrification rates in a freshwater marsh re-
ceiving secondarily treated sewage effluent were investigated in the
laboratory and in situ. Laboratory experiments measured these processes
both in water alone and in soil:water columns. In situ work consisted
of incubation of buried polyethylene bags containing ammonium or nitrate
amended soil.

Nitrification did not occur in the natural marsh water, but did
occur in marsh water receiving sewage effluent. Ammonium in the over-
lying water of soil:water columns was nitrified at a rate considerably
greater than in the absence of soil (Table 2, page 24). Water depth (15
cm vs. 30 cm) had little effect on nitrification rates, although the
water depth effect may have been negated by artificial aeration. Inclu-
sion of plants increased ammonium removal rates, likely due to plant
assimilation of ammonium in addition to nitrification. Addition of
calcium carbonate (agricultural limestone) increased the nitrification
rate to the point where nitrate accumulated in the overlying water. In
situ incubation of ammonium amended marsh soil indicated that some
nitrification would occur in the plant root zone, probably due to oxygen
diffusing from the roots. Nitrification was not observed below the root
zone.

Denitrification did not occur in marsh water or in oxidation pond
water in the absence of soil. Apparently there was not sufficient
available organic carbon in these waters to deplete oxygen supply and
allow denitrification to occur. Rapid, first-order denitrification was
observed in soil:water columns (Table 4, page 28). Average nitrate
removal rates, assuming a floodwater nitrate concentration of 10 mg N
liter-1, were 1.2 kg N ha-lday-l for columns without plants and 2.0 kg N
ha-lday-l with plants. As with nitrification, water depth did not
affect denitrification rate, again probably due to artificial aeration.
In contrast to the results for nitrification, liming had no effect on
denitrification rate. This may have been due to the fact that the lime
was applied to the soil surface while denitrification occurs below the
soil surface. Given sufficient time, it is anticipated that the lime
would move downward in the soil and increase the pH in the zone of
denitrification. Denitrification was rapid in nitrate amended soil
samples incubated in situ (at both 8 and 30 cm depths) and was accelerated
by increasing the soil pH with lime.

The above nitrification and denitrification rates represent our
best estimates based on primarily laboratory investigations. Further
field investigations are needed to ascertain the effects of natural
variations in water depth, dissolved oxygen concentration, plant density,
and temperature.


-36-









CHAPTER III


DENITRIFICATION IN FIFTEEN SELECTED FLORIDA WETLAND SOILS


Denitrification can be an important mechanism for the removal of
nitrate from wastewater applied to wetlands, as indicated in the pre-
vious chapter. The objective of this investigation was to measure
denitrification rates in a variety of flooded wetland soils and to
determine quantitative relationships between denitrification rate and
selected soil characteristics in order to develop a predictive equation
for denitrification rates in specific soils. Since nitrous oxide, one
of the gaseous end-products of denitrification, is known to deplete
ozone in the stratosphere, a quantitative evaluation of nitrous oxide
evolution from wetland soils was also included in this study.

METHODS AND MATERIALS

Soils and Soil Sampling

Soils

A total of 15 soils of widely varying organic matter content, re-
action (pH), and texture were selected for study. All are classified as
very poorly drained or poorly drained, and have water standing on the
surface for one to nine months per year. Geographic distribution,
ranging from Palm Beach and Lee Counties in the south to Walton County
in the northwest, is shown in Figure 11. A list of soil types, classifi-
cations, and descriptions* is given below. These are general descrip-
tions of the soil series, not of the actual pedons at sampling sites.
Soils are listed in declining order of denitrification rate, as deter-
mined in the present study.

Everglades muck--typic medihemist--very poorly drained organic soils
that formed in thick deposits of hydrophytic plant remains. They
occur in freshwater swamps and marshes. In a representative pro-
file, the surface layer is a black muck, about 12 inches thick.
Below it, to a depth of 46 inches, is a reddish-brown mucky peat.
Below a depth of 46 inches, and extending to below 78 inches, is
dark brown peat.

Floridana sand--arenic argiaquoll--a very poorly drained soil that
occurs on broad low flats and in depressions. A representative
profile has a black sand surface soil about 12 inches thick and
a grayish-brown sand subsurface layer over a dark gray sandy clay
loam subsoil that begins within a depth of 20 to 40 inches below
the soil surface. These soils are formed in sandy and loamy
marine sediments.



* Adapted from U.S.D.A. Soil Conservation Service soil interpretation
sheets (Soil Form #5) for the soils series described.


-37-
































1. Everglades muck
2. Floridana f.s.
3. Astor s.
4. Surrency s.
5. Samsula muck
6. Pickney f.s.
7. Pelham l.f.s.
8. Riviera f.s.
9. Chastain si.l.
10. Brighton peat
11. Iberia s.c.
12. Chobee f.s.l.
13. Delray f.s.
14. Eureka f.s.l.
15. Valkaria f.s.












Figure 11. Location of sampling sites.


-38-


~p~~








Astor sand--cumulic haplaquoll--very poorly drained soils that occur
in low flat areas, poorly defined drainageways, and floodplains
in the lower coastal plain. A typical profile has a black to
very dark grayish-brown sand surface layer about 32 inches thick.
Below this to a depth of 80 inches is dark grayish-brown sand.

Surrency sand--arenic paleaquult--very poorly drained nearly level
soils in Central and North Florida. In a representative pro-
file, the surface layer is sand about 22 inches thick. The
upper 16 inches is black and the lower 6 inches is dark gray.
The subsurface layer is light brownish-gray sand about 9 inches
thick. Between depths of 31 and 65 inches, the subsoil is gray
sandy clay loam.

Samsula muck--terric medisaprist--nearly level very poorly drained
organic soils. They occur in fresh water swamps and marshes.
In a representative profile, the surface 9 inches is a black
muck. The subsurface layer is a dark reddish-brown muck to a
depth of 36 inches. Next is 10 inches of dark grayish-brown
sand and then gray sand to below a depth of 55 inches. These
soils formed in hydrophytic plant remains.

Pickney fine sand--cumulic humaquept--very poorly drained soils in
depressions and drainageways of the lower coastal plain. The
typifying pedon has black loamy fine sand A horizons over dark
gray fine sand C horizons.

Pelham loamy fine sand--arenic paleaquult--poorly drained nearly level
soils on the coastal plain. Typically the surface layer is very
dark gray loamy fine sand about 5 inches thick. The subsurface
is gray loamy sand about 21 inches thick. The subsoil extends
below a depth of 68 inches. It is mostly light gray sandy clay
loam mottled with yellow and red.

Riviera fine sand--arenic glassaqualf--very poorly drained soils in
depressed areas of the coastal plain. In a representative pro-
file, the surface layer is dark grayish-brown sand about 6 inches
thick. The subsurface layer is white sand about 22 inches thick
and it extends into the underlying grayish-brown sandy clay loam
subsoil. They formed in unconsolidated marine sands and loams.

Chastain silt loam--typic haplaquept--poorly drained, slowly permeable
soils on flood plains of rivers on the coastal plain. Typically
these soils have yellowish-brown surface layers over gray clayey
subsoils.

Brighton peat--typic medifibrist--very poorly drained organic soils
that occur in depressions and freshwater marshes and swamps in
peninsular Florida. Typically they have a black peat surface
layer 9 inches thick. Beneath the surface layer dark yellowish-
brown and dark brown peat extends to 63 inches or more deep.

Iberia sandy clay--vertic haplaquoll--poorly drained, very slowly
permeable soils. They have a black clay surface and a dark
gray clay subsurface. These soils formed in alkaline clayey
alluvium. They occur on the lower part of the natural levees,
often in depressions.
-39-








Chobee fine sandy loam--typic argiaquoll--very poorly drained soils
that occur in small to large depressions, low areas, or river
flood plains. A representative profile has a black fine sandy
loam surface layer about 7 inches thick and a black to very dark
gray sandy loam subsoil that contains calcium carbonate nodules.

Delray fine sand--grossarenic argiaquoll--deep, very poorly drained
soils that occur on low, broad flats or depressions in South
and Central Florida. A representative profile has a black fine
sandy surface layer. Between a depth of 46 to 60 inches Ls a
dark grayish-brown fine sandy loam subsoil. These soils formed
in thick deposits of sandy and loamy marine sediments.

Eureka fine sandy loam--typic albaqualf--poorly drained soils that
occur in South and Central Florida. In a representative profile
the surface layer is black loamy fine sand about 4 inches thick.
The subsurface layer is grayish-brown loamy fine sand about 7
inches thick. The subsoil begins at depths of 11 inches and
extends to 72 inches and deeper. It is gray sandy clay.

Valkaria fine sand--spodic psammaquent--poorly drained soils that
occur on low, poorly defined sloughs and areas adjoining swamps in
the lower coastal plain. In a representative profile, the surface
layer is sand about 9 inches thick; the upper 5 inches is black
and the lower 4 inches is dark brownish-gray. Between depths of
9 to 15 inches is light gray sand, with brownish or grayish sand
to depths of 80 inches.


Collection of Samples

Samples for column studies were collected in 5 cm i.d. by 60 cm
long plexiglass columns. Where possible, sharpened columns were simply
driven into the soil to a depth of 25 cm with a, hammer. In cases where
this resulted in excessive soil compaction, a more involved method was
used. A rubber stopper exactly matching the inner diameter of the
column was inserted in the bottom of the column and suspended at the
soil surface from a tripod by a cord running down the center of the
column (Figure 12(a)). A small amount of water was poured into the
column to help maintain a seal between column and stopper. The column
was then driven into the soil while the stopper remained stationary,
creating a suction in the column to prevent the soil surface from
receding as the column moved downward (Figure 12(b)). To remove the
stopper, the cord was pulled upward at an angle, thereby tilting the
stopper and breaking the seal (Figure 12(c)). Columns were extracted
with a rocking, twisting motion after stoppering the top to prevent soil
loss out of the bottom. This sampling method was adapted from that of
Coleman (1979).

Bulk samples of the top 10 cm of soil were stored in polyethylene
bags at 40C for a maximum of 5 days before determination of mineral-
izable organic carbon and pH. Subsequently, samples were oven dried at
1040C for total organic carbon determination.


-40-


































Surface


Figure 12. Apparatus for obtaining soil cores with minimum soil compaction.


I-Column


,Stopper







Description of Experiments


Denitrification and Nitrous Oxide Evolution

A laboratory study was conducted to determine denitrification rates
and nitrous oxide (N20) evolution rates for 14 of the soil types col-
lected. Three columns of each soil type were flooded with deionized
water to a depth of 15 cm. Aquarium airstones (previously determined to
be carbonate and ammonia free) were used to aerate and mix floodwater.
Air was prehumidified and stripped of ammonia by use of an acidified
(H2S04) water trap. Air flow rate in each column was approximately 30
to 40 ml min-1. Water levels were maintained throughout the study by
periodic additions of deionized water. This and all other studies were
conducted within a temperature range of 260C20C.

Columns, prepared as described above, were allowed to stand for 7
days before initial nitrate application. Background floodwater nitrate
concentrations were measured during this time. Water samples for ni-
trate determination were pipetted from the floodwater surface. Nitrate-
N levels in the floodwater were then raised to 20 mg liter-1 by addition
of potassium nitrate solution. A 2 week incubation period followed, for
the purpose of allowing microbial adaptation to nitrate and establishment
of a relatively stable nitrate diffusion front in the soil. During this
period, concentrations of nitrate were monitored every 3 to 4 days and
additions made if nitrate-N dropped below 10 to 15 mg liter-1. No
additions of nitrate were made within 3 days of the end of the adapta-
tion period. Floodwater nitrate concentrations were measured at 0, 4,
8, and 12 days after the end of this period. Nitrate in two arbitrarily
selected extra colums (Surrency soil) was measured more frequently and
for a longer period of time for the purpose of obtaining detailed
nitrate disappearance curves. One column per soil type was selected for
determination of N20 evolution rate by the method described below. This
was done approximately 3 weeks after initial nitrate application.

Long Term Denitrification and Nitrous Oxide Evolution

Two soils, Samsula and Pelham, were selected for a more detailed
study to determine changes in denitrification and N20 evolution rates
over a 10 week period. Eight columns of each of the soils were treated
similarly to those in the previous experiment. Denitrification rates
(calculated as described in results and discussion) were determined at
2, 5, and 8 weeks, with nitrate-N maintained between 5 and 20 mg liter-1
throughout the study. Each denitrification rate determination consisted
of three nitrate concentration measurements made at 4 day intervals.
Nitrous oxide evolution from each of two Samsula and three Pelham
columns was determined at least three times during the study period by
the method described below.

Effect of Floodwater Nitrate Concentration on N20 Evolution

A short experiment was conducted to determine if N20 evolution was
proportional to nitrate concentration. Two Samsula and two Pelham
columns were treated as in previous experiments. Nitrous oxide


-42-








evolution was measured at least three times per column over a period of
several days as floodwater nitrate-N concentrations fell from 15 mg
liter-l to below 5 mg liter-1. Background evolution rates for these
columns were determined prior to initial addition of nitrate to the
floodwater.

Method for Measuring N20 Evolution

Nitrous oxide evolution from the columns was determined in a semi-
enclosed, continuous flow collection chamber. This involved sealing the
top of the column, except for small air entry and exit tubes. Aeration
continued, but at a reduced, controlled rate of 7 to 20 ml min-.
Floodwater oxygen concentration was periodically measured to insure
sufficient aeration. Air flow was measured with a soap-film flow meter
connected to the air exit tube (Figure 13).

In this semi-enclosed chamber, N20 concentration in the headspace
will rise slightly due to N20 production in the soil. In response to
this, N20 dissolved in the floodwater also increases. Eventually an
equilibrium is established, and N20 no longer accumulates in the flood-
water or headspace. At this time N20 entering the chamber plus N20
produced in the column will equal N20 exiting the chamber. If flow
rate, influent N20 concentration, and effluent N20 concentration are
known, N20 production can be calculated as follows:

R = (4.6xi0-14) (E A)(f)(4.9x106 )(28) [2]

where R is N20-N evolution rate, ha-lday-l; 4.6x10-14 is the number of
moles of gas in 1 ml at a concentration of I ppb (v:v) at standard
temperature and pressure; E and A are effluent and ambient (incoming
air) N 0 concentrations, ppb (v:v); f is air flow rate, ml day-1;
4.9xl0O converts the column surface area to hectares; and, 28 is the
mole to gram conversion factor for N20-N. In practice, a 12 hour
equilibration time was found to be more than sufficient to establish
constant N20 concentration in effluent air. Five, I ml samples of
effluent air and five samples of influent (ambient) air were collected
for each N20 evolution determination. A similar method was described by
Kaspar and Tiedje (1979) for use with unflooded soils.

In the long term study described above, N20 samples were stored in
plastic syringes for up to 3 days. It was found that concentrations
decreased with storage; i.e., the difference between sample and normal
ambient (300 ppb) N20 concentrations declined by approximately 24, 45,
and 65% after 1, 2, and 3 days, respectively. Measured N20 concen-
trations of stored samples were corrected by the appropriate loss
factor, and corrected values were used to determine evolution rates.


Analytical Methods

Nitrate

Nitrate concentration was measured with an Orion model 93-07
nitrate electrode in conjunction with an Orion 90-01 double junction
reference electrode and Orion 701 digital meter. Ionic strength of
sample and standard solutions were adjusted with 2N ammonium sulfate.


-43-


















Flow meter


* Effluent air


Figure 13. System for determining nitrous oxide evolution from flooded
soil cores.


-44-










Nitrous Oxide

A Tracor model 222 gas chromatograph with a Ni-63 electron capture
detector were used for N20 determinations. Carrier gas was 5% methane
in argon, with a flow rate of 28 cm3 min-1. Porapak Q was used in the
separation column. Column temperature was 50 C and detector temperature
was 3500C. Sample volume was 1.0 ml.

Bulk Density and Porosity

Bulk density and porosity of the upper 10 cm of soil in each col-
umn were determined gravimetrically. Saturated soil sections were
weighed, dried overnight at 104 C, and reweighed. Bulk density was
calculated as the dry weight per original soil volume, and porosity was
calculated as the volume of water lost per original soil volume.

Other Methods

Fresh soil samples (not dried) were used for pH determination.
Measurements were made in deionized water with a soil:liquid ratio of
1:1 (v:v). A combination pH electrode and Orion 701 meter were used.
Total N was determined by the micro-Kjeldahl procedure. Total organic
carbon was determined by the Walkley-Black method (Allison, 1965).
Efficiency of organic matter oxidation by dichromate was assumed to be
77%.

/


-45-










RESULTS AND DISCUSSION


Comparative Denitrification Rates

Denitrification rates were compared for 14 soil types encompassing
a wide range of organic matter contents and pH values. Values represent
an average of three replications. Measurements of nitrate concentration
in the overlying water were taken three or four times over a 12 day
period following a 2 week preincubation period. In addition, two columns
of Surrency soil were sampled ten times each in order to obtain more
detailed nitrate disappearance curves. Prior to rate comparison, the
data were analyzed to determine the rate-order of the denitrification
reaction.

Floodwater nitrate disappearance data are presented in Figures 14
and 15. Simple visual examination of the graphs indicates that the
process is not zero-order. If the reaction is first-order (as in the
1974 model of Bouldin et al.), the following relationship between
concentration and time exists:

In CO0 = k t
C 1 [3]

where CO is initial nitrate concentration, mg liter-1; C is nitrate
concentration at time t, mg liter-l; kI is first-order rate constant,
day-l; and t is time, days. If the reaction is half-order (i.e., rate
is proportional to the square root of concentration, as in Bouldin's
1968 model), the concentration versus time relationship is as follows
(Gucker and Siefert, 1966):

C CO =-kt [4]

where k' is half-order rate constant, day-', and other variables are as
in Equation 3.

The data in Figure 14 were fitted to Equations 3 and 4 by least-
squares linear regression. Due to differing initial floodwater nitrate
concentrations (made unavoidable by the adoption of the preincubation
period) and differing denitrification rates in each column, each was
treated separately, with separately calculated k values and regression
coefficients. Since a total of 38 columns were studied, 38 separate
regression procedures were performed for each equation, each with I or
2 degrees of freedom. Mean r2 values were 0.981 for Equation 3 and
0.978 for Equation 4. Data from 21 of the columns fit Equation 3
better, 12 columns fit Equation 4 better, and 7 columns fit both equa-
tions equally well. Thus, the data weakly support Equation 3, or first-
order kinetics, over Equation 4, or half-order kinetics.

Results for the two more intensively sampled columns of Surrency
soil (Figure 15) were ambiguous. Column A yielded r2 values of 0.988
and 0.986 for Equations 3 and 4, respectively, while column B yielded


-46-















Floridana fine sand


12 0
time, days


Astor sand


12 0

time, days


Surrency sand




















I I I p


Figure 14.


Changes in floodwater nitrate concentrations with time
in soil:water columns of 14 soil types.


10


E
z

0
z


16

14


E


0^
z


I I I i


i I I p


Everglades muck


-47-














- Samsula muck

*













0^


12 0

time,days


E

z
0
z

























z
o



0
z


Figure 14.


Pickney fine sand


A


0


4 8


Chastain silt loam


12 0

time, days


(Continued.)


-48-


Riviera fine sand


i l i I


I I I I


a i I


I


i i















Brighton peat


Iberia sandy clay


12 0
time, days


Chobee fine sandy loam


Figure 14.


Delray fine sand


0


10


12 0

time, days


(Conti nued.)


-49-


E
2F


z


I




0
2


I I I I


I I i i


I I i I














Eureka fine sandy loam


S Valkaria ine sand






:^^0


12 0

time, day s


Figure 14. (Continued.)


I

z


V I F


I I


-50-



















-- column A
*-a column B


time,days


Figure 15.


Changes in floodwater nitrate concentrations with
time in two soil:water columns of Surrency soil.







r2 values of 0.993 and 0.994 for the respective equations. Statistical
analysis of the data from these columns used ten observations, or eight
degrees of freedom per column.

Although the data appear to fit both first-order and half-order
apparent kinetics, first-order kinetics are assumed for the remainder of
this work, and only first-order rate constants are reported. This is
partly in order to be consistent with published work and partly because
the concept of first-order rate constant is more meaningful to most
readers than that of half-order rate constant. The idea of half-order
kinetics has been raised here to illustrate that a good fit of data to
first-order kinetics does not preclude other possibilities. It is
likely that the apparent kinetics of denitrification in flooded systems
are complex and lie somewhere between first-order and half-order reaction
rates, as suggested in Chapter 1.

Mean first-order denitrification rate constants for the 14 soil
types and results of Duncan's New Multiple Ran e Test are presented in
Table 5. Rate constants range from 0.192 day" for the Everglades soil
to 0.040 day-1 for the Valkaria soil. Significant differences exist
among the soils, but groups of non-significantly different soils are
broad and overlapping. This is due primarily to the high degree of
variation within soil types. Spacial variation in soil organic matter
content and availability may account for this. Factors such as the
presence of decomposing soil fauna may increase organic matter avail-
ability and, consequently, denitrification rates. In cases where
decomposing animals were visually apparent, columns were discarded.

The last column in Table 5 gives nitrate-N removal rates in g ha-1
day-1, assuming maintenance of 10 mg liter-1 nitrate-N in the flood-
water. The Everglades soil will remove 2,900 g ha-lday-1, while only
600 g ha-lday-l is removed by the Valkaria soil. All denitrification
data given here assume a 15 cm floodwater depth. As noted in Chapter 2,
first-order rate constants are dependent on floodwater depth, but
equivalent nitrate-N removal rates are not.

The experimental conditions of this study include well-aerated
floodwater and no downward flow of water into the soil. As a result,
floodwater nitrate must diffuse through a relatively thick aerobic soil
zone into the zone of denitrification, and denitrification is relatively
slow. Downward flow of water, which often occurs in the field, can be
expected to increase denitrification rate by accelerating nitrate
transport. Reduced floodwater oxygen content would also increase
denitrification rate by narrowing the aerobic soil zone. For these
reasons, denitrification rates presented here should be regarded as
minimum expected in situ rates.

Selected characteristics of 15 soil types (Table 6) were examined
for possible correlation with denitrification rate. The 14 soil types
discussed above, plus Pelham soil, discussed later, were studied.
Simple and multiple linear regression models were constructed using
various combinations of characteristics as independent variables and
denitrification rate as the dependent variable. Complex models were
analyzed by the General Linear Models procedure of the Statistical


-52-












Table 5. Comparative denitrification rates of 14 representative
Florida wetland soils.


Soil type


Everglades muck

Floridana f.s.

Astor s.

Surrency s.

Samsula muck

Pickney f.s.

Riviera f.s.

Chastain si.l.

Brighton peat

Iberia s.c.

Chobee f.s.l.

Delray f.s.

Eureka f.s.l.

Valkaria f.s.


Deni trifi cation
rate constant

------------day

0.192

0.149

0.108

0.105

0.102

0.087

0.076

0.067

0.066

0.063

0.062

0.061

0.049

0.040


Standard
error
of mean

-----------

0.017 --

0.020 ab

0.018 bc

0.001 c

0.009 c

0.025 cd

0.009 cd

0.014 cd

0.010 cd

0.012 cd

0.009 cd

0.021 cd


0.005 d
0.005 d


Equivalent
nitrate-N t
removal rate

g ha-l day-
2,900

2,200

1,600

1,600

1,500

1,300

1,100

1,100

990

950

930

920

740

600


Assuming a nitrate-N concentration of

t Omitted from statistical analysis due
Rates followed by the same letter are
a = 0.05.


10 mg liter-1.
to missing data.
not significantly different;


-53-











Table 6. Selected chemical and physical characteristics of surface horizons (top 10 cm) of the
soils studied.

enitriication Total organic carbon Bulk
Denitrification Bulk
Soil type rate constant By vol. By wt. pH density C:N ratio

----day-]----- mg cm-3 --%-- g cm-3 ---w w---

Everglades muck 0.191 59 46.0 6.2 0.13 25:1
Floridana f.s. 0.149 127 12.6 6.5 0.97 8.8:1
Astor s. 0.108 60 7.4 6.7 0.81 6.3:1
Surrency s. 0.105 80 42.0 4.9 0.19 27:1
Samsula muck 0.102 38 54.0 4.6 0.08 16:1
IPickney f.s. 0.087 149 27.0 5.0 0.55 33:1
Pelham 1.f.s. 0.080o N.D.t 6.5 5.6 N.D. N.D.
Riviera f.s. 0.076 52 5.0 4.5 1.04 13:1
Chastain si.l. 0.067 24 5.2 5.6 1.37 N.D.
Brighton peat 0.066 44 53.0 4.2 0.09 21:1
Iberia s.c. 0.063 40 4.0 6.1 1.00 N.D.
Chobee f.s.l. 0.062 59 7.4 6.8 0.79 7.5:1
Delray f.s. 0.061 65 6.4 6.9 1.01 6.0:1
Eureka f.s.l. 0.049 48 5.7 5.5 0.84 6.6:1
Valkaria f.s. 0.040 6.9 0.43 6.6 1.60 19:1


Rate determined in long term study, discussed in later section.

N.D. = not determined.









Analysis System (Barr et al., 1976). Since denitrification is dependent
on an available carbon source, total organic carbon content on a weight
per soil volume basis and on a weight per soil weight basis were tested
as independent variables. The literature suggests that soil pH has a
significant effect on denitrification rate only at values somewhat below
neutrality. For this reason, a pH-deficit variable was introduced. The
new variable was defined as follows:

if observed pH>X, pH-deficit = 0

if observed ph
where X is an arbitrary pH threshold. A number of values for X were
tested, ranging from 5.5 to 7.0; pH 6.5 was found to perform best in
regression equations.

Organic carbon content on a weight per volume basis showed a rather
poor correlation with denitrification rate. Theoretically, this para-
meter would be expected to perform as well as or better than organic
carbon content by weight, since microbial activities tend to depend on
volume-based, rather than weight-based substrate concentrations. A
possible explanation for the poor performance of this variable lies in
the fact that volume-based organic carbon values were calculated by
multiplying weight-based values by bulk density. Bulk densities were
determined for the top 10 cm of each soil column. Substantial vertical
variation in bulk density within individual columns was visually ap-
parent, with very low densities existing in the top few centimeters of
some columns. Average bulk densities (and therefore volumetric carbon
contents) for 10 cm sections might thus not reflect those very close to
the surface, where much or all of the denitrification process occurs.

As single independent variables, neither pH-deficit nor organic
carbon by weight correlated well with denitrification rate, but models
combining these variables fit the data quite well. The best performing
model (r2 = 0.864, d.f. = 13) is given in Equation 5:


kI = (8.9xl04) (OCW) (3.9x10- )(OCW)(pH6.5) + 0.002 [51

where kl is apparent first-order denitrification rate constant, day-];
OCW is organic carbon content by weight, %; and pH6.5 is pH-deficit, as
described above. Although Equation 5 is a somewhat empirical relation-
ship, it complies with certain theoretical considerations. Introduction
of the square root factor is appropriate, since both the 1968 model of
Bouldin and the 1974 model of Bouldin et al. indicate that denitrifi-
cation rate is proportional to the square root of biochemical electron
acceptor demand. The above combination of organic carbon content by
weight, a pH correction factor for acid soils, and a small constant
appears to be a fairly good estimation of electron acceptor demand. The
constant (0.002) in the equation presents a problem by causing the
predicted kI value of a soil with no organic matter to be 0.04 when such
a soil would probably have a kI value of zero. It is worth noting


-55-








that because of the square root factor, a 50% reduction in electron
acceptor demand only reduces denitrification rate by 29%. This helps
explain the fairly narrow range of denitrification rates among the soils
studied, in spite of the wide range of organic carbon contents and
expected electron acceptor demands.

It is felt that Equation 5 could act as a simple predictor of
denitrification capacities of wetland soils in Florida. Only two easily
determined soil characteristics need be measured in order to make a
prediction. Table 7 gives observed denitrification rates for each soil
type, rates predicted by Equation 5, and percent error of predictions.
Some errors are relatively high, particularly for some of the more
slowly denitrifying soils. In view of the wide variation in denitri-
fication rates within soil types (Tables 5 and 9), broad confidence
intervals should be assumed for any predictive equation. Due to the
considerations noted above, the equation should be regarded as invalid
for soils with organic carbon contents below 0.5%. Fortunately, most
wetland soils in Florida are fairly high in organic carbon content.

Nitrous Oxide Evolution

Nitrous oxide evolution rates were determined for one column of
each of 14 of the soil types discussed above. Observed values were
adjusted to account for variations in floodwater nitrate concentration
at time of measurement. Adjusted evolution rates were correlated with
denitrification rates of the columns on which the N20 evolution measure-
ments were performed.

The continuous flow, semi-enclosed system used to determine N20
evolution has certain advantages and disadvantages as compared to the
more commonly used totally enclosed system. Perhaps the most important
advantage is that N20 levels in the soil:water:air system remain fairly
close to those which would occur in the field, where evolved N20 can
freely escape into the atmosphere. The highest observed N20 concentra-
tion in effluent air from the semi-enclosed system was approximately 2.5
times the normal ambient concentration. In contrast, completely closed
systems often allow N20 concentrations to increase to more than 10 times
the ambient concentration (Denmead, 1979). Thus the microorganisms in
closed systems are exposed to a highly N20 enriched environment which
might have an effect on further production and reduction of N20. A
major disadvantage of the semi-enclosed system is that low N20 concen-
trations result in diminished measurement accuracy. In some cases,
effluent concentrations were only 10% greater than ambient levels, and
therefore accuracy suffered. Five replicate samples per evolution
determination were used in an attempt to offset this problem, resulting
in a 15 ppb (v:v) confidence interval about the mean (a = 0.05). Use
of airstones in the collection system offered the advantage of main-
taining aerobic floodwater such as would often occur in the field, but
also resulted in some difficulties. Pores of airstones tended to clog,
sometimes reducing airflow to nearly zero. This caused many delays in
data collection, since a 12 hour equilibration period at a relative
constant flow rate was required before sampling.


-56-










Table 7. Denitrification rate constants as predicted by Equation 5.


Soil type


Everglades muck

Floridana f.s.

Astor s.

Surrency s.

Samsula muck

Pickney f.s.

Riviera f.s.

Chastain si.l.

Brighton peat

Iberia s.c.

Chobee f.s.l.

Delray f.s.

Eureka f.s.l.

Valkaria f.s.


Total
organic
pH carbon



6.2 46.0

6.5 12.6

6.7 7.4

4.9 42.0

4.6 54.0

5.0 27.0

4.5 5.0

5.6 5.2

4.2 53.0

6.1 4.0

6.8 7.4

6.9 6.4

5.5 5.7

6.6 0.43


By volume.

[(predicted-observed)/observed] x 100.


-57-


Rate constant

Observed Predicted

--------day-------....

0.192 0.193

0.149 0.115

0.108 0.093

0.105 0.113

0.102 0.098

0.087 0.108

0.076 0.050

0.067 0.069

0.066 0.057

0.063 0.070

0.062 0.093

0.061 0.088

0.049 0.069

0.040 0.049


Errors



+ 1

-23

-14

+ 8

- 4

+24

-34

+ 3

-14

+11

+50

+44

+41

+23








Sampling delays made it necessary to make N20 evolution determin-
ations at different floodwater nitrate-N concentrations for each column.
Concentrations ranged from 4 to 9 mg liter-1. In order to allow com-
parison of evolution values, observed values were adjusted to reflect a
10 mg liter-1 nitrate-N concentration in all columns. This was done on
the assumption that N20 evolution is directly proportional to floodwater
nitrate concentration. In an experiment designed to verify this as-
sumption, evolution rates from two Samsula and two Pelham columns were
measured as floodwater nitrate-N concentrations declined from 15 mg
liter-1 to less than 5 mg liter1-. Results are given in Figure 16 as
graphs of nitrate-N concentration versus N20 evolution. Plots are
reasonably linear, indicating that within the concentration range in
question, the above assumption is approximately valid. There is some
indication, particularly in the case of Samsula, that the relationship
deviates from linearity, but this deviation does not appear to be
extreme. It has been reported that N20 evolution relative to nitrate
concentration can increase as nitrate concentration increases, due to
nitrate inhibition of N20 reduction (Blackmer and Bremner, 1979; Terry
and Tate, 1980). Such inhibition does not appear to have occurred to
any significant degree in this experiment. Since nitrate must diffuse
from floodwater through an aerobic soil zone, it is likely that nitrate
concentrations in the zone of active denitrification were somewhat lower
than those in the floodwater. Such extremely low concentrations may not
have been sufficient to cause significant inhibition of N20 reduction.

Nitrous oxide evolution values for the 14 soil types are presented
in Table 8. Since evolution was determined for only one column per soil
type, values presented here may vary from the mean evolution rates of
the individual soils. It is likely, however, that the data accurately
represent the range of N20 evolution rates to be expected in Florida
wetland soils. The data also allow examination of the correlation
between N20 evolution and various soil characteristics. Evolution rates,
based on a 10 mg liter-1 floodwater nitrate-N concentration, range from
less than 5 g N ha-lday-1 for the Floridana soil to 42 g N ha-lday-1 for
the Riviera soil, with a mean of 20.9 g N ha-lday-1. The percentage of
consumed nitrate-N accountable for as N20-N ranges from less than 0.2%
for Floridana to 6.5% for the Valkaria, with a mean of 2.1%. Mean
values for organic soils are 18.5 g ha-ld y-l and 1.1%, while mineral
soils have mean values of 22.6 g ha-lday and 2.3%. The lower means
for organic soils may reflect a higher electron acceptor demand, as
discussed below. Percentage figures are in fairly close agreement with
the 0.8 to 1.4% values reported for flooded soils by Denmead et al.
(1979).

Nitrous oxide evolution and percent of consumed nitrate-N account-
able for as N20-N were tested for possible correlation with nitrate
consumption rate (apparent first-order denitrification rate constant)
and soil pH. Neither of the N 0 parameters showed any correlation with
soil pH. Nitrous oxide evolution as g N ha 'day-l was weakly, nega-
tively correlated with nitrate consumption rate (r2 = 0.164). Percent
N20 showed fairly strong correlation with nitrate consumption rate,


-58-

























Samsula
o column
x column


5 10 15
NO3 -N, mgl -


Pelham
o column A
x column B


5 10 15

NO3- N) mg 1'
3


Figure 16.


Nitrous oxide evolution from flooded soil cores in
relation to floodwater nitrate concentrations.


o


0

IN
z


z
I


-59-











Table 8. Nitrous
soils.


oxide evolution rates from 14 representative Florida wetland


Soil type



Everglades muck

Floridana f.s.

Astor s.

Surrency s.

Samsula muck

Pickney f.s.

Riviera f.s.

Chastain si.l.

Brighton peat

Iberia s.c.

Chobee f.s.l.

Delray f.s.

Eureka f.s.l.

Valkaria f.s.

mean


Denitrification
rate constant

---- day-1-----


0.175

0.112

0.077

0.105

0.110

0.065

0.065

0.092

0.081

0.050

0.071

0.043

0.161

0.040


Floodwater
N03-N conc.

mg liter-1


4.0

4.0

7.0

5.0

8.0

6.5

5.0

6.0

7.5

9.0

7.0

9.0

3.5

8.5


N20-N

Observed

---------g


7.2

t

7.0

13.4

15.4

13.1

21 .0

15.5

7.9

24.1

10.9

18.5

5.0

32.5


evolution rate

Adjusted % N20O

ha-lday-1


18.0

t

10.0

26.8

19.3

20.2

42.0

25.8

10.7

26.8

15.6

20.6

14.3

38.2

20.9


0.7

0.2

0.9

1.7

1.2

2.1

4.4

1.4

0.9

3.6

1.5

3.2

0.6

6.5

2.1


Rate for individual column, not mean rate for soil type.

Adjusted to 10 mg liter-1 floodwater N03-N.
(N20-N evolved/N03-N consumed) x 100.


-60-









as indicated in Figure 17, and in the equation below (r2 = 0.706, d.f.
= 12):

%N20 = 0.25(0/kI) 1.26 [6]

where kI is the first-order denitrification rate constant, day-l, and
%N20 is (N20-N evolution rate/nitrate-N consumption rate) x 100.
Equation 6 agrees with the relationship between N20 release and electron
acceptor demand reported by Focht and Vertraete (1977). In this case,
denitrification rate can be considered an indication of electron ac-
ceptor demand. Apparently, in soils with high electron acceptor demand,
N20 produced by denitrification tends to be further reduced to N2, thus
acting as an additional sink for electrons. In soils with lower elec-
tron acceptor demands, less of the N20 produced is further reduced and
thus more is evolved.

Long Term Denitrification and Nitrous Oxide Evolution

Columns of Samsula and Pelham soils (eight per soil type) were
incubated for 10 weeks with floodwater nitrate-N concentrations main-
tained between 5 and 20 mg liter-1. Denitrification rates were deter-
mined at 2, 5, and 8 weeks. Nitrous oxide evolution from two Samsula
and three Pelham columns was also measured periodically during this
time.

Denitrification rates are given in Table 9. Mean denitrification
rates for the two soil types remained relatively constant throughout the
study period. Individual denitrification rates fluctuated somewhat,
however. In most columns, the change in rate appeared to be random.
This may reflect irregular transport of nitrate into the soil. On
several occasions, relatively large volumes of gases (presumably bio-
genic) were observed to be rapidly released from the soil. Release of
gases had a stirring effect on the upper layer of soil, resulting in a
loose, flocculent soil mass. In one case, a layer of soil material
approximately I cm thick was observed floating on top of the floodwater,
having been lifted by trapped gas. Such disruptions of the soil surface
could result in increased transport of floodwater nitrate into the soil
and thus increase the denitrification rate for short periods following
disruptions. It may be assumed that this effect also occurs in nature.
Temporal variation in decomposition rates of soil, plant and animal
matter may also have contributed to variations in individual column
denitrification rates.

In the case of Samsula columns 5 and 6, denitrification rates rose
drastically during the study period. The rise was accompanied by
darkening of the floodwater, probably due to accumulation of dissolved
organic substances. Darkening of floodwater occurred to a lesser degree
in all Samsula columns. Increased organic substrate availability is a
likely cause for the accelerated denitrification. It is possible that
columns 5 and 6 contained plant rhizomes or other organic materials
which decomposed only after a lag period. Such lag periods are not
uncommon in soil microbiological systems, and may be the result of such
factors as substrate recalcitrance and microbial succession (Stotzky,
1973).


-61-

















%/0N 20 =0.25 (I1) -126
2 I


r 2 =O.706


0
0
0 5-


E

o 4-
U
z

0
z
-z 3-

0


2-










0 5 10 15 20
I (days)
k






Figure 17. Relationship between reciprocal of denitrification rate
constant and percent nitrous oxide evolution for 14
soil types.


-62-









Table 9. Long term (8 week) denitrification rates in Samsula and Pelham
soil columns.

Denitrification rate constant

Soil type Replicate Week 2 Week 5 Week 8
---------------------day----------------

Samsula muck 1 0.093 0.080 0.073
2 0.109 0.062 0.074
3 0.067 0.077 0.054
4 0.049 0.080 0.105
5 0.112 0.091 0.208
6 0.105 0.209 0.237
7 0.060 0.062 0.059
8 0.101 0.064 0.074
avg. 0.0870.021 0.0910.041 0.0950.050

Pelham 1.f.s. 1 0.090 0.097 0.094
2 0.064 0.081 0.065
3 0.101 0.062 0.100
4 0.078 0.100 0.074
5 0.053 0.037 0.054
6 0.091 0.061 0.064
7 0.092 0.072 0.048
8 0.051 0.071 0.054
avg. 0.0780.016 0.0730.017 0.0690.016


-63-








Nitrous oxide evolution from Samsula (columns 3 and 8) and Pelham
(columns 5, 6, and 7) soils are presented in Figure 18. Evolution is
shown as percent of consumed nitrate-N accountable for as N20-N. Nitrate
consumption is based on the denitrification rate occurring in each
column at the time N20 evolution was determined. In three of the five
columns, evolution decreased during the study period. This agrees with
the observation of Guthrie and Duxbury (1979) that N20 evolution decreases
with increased time of soil exposure to nitrate. This decrease would be
expected to result from microbial adaptation and selection. With time,
the microbial community is likely to maximize use of N20 as an electron
acceptor, since electron acceptors are in short supply in anoxic environ-
ments. Nitrous oxide evolution increased in the case of column 7 of
Pelham soil. The first value for this column is extremely low, and may
be in error. Unusual population dynamics or soil conditions may be
responsible for the upward trend in evolution.

An interesting phenomenon was accidently observed in column 8 of
the Samsula soil (Figure 18). Air flow in the column was inadvertently
reduced to less than 3 ml min-1, and floodwater dissolved oxygen concen-
tration dropped to approximately 1 mg liter-1. Coinciding with this was
a peak in percent N20 evolution of 47 g ha-lday-1. Evolution subsided
to 35 g ha-lday-l one day later, after flow was increased to 7 ml min-l
and dissolved oxygen content increased to about 7 mg liter-1. A similar
effect occurred in column 4 of Samsula soil during an earlier, unreported
phase of this study. Aeration slowed and dissolved oxygen concentration
dropped to less than 1 mg liter-1, resulting in an evolution rate of 40
g N ha-lday-1. Five days earlier, evolution from this column had been
23 g N ha-lday-1 and 9 days later the value was 16 g N ha-lday-1, with
more adequate aeration in both cases. The observed increases in N20
evolution rate may be due to accelerated denitrification resulting from
low floodwater dissolved oxygen, a phenomenon reported by Van Kessell
(1977). No increase in denitrification rate could be detected in the
present study because the periods of anoxia were very short compared to
nitrate sampling intervals.

Regardless of the actual cause, variation in floodwater oxygen
content appears to have a marked effect on N20 evolution and perhaps on
denitrification. The problem arises as to how closely results presented
in this study reflect those occurring in nature. With the above excep-
tions, floodwater dissolved oxygen content remained relatively constant
throughout the study. In nature, dissolved oxygen content often under-
goes diurnal fluctuations due to daytime oxygen production during
photosynthesis. This difference may have resulted in lower observed N20
evolution and denitrification rates in this study than would occur in
the field.


-64-














Samsula
- column 3
column 8


4-








02-
1 \


Weeks after initial nitrate application


Pelham
o.o column 5
- column 7
- column 8


I..


Weeks after initial nitrate application


Figure 18.


Long term (10 week) nitrous oxide evolution from flooded
Samsula and Pelham soil columns.


-65-









SUMMARY AND CONCLUSIONS


Denitrification and N20 evolution rates in a wide variety of
Florida wetland soils were investigated in the laboratory using intact
soil columns. It was determined that denitrification in the flooded
soils followed apparent first-order kinetics, with rate constants
ranging from 0.040 day-1 to 0.192 day-]. Denitrification rate was
related to soil organic carbon content and pH by the following formula:


kI = (8.9xlO4)(OCW) (3.9xlO )(OCW)(pH6.5) + 0.002

where k1 is denitrification rate constant, day-1; OCW is organic carbon
content by weight, %; and pH6.5 is pH deficit below pH 6.5. Residence
times required for 50% removal of floodwater, assuming a floodwater
depth of 15 cm and irrespective of nitrate concentration, ranged from
3.6 to 17 days. For 90% removal, 12 to 58 days would be required. These
figures represent minimum removal rates. Either mass flow of floodwater
into the soil or low floodwater dissolved oxygen content would be likely
to result in higher removal rates. Assimilation of nitrate by vascular
plants would also increase removal rates significantly. It is felt that
many Florida wetland soils could function as efficient treatment systems
for nitrate-bearing wastewater. Soil organic carbon content and pH, as
well as hydrologic factors need to be considered in site selection.

A semi-enclosed chamber system was devised for determination of N20
evolution rates. This system functioned satisfactorily without exposure
of the soil:water environment to greatly elevated N20 concentrations, as
occurs in totally enclosed systems. Nitrous oxide accounted for from
less than 0.2 to 6.5% of consumed nitrate. Wetland denitrification thus
appears to be a relatively innocuous means of nitrate disposal, from the
viewpoint of N20 production and stratospheric ozone destruction. Soils
which denitrified rapidly tended to evolve less N20 than those with
slower denitrification rates. Contrary to many reports, N20 evolution
in the soils studied was not related to soil pH. This may be due to the
relatively low floodwater nitrate-N concentrations (less than 20 mg
liter-1) in the present study.

Two soil types were studied during a 10 week period of continuous
nitrate exposure. Denitrification rates of individual columns tended to
fluctuate randomly, but average rates remained relatively constant.
Average N20 evolution rate declined by approximately 60% between weeks
3 and 10, probably due to microbial adaptation and selection. Reduction
in floodwater dissolved oxygen content appeared to result in increased
N20 evolution.

Further study is needed to determine more precisely the effect of
naturally occurring fluctuations in floodwater oxygen content on de-
nitrification and N20 evolution rates. Application of high BOD waste-
water could be studied in conjunction with this. The effect of


-66-










downward water flow on the above rates also warrants investigation.
Denitrification and N20 evolution should be investigated in other
wetland ecosystems, including estuarine environments. In situ studies
and refinement of in situ methods, particularly for N20 evolution
measurement, are needed in order to confirm the results of laboratory
work.


-67-









LITERATURE CITED


Alexander, M. 1965. Nitrification. In W. V. Bartholomew and F. E.
Clark (eds.), Soil Nitrogen. Agronomy 10:307-343. American
Society of Agronomy, Madison, WI.

Alexander, M. 1977. The nitrogen cycle: Nitrification; Denitrifi-
cation. In Introduction to soil microbiology. 2nd ed. John
Wiley and Sons, Inc. p. 251-286.

Allison, L. E. 1965. Organic carbon. In C. A. Black (ed.), Methods
of soil analysis. Agronomy 9:1367-1378. American Society of
Agronomy, Madison, WI.

Andersen, J. M. 1976. Importance of the denitrification process for
the rate of degradation of organic matter in lake sediments. In
H. L. Golterman (ed.), Interactions between sediments and fresh
water. Dr. W. Junk B. V. Publishers, the Hague. p. 357-362.

Ardakani, M. S., L. W. Baker, and A. D. McLaren. 1975. Reduction
of nitrate in a soil column during continuous flow. Soil Sci.
Soc. Am. Proc. 30:290-296.

Armstrong, W. 1967. The oxidising activity of roots in waterlogged
soils. Physiol. Plant. 20:920-926.

Bailey, L. D. 1976. Effects of temperature and root on denitrifi-
cation in a soil. Can. J. Soil Sci. 56:79-87.

Bailey, L. D., and E. G. Beauchamp. 1973. Effects of temperature
on nitrate and nitrite reduction, nitrogenous gas production,
and redox potential in a saturated soil. Can. J. Soil Sci.
53:213-218.

Barr, A. S., J. H. Goodnight, J. P. Sail, and J. T. Helwig. 1976.
A user's guide to S.A.S. S.A.S. Institute, Raleigh, NC.

Blackmer, A. M., and J. M. Bremner. 1978. Inhibitory effect of
nitrate on reduction of nitrous oxide to nitrogen by soil micro-
organisms. Soil Biol. Biochem. 10:187-191.

Blackmer, A. M., and J. M. Bremner. 1979. Stimulatory effect of
nitrate on reduction of nitrous oxide to nitrogen by soil micro-
organisms. Soil Biol. Biochem. 11:313-315.

Bouldin, D. R. 1968. Models for describing the diffusion of oxygen
and other mobile constituents across the mud-water interface.
J. Ecol. 56:77-87.

Bouldin, D. R., R. L. Johnson, C. Burda, and C. Kao. 1974. Losses
of inorganic nitrogen from aqueous systems. J. Environ. Qual.
2:107-113.


-68-







Boyt, F. L., S. E. Bayley, and J. Zoltek. 1977. Removal of nutrients
from treated municipal wastewater by wetland vegetation. J. Water
Pollut. Control Fed. 48:789-799.

Bremner, J. M. 1965. Inorganic forms of nitrogen. In C. A. Black
(ed.), Methods of soil analysis. Agronomy 9:1179-1206. American
Society of Agronomy, Madison, WI.

Bremner, J. M., and A. M. Blackmer. 1979. Effects of acetylene and
soil water on emission of nitrous oxide from soil. In Agronomy
abstracts. American Society of Agronomy, Madison, WT. p 26.

Bremner, J. M., and K. Shaw. 1958. Denitrification in soil. II.
Factors effecting denitrification. J. Agric. Sci. 51:40-52.

Brezonik, P. L. 1977. Denitrification in natural water. Prog.
Wat. Tech. 8:373-392.

Broadbent, F. E., and F. E. Clark. 1965. Denitrification. In W. V.
Bartholemew and F. E. Clark (eds.), Soil nitrogen. Agronomy
10:347-362. American Society of Agronomy, Madison, WI.

Buresh, R. J., and W. H. Patrick. 1978. Nitrate reduction to ammonia
in anaerobic soil. Soil Sci. Soc. Am. J. 42:913-918.

Burford, J. R., and J. M. Bremner. 1975. Relationship between the
denitrification capacities of soils and total, water, soluble,
and readily decomposable soil organic matter. Soil Biol. Biochem.
7:389-394.

Burford, J. R., and R. C. Stephanson. 1973. Measurement of gaseous
loss of nitrogen from soils. Soil Biol. Biochem. 5:133-141.

Chen, R. L., D. R. Keeney, J. C. Konrad, A. J. Holding, and D. A.
Graetz. 1972. Gas production in sediments of Lake Mendota,
Wisconsin. J. Environ. Qual. 1:155-157.

Coleman, J. M. 1979. Present and past nutrient dynamics of a small
pond in Southwest Florida. Ph.D. Thesis. Univ. of Fla.

Cooper, G. S., and R. L. Smith. 1963. Sequence of products formed
during denitrification in some diverse western soils. Soil Sci.
Soc. Am. Proc. 27:659-662.

Council for Agricultural Science and Technology (C.A.S.T.). 1976.
Effect of increased nitrogen fixation on stratospheric ozone.
Report #53. C.A.S.T.

Craswell, E. T. 1978. Some factors influencing denitrification
and nitrogen immobilization in a clay soil. Soil Biol. Biochem.
10:241-245.

Day, P. R., H. E. Doner, and A. D. McLaren. 1978. Relations
among microbial populations and rates of nitrification and
denitrification in a Hanford soil. In D. R. Nielsen and J. G.
MacDonald (eds.), Nitrogen in the environment. V. 2: Soil-
plant-nitrogen relationships. Academic Press. New York.
p. 305-363.
-69-









Delwiche, C. C., and B. A. Bryan. 1976. Denitrification. Ann. Rev.
Microbiol. 30:241-262.

Denmead, 0. T. 1979. Chamber systems for measuring nitrous oxide
emission from soils in the field. Soil Sci. Soc. Am. J. 43:
89-95.

Denmead, 0. T., J. R. Freney, and J. R. Simpson. 1979a. Nitrous
oxide emission during denitrification in a flooded field. Soil
Sci. Soc. Am. J. 43:716-718.

Denmead, 0. T., J. R. Freney, and J. W. Simpson. 1979b. Studies of
nitrous oxide emission in a grass sward. Soil Sci. Soc. Am. J.
43:726-728.

Doner, H. E., and A. D. McLaren. 1978. Temperature dependence of
nitrate losses and denitrifier populations in soil columns.
In W. E. Krumbein (ed.), Environmental biogeochemistry and
geomicrobiology. V. 2: Geomicrobiology, biogeochemistry, and
energy flow. Ann Arbor Science, Ann Arbor, MI. p. 573-582.

Doner, H. E., M. G. Volz, and A. D. McLaren. 1974. Column studies
of denitrification in soils. Soil Biol. Biochem. 6:341-346.

Ekpete, D. M., and A. H. Cornfield. 1965. Effect of pH and addition
of organic matter on denitrification losses from soil. Nature.
208:1200.

Engler, R. M., and W. H. Patrick. 1974. Nitrate removal from flood-
water overlying flooded soils and sediments. J. Environ. Qual.
3:409-415.

Eno, C. F. 1960. Nitrate production in the field by incubating the
soil in polyethylene bags. Soil Sci. Soc. Am. Proc. 24:277-279.

EPA. 1973. Nitrification and Denitrification Facilities:Wastewater
Treatment. U. S. Environmental Protection Agency Tech. Transfer
Publication. (EPA-625/4-73-994a).

EPA. 1975. Process Design Manual for Nitrogen Control. U. S. Environ-
mental Protection Agency.

Erickson, N. L. 1978. Nitrification-denitrification in a freshwater
marsh receiving secondarily treated sewage effluent. M.S. Thesis.
Univ. of Fla.

Fetter, C. W., W. E. Sloey, and F. L. Spangler. 1978. Use of a
natural marsh for wastewater polishing. J. Water Pollut. Control
Fed. 50:290-307.

Focht, D. D. 1974. The effect of temperature, pH, and aeration on
the production of nitrous oxide and gaseous nitrogen--a zero-
order kinetic model. Soil Sci. 118:173-179.


-70-









Focht, D. D., and W. Verstraete. 1977. Biochemical ecology of
nitrification and denitrification. Adv. Microbiol. Eco. 1:
124-204.

Galsworthy, A. M., J. R. Burford, and D. J. Greenland. 1978. Model-
ing of rates of nitrification and denitrification in soil.
In D. R. Nielsen and J. G. MacDonald (eds.), Nitrogen in the
environment. V. 2: Soil-plant-nitrogen relationships. Academic
Press, New York. p. 377-392.

Gilliam, J. W., and R. P. Gambrell. 1978. Temperature and pH as
limiting factors in loss of nitrate from saturated Atlantic
Coastal Plain soils. J. Environ. Qual. 7:526-532.

Gucker, F. T., and R. L. Seifert. 1966. Chemical kinetics. In
Physical chemistry. W. W. Norton and Co., New York. p. Tl -
734.

Guthrie, T. F., and J. M. Duxbury. 1978. Nitrogen mineralization
and denitrification in organic soils. Soil Sci. Soc. Am. J.
42:908-912.

Howeler, R. H. 1972. The oxygen status of lake sediments. J. Environ.
Qual. 4:366-371.

Howeler, R. H., and D. R. Bouldin. 1971. The diffusion and consump-
tion of oxygen in submerged soils. Soil Scl. Soc. Am. Proc.
35:202-208.

Kaspar, H. F., and J. M. Tiedje. 1979. Nitrous oxide evolution
from soil cores. In Agronomy abstracts. American Society of
Agronomy, Madison, WI.

Keeney, D. R., and W. R. Gardner. 1970. The dynamics of nitrogen
transformations in the soil. In S. F. Singer (ed.), Global
Effects of Environmental Pollution. Springer-Verlag, New York.
p. 96-103.

Lance, J. F., and T. D. Whisler. 1972. Nitrogen balance in soil
columns intermittently flooded with secondary sewage effluent.
J. Environ. Qual. 2:180-186.

Liang, C. N., and M. A. Tabatabai. 1978. Effects of trace elements
on nitrification in soils. J. Environ. Qual. 7:291-293.

Matulewich, V. A., P. F. Strom, and H. S. Finsteln. 1975. Length
of incubation for enumerating nitrifying bacteria present in
various environments. Appl. Microblol. 29:265-268.

McGarity, J. W., and R. J. K. Meyers. 1968. Denitrifying activity
in solodized solonetz soils of Eastern Australia. Soil Sci.
Soc. Am. Proc. 32:812-817.


-71-








Moraghan, J. F., and R. J. Buresh. 1977. Chemical reduction of
nitrite and nitrous oxide by ferrous iron. Soil Sci. Soc. Am. J.
44:47-50.

Nash, C. L., and J. Bollag. 1974. Comparative denitrification of
selected microorganisms in a culture medium and in autoclaved
soil. Appl. Microbiol. 27:674-677.

Nelson, D. W., and J. M. Bremner. 1970. Role of soil minerals and
metallic cations in nitrate decompositon and chemodenitrifica-
tion in soils. Soil Biol. Biochem. 2:1-8.

Nommik, H. 1956. Investigations on denitrification in soil. Acta.
Agr. Scand. IV. 2:195-228.

Odum, H. T., and K. C. Ewel (eds.). 1977. "Cypress wetlands for water
management, recycling, and conservation. Fourth annual report
to the Rockefeller Foundation and N.S.F. Center for Wetlands,
Univ. of Fla.

Patrick, W. H., Jr., and R. D. Delaune. 1976. Nitrogen and phosphorus
utilization by Spartina alterniflora in a salt marsh in Barataria
Bay, Louisiana. Estuarine and Coastal Marine Sci. 4:59-64.

Patrick, W. H., and K. R. Reddy. 1976. Nitrificatlon-denitrification
reactions in flooded soils and water bottoms: Dependence on
oxygen supply and ammonium diffusion. J. Environ. Qual. 5:
469-472.

Patrick, W. H., Jr., and M. E. Tusneem. 1972. Nitrogen loss from
flooded soil. Ecology. 53:735-737.

Payne, W. J. 1973. Reduction of nitrogenous oxides by microorganisms.
Bacteriol. Rev. 37:409-452.

Phillips, R. E., K. R. Reddy, and W. H. Patrick. 1978. The role
of nitrate diffusion in determining the order and rate of deni-
trification in flooded soil: II. Theoretical analysis and
interrelation. Soil Sci. Soc. Am. J. 42:272-278.

Reddy, K. R. 1976. Nitrification-denltrification reactions in
flooded soils. Ph.D. Thesis. Louisiana State University.

Reddy, K. R., and D. A. Graetz. 1980. Use of shallow reservoirs
and flooded soil systems for agricultural drainage water treat-
ment: Inorganic nitrogen removal. J. Environ. Qual. (in
review).

Reddy, K. R., and W. H. Patrick. 1975. Effect of alternate aerobic
and anaerobic conditions on redox potential, organic matter
decomposition, and nitrogen loss in a flooded soil. Soil Biol.
Biochem. 7:87-94.


-72-









Reddy, K. R., W. H. Patrick, Jr., and R. E. Phillips. 1976. Ammonium
diffusion as a factor in nitrogen loss from flooded soils. Soil
Sci. Soc. Am. J. 40:528-533.

Reddy, K. R., W. H. Patrick, and R. E. Phillips. 1978. The role of
nitrate diffusion in determining the order and rate of denitri-
fication in flooded soil: I. Experimental results. Soil Sci.
Soc. Am. J. 42:268-272.

Reddy, K. R., P. D. Sacco, and D. A. Graetz. 1980. Nitrate reduction
in a soil-water system. J. Environ. Qual. (in press).

Requa, D. A., and E. D. Schroeder. 1973. Kinetics of packed bed
denitrification. J. Water Pollut. Control Fed. 45:2696-1707.

Rice, E. L., and S. K. Pancholy. 1973. Inhibition of nitrification
by climax ecosystems. II. Additional evidence and possible role
of tannins. Amer. J. Bot. 60:691-702.

Sherr, B. T., and W. J. Payne. 1978. Effect of Spartina alterniflora
root-rhizome system on salt marsh soil denitrifying bacteria.
Appl. Environ. Microbiol. 35:724-729.

Small, M. M. 1978. Artificial wetlands as non-point source waste-
water treatment systems. In M. A. Drew (ed.), Environmental
quality through wetland utilization. Coordinating Council on
the Restoration of the Kissimmee River Valley and Taylor Creek-
Nubbin Slough Basin, Tallahassee, FL. p. 171-181.

Sorensen, J. 1978. Capacity for denitrification and reduction of
nitrate to ammonia in a coastal marine sediment. Environ.
Microbiol. 35:301-325.

Stanford, G., J. 0. Legg, S. Dzienia, and E. C. Simpson. 1975.
Denitrification and associated nitrogen transformations in
soils. Soil Sci. 120:147-151.

Stanford, G., R. A. Vander Pol, and S. Dzienta. 1975. Denitrifi-
cation rates in relation to total and extractable soil carbon.
Soil Sci. Soc. Am. Proc. 39:284-289.

Starr, J. L., and J. Y. Parlange. 1975. Nonlinear denitrification
kinetics with continuous flow in soil columns. Soil Sci. Soc.
Am. Proc. 39:875-881.

Stensel, H. D., R. C. Loehr, and A. W. Lawrence. 1973. Biological
kinetics of suspended growth denitrification. J. Water Pollut.
Control Fed. 45:249-261.

Stefanson, R. C. 1972. Denitrification in sealed soil-plant systems.
Plant Soil. 33:113-149.

Stotzky, G. 1973. Activity, ecology, and population dynamics of
microorganisms in soil. C.R.C. Review of Microbiology. Chemical
Rubber Co., Cleveland. p. 57-124.


-73-









Struble, J. E. 1977. Nitrification, denitrification, and sedimenta-
tion rates in cypress ponds receiving secondarily treated sewage
effluent. M. S. Thesis. Department of Soil Science, Univ. of
Fla.

Tate, R. L., III. 1977. Nitrification in histosols: A potential
role for the heterotrophic nitriffer. Appl. and Environ.
Microblol. 33:911-914.

Terry, R. E., and R. L. Tate. 1980. The effect of nitrate on nitrous
oxide reduction in organic soils and sediments. Soil Sci. Soc.
Am. J. (in press).

Tiedje, J. M., N. V. Caskey, M. S. Smith, and R. B. Firestone. 1979.
Nitrous oxide production by bacteria that reduce nitrate to
ammonia. Agronomy abstracts. American Society of Agronomy,
Madison, WI. p. 165.

Tinsdale, S. L., and W. L. Nelson. 1975. Soil and Fertilizer nitrogen.
In Soil Fertility and Fertilizers. 3rd ed. MacMillan Pub. Co.,
Inc., New York. p. 122-188.

Tusneem, M. E., and W. H. Patrick. 1971. Nitrogen transformations
in waterlogged soils. Louisiana State Univ. Dept. of Agronomy.
Bulletin #657.

Van Kessel, J. F. 1977. Factors affecting denitrification rates in
two water-sediment systems. Water Res. 11:259-267.

Van Kleemput, 0., W. H. Patrick, and R. C. McIlhenny. 1975. Forma-
tion of chemical and biological denitrification products at
controlled pH and redox potential. Soil Biol. Biochem. 7:
329-332.

Van Kleemput, 0., W. H. Patrick, and R. C. Mcllhenny. 1976. Nitrate
decomposition in flooded soils under different pH and redox
conditions. Soil Sci. Soc. Am. J. 40:55-60.

Verhoeven, W. 1956. Some remarks on nitrate and nitrate metabolism
in microorganisms. In W. D. McElroy and B. Glass (eds.), A
symposium on inorganic nitrogen metabolism. Johns Hopkins
Press, Baltimore.

Verstraete, W., and M. Alexander. 1973. Heterotrophic nitrification
in samples of natural ecosystems. Environ. Sci. Tech. 7:39-42.

Volz, M. G., M. S. Ardakani, R. K. Scholz, L. H. Stolzy, and A. D.
McLaren. 1976. Soil nitrate loss during irrigation: Enhance-
ment by plant roots. Agron J. 68:621-627.

White, J. P., D. P. Schwert, J. P. Ondrako, and L. L. Morgan. 1977.
Factors affecting nitrification in situ in a heated stream.
Appl. and Environ. Microbiol. 33:918-925.


-74-









Wijier, J., and C. C. Delwiche. 1954. Investigation on the denitrify-
ing process in soil. Plant Soil. 5:155-169.

Yuan, T. L. 1974. A double buffer method for the determination of
lime requirement of acid soils. Soil Sci. Soc. Am. Proc. 38:
437-440.

Zoltek, J., S. E. Bayley, S. J. Hermann, L. R. Tortora, and T. J.
Dolan. 1979. Removal of nutrients from treated municipal
wastewater by freshwater marshes. Final report to city of
Clermont, Florida. Center for Wetlands, Univ. of Fla.


-75-




Full Text

PAGE 1

WATER IiRESOURCES .... researc center Publication No. 48 DENITRIFICATION IN WETLANDS AS A MEANS OF WATER QUALITY IMPROVEMENT By D. A. Graetz (Principal Investigator) Soil Science Department, IFAS University of Florida Gai nesvi 11 e UNIVERSITY OF FLORIDA

PAGE 2

Denitrification in Wetlands as a Means of Water Quality Improvement By D. A. Graetz, P. A. Krottje, N. L. Erickson, J. G. A. Fiskell and D. F. Rothwell PUBLICATION NO. 48 FLORIDA WATER RESOURCES RESEARCH CENTER RESEARCH PROJECT TECHNICAL COMPLETION REPORT OWRT Project Number B-035-FLA Matching Grant Agreement Number 14-34-0001-8073 Report Submitted: May, 1980 The work upon which this report is based was supported in part by funds provided by the United States Department of the Interior, Office of Water Research and Technology as Authorized under the Water Resources Research Act of 1964 as amended.

PAGE 4

ACKNOWLEDGEMENTS The authors wish to extend appreciation to the Office of Water Research and Technology, United States Department of the Interior, for financial support of this work. administrative assistance of Dr. W. H. Morgan {former Director} and Dr. James Heaney {present Director} of the Florida Water Resources Research Center is greatly appreciated. Thanks are also due Ms. Mary Robinson for accounting assistance. Sup plemental financial support of the Center of Environmental Programs, Institute of Food and Agricultural Sciences, University of Florida is gratefully acknowledged. Thanks are also due Drs. S. E. Bayley and J. Zoltek, Jr. for al low ing us to cooperate on their study of the Clermont marsh, and to Dr. K. R. Reddy for his constructive advice and assistance. The assistance of U.S.D.A. Soil Conservation Service personnel in locating sampl ing sites was extremely helpful. Special thanks go to Candace Cantlin and Diane Sompongse for their assistance in the field and in the laboratory.

PAGE 6

TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES LIST OF FIGURES ABSTRACT INTRODUCTION CHAPTER I LITERATURE REVIEW Wetlands as Wastewater Treatment Systems Nitrification Denitrification Denitrification Kinetics: Real and Apparent Additional Mechanisms for Nitrate Loss Nitrous Oxide Production Summary CHAPTER I I NITRIFICATION AND DENITRIFICATION IN A MARSH SOIL iv v vii 2 2 2 5 7 9 11 12 RECEIVING SECONDARILY TREATED SEWAGE EFFLUENT 13 METHODS AND MATERIALS 13 Soil and Water Sampling 13 Description of Experiments 14 Analytical Procedures 17 RESULTS AND DISCUSSION 18 Nitrification 18 Denitrification 27 SUMMARY AND CONCLUSIONS 36 CHAPTER I I I DENITRIFICATION IN FIFTEEN FLORIDA WETLAND SOILS 37 METHODS AND MATERIALS 37 i i ') )

PAGE 7

Soi ls and Soil Sampling Descriptions of Experiments Method of Measuring N20 Evolution Analytical Methods.. .. RESULTS AND DISCUSSION Comparative Denitrification Rates Nitrous Oxide Evolution ..... Long Term Denitrification and Nitrous Oxide Evolution SUMMARY AND CONCLUSIONS LITERATURE CITED iii Page 37 42 43 43 46 46 56 61 66 68

PAGE 8

1. LIST OF TABLES Effect of charcoal fi ltration and inoculation w'i th soi 1 water extract on nitrification in marsh and lake waters wi thout soi 1. 2. Ammonium removal rates from overlying water of soil:water columns containing ammonium amended sewage effluent. 3. Nitrate removal as affected by water source, associated cation and presence of soil. 4. Nitrate removal rates from overlying water of soil:water columns containing nitrate amended sewage effluent. 5. Comparative denitrification rates of 14 representative Florida wetland soils. 6. Selected chemical and physical characteristics of surface horizons (top 10 em) of the soils studied. 7. Denitrification rate constants as predicted by Equation 5. 8. Nitrous oxide evolution rates from 14 representative Florida wetland soils. 9. Long term (8 week) denitrification rates in Samsula and Pelham soil columns. iv 19 24 28 29 53 54 57 60 \ /1 63

PAGE 10

LIST OF FIGURES Page 1. Nitrate diffusion fronts in flooded soil systems (a) with 8 no aerobic zone, and (b) wIth aerobic zone. 2. Experimental soil :water columns for nitrification and denitrification studies. 3. Nitrification in soil:water columns flooded with 15 cm of 21 ammonium amended sewage effluent, with [0] and without [e] plants. 4. Nitrification in soil :water columns flooded with 30 cm of 22 ammonium amended sewage effluent, with [0] and without [e] plants. 5. Nitrification in limed soil :water columns flooded with 23 15 cm of ammonium amended sewage effluent, with [0] and without [e] plants. 6. Ammonium and nitrate concentrations in ammonium amended 26 marsh soil during situ incubation in polyethylene bags. 7. Nitrate removal from overlying water of soil :water columns 30 (with [0] and without [e) plants) containing nitrate amended sewage effluent to a depth of 15 cm. Nitrate was added on day 0 and day 18. 8. Nitrate removal from overlying water of soil :water columns 31 (with [0] and without [e) plants) containing nitrate amended sewage effl uent to a depth of 30 cm. Nitrate was added on day 0 and day 18. 9. Nitrate removal from overlying water of limed soil :water 32 columns (with [0] and without [e) plants) containing nitrate amended sewage effluent to a depth of 15 cm. Nitrate was added on day 0 and day 18. 10. Nitrate and ammonium concentrations in nitrate amended 34 marsh soil during situ incubation in polyethylene bags. 11. Location of soil sampling sites. 38 12. Apparatus for obtaining soil cores with minimum soi 1 41 compaction. 13. System for determining nitrous oxide evolution from flooded 44 soi 1 cores. v

PAGE 11

LI ST OF FIGURES (continued) 14. Changes in floodwater nitrate concentrations with time 47 in soil:water columns of 14 soil types. 15. Changes in floodwater nitrate concentrations with time 51 in two soil:water columns of. Surrency soil. 16. Nitrous oxide evolution from flooded soil cores in 59 relation to floodwater nitrate concentrations. 17. Relationship between reciprocal of denitrification rate 62 constant and percent nitrous oxide evolution for 14 soil types. 18. Long term (10 week) nitrous oxide evolution from flooded 65 Samsula and Pelham soil columns. vi '-",_ ... )

PAGE 12

ABSTRACT In recent years, much attention has been focused on the use of wetlands for nutrient removal from polluted waters such as secondarily treated sewage effluent and agricultural drainage waters. In the case of non-point sources such as agricultural runoff, wetland treatment may be the only feasible means of pollution abatement. Nitrification and subsequent denitrification can be an important route for the removal of nitrogen from water flowing through a wetland ecosystem. The objectives of this research were (I) to determine the extent of nitrification and denitrification occurring in a marsh ecosystem receiving secondarily treated sewage effluent, and (2) to measure denitrification and nitrous oxide evolution rates in a variety of flooded wetland soils. Nitrification and denitrification rates were measured in simulated marsh ecosystems, i.e., soil :water columns containing approximately 30 cm of marsh soil and either IS or 30 cm of overlying water (secondarily treated sewage effluent). Nitrification was apparently rate limiting when nitrogen was supplied in the form of ammonium. Therefore, efficiency of nitrogen removal would be increased by allowing the nitrogen in wastewater to nitrify prior to entering the marsh. First-order denitrification was observed in soil :water columns. Average nitrate removal rates, assum ing a floodwater nitrate concentration of 10 mg N liter-I, were 1.2 kg N ha-Iday-I without plants and 2.0 kg N ha-Iday-l with plants. Denitrification potentials of Florida wetland soils were determined using 15 soils with a wide range of texture, organic matter content, and soil reaction (pH). Denitrification followed apparent first-order kinetics, with rate constants varying from 0.040 day-l to 0.192 day-I. Nitrate re rates, assuming a 10 mg N liter-I floodwater nitrate concentration, ranged from 0.6 to 2.9 kg N ha-Iday-l. Denitrification rate was significantly correlated with soil organic carbon content and with soi I pH. Nitrous oxide accounted for from less than 0.2 to 6.5% of nitrate consumed. The experimental data obtained in this work indicate that many Florida wetlands could function as efficient treatment systems for nitratebearing wastewater and drainage water. Soils high in organic matter with a pH range of 5 to 7 would provide the most rapid removal rates. vi i

PAGE 14

INTRODUCTION Considerable concern exists over detrimental effects of human activities on natural waters. One major problem is excessive nitrogen input from such sources as domestic sewage, urban and agricultural runoff, and certain industrial wastes. In waters where nitrogen is the limiting nutrient, excessive input can accelerate eutrophication, thereby detracting from the aesthetic, recreational, and commercial value of the waters. Nitrogen pollution may also be a health hazard, as nitrate-N levels above 10 ppm in water for human or animal consumption are known to result in methemoglobinemia, a potentially fatal disorder. A variety of methods exist for removal of nitrogen from wastewater with tertiary treatment, including ammonia stripping and induced nitrification-denitrification. These generally require substantial capital investment and energy expenditure. Diversion of wastewater into wetlands has recently attracted interest as a low-cost, effective alternative for removal of nitrogen and other undesirable materials. In the case of non-point sources of pollution such as agricultural runoff, wetland treatment may be the only feasible means of pollution abatement. Nitrate is an important form of nitrogen in wastewater .. Denitrification, the microbial reduction of nitrate to gaseous products (primarily denitrogen and nitrous oxide), is among the mechanisms for removal of nitrate by wetlands. An undesirable aspect of denitrification is release of nitrous oxIde into the atmosphere, as this compound reacts with and destroys ozone in the stratosphere. Ozone is important in shielding the earth from solar ultraviolet radiation. Depletion of the stratospheric "ozone layerll and consequent high levels of ultraviolet radiation would be 1 ikely to have a number of negative effects on terrestrial 1 ifeforms, including increased incidence of skin cancer in humans. Knowledge of wetland denitrification rates and soil characteristics affecting these rates is incomplete, and knowledge of wetland nitrous oxide production is almost nonexistent. It is hoped that the results presented here may be useful in the evaluation of wetland sites for wastewater treatment, and perhaps contribute to the development of regional or global nitrogen budgets. Specific objectives include: (1) To determine nitrification and denitrification rates in a marsh soil receiving secondarily treated sewage effluent. (2) To measure denitrification rates in a variety of flooded wetland soils and to determine quantitative relationships between denitrification rate and selected soil characteristics in order to develop a predictive equation for denitrification rates in specific soils. (3) To Investigate nitrous oxide evolution from wetland soils in terms of amounts evolved and factors influencing evolution rates. -1-

PAGE 15

CHAPTER I LITERATURE REVIEW Wetlands as Wastewater Treatment Systems A variety of wetland ecosystems, including freshwater marshes (Fetter et al., 1978; Zoltek at al., 1979), hardwood swamps (Boyt et al., 1977), and cypress domes (Odum and Ewel, 1977) have been investigated as wastewater treatment systems. Studies have included systems with primarily downward percolation (Odum and Ewel, 1977), primarily lateral flow (Boyt et al., 1977), and intermediate situations (Zoltek et al., 1979). In general, wetlands have been quite efficient in removal of nitrogen, phosphorus, B.O.D., and col iform bacteria from both secondarily treated sewage and agricultural and urban runoff. Removal of 90% or more of incoming nitrogen and phosphorus is not uncommon, although much lower efficiencies have been reported (Fetter et al., 1978). Denitrification, plant uptake, and microbial assimilation (immobilization) are possible nitrogen removal mechanisms. The last two processes can remove both ammonia and nitrate, the primary inorganic nitrogen forms in wastewater. By definition, denitrification involves only nitrate. Ammonia entering a wetland must therefore be nitrified (biologically oxidized to nitrate) before it can be denitrified. Nitrification can occur in the aerobic floodwater or surface soil zone and the nitrate produced is subject to denitrification after diffusion into the deeper, anaerobic soi I (Patrick and Reddy, 1976). The kinetics of this situation are quite complex, and either nitrification or denitrification may be rate I imiting (Reddy and Graetz, 1980). A significant portion of the nitrogen entering a wetland may already be in the nitrate form (Fetter et al., 1978; Boyt et al., 1977). In one study, an artificial wetland treatment system was specifically designed to maximize nitrification during flow of runoff through pasture prior to entering the actual wetland (Small, 1978). Thus, the nitrogen burden on the wetland was mostly nitrate, with denitrification the primary removal mechanism. Nitrification Nitrification is a biologically mediated two-step reaction In which ammonium is first oxidized to nitrite and then to nitrate (Alexander, i977). The organisms responsible for these reactions are chemoautotrophs and therefore are not dependent upon organic carbon, but rather use carbon dioxide as the carbon source. Nitrifier numbers have been found to vary from zero to a mi 11 ion organisms per gram of soil, depending on conditions. Their numbers are generally less in water samples, with ammonium oxidizers ranging from zero to a few thousand and nitrite oxidizers from zero to a few hundred (Matulewich et al., 1975). Nitrifiers are reported to decrease to very low numbers in cl imax ecosystems (Rice and Pancholy, 1973). These low numbers have been partially explained by increasing concentrations of tannins (Rice and Pancholy, 1978) and other carbon containing compounds (EPA, 1973; EPA, 1975) that have been reported to have an inhibitory effect, although this has -2-

PAGE 16

been debated. Alexander (1977) attributed this "inhibition" to the depletion of inorganic nitrogen (i.e., ammonium and nitrate) which is assimilated by the prol iferating heterotrophic microorganisms decomposing the carbonaceous compounds. Free ammonia at high pH values has been found inhibitory (Alexander, 1977) as well as accumulation of certain trace metals (Liang and Tabatabai, 1978) and certain pesticides (Alexander, 1965). Although autotrophic microorganisms are most often responsible for nitrification, heterotrophic nitrification has recently been observed. Results indicate that under appropriate conditions of pH, sufficient carbon and ammoniacal nitrogen (and low competition), heterotrophic bacteria (and actinomycetes) can produce small quantities of nitrite as well as other more oxidized inorganic and organic nitrogen compounds, i.e., hydroxylamine, l-nitrosoethanol (Alexander, 1977; Verstraete and Alexander, 1973). Heterotrophic nitrate formation is usually minimal but it has been shown that certain fungi (in culture) are capable of oxidizing nitrite to nitrate. Recently, work done by Tate (1977) on an organic soil in South Florida with a pH of 7.1, showed that the populations of Nitrosomonas and Nitrobacter (autotrophs) were only sufficient to account for 0.1% of the nitrate present. With selective culturing techniques, large populations of Arthrobacter (heterotrophs) were found. When these bacteria were inoculated into sterile soil, nitrite was produced. Thus, it appears that in some cases, heterotrophic nitrification can be an important source of nitrate, but the general significance of this process has not yet been determined. Nitrification rates are very sensitive to environmental conditions. The most obvious consideration is that the nitrifiers exist under aerobic conditions with ammonium present. In flooded soils, oxygen supply is a potential 1 imiting factor for nitrification since few nitrifiers are present in the overlying aerobic waters and oxygen diffusion into the soil is slow. White et al. (1977) suggested the coupling of photosynthetic oxygen production in surface water with nitrification since nitrification rates were higher at midday than morning and higher in July and August than any other months. Other work by White et al. (1977) showed daytime nitrate production three times that at night. Consequently, development of an oxidized surface soil layer is crucial to the nitrification process which requires 4.3 mg of O 2 to oxidize one mill igram of ammonium nitrogen to nitrate (White et aT., 1977). Reddy et al. (1976) described the development of an oxidized layer in initially anaerobic soil :water columns amended with 200 mg NH4-N per gram of soil. In the first week a thin aerobic soil layer developed on top of the anaerobic soil, which increased to 0.5, 1.25, 1 1.6, and 2.0 cm after 7, 15, 30, 60, and 120 days respectively, at 28 C. Patrick and Delaune (1972) also found the aerobic layer to increase to 1.5 and 2.0 cm with time. The rate of development and thickness of this layer is determined by the oxygen supply and amount of readily decomposable organic matter, being slower and thinner with more organic matter due to increased microbial activity and higher oxygen demand (Engler and Patrick, 1974; Patrick and Reddy, 1976). The -3, )

PAGE 17

small size of this aerobic layer has been implicated as the limiting factor in nitrification (and consequently denitrification) in many flooded soils (Patrick and Tusneem, 1972). Ammonium is supplied to the oxidized soil layer by diffusion from eutrophic surface waters or from the anaerobic subsurface soil by diffusion and physical disruption (by benthic fauna and gas ebullition) (Chen et al., 1972). In experiments using flooded columns packed with ammonium amended soils, it was reported (Patrick and Reddy, 1976) that approximately half of the ammonium nitrified was initially present in the aerobic surface layer of soil and the remaining half diffused to the aerobic layer from the anaerobic subsurface soil. The diffusion coefficients calculated in the above experiments for ammonium moving in a saturated Crowley silt loam were 0.216 cm2/day (Patrick and Reddy, 1976; Reddy et al., 1976). Reddy et al. (1976) reported that nitrification foJlowed zero-order kinetics (based on ammonium loss). Other work (based on nitrite plus nitrate production) indicated a sigmoid relationship where the peak production varied with the initial concentration of ammonium (Alexander, 1977). Alexander (1975, 1977) reported a significant relationship between nitrate production and pH. In neutral to slightly alkaline soils, nilrifiers,are present in the largest numbers. The optimum range for nitrate production has been reported to vary from 6.6 to 9.0 (Alexander, 1977; Keeney and Gardner, 1970; EPA, 1975). At high pH values, ammonium changes to ammonia which is detrimental to Nltrobacter populations, and as a result, oxidation of nitrite to nitrate is inhibited. Lime has been reported to stimulate nitrification (Alexander, 1977; Tilsdale and Nelson, 1975) by serving as a pH buffer and a carbon source. On the other end of the pH scale, nitrification has been reported below pH 4 but generally does not occur at a significant rate until a pH of 5.0 to 5.5 is reached (Alexander, 1977; Tilsdale and Nelson, 1975). The nitrification process also lowers pH due to the release of hydrogen ions as ammonium is converted stepwise to nitrate (Broadbent, 1973). The extent of the lowering depends on the amount of ammonium nitrified and the buffering capacity of the system. Temperature has also been shown to have a marked effect on nitrification with an optimum range between 30 and 350C (Alexander, 1977; Broadbent, 1973). Above 400c and near freezing, the amount of nitrate or nitrite produced becomes negligible (Tilsdale and Nelson, 1975). White et al. (1977) reported nitrification (in stream bed sediments) to o cease when the water temperatures dropped below 17 C. Since the nitrlfiers are obi igate aerobes, they are more sensitive to excess moisture than dry conditions, although nitrate is not produced at very low moisture levels. Chemical breakdown of nitrite (and release as gaseous nitrous oxide) has also been impl icated under well aerated, -4-

PAGE 18

low moisture conditions. Lance and Whisler (1972) showed the effect of varying cycles of flooding and drying on nitrification of ammonium in wastewater and found essentially all the ammonium was nitrified with short frequent cycles: 2 days flooded followed by 5 days of drying. Other factors such as season or depth the profile are important only as they are related to factors such as temperature, pH, oxygen content, and nutrient availabil ity. Denitrification Biological denitrification is generally regarded as the use of nitrate and other oxidized forms of nitrogen as alternatives to oxygen in the respiratory process. A large number of heterotrophic bacteria, and a smaller number of autotrophs are capable of this process (Payne, 1973). Virtually all denitrifiers can respire aerobically, and some can obtain energy through fermentation. Because of the diversity of denitrifying organisms, denitrifier populations are relatively high in most soil habitats. With a few possible exceptions (Delwiche and Bryan, 1976), denitrification occurs only under at least partially anaerobic conditions, since oxygen is used preferentially to nitrate when both are present. In flooded systems, anaerobic conditions exist below a surface layer of aerobic soil, the thickness of which is controlled by oxygen concentration in the overlying water and oxygen demand of the soil (Bou I din, 1968). As with most biochemical processes, denitrification rate tends to increase with rising temperature until heat induced enzyme deactivation begins to occur (Focht, 1974). Slow denitrification has been reporded at 30C (Nommik, 1956), with rates increasing rapidly up to about 35 C (Bremner and Shaw, 1958; Bailey and Beauchamg, 1973). Rates increase somewhat more slowly from 350 to about 60-65 C, dropping off rapidly at higher temperatures (Nommik, 1956; Bremner and Shaw, 1958). Amount and source of available carbon is perhaps the most important soil characteristic determining denitrification rate. Many reports I ink rapid denitrification with high levels of organic matter, in both flooded and unflooded systems (McGarity and Meyers, 1968; Stefanson, 1972; Engler and Patrick, 1974). Added organic material tends to stimulate the process more than endogenous organic matter because the latter is generally humified and thus resistant to degradation (Bremner and Shaw, 1958; Reddy et al., 1978). Alternation of aerobic and bic conditions has been found to increase denitrification by incrl'asinq carbon decomposition (Lance and Whisler, 1972; Reddy and 1975). At least three studies have attempted to establ ish a quantitative relationship between organic carbon and denitrification. Stanford et al. (1975a) related hot water extractable carbohydrate to denitrification rate for 30 widely varying soils. An r2 value of 0.82 was obtained for carbohydrate vs. apparent first-order denitrification rate, while total organic carbon vs. apparent first-order denitrification rate gave an r2 of 0.69. Burford and Bremner (1975) studied 17 soils and obtained a 0.99 r2 value for zero-order denitrification rate vs. "miner al izable" carbon as determined by CO2 evolution. Total organic carbon -5-

PAGE 19

yielded only a 0.59 r2 value. In both of these studies, denitrification was assessed in batch studies under total anaerobiosis. The relevance to in situ denitrification rates may therefore be limited because oxygen is usually present to some degree in natural systems. Andersen (1976) studied nitrate loss from aerobic floodwater overlying five lake sediment cores. Correlating "available carbon" (determined by 02 consumption) with denitrification, he obtained r2 values of' 0.92 and 0.70 for low (2-3 ppm) and high (10 ppm) nitrate-N concentrations, respectively. These lower correlations than those obtained by Burford and Bremner (1975) may reflect such complicating factors as nitrate transport and the presence of oxygen. Plant roots have been reported to enhance denitrification, particularly in soils with low endogenous carbon (Stefanson, 1972; Bailey, 1976). This has been attributed to increased organic matter or lower oxygen levels in the root zone, or both (Volz et al., 1976; Alexander, 1977). Sherr and Payne (1978) conducted an interesting study on the effect of Spartina alterniflora roots on denitrifying activity. Above ground parts of the plants were cut and roots pruned in order to terminate the below ground effects of living plants. Denitrifying activity was found to be unaffected five months after perturbation, but significantly lowered after 18 months. This was taken to indicate that root exudates were relatively unimportant and that the main effect of roots was due to long term maintenance of underground biomass. Another important implication of this study is that root pruning, such as would occur in the taking of soil cores, may have 1 ittle short term effect on denitrification. There is a definite relationship between soil pH and denitrification, but the strength of this relationship is subject to question. A number of reports indicate that nitrate loss rates are highest between pH 7 and 8, and drop off considerably below pH 6 (Wijler and Delwiche, 1954; Nommik, 1956; Bremner and Shaw, 1958). Some of these reports (Nommik, 1956; Bremner and Shaw, 1958) are suspect because pH variation was accompl ished by amendment with acid or base, resulting in nonadapted microbial populations. Brezonik (1977) feels that the variety and adaptability of denitrifying organisms reduces the importance of pH in determining nitrate consumption rates. There are several reports of relatively rapid denitrification at pH values approaching 5.0 (Van Cleemput et al., 1975) and even below 5.0 (Ekpete and Cornfield, 1965; Erickson, 1978; Gilliam and Gambrell, 1978). A factor which has not been widely investigated is duration of exposure of the soil to nitrate. As with any microbial substrate, addition of nitrate to soil results in certain adaptive and selective changes in the soil population (Nash and Bollag, 1974; Doner and McLaren, 1978). Insufficient nitrate adaptation time has been invoked as the cause of irregular nitrate consumption rates in short term studies (Starr and Parlange, 1975). In one of the very few truly long term denitrification experiments, a soil column was leached with a solution of 100 ppm nitrate-N for over 1,000 days (Day et al., 1978). Nitrate removal rate in the top 12.5 cm was found to decrease several fold over the first 30 days, and more slowly thereafter. This effect was -6-

PAGE 20

attributed to depletion of available organic matter. Such a dramatic effect might not occur under other experimental conditions. Leaching tends to maximize organic matter loss through removal of the soluble fraction (Doner et al., 1974) and would be expected to maximize oxidation by continually and rapidly resupplying nitrate. Reddy found relatively 1 ittle change in nitrate removal rate from floodwater overlying an organic soil during a 25 day period (K. R. Reddy, personal commun i cat ion) Denitrification Kinetics: Real and Apparent The literature regarding denitrification kinetics is somewhat confusing, with zero-order, first-order, and Michaelis-Henton kinetics reported. Zero-order kinetics are reported most frequently (Nommik, 1956; Broadbent and Clark, 1965; Cooper and Smith, 1965; Van Cleemput et al., 1975). Some authors (Focht, 1974; Doner et al., 1974) regard these reports to be indicative of Hichaelis-Menten kinetics with very low Km va 1 ues, rather than t rue zero-order k i net i cs. I n sewage treatment systems, Michaelis-Menten kinetics with l
PAGE 21

-N .s: c. Q) 0 -N .s: c. Q) o Nitrate concentration water -----soil Nitrate concentration water aerobic soil anaerobic soil (a) ( b) Figure 1. Nitrate diffusion fronts in flooded soi I systems (a) with no aerobic zone, and (b) with aerobic zone. -8-

PAGE 22

consumption is proportional to depth of penetration. Both are proportional to the square root of oxygen concentration in the floodwater and also to the square root of volumetric sediment oxygen demand, as in the following equation: dQ = dt [1] where Q is quantity of oxygen (mg) crossing the soil:water interface; C is floodwater oxygen concentration, mg liter-l ; A is rate of oxygen consumption, mg cm-3day-l; 0 is oxygen diffusion coefficient, cm2day-l; and t is time, in days. These relationships were later verified experimentally (Howe1er and Bouldin, 1972). The above model would be expected to apply to denitrification if it were assumed that nitrate concentration in the floodwater decreased fairly slowly (as would be the case with all but very shallow water depths) and that nitrate consumption is uniform from the soil :water interface downward. The second assumption is not valid when oxygen is present in the water, since denitrification is theoretically zero in the aerobic soi 1 layer. Bouldin et a1. (1974) addressed this situation by developing a model which assumes no nitrate consumption in the aerobic zone and maintenance of nitrate concentration at zero at the aerobic: anaerobic interface. In this case, the second derivation of concentration with respect to depth is zero, and the diffusion front is as shown in Figure 1 (b). The total consumption is now directly proportional to floodwater concentration (i.e., apparent first-order) and directly proportional to the thickness of the aerobic zone. The assumption that nitrate reduction is rapid enough to maintain zero concentration at the aerobic:anaerobic interface is obviously and admittedly (Bouldin et al., 1974) erroneous, and is adopted for the sake of simpl ifying the model. As a result, there is no allowance in the model for volumetric nitrate consumption rate. This problem may be insignificant since oxygen consumption rate and denitrification rate are closely related. A truly accurate model would involve a combination of Figures 1 (a) and l(b) and would be somewhat complex, mathematically. A number of investigators (Van Kessel, 1976; Reddy et al., 1979) have obtained high r2 values when applying apparent first-order denitrification kinetics to nitrate consumption data under boundary conditions similar to those assumed by the model of Bouldin et al. (1974), indicating that this model may be adequate. Apparently no attempts to correlate nitrate loss to square root of floodwater concentration (as in the 1968 Bouldin model) have been publ ished. Additional Mechanisms for Nitrate Loss Although reduction of nitrate to nitrite in nature is thought Lo be a solely biological process, under certain conditions nitrite can be chemically decomposed into gaseous products. Four major types of chemoden i tr i f i cat i on are genera 11 y recogn i zed: (l) se 1 f-decoll1pos i t ion of nitrite into NO; (2) reaction of nitrite with ammonia, urea or amino acids yielding N2 (Van Slyke reaction); (3) reaction with organic matter yielding mostly N 2 ; and (4) reaction with heavy metals reSUlting -9-i /

PAGE 23

in NO, N20 or N2 Clark, 1965). The first three require a pH in the vicinity of 5.0 or lower. The last can occur at a pH as high as 8.0 (Moraghan and Buresh, 1977). There is evidence for the occurrence of all of these reactions in sterile, nitrite amended soils (Van Cleemput et al., 1976; Nelson and Bremner, 1970). Assessment of the natural significance of chemodenitrification is difficult due to inability to inhibit biological nitrite decomposition without inhibiting biological nitrite production (Van Cleemput et aI., 1975). However, it is generally felt that chemodenitrification is insignificant in nature (Burford and 1973; Focht, 1974). Dissimilatory reduction of nitrate to ammonia is analogous to deni tri fication, but the end product is ammonia rather than gaseous N' compounds. As in denitrification, the reduction is for the purpose of obtaining respiratory energy rather than obtaining cell nitrogen. This process was first described over 20 years ago (Verhoeven, 1956) but has only recently received much attention. Although populations capable of this process may be large (Tiedje et al., 1979), the process tends to occur only under intensely reduced conditions (Buresh and Patrick, 1978). Since nitrate is produced under oxidizing conditions, the majority of nitrate would likely be denitrified before reaching conditions sufficiently reducing for ammonia production. For this reason, Buresh and Patrick (1978) feel that dissimilatory reduction to ammonia is insignificant in flooded soils. Sorensen (1978) feels that this process may be of equal significance to denitrification in marine sediments. A significant portion of the nitrate entering soils can be assimi lated by microorganisms and converted to organic nitrogen. Assimilation of added nitrate is usually low when soil C:N ratio is narrow, since nitrogen assimilation needs would be met by mineral ized N. For this reason, assimi lation is lower with a well humified carbon source than with fresher material (Alexander, 1977). Reports of assimilation under saturated conditions range from about 10% to 37% of applied nitrate (Andersen, 1976; Chen et al., 1972; Chen and Keeney, 1974). There is some indication that percent assimilation is related to concentration of added nitrate, 37%, 14-26%i and 17% having been reported for concentrations of 2 ppm-N (Chen et al., 1972), 10 ppm-N (Chen and Keeney, 1974), and 100 ppm-N (Stanford et al., 1975b), respective'ly. This would be expected in view of the relatively low assimilation needs and high respiration needs of anoxic communities (Tusneem and Patrick, 1971). Doner et al. (1974) found that net assimilation was very low in relatively long term column studies with continuous nitrate supply. It was felt that once rapid increase in microbial biomass ceased and stable population developed, assimilation became nearly zero. An increase in assimilation relative to denitrification at low temperatures has been reported (Craswell, 1978), possibly indicating that denitrification is more affected by cold than is assimilation. -10-

PAGE 24

Nitrous Oxide Production The amount of nitrous oxide (N2 0) evolved during denitrification varies greatly. Ratios of N2:N20 can range from greater than 100:1 (Denmead, 1979) to less than 1:10 (Stefanson, 1972). The majority of reports are in the vicinity of 10:1, and the Council for Agricultural Science Technology (C.A.S.T., 1975) has estimated an average worldwide ratio of 16:1 for agricultural soils. Such factors as high organic matter content, high moisure content, fine soil texture, and low oxygen concentration have all been shown to decrease N20 emission (Nommik, 1956; Stefanson, 1972; Focht and Verstraete, 1977), indicating that of high electron acceptor demand result in more complete reduction of nitrate to N2. Guthrie and Duxbury (1978) reported an exception to this trend with lower N2:N20 ratios occurring in flooded histosols than in the same soil when drained. Ratios of N2:N20 have been reported to decrease atlow temperatures (Bailey and Beauchamp, 1973). Nitrous oxide production tends to increase with decreasing pH (Wijler and Delwiche, 1954; Nommik, 1956; Burford and Bremner, 1978). Blackmer and Bremner (1978) have shown that this may be due to an indirect effect of pH. These investigators reported that high nitrate inhibited N20 reduction, thereby increasing emission of this gas. The effect was greater at low pH than at high pH. In the absence of nitrate, N20 was reduced as rapidly at low pH as at high pH. It was concluded that the effect of pH was to control the degree of nitrate inhibition of N20 reduction. Prior exposure to denitrifying conditions (i .e., anoxia and nitrate) can also influence N2:N20 ratios. Blackmer and Bremner (1979) observed that nitrate application resulted in an increase in N2:N20 ratio over a 48 hour period, presumedly due to induction of N20 reduction activity in the denitrifier population. Guthrie and Duxbury (1978) noted indications of a more long term increase in N2:N20 ratio due to population selection in nitrate-amended soil. Nitrous oxide can be produced from processes other than denitrification. As mentioned above, chemodenitrification can potentially result in N20 production, but the extent of this in nature is unknown. Yoshida and Alexander (1970) reported that axenic cultures of Nitrosomonas can produce N20 from nitrite during nitrification of ammonia. This had been thought to be insignificant in nature (Focht, 1974) until recently when it was found that considerable N 2 0 production can occur during nitrification in aerobic soils (Bremner and Blackmer, 1979; Denmead et aI., 1979b). Dissimilatory reduction of nitrate to qmmonia can also result in N20 production in pure culture (Tiedje et al., 1979), but this has not been demonstrated in situ. -11-

PAGE 25

Summary Pertinent literature on nitrification, denitrification and N20 production in soils can be summarized as follows: (1) Many of the factors affecting nitrification and denitrification in soils are known in a general, qualitative way. Nitrification is favored by high oxygen availability and near neutral pH, and may be inhibited by certain organic substances. Poor aeration, high organic carbon availability and neutral pH tend to accelerate denitrification. In flooded systems, transport of nitrate into the anerobic zone is an important rate-determining factor which many studies have ignored. (2) In wetland systems, the potential exists for rapid nitrification of incoming ammonia in the upper, aerobic soil zone, and possibly in the water column. (3) It is likely that denitrification accounts for the majority of nitrate consumption in wetland soils, particularly under conditions of heavy nitrate enrichment. (4) Relatively little is known about rates of N20 production during denitrification. Average reported N 2:N2 0 ratios are in the vicinity of 10:1. Low electron acceptor demand and low pH tend to favor N 2 0 production. Soil processes other than denitrification, including nitrification and dissimilatory reduction of nitrate to ammonia, may be important N20 sources, at least under certain conditions. (5) The data base for nitrification, denitrification, and N20 production in wetlands is at present quite narrow. -12-

PAGE 26

CHAPTER II NITRIFICATION AND DENITRIFICATION IN A MARSH SOIL RECEIVING SECONDARILY TREATED SEWAGE EFFLUENT The objective of this study was to determine the feasibility of using natural wetlands to remove nitrogen via denitrification from secondari ly treated municipal sewage effluent. Since nitrogen in the form of nitrate is a prerequisite for denitrification, nitrification was included as part of the study. Nitrogen removal was investigated in laboratory and in situ experiments on a freshwater marsh in Central Florida. --METHODS AND MATERIALS Soil and Water Sampling Soil samples were obtained from a 32 hectare freshwater marsh located in the town of Clermont, Florida, about 100 mi les (160 kilometers) south of Gainesville on Route 27. The soil was classified as a variant of Brighton peat (typic medifibrist) characterized by 1.5 meters of relatively undecomposed, acidic, organic soil overlying 1.5 meters of sand over clay of an undetermined depth. Soil used in Ilbatch" type studies was collected in bulk from a zone between 8 and 45 cm deep. Soil was not taken from the top 8 cm because of the dense network of roots near the surface. Bulk samples were stored in a polyethylene pail (with lid) until used. Soil used in the column nitrification and denitrification studies was collected as intact columns, some with plants and some without. This was done by selecting an appropriate site (i .e., a young, healthy plant, or no plant), using a shovel to make the initial circular cut through the peat, roots, and/or rhizome, then placing polyvinyl chloride (PVC) pipe (70 x 10 cm, length x diameter) over the plant and pushing it into the soil as far as the cut was made with the shovel. This was followed by hammering the column about 45 to 50 cm into the soil. Compaction was less than 10%. The next step involved rocking the column back and forth in place to break the bottom loose and then pull ing the column up with a hand under the bottom to prevent the soi 1 from fall ing out. A "knock-out test capll was then pushed into the bottom of the column. Sil icone rubber sealant was used around the cap to stop leaks. In the lab, the columns were allowed to stabil ize for two to four weeks (with those containing plants placed under a window) while receiving only enough deionized water to keep the soil saturated. Water used in these investigations was obtained from the following sources: marsh surface water, oxidation pond water and secondarily treated sewage effluent from the Clermont sewage treatment plant, and Lake AI ice water (University of Florida campus). All water samples were collected and stored in polyethylene containers. Lake and marsh water -13-

PAGE 27

o was used within 24 hours. Effluent was stored at 4 C for a maximum of three days. Water samples containing particulate matter were allowed to settle and the water used was siphoned off the top. Plant species were in the process of change during the study period, due to changing weather patterns and lower water levels. Sagittaria lancifol ia (arrowhead) and Pontederia cordata var. Lancifol ia (pickerel weed) were the original dominant species which were invaded by Panicum sp. (maidencane) and Hibiscus sp. (marsh hibiscus). Description of Experiments Prel iminary Nitrification Study An initial study was designed to determine the effects of dissolved organic compounds (tannins) and bacterial inoculation on nitrification in marsh water. Yellow colored, dissolved organic compounds were removed from the marsh water by adding powdered activated charcoal, mixing for one hour, and filtering (#42 Whatman filter paper). A suspension of nitrifiers was obtained by shaking a mixture containing 100 grams of soil known to contain nitrifiers with 150 ml of deionized water for one hour and allowing the soil to settle out. Ten mill i 1 iter al iquots were used as inoculum for the appropriate treatment. All samples were amended with ammonium (as ammonium sulfate) to a concentration of about 50 ppm NH4-N, placed in polyethylene containers (in triplicate) and incubated at 250C for 30 days. After 25 days, it was noted that the pH had dropped considerably so it was adjusted up to 7.2 (O. I) with 0.5N NaOH. Twenty millil iter samples were treated with phenyl mercuric acetate (PMA) and stored at 4c until pH, ammonium and nitrate were determined, usually within 24 hours. Prel iminary Denitrification Study An initial laboratory study using test tubes of nitrate-amended solutions and marsh soil was designed to ascertain if denitrification would occur. Solutions of potassium nitrate in deionized water, in marsh water, and in oxidation pond water and of calcium nitrate in deionized water were prepared to a concentration of about 45 ppm N03-N. A 10 ml al iquot of each solution was added to a test tube (eight repl icates) containing 6.0 (.2) grams of marsh soil and incubated at 250C for I I days. The same design without the soil was employed to ascertain if denitrification would occur in the water alone. Sampl ing involved pouring off the solution and analyzing it for ammonium and nitrate. Nitrification and Denitrification Studies Using Soil :Water Columns In order to better simulate actual marsh conditions, additional experiments for both nitrification and denitrification were conducted using intact soil:water columns, half with and half without Sagittaria lancifol ia (arrowhead). Each column consisted of 45 cm of soil and either 15 or 30 cm of overlying effluent. Treatments in both studies -14-

PAGE 28

were the same (Figure 2). Four replications, two with and two without plants, of each of the following treatments were used: 1. Control -effluent to lS-cm depth. 2. N amended effluent to a lS-cm depth. 3. N amended effl uent to a 30-cm depth. 4. N amended effluent to a lS-cm depth with 10 grams CaC03 added to adjust pH. The amount of calcium carbonate was calculated using Yuan's (1974) double buffer method to bring the top 15 cm of soil to a pH of 7.0. It was applied on the surface of the soil, to the appropriate columns, several days before a study began. A second and third application of the appropriate form of N was pipetted into the individual soil:water columns as a more concentrated solution. In all cases, evapotranspiration was accounted for (two to four times weekly) by adding deionized water to maintain the appropriate depth. Columns in both studies .were aerated to maintain dissolved oxygen near saturation values. The overlying water in the columns was sampled (about 15 ml) at about half the water depth, i.e., 7.5 or 15 cm. Phenyl mercuric acetate was added and the sample was stored at 4c until ammonium and nitrate were determined, usually within 24 hours. In Situ Nitrification and Denitrification Studies For the purpose of obtaining natural nitrification and denitrification rates, an in situ study was conducted. Eno (1960) and more recently Struble (1977r-fOund polyethylene bags acceptable for this type of study due to their permeability to gases (specifically oxygen and carbon dioxide) and impermeability to ions such as nitrate and ammonium. Polyethylene bags were filled with 100 grams marsh soi 1 and subjected to the following treatments: 1. Control, shallow -soi 1 in bags placed at 8 cm. 2. Control, deep, lime -soil plus lime in bags buried at 30 cm. 3. Nitrification, shallow -soil amended with ammonium in bags buried at 8 cm. 4. Nitrification, deep -soil amended with ammonium in bags buried at 30 cm. 5. Denitrification, shallow -soil amended with nitrate in bags buried at 8 cm. 6. Denitrification, deep -soil amended with nitrate in bags bur i ed at 30 cm. 7. Denitrification, deep, lime -soil amended with ritrate plus 1 ime in bags buried at 30 cm. Bags were flattened to surface area and heat sealed. For the nitrification study, soil was amended with ammonium (as ammonium chloride) to bring the final concentration to about 35 ppm NH4 -N. -15I -->"

PAGE 29

TREATMENT AIR PUMP HUMIDIFIER STONE ROOTS SOil Figure 2. Experimental soil:water columns for nitrification and denitrification studies. -16-

PAGE 30

For the denitrification study, soil was similarly amended with nitrate (as potassium nitrate). Control soils received no nitrogen. Lime used in all cases was 0.85 gram of finely powdered calcium carbonate per bag which was the amount extrapolated from an earlier application of Yuan1s double buffer method and was intended to bring the pH of the soil to 7.0. Bags were retrieved at selected intervals. Temperature, pH, and dissolved oxygen readings were taken at each site. After bags were removed, sample pH was determined, and the contents of each bag were extracted with 2N KC1. The extractant was analyzed for ammonium and nitrate. Analytical Procedures Extraction Procedure Ammonium, nitrite, and nitrate were extracted from the soil by transferring the 100 grams of wet organic soil into a polyethylene sample bottle, adding 100 ml of 2N KCl solution, shaking briefly and' then allowing the suspension to stand for two hours. The mixture was shaken again before filtration (#40 Whatman filter paper) and finally stored, with PMA, at 4c until inorganic nitrogen was determined (Bremner, 1965). Determination of Ammonium and Nitrate Ammonium and nitrite plus nitrate values were obtained through steam distillation of water samples or potassium chloride extractants (Bremner, 1965). Determination of pH An Orion 701 or 401 meter was used in conjunction with a combination pH electrode. Dissolved Oxygen Determination A YSI Model 54 oxygen meter with a YSI Model 5419 probe was used to determine concentrations of dissolved oxygen. -17) /

PAGE 31

RESULTS AND DISCUSSION Nitrification Laboratory Investigations Since nitrification has been reported to be predominently an autotrophic process (Alexander, 1977) and therefore not in need of organic carbon compounds, initial studies were designed to determine if nitrification would occur in the marsh water alone, i.e., without soil. Preliminary results indicated that nitrification did not occur in the marsh water even though dissolved oxygen and pH were at acceptable levels. Thus, further studies were conducted to identify the reasons for the lack of nitrification. Since nitrification is reportedly inhibited by certain types of dissolved organic compounds (Rice and Pancholy, 1973; EPA, 1973; EPA, 1975), marsh water was filtered through charcoal to remove dissolved organics. Since it was also possible that nitrifying organisms simply were not present in the marsh water, an inoculant containing nitrifiers was added as a second treatment. Water from a eutrophic lake on the University of Florida campus (Lake Alice) was included for comparison purposes. All samples were aerated, amended with ammonium, and incubated in polyethylene containers for 30 days. Results of charcoal filtration and nitrifier inoculation are given in Table 1. Evaporation caused an increase in ammonium concentration so the effects of the two treatments can best be compared using the nitrate '""" \ data. Charcoal filtration had little effect on nitrification rate, however, inoculation increased the rate in both the marsh and lake ) water. These results indicate that nitrification will occur in over-lying marsh water if nitrifiers are present, but at least in this marsh, nitrification is minimal due to a lack of nitrifiers. The nitrification that occurred in the uninoculated lake water indicated that some nitri-fiers were already present which is a likely consequence of Lake Alice receiving sewage effluent. Having this nitrifying population already present would also explain why nitrification started sooner in the lake water than in the marsh water. Since the nitrification process releases hydrogen ions and water has a low buffering capacity, the pH dropped substantially during the course of the experiment. After 25 days, the pH had dropped approximately two units and was below that normally acceptable for nitrification which necessitated artificially adjusting it upward. Five days after adjustment, it had again dropped two to three pH units in actively nitrifying samples. In a natural aquatic system, this could rapidly limit nitrification unless the system was adequately buffered to maintain an acceptable pH range. Dissolved oxygen was not limiting throughout the study period. The above experiment suggested that nitrification did not occur in the overlying natural water of this marsh, but that nitrification would occur in the water if sewage effluent were added. Publ ished, work has indicated that nitrification will also occur in the thin, aerobic surface zone of a submerged soil. Thus, the next phase of the present study was designed using soil:water columns to simulate marsh conditions. -18\

PAGE 32

Table 1. Effect of charcoal f i 1 trat i on and inoculation with so i 1 water extract on nitrification in marsh and lake waters without so i 1 Treatments Days Water Char. .l. ',ource f i 1 t. Inoc. 0 3 7 11 18 25 25' I 30 ------------------NH4N mg 1 iter-I-------------------Marsh yes yes 52.4 49.0 51.5 51.3 52.2 59.7 54.7 Marsh no yes 50.1 46.5 49.6 47. 1 50.8 50. 1 52.9 Marsh yes no 52.4 51.9 52. 1 54. 1 57.6 61.6 65.6 Marsh no no 50. 1 51.1 48.7 52.2 54.0 60.1 63.8 Lake no yes 52.3 45.9 38.6 33.0 32.3 33.0 27.0 Lake no no 52.3 46.5 40.6 35.5 36.8 39. 1 40.8 ------------------NO -N 3 mg liter-1-------------------Marsh yes yes 0 0.58 1.0 3.6 4.9 7.9 17.0 Marsh no yes 0 0.70 1.0 2.7 4.9 11.9 13.4 Marsh yes no 0 0.10 1.0 0.18 0.46 0.10 0.8 Marsh no no 0 0 1 .0 0.50 0.28 1.2 0.8 Lake no yes 0.35 0 6.2 12.3 12.8 23.2 Lake no no 0.35 1.1 3.4 9.5 10.3 11.5 11.5 ------------Dissolved oxygen, mg 1 iter-1-------------Marsh yes yes 8.0 8.0 8. 1 8. 1 8.2 7.4 7.3 Ma rsh no yes 7.8 8. 1 8. 1 8.2 8.0 7.3 7.2 Marsh yes no 8.0 8.2 8.2 8.2 8.2 7.2 7.2 Marsh no no 7.8 7.8 8. 1 8.0 8.3 7.5 7.3 Lake no yes 8.3 7.7 7.3 8.0 8.0 7.2 7.2 Lake no no 8.3 7.9 7.4 8. 1 8.1 7.2 7.0 -------------------------pH---------------------------Marsh yes yes 6.6 6.6 6.2 5. 1 4.2 4. 1 7.2 4.4 Marsh no yes 6.5 6.6 6. 1 4.9 4.0 3.9 7. 1 4.2 Marsh yes no 6.6 6.7 6.6 6.5 5.9 5.6 7.2 6.0 Marsh no no 6.5 6.5 6.5 6.1 5.8 4.5 7.2 5.5 Lake no yes 7.9 7.9 '7.4 5.0 4.4 4.3 7 1 4.2 Lake no no 7.9 8.0 7.7 5.4 5.6 5.3 7.1 5.2 t pH adjusted to 7.2 (O. 1 ) with 0.5 N NaOH after 25 day samples taken. -19-

PAGE 33

Plants play an important role in control 1 ing nitrogen forms and concentrations in marsh systems through plant uptake and decomposition as well as the introduction of root exudates and oxygen into the rhizosphere soil. For this reason, they were included in this study with water depth and pH as variables. Soil:water columns (10 cm 1.0.) containing 45 cm of soil and either 15 or 30 cm of overlying secondarily treated sewage effluent were used. Treatmen.ts consisted of a control with sewage effluent to a depth of 15 cm, ammonium-amended effluent to a depth of 15 em, ammonium-amended effluent to a depth of 30 cm, and ammonium-amended effluent to a depth of 15 cm and pH adjusted with CaC03. Four repl ications, two with two without plants, of the above treatments were used. Concentrations of ammonium and nitrate observed over a 52 day period in the overlying water are depicted graphically in Figures 3 through 5. Overlying water in the columns was not aerated initially and oxygen diffusion was not sufficient to provide adequate aeration. Con sequently oxygen became depleted within the first few days of the study, curtailing nitrification. This resulted in the leveling off or increase of ammonium concentrations in columns without plants. The increase in ammonium concentration was probably due to ammonification, which does not require oxygen. This increase was especially evident in the limed columns as would be expected since ammonification is favored at neutral pH levels (Hubbell, 1971). The effect was not observed in columns with plants because the plants continued to assimilate ammonium. Subsequently, (day 9), all columns were artificially aerated. Nitrification rate appears to be related to ammonium concentration; however, due to the 1 imited amount of data, first-order kinetics cannot be assumed. No attempt will be made to quantitatively evaluate the kinetics of nitrification in flooded systems. It is likely that nitrification kinetics under the present experimental conditions are quite complex, since a portion of the nitrification is occurring in the water column or at the soil:water interface, while a portion occurs at some small depth within the soil. The former would be expected to be a zeroorder process, while the latter is 1 ikely non-zero-order because of diffusion dependence (as discussed for denitrification in Chapter 1). Ammonium removal rates given In Table 2 are derived from the slopes of removal curves at approximately 15 to 25 mg 1 iter-l ammonium-N concentration. In the 15 cm water depth columns with plants, all added ammonium was depleted in approximately 15 days despite the above mentioned aeration problem (Figure 3). At 18 days, additional ammonium was added to both the treatments with and without plants. Ammonium removal continued to be rapid with plants and, in contrast to the earl ier days the study, ammonium removal was also relatively rapid in the columns without plants. Only small amounts of nitrate were detected in either column (Figure 3). This can be explained partly by the fact that some nitrate was removed by the plants and, more importantly, by nitrate diffusion into the anaerobic sediment and subsequent denitrification. Ammonium removal rate (Table 2) was greater with plants (2.9 kg N ha-l day-I) than without plants (1.5 kg N ha-lday-l). -20-

PAGE 34

Q) I Q) 50 40 20 E 10 o 10 o 10 Water depth 15 em Ammonium -N 20 30 40 50 days Nitrate -N 20 30 40 50 days Figure 3. Nitrification in soi 1 :water columns flooded with 15 cm of ammonium amended sewage effluent, with [0) and without [e) plants. -21"

PAGE 35

50 40 30 Q) 20 10 0 20 I Q) C' 10 E 0 Fi gure 4. 0 10 0 10 Nitrification in ammonium amended plants. Water depth 30 em Ammonium -N 20 30 40 50 days Nitrate -N 20 30 40 50 days soil :water columns flooded with 30 cm of sewage effluent, with [0] and wi thout [. ] -22-

PAGE 36

pH Adjusted, water depth 15 em 50 Ammonium -N 40 30 01 E 20 10 10 20 30 40 50 days Nitrate -N I Q) 01 E 20 o 10 20 30 40 days Figure 5. Nitrification in limed soil:water columns flooded with 15 em of ammonium amended sewage effluent, with [0] and wi thout [.] plants. 2150

PAGE 37

Table 2. Ammonium removal rates from overlying water of soil :water columns containing ammonium amended sewage eff 1 uent. Treatment Without Plant With Plant ---------kg N ha-1day-l---------15 cm water depth 1.5t 2.9 30 cm water depth 1.2 5.3 15 cm water depth, pH adj. 11.9 11.2 t -1 Rates assume an ammonium-N concentration of 20 mg liter -24-

PAGE 38

Increasing the water depth to 30 cm did not greatly affect the ammonium removal rate (Table 2) but because the total amount of added ammonium was greater, the removal time was somewhat longer (Figure 4). As with the 15 cm water depthsj the ammonium removal rate was greater with plants (5.3 kg N ha-lday-) than without plants (1.2 kg N ha-l day-I). Nitrate concentrations (Figure 4) were low throughout the study period for the same reasons noted for the 15 cm water depth columns. Increasing the pH significantly increased ammonium removal rate (Figure 5 and Table 2). Ammonium concentrations decreased from about 40 ppm N to less than 5 ppm within 3 days. The fact that nitrification was responsible for much of this decrease is indicated by the high (20 ppm N) nitrate concentrations observed in the overlying water. Nitrifying bacteria require a pH near neutral ity to thrive and are considerably inhibited by pH values below 5. There was little difference in ammonium removal rates in pH adjusted treatments with or without plants possibly because the bacteria were able to effectively compete with the plants for the ammonium. The effect of the plants is, however, very evident from the nitrate concentrations. Nitrate concentrations were consistently lower in the presence of plants. Control columns consisted of secondarily treated effluent to a 15-cm depth overlying marsh soil with no treatment except aeration. Concentrations of ammonium and nitrate in the controls with and without plants remained low throughout the study period. When columns were dismantled at the end of the 52 days, pH in the top 15 cm of soil for all except the limed columns averaged around 5.5, dropping down to about 4.5 over the next 30 cm. Surface pH of the limed columns approached neutral ity but decreased with depth (within 10 cm) to values simi lar to the other treatments since the 1 ime had not been incorporated into the soil. In Situ Investigations Since nitrification did not appear to occur naturally in the water alone but did occur when in contact with soil in the column experiments, an in situ study was performed to determine location of and rates at which this oxidative process would occur in the soil profi Ie. Polyethylene bags containing marsh soil amended with ammonium were placed at shallow (8 cm) and deep (30 cm) positions in the marsh soil. Ammonium and nitrate concentrations in the sample bags are shown in Figure 6. Ammonium values decl ined with time at the 8 cm depth but remained relatively constant at the 30 cm depth. This indicates that nitrification was occurring at the 8 cm level but not at 30 cm. During the time of this in situ study, the water table flucuated at and slightly below the soil Under these conditions, the 8 cm depth may have been under aerated conditions periodically due to the depth of the water table. However, oxygen has been reported to diffuse from the roots of a number of marsh plants into the surrounding waterlogged soil (Armstong, 1967) and this mechanism is 1 ikely responsible for providing much of the oxygen required by nitrification. Roots were largely absent at the -25"

PAGE 39

I I Q) at E Q) at E Ammonium -N 40 30 20 shallow )I-tC deep control, shallow o-a control, deep 10 days Nitrate -N 10 5 5 10 15 20 25 days Figure 6. Ammonium and nitrate concentrations in ammonium amended marsh soil during situ incubation in polyethylene bags. -26-

PAGE 40

30 cm depth of the deep samples. Small increases in nitrate concentration were observed at both depths. At the 8 cm depth the increase in nitrate concentration did not correspond in quantity to the decrease in ammonium concentration, indicating that some denitrification and/or plant assimilation was also occurring at this depth. The presence of nitrate at the 30 cm depth is 1 ikely an artifact of the sample extraction technique after removed from the soil. Denitrification Laboratory Investigations Prel iminary laboratory denitrification studies were conducted to obtain approximate denitrification rates and ascertain how these rates would be affected by the presence or absence of soil, associating cation, and water source. This entailed a "batchll type study which excluded plants. Water samples from several sources were amended with potassium or calcium nitrate salts and incubated with and without soil. The data (Table 3) shows a steady decl ine in nitrate concentrations with time, for all soil-water treatments. No clear effect of water source nor corresponding cation on nitrate removal rate was observed. The most 1 ikely nitrate removal mechanisms are denitrification and microbial immobil ization, with the former being the dominant reaction. Solutions not in contact with soil showed no appreciable change (sl ight increase due to evaporation) which indicates that denitrification did not occur in these waters in the absence of soil. This is in agreement with similar work done by Engler et al. (1976) and 1 ikely was due to insufficient readily available organic carbon, even in oxidation pond water, to support denitrification. After finding that water source and associating cation had 1 ittle effect on the denitrification rate, an investigation was initiated using soil:water columns to simulate the marsh environment. Potassium nitrate and secondary sewage effluent were used as the nitrate and water source, respectively. Water depth, pH, and plants were used as variables in a manner similar to that used in the nitrification study. Nitrate concentrations in the overlying water of each treatment are presented graphically in Figures 7 through 9. Nitrate removal appeared to follow first-order kinetics. (Denitrification kinetics, including methods for determining reaction order, are discussed in Chapter 3.) Both first-order rate constants (day-I) and nitrate removal rates (kg N ha-lday-l, assuming a floodwater nitrate concentration of 10 mg N 1 iter-I) are given in Table 4. Although useful for theoretical and data manipulation purposes, first-order rate constants can be misleading because they express nitrate removal as a function of the amount of nitrate present. That is, if floodwater depth (and therefore the amount of nitrate in the system) is doubled while removal rate remains the same, the first-order constant will be reduced by one half. For this reason, when floodwater depth is a variable, nitrate removal rates are of more value for comparison purposes than rate constants. -27-

PAGE 41

Table 3. Nitrate removal as affected by water source, associated cation and presence of soil. Days Treatment 0 2 5 1 1 ------NO -N 3 mg 1 -1 Iter -----Ca(N03)2 in deionized water (no so i 1 ) 40.5 41.8 41.3 42.3 KN03 in deionized water (no so i 1 ) 45.0 45.2 46.6 47.3 KN03 in oxid pond water (no so i 1 ) 51.6 53.0 54.7 58.0 KN03 in marsh water (no so i 1 ) 43.2 44.9 44.8 49.2 Ca(N03)2 in deionized water (over so i 1 ) 40.5 30.9 25.5 12.2 KN03 in deionized water (over so i 1) 45.0 33.9 26.3 15.0 KN03 in oxid pond water (over soil) 51.6 43.0 30.0 21.5 KN03 in marsh water (over so i 1 ) 43.2 35.0 26.5 14.3 ------% NOrN remaining-----Ca(N03)2 in deionized water (over soi 1) 100 76.3 63.0 30.1 KN03 in deionized water (over soi 1) 100 75.2 58.4 33.4 KN03 in oxid pond water (over soi 1) 100 83.3 58.2 41.6 KN03 in marsh water (over soil) 100 81.0 61.3 34.3 -28-

PAGE 42

Table 4. Nitrate removal rates from overlying water of soil :water columns containing nitrate amended sewage effluent. Treatment 15 cmwater depth wi thout plant with plant 30 cm water depth wi thout plant with plant 15 cm water depth, wi thout plant with plant t Nitrate removal rate pH adj. 1.9 1.5 0.8 2.9 1.0 1.6 Denitrification rate constant -1 ------------day ------------o. 127 0.102 0.026 0.098 0.067 0.105 t Assuming floodwater nitrate-N concentration of 10 mg liter-l -29-

PAGE 43

. I I ... Q) 01 E Water depth 15 em 40 Nitrate -N 30 20 10 20 30 40 50 days Figure 7. Nitrate removal from overlying water of soi I :water columns (with [0] and without [e] plants) containing nitrate amended sewage effluent to a depth of 15 em. Nitrate was added on day 0 and day 18. -30-

PAGE 44

I .... Q) -at E Water depth 30 em 40 Nitrate -N 30 20 10 1.0 20 30 40 50 days Figure 8. Nitrate removal from overlying water of soi 1 :water columns (with [0] and without [e] plants) containing nitrate amended sewage effluent to a depth of 30 cm. Nitrate was added on day 0 and day 18. -31"\

PAGE 45

pH Adjusted. water depth 15 em 40 Nitrate -N 30 I ... cP at E 20 10 o 10 20 30 40 days Figure 9. Nitrate removal from overlying water of limed soil :water columns (with [0] and without [e] plants) containing nitrate amended sewage effluent to a depth of 15 cm. Nitrate was added on day 0 and day 18. -3250

PAGE 46

Removal rates for the treatments without plants ranged from 0.8 to 1.9 kg N ha-lday-l. Results of the three treatments are difficult to compare because of the low number of replications and the high spacial variation found in wetland soils. (Chapter 3 contains further discussion of spacial variation found in a number of wetland soils.) The two water levels were originally utilized because it was felt that water depth would influence the aeration status of the system, with possibly less oxygen reaching the soil with the deeper water depth, thereby increasing the rate of denitrification. However, the need to artificially aerate the water column negated the water depth effect. It was also anticipated that pH adjustment would increase denitrification rate. Since the pH adjusted treatment had the lowest rate, it was assumed that this value reflected natural variation rather than a true pH effect. (Lack of pH effect is probably the result of lime having been applied to the soil surface while denitrification is a subsurface process. Incorpora tion of lime into the soil would be impractical in the field due to plant destruction, etc.) Thus, the overall denitrification rate in these systems was taken as the average of the three treatments. Therefore, a removal rate of 1.2 kg N ha-lday-l (at 10 mg nitrate-N liter-I) would be expected for this soil. Removal rates at other nitrate concentrations would vary in direct proportion to the concentration, since the reaction is first-order. Inclusion of plants in the system substantially increased nitrate removal rates. A number of factors may have been responsible for this. Plant roots may have provided more favorable conditions for denitrification via root exudates or other influences. Downward movement of floodwater into the soil due to transpirational pumping may have increased denitrification rates by accelerating nitrate transport to denitrification sites. Plants may also have assimilated nitrate, augmenting denitrification as a removal mechanism. Because plants in different columns varied both in size and vigor, as well as for reasons noted for the columns without plants, removal rates for columns with plants, averaged over the three treatments, were used as an indication of typical removal rates. Values (again 10 mg nitrate-N 1 iter-I) ranged from 1.5 to 2.9 kg N ha-lday-with an average removal rate of 2.0 kg N ha-lday-l. Ammonium levels in the overlying water of all treatments were less than 1 mg N liter-l at all except the initial sampling date. The control treatment also had low nitrate values. This indicates that either very 1 ittle ammonium was being released from the soil, or that it was being nitrified and then rapidly denitrified. In Situ Investigations Nitrate removal as a function of depth and incorporated lime was investigated by in situ incubation of nitrate and lime amended marsh soil in polyethylene bags. Soils amended with nitrate showed rapid nitrate removal, with 25 mg 1 iter-l nitrate removed in 2 to 3 days, depending on treatment (Figure 10). The effect of depth on nitrate removal is shown by comparing data for unl imed, shallow site (3. 1 N per gram wet soil per day) with ynlimed, deep samples (5. 1 N g-lday-l). -33-

PAGE 47

I Q) I Ct E Q) 30 Nitrate-N shallow )I---fC deep C)-...O deep, lime o---a control, shallow contro I, deep 10 o 5 10 15 20 days 20 Ammonium-N ________ L________ ________ ________ o 5 10 days 15 20 Figure 10. Nitrate and ammonium concentrations in nitrate amended marsh soil during situ incubation in polyethylene bags. -34-

PAGE 48

It appears that at the shallow depth there are both aerobic (likely in close proximity to plant roots) and anaerobic sites. Thus, nitrification and denitrification were occurring simultaneously at this depth. This is also indicated by the small amount of ammonium accumulation and slower nitrate removal compared to the 30 cm depth. Limed soils ame?ded with nitrate and placed deep in the profile removed 12.8 N gday-I. When this is compared to a similarly placed, unlimed sample in which 5. 1 N g-lday-l was removed, it is obvious that liming enhanced the removal rate. Apparently, denitrification is 1 imited by the natural soil pH of 4.7. The immediate stimulation of denitrification by raising the pH is indirect evidence that most nitrate removal occurs biologically rather than chemically, since the latter is inhibited at higher pH values. Another interesting aspect of I iming is the increased ammonification shown by limed treatments (Figure 10). This ammonium accumulation did not occur in shallow samples, probably due to the presence of small amounts of oxygen, but occurred in the deep, more anaerobic samples. -35\ ..

PAGE 49

SUMMARY AND CONCLUSIONS Nitrification and denitrification rates In a freshwater marsh receiving secondarily treated sewage effluent were investigated in the laboratory and in situ. Laboratory experiments measured these processes both in water aTOne and in soil:water columns. In situ work consisted of incubation of buried polyethylene bags containing ammonium or nitrate amended soil. Nitrification did not occur in the natural marsh water, but did occur in marsh water receiving sewage effluent. Ammonium in the oVerlying water of soil:water columns was nitrified at a rate considerably greater than in the absence of soil (Table 2, page 24). Water depth (15 cm vs. 30 cm) had 1 ittle effect on nitrification rates, although the water depth effect may have been negated by artificial aeration. Inclu sion of plants increased ammonium removal rates, I ikely due to plant assimilation of ammonium in addition to nitrification. Addition of calcium carbonate (agricultural I imestone) increased the nitrification rate to the point where nitrate accumulated in the overlying water. In situ incubation of ammonium amended marsh soil indicated that some nitrification would occur in the plant root zone, probably due to oxygen diffusing from the roots. Nitrification was not observed below the root zone. Denitrification did not occur in marsh water or in oxidation pond water in the absence of soil. Apparently there was not sufficient available organic carbon in these waters to deplete oxygen supply and allow denitrification to occur. Rapid, first-order denitrification was observed in soil:water columns (Table 4, page 28). Average nitrate removal rates, assuming a floodwater nitrate concentration of 10 mg N I iter-I, were 1.2 kg N ha-Iday-l for columns without plants and 2.0 kg N ha-Iday-I with plants. As with nitrification, water depth did not affect denitrification rate, again probably due to artificial aeration. In contrast to the results for nitrification, I iming had no effect on denitrification rate. This may have been due to the fact that the 1 ime was appl ied to the soil surface while denitrification occurs below the soil surface. Given sufficient time, it is anticipated that the lime would move downward in the soil and increase the pH in the zone of denitrification. Denitrification was rapid in nitrate amended soil samples incubated in situ (at both 8 and 30 cm depths) and was accelerated by increasing the soir-pH with lime. The above nitrification and denitrification rates represent our best estimates based on primarily laboratory investigations. Further field investigations are needed to. ascertain the effects of natural variations in water depth, dissolved oxygen concentration, plant density, and temperature. -36-

PAGE 50

CHAPTER III DENITRIFICATION IN FIFTEEN SELECTED FLORIDA WETLAND SOILS Denitrification can be an important mechanism for the removal of nitrate from wastewater applied to wetlands, as indicated in the previous The objective of this investigation was to measure denitrification rates in a variety of flooded wetland soils and to determine quantitative relationships between denitrification rate and selected soil characteristics in order to develop a predictive equation for denitrification rates in specific soils. Since nitrous oxide, one of the gaseous end-products of denitrification, is known to deplete ozone in the stratosphere, a quantitative evaluation of nitrous oxide evolution from wetland soils was also included in this study. METHODS AND MATERIALS Soils and Soil Sampling So i 1 s A total of 15 soils of widely varying organic matter content, reaction (pH), and texture were selected for study. All are classified as very poorly drained or poorly drained, and have water standing on the surface for one to nine months per year. Geographic distribution, ranging from Palm Beach and Lee Counties in the south to Walton County in the northwest, is shown in Figure 11. A list of soil types, classifications, and descriptions* is given below. These are general descriptions of the soil series, not of the actual pedons at sampling sites. Soils are 1 isted in declining order of denitrification rate, as determined in the present Everglades muck--typic medihemist--very poorly drained organic soils that formed in thick deposits of hydrophytic plant remains. They occur in freshwater swamps and marshes. I n a representat i ve profile, the surface layer is a black muck, about 12 inches thick. Below it, to a depth of 46 inches, is a reddish-brown mucky peat. Below a depth of 46 inches, and extending to below 78 inches, is dark brown peat. Floridana sand--arenic argiaquoll--a very poorly drained soil that occurs on broad low flats and in depressions. A representative profile has a black sand surface soil about 12 inches thick and a grayish-brown sand subsurface layer over a dark gray sandy clay loam subsoil that begins within a depth of 20 to 40 inches below the soil surface. These soils are formed in sandy and loamy marine sediments. Adapted from U.S.D.A. Soil Conservation Service soil interpretation sheets (Soil Form #5) for the soils series described. -37,-} "".--

PAGE 51

1. Everglades muck 2. Floridana f.s. 3. Astor s. 4. Surrency s. 5. Samsula muck 6. Pickney f. s 7. Pelham 1.f.s. 8. Riviera f.s 9. Chastain si. 1. 10. Brighton peat 11. I be ria s. c. 12. Chobee f.s.l. 1 3. De 1 ray f. s 14. Eureka f.s.l. 15. Valkaria f.s. Figure 11. Location of sampling sites. -38-

PAGE 52

Astor sand--cumulic haplaquoll--very poorly drained soils that occur in low flat areas, poorly defined drainageways, and floodplains in the lower coastal plain. A typical profile has a black to very dark grayish-brown sand surface layer about 32 inches thick. Below this to a depth of 80 inches is dark grayish-brown sand. Surrency sand--arenic paleaquult--very poorly drained nearly level soils in Central and North Florida. In a representative profile, the surface layer is sand about 22 inches thick. The upper 16 inches is black and the lower 6 inches is dark gray. The subsurface layer is light brownish-gray sand about 9 inches thick. Between depths of 31 and 65 inches, the subsoil is gray sandy clay loam. Samsula muck--terric medisaprist--nearly level very poorly drained organic soils. They occur in fresh water swamps and marshes. In a representative profile, the surface 9 inches is a black muck. The subsurface layer is a dark reddish-brown muck to a depth of 36 inches. Next is 10 inches of dark grayish-brown sand and then gray sand to below a depth of 55 inches. These soils formed in hydrophytic plant remains. Pickney fine sand--cumulic humaquept--very poorly drained soils in depressions and drainageways of the lower coastal plain. The typifying pedon has black loamy fine sand A horizons over dark gray fine sand C horizons. Pelham loamy fine sahd--arenic paleaquult--poorly drained nearly level soils on the coastal plain. Typically the surface layer is very dark gray loamy fine sand about 5 inches thick. The subsurface is gray loamy sand about 21 inches thick. The subsoil extends below a depth of 68 inches. It is mostly light gray sandy clay loam mottled with yellow and red. Riviera fine sand--arenic glassaqualf--very poorly drained soils in depressed areas of the coastal plain. In a representative profile, the surface layer is dark grayish-brown sand about 6 inches thick. The subsurface layer is white sand about 22 inches thick and it extends into the underlying grayish-brown sandy clay loam subsoil. They formed in unconsolidated marine sands and loams. Chastain silt loam--typic haplaquept--poorly drained, slowly permeable soils on flood plains of rivers on the coastal plain. Typically these soils have yellowish-brown surface layers over gray clayey subsoils. Brighton peat--typic medifibrist--very poorly drained organic soils that occur in depressions and freshwater marshes and swamps in peninsular Florida. Typically they have a black peat surface layer 9 inches thick. Beneath the surface layer dark yellowishbrown and dark brown peat extends to 63 inches or more deep. Iberia sandy clay--vertic haplaquoll--poorly drained, very slowly permeable soils. They have a black clay surface and a dark gray clay subsurface. These soils formed in alkaline clayey alluvium. They occur on the lower part of the natural levees, often in depressions. -39"",J', -,'. .' .. .. ,......}

PAGE 53

Chobee fine sandy loam--typic argiaquoll--very poorly drained soils that occur in small to large depressions, low areas, or river flood plains. A representative profile has a black fine sandy loam surface layer about 7 inches thick and a black to very dark gray sandy loam subsoil that contains calcium carbonate nodules. Delray fine sand--grossarenic argiaquoll--deep, very poorly drained soils that occur on low, broad flats or depressions in South and Central Florida. A representative profile has a black fine sandy surface 1 ayer. Between a depth of 46 to 60 inches Ls a dark grayish-brown fine sandy loam subsoil. These soils formed in thick deposits of sandy and loamy marine sediments. Eureka fine sandy loam--typic albaqualf--poorly drained soils that occur in South and Central Florida. In a representative profile the surface layer is black loamy fine sand about 4 inches thick. The subsurface layer is grayish-brown loamy fine sand about 7 inches thick. The subsoil begins at depths of 11 inches and extends to 72 inches and deeper. It is gray sandy clay. Valkaria fine sand--spodic psammaquent--poorly drained soils that occur on low, poorly defined sloughs and areas adjoining swamps in the lower coastal plain. In a representative profile, the surface layer is sand about 9 inches thick; the upper 5 inches is black and the lower 4 inches is dark brownish-gray. Between depths of 9 to 15 inches is 1 ight gray sand, with brownish or grayish sand to depths of 80 inches. Collection of Samples Samples for column studies were collected in 5 cm i.d. by 60 cm long plexiglass columns. Where possible, sharpened columns were simply driven into the soil to a depth of 25 cm with a. hammer. In cases where this resulted in excessive soil compaction, a more involved method was used. A rubber stopper exactly matching the inner diameter of the column was inserted in the bottom of the column and suspended at the soil surface from a tripod by a cord running down the center of the column (Figure 12(a)). A small amount of water was poured into the column to help maintain a seal between column and stopper. The column was then driven into the soil while the stopper remained stationary, creating a suction in the column to prevent the soil surface from receding as the column moved downward (Figure 12(b)). To remove the stopper, the cord was pulled upward at an angle, thereby tilting the stopper and breaking the seal (Figure 12(c)). Columns were extracted with a rocking, twisting motion after stoppering the top to prevent soil loss out of the bottom. This sampling method was adapted from that of Coleman (1979). Bulk samples of the top 10 em of soil were stored in polyethylene bags at 4c for a maximum of 5 days before determination of mineralizable organic carbon and pH. Subsequently, samples were oven dried at 104c for total organic carbon determination. -40-

PAGE 54

I .p-, Eyebolt I Stopper Surfoce (a) ( b) (e) Figure 12. Apparatus for obtaining soil cores with minimum soil compaction. "-...--,,<

PAGE 55

Description of Experiments Denitrification and Nitrous Oxide Evolution A laboratory study was conducted to determine denitrification rates and nitrous oxide (N2 0) evolution rates for 14 of the soil types collected. Three columns of each soil type were flooded with deionized water to a depth of 15 cm. Aquarium airstones (previously determined to be carbonate and ammonia free) were used to aerate and mix floodwater. Air was prehumidified and stripped of ammonia by use of an acidified (H2 S04) water trap. Air flow rate in each column was approximately 30 to 40 ml min-I. Water levels were maintained throughout the study by periodic additions of deionized water. This and all other studies were conducted within a temperature range of 26C0C. Columns, prepared as described above, were allowed to stand for 7 days before initial nitrate application. Background floodwater nitrate concentrations were measured during this time. Water samples for nitrate determination were pipetted from the floodwater surface. NitrateN levels in the floodwater were then raised to 20 mg liter-l by addition of potassium nitrate solution. A 2 week incubation period followed, for the purpose of allowing microbial adaptation to nitrate and establishment of a relatively stable nitrate diffusion front in the soil. During this period, concentrations of nitrate were monitored every 3 to 4 days and additions made if nitrate-N dropped below 10 to 15 mg liter-I. No additions of nitrate were made within 3 days of the end of the adaptation period. Floodwater nitrate concentrations were measured at 0, 4, 8, and 12 days after the end of this period. Nitrate in two arbitrarily selected extra colums (Surrency soil) was measured more frequently and for a longer period of time for the purpose of obtaining detailed nitrate disappearance curves. One column per soil type was selected for determination of N20 evolution rate by the method described below. This was done approximately 3 weeks after initial nitrate application. Long Term Denitrification and Nitrous Oxide Evolution Two soils, Samsula and Pelham, were selected for a more detailed study to determine changes in denitrification and N 2 0 evolution rates over a 10 week period. Eight columns of each of the soils were treated similarly to those in the previous experiment. Denitrification rates (calculated as described in results and discussion) were determined at 2, 5, and 8 weeks, with nitrate-N maintained between 5 and 20 mg liter-l throughout the study. Each denitrification rate determination consisted of three nitrate concentration measurements made at 4 day intervals. Nitrous oxide evolution from each of two Samsula and three Pelham columns was determined at least three times during the study period by the method described below. Effect of Floodwater Nitrate Concentration on NZO Evolution A short experiment was conducted to determine if N20 evolution was proportional to nitrate concentration. Two Samsula and two Pelham columns were treated as in previous experiments. Nitrous oxide -42-

PAGE 56

evolution was measured at least three times per column over a period of several days as floodwater nitrate-N concentrations fell from 15 mg 1 iter-l to below 5 mg 1 iter-I. Background evolution rates for these columns were determined prior to initial addition of nitrate to the floodwater. Method for Measuring N20 Evolution Nitrous oxide evolution from the columns was determined in a semienclosed, continuous flow collection chamber. This involved sealing the top of the column, except for small air entry and exit tubes. Aeration continued, but at a reduced, controlled rate of 7 to 20 ml min-I. Floodwater oxygen concentration was periodically measured to insure sufficient aeration. Air flow was measured with a soap-film flow meter connected to the air exit tube (Figure 13). In this semi-enclosed chamber, N20 concentration in the headspace will rise slightly due to N20 production in the soil. In response to this, N20 dissolved in the floodwater also increases. Eventually an equilibrium is established, and N 2 0 no longer accumulates in the floodwater or headspace. At this time N20 entering the chamber plus N20 produced in the column will equal N20 exiting the chamber. If flow rate, influent N20 concentration, and effluent N20 concentration are known, N20 production can be calculated as follows: R = (4.6xlO-14) (E -A)(f) (4.9xl06 ) (28) [2 ] where R is N20-N evolution rate, ha-lday-l; 4.6xlO-14 is the number of moles of gas in 1 mf at a concentration of 1 ppb (v:v) at standard temperature and pressure; E and A are 'effluent and ambient (incoming air) N20 concentrations, ppb (v:v); f is air flow rate, ml day-I; 4.9xl06 converts the column surface area to hectares; and, 28 is the mole to gram conversion factor for N20-N. In practice, a 12 hour equilibration time was found to be more than sufficient to establish constant N20 concentration in effluent air. Five, 1 ml samples of effluent air and five samples of influent (ambient) air were collected for each N20 evolution determination. A similar method was described by Kaspar and Tiedje (1979) for use with unflooded soils. In the long term study described above, N20 samples were stored in plastic syringes for up to 3 days. It was found that concentrations decreased with storage; i.e., the difference between sample and normal ambient (300 ppb) N20 concentrations decl ined by approximately 24, 45, and 65% after 1, 2, and 3 days, respectively. Measured N20 concentrations of stored samples were corrected by the loss factor, and corrected values were used to determine evolution rates. Analytical Methods Nitrate Nitrate concentration was measured with an Orion model 93-07 nitrate electrode in conjunction with an Orion 90-01 double junction reference electrode and Orion 701 digital meter. Ionic strength of sample and standard solutions were adjusted with ammonium sulfate. -43\

PAGE 57

/Flow meter Effluent air/ Water Soil Airstone Air Pump Figure 13. System for determining nitrous oxide evolution from flooded soi 1 cores. -44-

PAGE 58

Nitrous Oxide A Tracor model 222 gas chromatograph with a Ni-63 electron capture detector were used for N20 determinations. Carrier gas was 5% methane in argon, with a flow rate of 28 cm3 min-I. Porapak Q was used in the separation column. Column temperature was 500C and detector temperature o was 350 C. Sample volume was 1.0 mI. Bulk Density and Porosity Bulk density and porosity of the upper 10 cm of soil in each column were determined gravimetrically. Saturated soil sections were weighed, dried overnight at 104C, and reweighed. Bulk density was calculated as the dry weight per original soil volume, and porosity was calculated as the volume of water lost per original soil volume. Other Methods Fresh soil samples (not dried) were used for pH determination. Measurements were made in deionized water with a soil :liquid ratio of 1:1 (v:v). A combination pH electrode and Orion 701 meter were used. Total N was determined by the micro-Kjeldahl procedure. Total organic carbon was determined by the Walkley-Black method (Allison, 1965). Efficiency of organic matter oxidation by dichromate was assumed to be 77%. -45-

PAGE 59

RESULTS AND DISCUSSION Comparative Denitrification Rates Denitrification rates were compared for 14 soil types encompassing a wide range of organic matter contents and pH values. Values represent an average of three repl ications. Measurements of nitrate concentration in the overlying water were taken three or four times over a 12 day period following a 2 week preincubation period. In addition, two columns of Surrency soil were sampled ten times each in order to obtain more detailed nitrate disappearance curves. Prior to rate comparison, the data were analyzed to determine the rate-order of the denitrification reaction. Floodwater nitrate disappearance data are presented in Figures 14 and 15. Simple visual examination of the graphs indicates that the process is not zero-order. If the reaction is first-order (as in the 1974 model of Bouldin et al.), the following relationship between concentration and time exists: I n Co = k t c1 [3] where Co is initial nitrate concentration, mg liter-l ; C is nitrate concentration at time t, mg liter-l ; kl is first-order rate constant, day-l; and t is time, days. If the reaction is half-order (i.e., rate is proportional to the square root of concentration, as in Bouldin's 1968 model), the concentration versus time relationship is as follows (Gucker and Siefert, 1966): 1: 1: C2 -C [4] o '"l 1, where is half-order rate constant, and other variables are as in Equation 3. The data in Figure 14 were fitted to Equations 3 and 4 by leastsquares 1 inear regression. Due to differing initial floodwater nitrate concentrations (made unavoidable by the adoption of the preincubation period) and differing denitrification rates in each column, each was treated separately, with separately calculated k values and regression coefficients. Since a total of 38 columns were studied, 38 separate regression procedures were performed for each equation, each with 1 or 2 degrees of freedom. Mean r2 values were 0.981 for Equation 3 and 0.978 for Equation 4. Data from 21 of the columns fit Equation 3 better, 12 columns fit Equation 4 better, and 7 columns fit both equations equally well. Thus, the data weakly support Equation 3, or firstorder kinetics, over Equation 4, or half-order kinetics. Results for the two more intensively sampled columns of Surrency soil (Figure 15) were ambiguous. Column A yielded r2 values of 0.988 and 0.986 for Equations 3 and 4, respectively, while column B yielded -46-

PAGE 60

-47-

PAGE 61

""j at E -Z I 0 z z I o z 8 6 4 2 1 muck f e sand Pickney In 0 4 0 8 12 112 time, days Riviera In f e sand 2 8 12 o time, days (Continued.) Figure 14. -48tal n silt loam Chas x -. __ o 1:/2 4 .. /

PAGE 62

16 Brighton peat Iberia sandy clay 14 12 I 10 01 .------E 8 z t "' 0 z 6 4 0_ -0 2 0 0 4 8 12 0 4 8 12 time, days Chobee fine sandy loam Delray fine sand I 01 o E 8 Z t 0"' 6 z 0 0 _____ 0 0 4 8 12 0 4 8 12 time, days Figure 14. (Continued.) -49-

PAGE 63

16 Eureka fine sandy loam Val karia fine sand 14 12 10 I 0 e "'------0_ If z I -------. -0 cr Z o __ ______ ______ ______ o 4 8 12 o 4 8 12 time, days Figure 14. (Continued.) -50-

PAGE 64

I V1 \. 'j 01 E z 10: ''''' o z o 2 4 6 8 tlml,days --column A column B 12 14 16 Figure 15. Changes in floodwater nitrate concentrations wtth time in two' soil:water columns of Surrency soil. \ 18 < J

PAGE 65

r2 values of 0.993 and 0.994 for the respective equations. Statistical analysis of the data from these columns used ten observations, or eight degrees of freedom per column. Although the data appear to fit both first-order and half-order apparent kinetics, first-order kinetics are assumed for the remainder of this work, and only first-order rate constants are reported. This is partly in order to be consistent with published work and partly because the concept of first-order rate constant is more meaningful to most readers than that of half-order rate constant. The idea of half-order kinetics has been raised here to illustrate that a good fit of data to first-order kinetics does not preclude other possibilities. It is likely that the apparent kinetics of denitrification in flooded systems are complex and lie somewhere between first-order and half-order reaction rates, as suggested in Chapter 1. Mean first-order denitrification rate constants for the 14 soil types and results of Duncan's New Multiple Test are presented in Table 5. Rate constants range from 0.192 day-for the Everglades soil to 0.040 day-l for the Valkaria soil. Significant differences exist among the soils, but groups of non-significantly different soils are broad and overlapping. This is due primarily to the high degree of variation within soil types. Spacial variation in soil organic matter content and availabil ity may account for this. Factors such as the presence of decomposing soil fauna may increase organic matter availability and, consequently, denitrification rates. In cases where decomposing animals were visually apparent, columns were discarded. The last column in Table 5 gives nitrate-N removal rates in g ha-l day-I, assuming maintenance of 10 mg liter-l nitrate-N in the floodwater. The Everglades soil will remove 2,900 g ha-lday-l, while only 600 g ha-lday-l is removed by the Valkaria soil. All denitrification data given here assume a 15 cm floodwater depth. As noted in Chapter 2, first-order rate constants are dependent on floodwater depth, but equivalent nitrate-N removal rates are not. The experimental conditions of this study include well-aerated floodwater and no downward flow of water into the soil. As a result, floodwater nitrate must diffuse through a relatively thick aerobic soil zone into the zone of denitrification, and denitrification is relatively slow. Downward flow of water, which often occurs in the field, can be expected to increase denitrification rate by accelerating nitrate transport. Reduced floodwater oxygen content would also increase denitrification rate by narrowing the aerobic soil zone. For these reasons, denitrification rates presented here should be regarded as minimum expected situ rates. Selected characteristics of 15 soil types (Table 6) were examined for possible correlation with denitrification rate. The 14 soil types discussed above, plus Pelham soil, discussed later, were studied. Simple and multiple 1 inear regression models were constructed using various combinations of characteristics as independent variables and denitrification rate as the dependent variable. Complex models were analyzed by the General Linear Models procedure of the Statistical -52-

PAGE 66

Table 5. Comparative denitrification rates of 14 representative Florida wetland soils. Soil type Everglades muck Floridana f.s. Astor s. Surrency s. Samsu1a muck Pickney f. s. Riviera f.s. Chastain si.1. Br i ghton peat I be ria s. c Chobee f.s.1. De 1 ray f. s Eureka f.s.l. Val ka ria f. s Denitrification rate constant Standard error of mean -1 ------------day -----------0.192 0.017 __ f 0.149 0.108 0.105 0.102 0.087 0.076 0.067 0.066 0.063 0.062 0.061 0.049 0.040 0.020 ab 0.018 bc 0.001 c 0.009 c 0.025 cd 0.009 cd 0.014 cd 0.010 cd 0.012 cd 0.009 cd 0.021 cd 0.005 d t Assuming a nitrate-N concentration of 10 mg 1 iter-I. Equivalent nitrate-N t remova 1 rate -1 -1 g ha day 2,900 2,200 1 ,600 1 ,600 1 ,500 1 ,300 1 ,100 1 ,100 990 950 930 920 740 600 f Omitted from statistical analysis due to missing data. Rates followed by the same letter are not significantly different; CI. = 0.05. -53-

PAGE 67

Table 6. Selected chemical and physical characteristics of surface horizons (top 10 cm) of the soils studied. Denitrification Total organic carbon Bulk So i 1 type rate constant By vol. By wt. pH density C:N ratio ----day-l-----mg cm-3 --%--g cm-3 ---w:w---Everglades muck o. 191 59 46.0 6.2 0.13 25: 1 Floridana f.s. 0.149 127 12.6 6.5 0.97 B. B: 1 Astor s. o. loB 60 7.4 6.7 O.Bl 6.3: 1 Surrency s. 0.105 Bo 42.0 4.9 0.19 27: 1 Samsula muck 0.102 3B 54.0 4.6 0.08 16: 1 I Pickney f. s 0.OB7 149 5.0 0.55 33: 1 Vl 27.0 .t:-I Pelham l.f.s. O.OBOt f 6.5 5.6 N.D. N.D. N.D. Riviera f.s. 0.076 52 5.0 4.5 1.04 1 3: 1 Chastain si.1. 0.067 24 5.2 5.6 1. 37 N.D. Brighton peat 0.066 44 53.0 4.2 0.09 21 : 1 I be ria s. c. 0.063 40 4.0 6.1 1.00 N.D. Chobee f.s.l. 0.062 59 7.4 6.8 0.79 7.5:1 Delray f.s. 0.061 65 6.4 6.9 1 .01 6.0: 1 Eureka f.s.l. 0.049 4B 5.7 5.5 0.84 6.6: 1 Valkaria f.s. 0.040 6.9 0.43 6.6 1.60 19: 1 t Rate determined in long term study. discussed in later section. f N.D. ; not determined.

PAGE 68

Analysis System (Barr et al., 1976). Since denitrification is dependent on an available carbon source, total organic carbon content on a weight per soil volume basis and on a weight per soil weight basis were tested as independent variables. The literature suggests that soil pH has a significant effect on denitrification rate only at values somewhat below neutrality. For this reason, a pH-deficit variable was introduced. The new variable was defined as follows: if observed pH-deficit = 0 if observed ph
PAGE 69

that because of the square root factor, a 50% reduction in electron acceptor demand only reduces denitrification rate by 29%. This helps explain the fairly narrow range of denitrification rates among the soils studied, in spite of the wide range of organic carbon contents and expected electron acceptor demands. It is felt that Equation 5 could act as a simple predictor of denitrification capacities of wetland soils in Florida. Only two easily determined soil characteristics need be measured in order to make a prediction. Table 7 gives observed denitrification rates for each soil type, rates predicted by Equation 5, and percent error of predictions. Some errors are relatively high, particularly for some of the more slowly denitrifying soils. In view of the wide variation in denitrification rates within soil types (Tables 5 and 9), broad confidence intervals should be assumed for any predictive equation. Due to the considerations noted above, the equation should be regarded as invalid for soils with organic carbon contents below 0.5%. Fortunately, most wetland soils in Florida are fairly high in organic carbon content. Nitrous Oxide Evolution Nitrous oxide evolution rates were determined for one column of each of 14 of the soil types discussed above. Observed values were adjusted to account for variations in floodwater nitrate concentration at time of measurement. Adjusted evolution rates were correlated with denitrification rates of the columns on which the N20 evolution measurements were performed. The continuous flow, semi-enclosed system used to determine N20 evolution has certain advantages and disadvantages as compared to the more commonly used totally enclosed system. Perhaps the most important advantage is that N20 levels in the soil :water:air system remain fairly close to those which would occur in the field, where evolved N20 can freely escape into the atmosphere. The highest observed N20 concentration in effluent air from the semi-enclosed system was approximately 2.5 times the normal ambient concentration. In contrast, completely closed systems often allow N20 concentrations to increase to more than 10 times the ambient concentration (Denmead, 1979). Thus the microorganisms in closed systems are exposed to a highly N20 enriched environment which might have an effect on further production and reduction of N20. A major disadvantage of the semi-enclosed system is that low N20 concentrations result in diminished measurement accuracy. In some cases, effluent concentrations were only 10% greater than ambient levels, and therefore accuracy suffered. Five repl icate samples per evolution determination were used in an attempt to offset this problem, resulting in a ppb (v:v) confidence interval about the mean (a = 0.05). Use of airstones in the collection system offered the advantage of maintaining aerobic floodwater such as would often occur in the field, but also resulted in some difficulties. Pores of airstones tended to clog, sometimes reducing airflow to nearly zero. This caused many delays in data collection, since a 12 hour equi libration period at a relative constant flow rate was required before sampling. -56-

PAGE 70

Table 7. Denitrification rate constants as predicted by Equation 5. Total Rate constant organic Errorf Soi 1 type pH carbont Observed Predicted ---%-----------day-1-----------%--Everglades muck 6.2 46.0 0.192 0.193 + 1 F10ridana f. s. 6.5 12.6 0.149 0.115 -23 Astor s. 6.7 7.4 0.108 0.093 -14 Surrency s. 4.9 42.0 0.105 0.113 + 8 Samsu1a muck 4.6 54.0 0.102 0.098 -4 Pickney f. s. 5.0 27.0 0.087 0.108 +24 Riviera f. s. 4.5 5.0 0.076 0.050 -34 Chastain s i .1. 5.6 5.2 0.067 0.069 + 3 Brighton peat 4.2 53.0 0.066 0.057 -14 i / Iberia s.c. 6.1 4.0 0.063 0.070 +11 Chobee f. s. 1. 6.8 7.4 0.062 0.093 +50 De 1 ray f. s. 6.9 6.4 0.061 0.088 +44 Eureka f. s. 1. 5.5 5.7 0.049 0.069 +41 Va 1 kari a f. s. 6.6 0.43 0.040 0.049 +23 t By volume. f [(predicted-observed)/observed] x 100. -57-

PAGE 71

Sampling delays made it necessary to make N 2 0 evolution determinations at different floodwater nitrate-N concentrations for each column. Concentrations ranged from 4 to 9 mg 1 iter-I. In order to allow comparison of evolution values, observed values were adjusted to reflect a 10 mg llter-l nitrate-N concentration in all columns. This was done on the assumption that N 2 0 evolution is directly proportional to floodwater nitrate concentration. In an experiment designed to verify this assumption, evolution rates from two Samsula and two Pelham columns were measured as floodwater nitrate-N concentrations decl ined from 15 mg liter-l to less than 5 mg liter-I. Results are given in Figure 16 as graphs of nitrate-N concentration versus N20 evolution. Plots are reasonably linear, indicating that within the concentration range in question, the above assumption is approximately val ide There is some indication, particularly in the case of Samsula, that the relationship deviates from 1 inearity, but this deviation does not appear to be extreme. It has been reported that N20 evolution relative to nitrate concentration can increase as nitrate concentration increases, due to nitrate inhibition of N20 reduction (Blackmer and Bremner, 1979; Terry and Tate, 1980). Such inhibition does not appear to have occurred to any significant degree in this experiment. Since nitrate must diffuse from floodwater through an aerobic soil zone, it is likely that nitrate concentrations in the zone of active denitrification were somewhat lower than those in the floodwater. Such extremely low concentrations may not have been sufficient to cause significant inhibition of N 2 0 reduction. Nitrous oxide evolution values for the 14 soil types are presented in Table 8. Since evolution was determined for only one column per soil type, values presented here may vary from the mean evolution rates of the individual soils. It is 1 ikely, however, that the data accurately represent the range of N20 evolution rates to be expected in Florida wetland soils. The data also allow examination of the correlation between N 2 0 evolution and various soil characteristics. Evolution rates, based on a 10 mg 1 iter-l floodwater nitrate-N concentration, range from less than 5 g N ha-lday-l for the Floridana soi 1 to 42 g N ha-lday-l for the Riviera soil, with a mean of 20.9 g N ha-lday-l. The percentage of consumed nitrate-N accountable for as N20-N ranges from less than 0.2% for Floridana to 6.5% for the Valkaria, with a mean of 2.1%. Mean values for organic soi Is are 18.5 g ha-1 9Qy-l and 1.1%, while mineral soils have mean values of 22.6 g ha-lday I and 2.3%. The lower means for organic soils may reflect a higher electron acceptor demand, as discussed below. Percentage figures are in fairly close agreement with the 0.8 to 1.4% values reported for flooded soils by Denmead et al. (1979). Nitrous oxide evolution and percent of consumed nitrate-N accountable for as N20-N were tested for possible correlation with nitrate consumption rate (apparent first-order denitrification rate constant) and soil pH. Neither of the NZO parameters showed any correlation with soi 1 pH. Nitrous oxide evolution as g N ha-1day-l was weakly, negatively correlated with nitrate consumption rate (r2 = 0.164). Percent N20 showed fairly strong correlation with nitrate consumption rate, -58-

PAGE 72

I '1:J Ie ..c 0' z I 0 ZN .. Ie ..c 0' -Z I 0 N Z 30 20 10 30 20 10 5 5 10 I-I N0 3 N, mg Samsula o column A x column B 15 Pelham o column A x column B 15 Figure 16. Nitrous oxide evolution from flooded soil cores in relation to floodwater nitrate concentrations. -59-

PAGE 73

Table 8. Nitrous oxide evolution rates from 14 representative Florida wetland so i 1 s. Soil type Everglades muck Floridana f.s. Astor s. Surrency s. Samsula muck Pickney f.s. Riviera f.s. Chastain si.1. Brighton peat I be ria s. c. Chobee f.s.l. De 1 ray f. s. Eureka f.s.l. Val ka ria f. s mean Denitrification rate constantt ----day-l----0.175 O. 112 0.077 0.105 0.110 0.065 0.065 0.092 0.081 0.050 0.071 0.043 0.161 0.040 Floodwater NOrN conc. mg 1 i ter-1 4.0 4.0 7.0 5.0 8.0 6.5 5.0 6.0 7.5 9.0 7.0 9.0 3.5 8.5 N20-N evolution rate Observed 7.2 t 7.0 13.4 15.4 13. 1 21.0 15.5 7.9 24.1 10.9 18.5 5.0 32.5 18.0 t 10.0 26.8 19.3 20.2 42.0 25.8 10.7 26.8 15.6 20.6 14.3 38.2 20.9 t Rate for individual column, not mean rate for soil type. f Adjusted to 10 mg 1 iter-1 floodwater N03-N. (N20-N evolved/N03-N consumed) x 100. -60-0.7 0.2 0.9 1.7 1.2 2.1 4.4 1.4 0.9 3.6 1.5 3.2 0.6 6.5 2.1

PAGE 74

as indicated in Figure 17, and in the equation below (r2 = 0.706, d.f. = 12): [6] where kl is the first-order denitrification rate constant, day-I, and %N20 is (N20-N evolution rate/nitrate-N consumption rate) x 100. Equation 6 agrees with the relationship between N20 release and electron acceptor demand reported by Focht and Vertraete (1977). In this case, denitrification rate can be considered an indication of electron acceptor demand. Apparently, in soils with high electron acceptor demand, N20 produced by denitrification tends to be further reduced to N 2 thus acting as an additional sink for electrons. In soils with lower electron acceptor demands, less of the N20 produced is further reduced and thus more is evolved. Long Term Denitrification and Nitrous Oxide Evolution Columns of Samsula and Pelham soils (eight per soil type) were incubated for 10 weeks with floodwater nitrate-N concentrations maintained between 5 and 20 mg 1 iter-I. Denitrification rates were determined at 2, 5, and 8 weeks. Nitrous oxide evolution from two Samsula and three Pelham columns was also measured periodically during this time. Denitrification rates are given in Table 9. Mean denitrification rates for the two soil types remained relatively constant throughout the study period. Individual denitrification rates fluctuated somewhat, however. In most columns, the change in rate appeared to be random. This may reflect irregular transport of nitrate into the soil. On several occasions, relatively large volumes of gases (presumably biogenic) were observed to be rapidly released from the soil. Release of gases had a stirring effect on the upper layer of soil, resulting in a loose, flocculent soil mass. In one case, a layer of soil material approximately 1 cm thick was observed floating on top of the floodwater, having been 1 if ted by trapped gas. Such disruptions of the soil surface could result in increased transport of floodwater nitrate into the soil and thus increase the denitrification rate for short periods following disruptions. It may be assumed that this effect also occurs in nature. Temporal variation in decomposition rates of soil, plant and animal matter may also have contributed to variations in individual column denitrification rates. In the case of Samsula columns 5 and 6, denitrification rates rose drastically during the study period. The rise was accompanied by darkening of the floodwater, probably due to accumulation of dissolved organic substances. Darkening of floodwater occurred to a lesser degree in all Samsula columns. Increased organic substrateavailabi lity is a 1 ikely cause for the accelerated denitrification. It is possible that columns 5 and 6 contained plant rhizomes or other organic materials which decomposed only after a lag period. Such lag periods are not uncommon in soil microbiological systems, and may be the result of such factors as substrate recalcitrance and microbial succession (Stotzky, 1973) -61-

PAGE 75

0 0 J( Q) e 1/1 C 0 u Z Oft') Z "-" Q) 0 > Q) Z 0 N Z 6 5 4 o % N 2 0 = 0.25 (.!) -1.26 k, 2 r =0.706 .1.. (days) kJ Figure 17. Relationship between reciprocal of denitrification rate constant and percent nitrous oxide evolution for 14 soil types. -62

PAGE 76

Table 9. Long term (8 week) denitrification rates in Samsula and Pelham --", so i 1 columns. Den i t r i fica t ion rate constant So i 1 type Rep 1 icate Week 2 Week 5 Week 8 Samsula muck 0.093 0.080 0.073 2 0.109 0.062 0.074 3 0.067 0.077 0.054 4 0.049 0.080 0.105 5 O. 112 0.091 0.208 6 0.105 0.209 0.237 7 0.060 0.062 0.059 8 O. 101 0.064 0.074 avg. 0.087.021 0.091.041 0.095.050 Pe 1 ham 1. f. s. 0.090 0.097 0.094 2 0.064 0.081 0.065 3 O. 101 0.062 0.100 4 0.078 0.100 0.074 5 0.053 0.037 0.054 6 0.091 0.061 0.064 7 0.092 0.072 0.048 8 0.051 0.071 0.054 avg. 0.078.016 0.073.017 0.069.016 -63-

PAGE 77

Nitrous oxide evolution from Samsula (columns 3 and 8) and Pelham (columns 5, 6, and 7) soils are presented in Figure 18. Evolution is shown as percent of consumed nitrate-tJ accountable for as N20-N. Nitrate consumption is based on the denitrification rate occurring in each column at the time N20 evolution was determined. In three of the five columns, evolution decreased during the study period. This agrees with the observation of Guthrie and Duxbury (1979) that N 2 0 evolution decreases with increased time of soi 1 exposure to nitrate. This decrease would be expected to result from microbial adaptation and selection. With time, the microbial community is 1 ikely to maximize use of N20 as an electron acceptor, since electron acceptors are in short supply in anoxic environments. Nitrous oxide evolution increased in the case of column 7 of Pelham soil. The first value for this column is extremely low, and may be in error. Unusual population dynamics or soil conditions may be responsible for the upward trend in evolution. An interesting phenomenon was accidently observed in column 8 of the Samsula soil (Figure 18). Air flow in the column was inadvertently reduced to less than 3 ml min-I, and floodwater dissolved oxygen concentration dropped to approximately I mg 1 iter-I. Coinciding with this was a peak in percent H20 evolution of 47 g ha-lday-l. Evolution subsided to 35 9 ha-lday-l one day later, after flow was increased to 7 ml minl and dissolved oxygen content increased to about 7 mg liter-I. A similar effect occurred in column 4 of Samsula soil during an earl ier, unreported phase of this study. Aeration slowed and dissolved oxygen concentration dropped to less than 1 mg 1 iter-I, resulting in an evolution rate of 40 9 N ha-lday-l. Five days earl ier, evolution from this column had been 23 g N ha-lday-l and 9 days later the value was 16 g N ha-lday-l, with more adequate aeration in both cases. The observed increases in N20 evolution rate may be due to accelerated denitrification resulting from low floodwater dissolved oxygen, a phenomenon reported by Van Kessell (1977). No increase in denitrification rate could be detected in the present study because the periods of anoxia were very short compared to nitrate sampling intervals. Regardless of the actual cause, variation in floodwater oxygen content appears to have a marked effect on N20 evolution and perhaps on denitrification. The problem arises as to how closely results presented in this study reflect those occurring in nature. With the above exceptions, floodwater dissolved oxygen content remained relatively constant throughout the study. In nature, dissolved oxygen content often undergoes diurnal fluctuations due to daytime oxygen production during photosynthesis. This difference may have resulted in lower observed N20 evolution and denitrification rates in this study than would occur in the field. -64-

PAGE 78

0 2 ;: c: 0 u z 0"" Z "-" .. > "0 > .. z 0 ZN 0 2 .. c: 0 u Z 01<'1 Z "-" .. > "0 > .. z 0 ZN Samsula 6 column 3 ..... column 8 4 ,/ \ // \ ,/ \ ,/ \ / ,/ 2 3 4 5 6 7 8 9 W .. ks aft.r initio I nitrate application !5 4 3 2 ____ ________ __________ ________ __________ ________ __________ ______ __ 3 4 5 Fi gure 18. 6 7 8 W kl aft.r initial nitrate application Long term (10 week) nitrous oxide evolution from flooded Samsula and Pelham soil columns. -65-9

PAGE 79

SUMMARY AND CONCLUSIONS Denitrification and N 2 0 evolution rates in a wide variety of Florida wetland soils were investigated in the laboratory using intact soi 1 columns. It was determined that denitrification in the flooded soils followed apparent first-order kinetics, with rate constants ranging from 0.040 day-l to 0.192 day-I. Denitrification rate was related to soil organic carbon content and pH by the following formula: .; -4 -4 kl = (8.9xlO ) (OCw) -(3.9xlO ) (OCW) (pH6 5 ) + 0.002 where kl is denitrification rate constant, day-I; OCW is organic carbon content by weight, %; and pH6.5 is pH deficit below pH 6.5. Residence times required for 50% removal of floodwater, assuming a floodwater depth of 15 cm and irrespective of nitrate concentration, ranged from 3.6 to 17 days. For 90% removal, 12 to 58 days would be required. These figures represent minimum removal rates. Either mass flow of floodwater into the soil or low floodwater dissolved oxygen content would be likely to result in higher removal rates. Assimilation of nitrate by vascular plants would also increase removal rates significantly. It Is felt that many Florida wetland soils could function as efficient treatment systems for nitrate-bearing wastewater. Soil organic carbon content and pH, as well as hydrologic factors need to be considered in site selection. A semi-enclosed chamber system was devised for determination of N20 evolution rates. This system functioned satisfactorily without exposure of the soil:water environment to greatly elevated N20 concentrations, as occurs in totally enclosed systems. Nitrous oxide accounted for from less than 0.2 to 6.5% of consumed nitrate. Wetland denitrification thus appears to be a relatively innocuous means of nitrate disposal, from the viewpoint of N20 production and stratospheric ozone destruction. Soils which denitrified rapidty tended to evolve less N20 than those with slower denitrification rates. Contrary to many reports, N20 evolution in the soils studied was not related to soil pH. This may be due to the relatively low floodwater nitrate-N concentrations (less than 20 mg liter-I) in the present study. Two soil types were studied during a 10 week period of continuous nitrate exposure. Denitrification rates of individual columns tended to fluctuate randomly, but average rates remained relatively constant. Average N 2 0 evolution rate decl ined by approximately 60% between weeks 3 and 10, probably due to microbial adaptation and selection. Reduction in floodwater dissolved oxygen content appeared to result in increased N20 evolution. Further study is needed to determine more precisely the effect of naturally occurring fluctuations in floodwater oxygen content on denitrification and N20 evolution rates. Appl ication of high BOD wastewater could be studied in conjunction with this. The effect of -66-

PAGE 80

downward water flow on the above rates also warrants investigation. Denitrification and N20 evolution should be investigated in other wetland ecosystems, including estuarine environments. situ studies and refinement of situ methods, particularly for N20 evolution measurement, are needed in order to confirm the results of laboratory work. -67-

PAGE 81

LITERATURE CITED Alexander, M. 1965. Nitrification. In \J. V. Bartholomew and F. E. Clark (eds.), Soil Nitrogen. Agronomy 10:307-343. American Society of Agronomy, Madison, WI. Alexander, M. 1977. The nitrogen cycle: Nitrification; Denitrifi-cation. In Introduction to soi 1 microbiology. 2nd ed. John Wi ley and-Sons, Inc. p. 251-286. All ison, L. E. 1965. Organic carbon. In C. A. Black (ed.), Methods of soil analysis. Agronomy 9:1367-1378. American Society of Agronomy, Madison, WI. Andersen, J. M. 1976. Importance of the denitrification process for the rate of degradation of organic matter in lake sediments. In H. L. Go1terman (ed.), Interactions between sediments and water. Dr. W. Junk B. V. Pub1 ishers, the Hague. p. 357-362. Ardakani, t1. 5., L. W. Baker, and A. D. McLaren. 1975. of nitrate in a soil column during continuous flow. Soc. Am. Proc. 30:290-296. Reduction Soi 1 Sci. Armstrong, W. 1967. The oxidising activity of roots in waterlogged soils. Physiol. Plant. 20:920-926. Bai ley, L. D. 1976. Effects of temperature and root on denitrifi-cat ion ina so i 1 Can. J. So i 1 Sc i. 56: 79-87 Bailey, L. D., and E. G. Beauchamp. 1973. Effects of temperature on nitrate and nitrite reduction, nitrogenous gas production, and redox potential in a saturated soi 1. Can. J. Soil Sci. 53:213-218. Barr, A. 5., J. H. Goodnight, J. P. Sail, and J. T. Helwig. 1976. A user's guide to S.A.S. S.A.S. Institute, Raleigh, NC. Blackmer, A. M., and J. M. Bremner. 1978. Inhibitory effect of nitrate on reduction of nitrous oxide to nitrogen by soi 1 micro-organisms. Soi 1 Bioi. Biochem. 10: 187-191. Blackmer, A. M., and J. M. Bremner. 1979. Stimulatory effect of nitrate on reduction of nitrous oxide to nitrogen by soil micro-organisms. Soil Bioi. Biochem. 11:313-315. Bouldin, D. R. 1968. Hode1s for describing the diffusion of oxygen and other mobile constituents across the mud-water interface. J. Eco1. 56:77-87. Bouldin, D. R., R. L. Johnson, C. Burda, and C. Kao.' 1974. Losses of inorganic nitrogen from aqueous systems. J. Environ. Qual. 2:107-113. -68-

PAGE 82

Boyt, F. L., S. E. Bayley, and J. Zoltek. 1977. Removal of nutrients from treated municipal wastewater by wetland vegetation. J. Water Pollut. Control Fed. 48:789-799. Bremner, J. M. 1965. Inorganic forms of nitrogen. In C. A. Black (ed.), Methods of soil analysis. Agronomy American Society of Agronomy, Madison, WI. Bremner, J. M., and A. M. Blackmer. 1979. Effects of acetylene and soil water on emission of nitrous oxide from soil. In Agronomy abstracts. American Society of Agronomy, Madison, wl7 p 26. Bremner, J. M., and K. Shaw. 1958. 'Denitrification in soi 1. II. Factors effecting denitrification. J. Agric. Sci. 51 :40-52. Brezonik, P. L. Wat. Tech. 1977. Denitrification in natural water. 8:373-392. Prog. Broadbent, F. E., and F. E. Clark. 1965. Denitrification. In W. v. Bartholemew and F. E. Clark (eds.), Soil nitrogen. Agronomy 10:347-362. American Society of Agronomy, Madison, WI. Buresh, R. J., and W. H. Patrick. 1978. tJitrate reduction to ammonia in anaerobic soil. Soil Sci. Soc. Am. J. 42:913-918. Burford, J. R., and J. M. Bremner. denitrification capacities of and readily decomposable soil 7:389-394. 1975. Relationship between the soils and total, water, soluble, organic matter. Soil Biol. Biochem. Burford, J. R., and R. C. Stephanson. 1973. Measurement of gaseous loss of nitrogen from soils. Soil Biol. Biochem. 5:133-141. Chen, R. L., D. R. Keeney, J. C. Konrad, A. J. Holding, and D. A. Graetz. 1972. Gas production in sediments of Lake Mendota, Wisconsin. J. Environ. Qual. 1:155-157. Coleman, J. M. 1979. Present and past nutrient dynamics of a small pond in Southwest Florida. Ph.D. Thesis. Univ. of Fla. Cooper, G. S., and R. L. Smith. 1963. Sequence of products formed during denitrification in some diverse western soils. Soil Sci. Soc. Am. Proc. 27:659-662. Council for Agricultural Science and Technology (C.A.S.T.). 1976. Effect of increased nitrogen fixation on stratospheric ozone. Report #53. C.A.S.T. Craswell, E. T. 1978. Some factors influencing denitrification and nitrogen immobilization in a clay soil. Soil Biol. Biochem. 10:241-245. Day, P. R., H. E. Doner, and A. D. McLaren. 1978. Relations among microbial populations and rates of nitrification .and denitrification in a Hanford soil. In D. R. Nielsen and J. G. MacDonald (eds.), Nitrogen in the enVironment. V. 2: Soil plant-nitrogen relationships. Academic Press. New York. p. 305-363. -69,,'.'"

PAGE 83

Delwiche, C. C., and B. A. Bryan. 1976. Denitrification. Ann. Rev. Microbiol. 30:241-262. Denmead, O. T. 1979. Chamber systems for measuring nitrous oxide emission from soils in the field. Soil Sci. Soc. Am. J. 43: 89-95. Denmead, O. T., J. R. Freney, and J. R. Simpson. 1979a. Nitrous oxide emission during denitrification in a flooded field. Soil Sci. Soc. Am. J. 43:716-718. Denmead, O. T., J. R. Freney, and J. W. Simpson. 1979b. Stud i es of nitrous oxide emission in a grass sward. Soil Sci. Soc. Am. J. 43:726-]28. Doner, H. E., and A. D. McLaren. 1978. Temperature dependence of nitrate losses and denitrifier populations in soil columns. In W. E. Krumbein (ed.), Environmental biogeochemistry and geomicrobiology. V. 2: Geomicrobiology, biogeochemistry, and energy flow. Ann Arbor Science, Ann Arbor, HI. p. 573-582. Doner, H. E., M. G. Volz, and A. D. McLaren. 1974. Column studies of denitrification in soi Is. Soi 1 BioI. Biochem. 6:341-346. Ekpete, D. M., and A. H. Cornfield. 1965. Effect of pH and addition of organic matter on denitrification losses from soil. Nature. 208:1200. Engler, R. M., and W. H. Patrick. 1974. removal from floodwater overlying flooded soils and sediments. J. Environ. Qual. 3:409-415. Eno, C. F. 1960. Nitrate production in the field by incubating the soil in polyethylene bags. Soil Sci. Soc. Am. Proc. 24:277-279. EPA. 1973. Nitrification and Denitrification Facil ities:Wastewater Treatment. U. S. Environmental Protection Agency Tech. Transfer Publ ication. (EPA-625/4-73-994a). EPA. 1975. Process Design Manual for Nitrogen Control. U. S. Environmental Protection Agency. Erickson, N. L. 1978. Nitrification-denitrification in a freshwater marsh receiving secondarily treated sewage effluent. M.S. Thesis. Univ. of Fla. Fe t t e r, C. 'vi., 'vi. E. Sloe y, and F. L. Spa n g I e r 1 978 Use 0 f a natural marsh for wastewater pol ishing. J. Water Pollut. Control Fed. 50:290-307. Focht, D.O. 1974. The effect of temperature, pH, and aerat i on on the production of nitrous oxide and gaseous nitrogen--a zeroorder kinetic model. Soil Sci. 118:173-179. -70-

PAGE 84

Focht, D. D., and W. Verstraete. 1977. nitrification and denitrification. 124-204. Biochemical ecology of Adv. Microbiol. Eco. 1: Galsworthy, A. M., J. R. Burford, and D. J. Greenland. 1978. Model ing of rates of nitrification and denitrification in soil. In D. R. Nielsen and J. G. MacDonald (eds.), Nitrogen in the environment. V. 2: Soil-plant-nitrogen relationships. Academic Press, New York. p. 377-392. Gilliam, J. W., and R. P. Gambrell. 1978. Temperature and pH as 1 imitlng factors in loss of nitrate from saturated Atlantic Coastal Plain soils. J. Environ. Qual. 7:526-532. Gucker, F. T., and R. L. Seifert. 1966. Chemical kinetics. In Physical chemistry. W. W. Norton and Co., New York. p.6[l734. Guthrie, T. F., and J. M. Duxbury. 1978. and denitrification in organic soils. 42:908-912. Nitrogen mineralization Soil Sci. Soc. Am. J. Howeler, R. H. 1972. The oxygen status of lake sediments. J. Environ. Qual. 4:366-371. Howeler, R. H., and D. R. Bouldin. 1971. The diffusion and consumption of oxygen in submerged soils. Soil Sci. Soc. Am. Proc. 35:202-208. Kaspar, H. F., and J. M. Tledje. 1979. Nitrous oxide evolution from soil cores. In Agronomy abstracts. American Society of Ag ronomy, Mad i son, WI Keeney, D. R., and W. R. Gardner. 1970. The dynamics of nitrogen transformations in the soil. In S. F. Singer (ed.), Global Effects of Environmental Pollution. Springer-Verlag, New York. p.96-103. Lance, J. F., and T. D. Whis)er. 1972. Nitrogen balance in soil columns intermittently flooded with secondary sewage effluent. J. Env iron. Qua 1 2: 180-186. Liang, C. N., and M. A. Tabatabai. 1978. Effects of trace elements on nitrification in soils. J. Environ. Qual. 7:291-293. Matulewich, V. A., P. F. Strom, and t1. S. Finsteln. 1975. Length of incubation for enumerating nitrifying bacteria present in various environments. Appl. Microblol. 29:265-268. McGarity, J. W., and R. J. K. Meyers. 1968. Denltrifylng activity in solodized solonetz soils of Eastern Australia. Soil Sci. Soc. Am. Proc. 32:812-817. -71-

PAGE 85

Moraghan, J. F., and R. J. Buresh. 1977. Chemical reduction of nitrite and nitrous oxide by ferrous iron. Soil Sci. Soc. Am. J. 44:47-50. Nash, C. L., and J. Bollag. 1974. Comparative denitrification of selected microorganisms in a culture medium and inautoclaved soil. Appl. Microbiol. 27:674-677. Nelson, D. W., and J. M. Bremner. 1970. Role of soi I minerals and metallic cations in nitrate decompositon and chemodenitrificat Ion I n so i 1 s. So i 1 B i 01. B i ochem 2 : 1 8. Nommik, H. 1956. Investigations on denitrification in soi 1. Acta. Agr. Scand. IV. 2:195-228. Odum, H. T., and K. C. Ewel (eds.). 1977. 'Cypress wetlands for water management, recycl lng, and conservation. Fourth annual report to the Rockefeller Foundation and N.S.F. Center for Wetlands, Univ. of Fla. Patrick, W. H., Jr., and R. D. Delaune. 1976. Nitrogen and phosphorus utilization by Spartina alterniflora in a salt marsh in Barataria Bay, Louisiana. Estuarine and Coastal Marine Sci. 4:59-64. Patrick, W. H., and K. R. Reddy. 1976. Nitrification-denitrification reactions in flooded soils and water bottoms: Dependence on oxygen supply and ammonium diffusion. J. Environ. Qual. 5: 469-4]2. Patrick, W. H., Jr., and M. E. Tusneem. 1972. Nitrogen loss from flooded soil. Ecology. 53:735-737. Payne, W. J. 1973. Bacteriol. Rev. Reduction of nitrogenous oxides by microorganisms. 37:409-452. Phillips, R. E., K. R. Reddy, and W. H. Patrick. 1978. The role of nitrate diffusion in determining the order and rate of denitrification in flooded soi I: II. Theoretical analysis and interrelation. Soil Sci. Soc. Am. J. 42:272-278. Reddy, K. R. 1976. flooded so i 1 s Nitrification-denitrification reactions in Ph.D. Thesis. Louisiana State University. Reddy, K. R., and D. A. Graetz. 1980. Use of shallow reservoirs and flooded soil systems for agricultural drainage water treatment: I norgan i c nit rogen remova 1 J. Env iron. Qua I (i n review). Reddy, K. R., and W. H. Patrick. and anaerobic conditions on decomposition, and nitrogen Biochem. 7:87-94. 1975. Effect of alternate aerobic redox potential, organic matter loss in a flooded soi 1. Soi 1 BioI. -72-

PAGE 86

Reddy, K. R., W. H. Patrick, Jr., and R. E. Phill ips. 1976. Ammonium diffusion as a factor in nitrogen loss from flooded soils. Soil Sci. Soc. Am. J. 40:528-533. Reddy, K. R., W. H. Patrick, and R. E. Phillips. 1978. The role of nitrate diffusion in determining the order and rate of denitrification in flooded soil: I. Experimental results. Soil Sci. Soc. Am. J. 42:268-272. Reddy, K. R., P. D. Sacco, and D. A. Graetz. 1980. Nitrate reduction in a soil-water system. J. Environ. Qual. (in press). Requa, D. A., and E. D. Schroeder. 1973. Kinetics of packed bed denitrification. J. Water Pollut. Control Fed. 45:2696-1707. Rice, E. L., and S. K. Pancholy. by cl imax ecosystems. I I. of tannins. Amer. J. Bot. 1973. Inhibition of nitrification Additional evidence and possible role 60:691-702. Sherr, B. T., and W. J. Payne. 1978. Effect of Spartina alterniflora root-rhizome system on salt marsh soil denitrifying bacteria. Appl. Environ. Microbiol. 35:724-729. Small, M. M. 1978. Artificial wetlands as non-point source wastewater treatment systems. In M. A. Drew (ed.), Environmental qual ity through wetland utTTization. Coordinating Council on the Restoration of the Kissimmee River Valley and Taylor Creek Nubbin Slough Basin, Tallahassee, FL. p. 171-181. Sorensen, J. 1978. Capacity for denitrification and reduction of nitrate to ammonia in a coastal marine sediment. Environ. Microbiol. 35:301-325. Stanford, G., J. O. Legg, S. Dzienia, and E. C. Simpson. 1975. Denitrification and associated nitrogen transformations in soi Is. Soi 1 Sci. 120: 147-151. Stanford, G., R. A. Vander Pol, and S. Dzienia. 1975. Denitrification rates in relation to total and extractable soil carbon. Soil Sci. Soc. Am. Proc. 39:284-289. Starr, J. L., and J. Y. Parlange. 1975. Nonl inear denitrification kinetics with continuous flow in soil columns. Soil Sci. Soc. Am. Proc. 39:875-881. Stensel, H. D., R. C. Loehr, and A. W. Lawrence. 1973 .. Biological kinetics of suspended growth denitrification. J. Water Pollut. Control Fed. 45:249-261. Stefanson, R. C. Plant Soil. 1972. Denitrification in sealed soil-plant systems. 33: 113-149. Stotzky, G. 1973. Activity, ecology, and population dynamics of microorganisms in soil. C.R.C. Review of Microbiology. Chemical Rubber Co., Cleveland. p. 57-124. -73-

PAGE 87

Struble, J. E. tion rates effl uent. Fla. 1977. Nitrification, denitrification, and sedimentain cypress ponds receiving secondarily treated sewage M. S. Thesis. Department of Soil Science, Univ. of Tate, R. L., III. 1977. Nitrification in histosols: A potential role for the heterotrophic nltrifler. Appl. and Environ. Microbiol. 33:911-914. Terry, R. E., and R. L. Tate. 1980. The effect of nitrate on nitrous oxide reduction in organic soils and sediments. Soil Sci. Soc. Am. J. (in press). Tiedje, J. M., N. V. Caskey, M. s. Smith, and R. B. Firestone. 1979. Nitrous oxide production by bacteria that reduce nitrate to ammonia. Agronomy abstracts. American Society of Agronomy, Mad i son, WI. p. 165. Tinsdale, s. L., and W. L. Nelson. 1975. Soil and Fertilizer nitrogen. In Soil Fertility and Fertilizers. 3rd ed. MacMillan Pub. Co., Inc., New York. p. 122-188. Tusneem, M. E., and W. H. Patrick. 1971. Nitrogen transformations in waterlogged soils. Louisiana State Univ. Dept. of Agronomy. Bulletin #657. Van Kessel, J. F. 1977. Factors affecting denitrification rates in two water-sediment systems. Water Res. 11 :259-267. Van Kleemput, 0., W. H. Patrick, and R. C. Mcilhenny. 1975. Formation of chemical and biological denitrification products at controlled pH and redox potential. Soil BioI. Biochem. 7: 329-332. Van Kleemput, 0., W. H. Patrick, and R. C. Mcilhenny. 1976. Nitrate decomposition in flooded soils under different pH and redox conditions. Soil Sci. Soc. Am. J. 40:55-60. Verhoeven, W. 1956. Some remarks on nitrate and nitrate metabolism in microorganisms. In W. D. McElroy and B. Glass (eds.), A symposium on inorganiC nitrogen metabol ism. Johns Hopkins Press, Baltimore. Verstraete, W., and M. Alexander. 1973. Heterotrophic nitrification in samples of natural ecosystems. Environ. Sci. Tech. 7:39-42. Volz, M. G., M. s. Ardakani, R. K. Scholz, L. H. Stolzy. and A. D. McLaren. 1976. Soil nitrate loss during irrigation: Enhance ment by plant roots. Agron J. 68:621-627. White, J. P., D. P. Schwert, J. P. Ondrako, and L. L. Morgan. 1977. Factors affecting nitrification in situ in a heated stream. Appl. and Environ. Microbiol. 33:9llr-925. -74-

PAGE 88

Wijler, J., and C. C. Delwiche. 1954. ing process in soil. Plant Soil. Investigation on the denitrify-5:155-169. Yuan, T. L. 1974. A double buffer method for the determination of lime requi rement of acid soi Is. Soi I Sci. Soc. Am. Proc. 38: 437-440. Zoltek, J., S. E. Bayley, S. J. Hermann, L. R. Tortora, and T. J. Dolan. 1979. Removal of nutrients from treated municipal wastewater by freshwater marshes. Final report to city of Clermont, Florida. Center for Univ. of Fla. -75-