Citation
Nitrogen Dynamics in a Constructed Wetland Receiving Plant Nursery Runoff in Southeastern United States

Material Information

Title:
Nitrogen Dynamics in a Constructed Wetland Receiving Plant Nursery Runoff in Southeastern United States
Creator:
Govindarajan, Bhuvaneswari
Place of Publication:
[Gainesville, Fla.]
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (105 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Interdisciplinary Ecology
Committee Chair:
Sharma, Jyotsna
Committee Members:
Clark, Mark W.
Reddy, Konda R.
Sharma, Kanika

Subjects

Subjects / Keywords:
Constructed wetlands ( jstor )
Nitrates ( jstor )
Nitrogen ( jstor )
Nutrients ( jstor )
Oxygen ( jstor )
Rhizosphere ( jstor )
Surface runoff ( jstor )
Surface water ( jstor )
Water depth ( jstor )
Wetlands ( jstor )
Interdisciplinary Ecology -- Dissertations, Academic -- UF
constructed, denitrification, nitrogen, nursery, runoff, wetlands
Genre:
Electronic Thesis or Dissertation
bibliography ( marcgt )
theses ( marcgt )
Interdisciplinary Ecology thesis, M.S.

Notes

Abstract:
Constructed wetlands are a cost effective method for on-site nutrient removal from polluted water from nonpoint sources. Nitrogen (N) concentration and denitrification potential were studied in a constructed wetland receiving plant nursery runoff over five months in southeastern US. Nursery runoff, before entering the wetland, contained an average total nitrogen concentration of 34 mg L-1. Majority of nitrogen was in the form of nitrate nitrogen (NO3--N; 30 mg L-1) followed by ammonium nitrogen (NH4+-N; 2 mg L-1) and organic nitrogen (2 mg L-1). Total phosphorus concentration was 1 mg L-1 and dissolved organic carbon was 11 mg L-1. Mean nutrient removal efficiency of the constructed wetland was 40% for TN, 40% for NO3--N, 59% for NH4+-N, and 16% for TP. Nitrate nitrogen removal efficiency was inversely related to the loading rate of NO3--N; and directly related to temperature of the water column. The gradual increase in temperature over months might have enhanced NO3- - N removal by influencing plant growth and activity of microorganisms, which is also evidenced by the gradual decrease in the concentration of DO in the water column during the study period. The theoretical hydraulic residence time was estimated to be 3-5 days in the model. The nutrients (N and P) analyzed in this study exhibited nutrient gradient from the inflow to outflow indicating removal of these nutrients within the wetland. The denitrification potential of the water column was relatively low (0.01 to 0.03 mg N2O-N L-1 hr-1), and could be due to carbon limitation in the wetland water column as evident by low DOC concentration and low DOC: NO3--N ratio. The presence of few attachment sites for bacteria and lack of particulate substances to support the microbial activity possibly resulted in low denitrification rates in the water column. The mean denitrification potential in the rhizosphere soil of Typha latifolia (3 mg N2O-N kg -1 hr-1) and Canna flaccida (4 mg N2O-N kg -1 hr-1) over five months were statistically similar. As the combined denitrification potential of the water column (0.72 mg N2O-N kg-1 day-1) and the rhizosphere soil (96 mg N2O-N kg-1 day-1) in this study was less than the total removal (1.75 g m-2 day-1), it is likely that other NO3--N processes are contributing to the observed NO3--N removal in this constructed wetland. Plant uptake and microbial denitrification in the sediments and/or soil are considered as major NO3--N removal mechanisms in constructed wetlands. Hence, it is concluded that rhizosphere denitrification is significantly contributing to NO3- -N removal in the constructed wetland receiving plant nursery runoff. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2008.
Local:
Adviser: Sharma, Jyotsna.
Statement of Responsibility:
by Bhuvaneswari Govindarajan

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright by Bhuvaneswari Govindarajan. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/11/2008
Classification:
LD1780 2008 ( lcc )

Downloads

This item has the following downloads:


Full Text





areas and forestry operations are the primary reason for the lack of improvement in controlling

nonpoint source pollution.

Water Quality in the United States

At the beginning of the twenty-first century, nonpoint source pollution stands as the

primary cause of water quality impairment within the United States. According to the USEPA

(2002a), nonpoint source pollution is the main reason that approximately 39% of surveyed rivers,

45% of lakes and 51% of estuaries are not clean enough to meet basic uses such as fishing or

swimming. The report from National Coastal Condition (USEPA, 2002b) proj ected that the 70%

of the US estuary conditions would worsen by 2020 due to eutrophication caused by nonpoint

source pollutants. The National Water Quality Assessment (NAWQA) program [United States

Geological Survey (USGS), 1998] of the US assessed the quality of water resources, especially

streams, river basins, groundwater, and aquifer systems in the Unites States from 1991-2001 and

reported that the contamination of streams and groundwater is widespread in agricultural and

urban areas due to nonpoint source pollutants.

States and other jurisdictions reported in the National Water Quality Inventory (NWQI,

1998) that agriculture and urban runoff are among the leading contributors to deteriorate water

quality nationwide. The most common nonpoint source pollutants causing water quality

impairment include nutrients [nitrogen (N) and phosphorus (P)], pesticides, chemicals, siltation

(soil particles), metals, and pathogens (bacteria and viruses). According to Faeth (2000),

agriculture is the largest source of pollution in the US degrading the quality of surface waters

such as rivers and lakes, with croplands alone accounting for nearly 40% of the N and 30% of

the P pollution. A maj or nonpoint source pollutant from these activities is an excess of nutrients,

which can occur through applications of crop fertilizers. For example, in agricultural systems,


























F~s;a~s;~i~~


Figure 2-2. Runoff generated from the microirrigated nursery beds at Monrovia Nursery









V = EAh, in which
V wetland water volume in m3;
E wetland porosity in m3/ m3;
A = wetland area in m2;
h = mean water depth in m.

Wetland porosity is a fraction of total wetland volume available through which water can

flow (USEPA, 2000). It is the amount of wetland water volume not occupied by plants and

expressed as a decimal (NRCS, 2002). Therefore, wetland porosity and plant density are

inversely related; the lower the wetland porosity values, higher the plant density. As it is

difficult to accurately measure wetland porosity in the Hield, highly variable porosity values were

reported in the literature. For example, Reed et al. (1995) reported values ranging from 0.65 to

0.75 for fully vegetated wetlands, and for dense to less mature wetlands, respectively; Kadlec

and Knight (1996) reported average wetland porosity values ranging from 0.95 to 1.0.

According to USEPA (2000), it was suggested to use the wetland porosity value of 0.65 to 0.75

for fully vegetated wetlands, while considering design of constructed wetlands.

Denitrification potential (Water). Denitrifieation potential in water samples was

determined by measuring the denitrification enzyme activity (DEA) within a week of sample

collection (Tiedje, 1982; White and Reddy, 1999). Twenty ml of water sample (unfiltered) was

taken into either 120 ml or 160 ml glass serum bottle. The bottles were capped and crimped

ensuring proper seal. The bottles were purged with nitrogen gas (N2) for approximately 5

minutes to achieve anaerobic conditions. The initial pressure of less than 20 psi was maintained.

High grade acetylene (C2H2) gaS was generated by adding N2 purged water to a separate bottle

filled with Calcium Carbide (CaC2) TOcks, which was already capped, crimped and subj ected to

N2 purging to remove any oxygen. Nine ml of acetylene gas was inj ected into each sample bottle

from which 9 ml headspace was removed before adding acetylene. Samples were then kept in a










LIST OF TABLES


Table page

3-1 Mean load-in, load-out, removal rate (g m-2 day l) of nutrients and percentage
nutrient removal efficiency of the constructed wetland over five months (April and
August 2007). ............ ...............61.....

3-2 Mean concentrations of nutrients (mg L 1) as affected by location of sample
collection, depth within location, and months (April and August 2007) in the
constructed wetland receiving nursery runoff (n=18). ............ ...... ............... 6

3-3 Physiochemical parameters (temperature, dissolved oxygen, pH) as affected by
location of sample collection, depth within location, and months (April and August
2007) in the constructed wetland receiving nursery runoff (n=3). ........... ...................63

3-4 Denitrification potential (mg N20-N kg-l hr- ) as affected by macrophyte rhizosphere
soil in three sections (east, middle, and west) between May 2007 and August 2007 in
the constructed wetland receiving nursery runoff. (n =3) ................. ........... ...........64

4-1 Nitrate removal rates in constructed wetlands, under different hydraulic residence
periods, as reported by other researchers. .............. ...............89....

4-2 Denitrification potentials of different substrate reported by other researchers in
constructed wetlands. ............ ...............90.....





5 0.75:


r


1500


1754C


O


I 0.95Rnudly


1.Rundl~y


Figure 3-10. Theoretical hydraulic residence time (in days) as affected by hydraulic loading rate (m3 day l) and three estimates of
wetland porosity values (0.75, 0.95 and 1.0) in the study cell of a constructed wetland receiving plant nursery runoff









Table 3-2. Mean concentrations of nutrients (mg L 1) as affected by location of sample collection, depth within location, and months
(April and August 2007) in the constructed wetland receiving nursery runoff(n=18). Capital letters indicate statistical
difference between depths (top and bottom) at inlet and outlet over five months; lower-case letters indicate statistical
difference between locations (inlet and outlet) at top and bottom depths over five months; asterisks(*, **) indicate
statistical difference between locations at top and bottom for every month. Means followed by the same letter and same
number of asterisks are statistically similar.

Location Depth Month NO3--N mg L NH4--N mg L TP mg L^1 DOC mg L1
Inlet Top April 46 Aa 4.0 Aa 1.8 Aa 11 Aa *
May 42 B a 0.2 BC a 2.0 Aa 13 Ba *
June 8 D a 2.3 AB a 1.4 Aa 10 Aa *
July 15 Ca 0.2 B a 1.3 Aa 11 Aa *
Aug 11 CD a 1.3 ABC a 0.9 A a 11 A a *
Bottom April 42 Aa 2.2 AB a 2.1 A a 11 AB a *
May 36 B b 0.2 B a 2.4 A a 12 AB a *
June 7 D a 2.0 AB a 3.5 Ab 10 Ba *
July 11 C a 0.9 BC a 2.3 A a 11 AB a
Aug 7 CD b 3.6 Aa 3.0 Ab 13 A a *
Outlet Top April 43 Aa 1.0 Aa 1.6 Aa 11 Aa *
May 32 Ba ** 0.1 Aa 1.2 Aa 11 Aa *
June 4 Ca ** 1.7 Aa 0.9 Aa 9Aa *
July 7 Ca ** 0.1 Aa 1.2 Aa 9Aa *
Aug 4 Ca ** 0.1 Aa 0.7 Aa 10 Aa *
Bottom April 42 Aa 2.3 Aa 1.5 Aa ** 10 Aa *
May 26 Ba ** 0.1 Aa 1.3 Aa 11 Aa *
June 4 Ca 1.8 Aa 0.9 Aa 10 Aa *
July 5 Ca ** 0.1 Aa 1.5 Aa ** 9Aa *
Aug 3 Cb ** 0.1 Aa ** 1.0 Aa 10 Aa **









concentration/low DOC: NO3- -N, which limits denitrification. The presence of few attachment

sites for bacteria possibly resulted in low denitrification rates in the water column.

Mean rhizosphere denitrification potential over five months for Can2na flaccida and Typha

latifolia was 3 mg and 4 mg N20-N kg-l hr- respectively. The mean denitrification potential of

two species was statistically similar. Regardless, the denitrification values obtained in this study

are relatively high when compared to several other study studies. Hunt et al. (2003) obtained

higher denitrification potential of 0.21 mg N kg-l hr- in Typha dominated wetland soils, as

compared to 0.52 mg N kg-l hr- in Juncus dominated wetlands used for treatment of swine

wastewater. Rhizosphere denitrif ication potential values of Canna flaccida~ and Typha latifolia in

this study were comparatively higher than the rhizosphere denitrification potential values of

other plant species Lolium perenne, Zostera marina, Potamoge~~~~tttt~~~~tttto n pectinatus, Lobelia

dortmanna and Littorella uniflora. Results from this study show that denitrification rates

comparable to or higher than those reported from bulk soil of vegetated plots can be obtained in

rhizosphere soil closely adhered to the roots.

Nitrate nitrogen removal efficiency of the constructed wetland receiving plant nursery

runoff was observed be 40% over five month period, with an average removal rate of 1.75 g m-2

day- The mean loading rate of NO3- -N was 4.43 g m-2 day l. Compared to mean removal rate

of NO3- -N (1.75 g m-2 day )~, the amount of NO3- -N removed through coupled (0. 1 g N20-N kg-

Sday- ) denitrification potential [mean denitrification potential of water column (0.72 mg N20-N

kg-l day- ) plus mean rhizosphere denitrification potential (96 mg N20-N kg-l day- )] was lower.

Therefore, it is likely that the rest of NO3- -N would be removed by other pathways contributing

to removal of NO3- -N. The maj or NO3- -N removal pathways are plant and microbial









sedimentation is a significant problem in Florida, where approximately one-third of landforms

are wetlands, and drinking water is mostly drawn from underground aquifers which are close to

the land surface (Marella and Fanning, 1996).

The generally temperate climate in Georgia is influenced by the proximity to the Atlantic

Ocean and to the Gulf of Mexico. The coastal plains, underlain by sand and limestone, occupy

over 60% of the state. Because of the sandy soil and flat topography in the coastal areas,

infiltration is a larger concern rather than runoff. Nearly one third of the state is underlain by

sandy/clay loam soils, which support variety of agricultural crops. The state typically receives

100-125 cm of rainfall every year (www.georgia. org; www. coastgis.marsci .uga. edu, October

2007).

Florida is the second leading producer of ornamental plants in the US. The greenhouse and

nursery industry comprises 50,992 hectares of production area in Florida (Hodges and Haydu,

2002). According to the United States Department of Agriculture (USDA) estimates, the value

of nursery and greenhouse crops in Florida was approximately $1.63 billion in 2004 (Jerardo,

2005). Georgia is among the top ten producers of floriculture crops, ranking tenth in the US.

The value of nursery and greenhouse crops in Georgia was approximately $400 million in 2006

[National Agricultural Statistics Service (NASS), http://www.nass.usda.gov/, accessed on

October, 2007].

While greenhouse crop production generates significant runoff, this problem is most

noticeable in the container nursery industry where large numbers of containerized plants are

grown outdoors. Generally, for production of containerized plants outdoors, plants are grown in

containers placed on gravel beds which are lined with plastic to prevent weed growth, and

irrigation is achieved through overhead sprinklers (USDA, 1998a; Lea-cox and Ross, 2001). The









such as Watershed Restoration Act and Total Maximum Daily Load (TMDLs), imposed by the

environmental protection agencies against contamination of water resources necessitate

recycling/treatment of nursery runoff before discharging into nearby waterbodies (Headley et al.,

2001; Lea-cox et al., 2002; Huett et al., 2005).

Management Strategies to Control Nonpoint Source Pollution

The control of nonpoint source pollution centers on adoption of Best Management

Practices (BMPs) to manage land and to control the release of pollutants into the atmosphere;

further, establishment of threshold levels of contaminants (e.g., TMDLs) is also a strategy for

limiting nonpoint source pollution (Carpenter et al., 1998).

Best Management Practices can be grouped into two categories; structural and

nonstructural (Horner et al., 1994). Non structural BMPs include practices such as preservation

of natural areas and drainage systems, land use methods, and efficient use of fertilizers and

irrigation water. Structural BMPs include establishment of detention/retention and recycling

ponds, grass swales, filter strips, riparian buffer strips, floodplains, and wetlands. The

commonly used structural BMPs are retention and recycling ponds, riparian buffers strips, and

grass swales to control nonpoint source pollution from agricultural operations, including plant

nurseries. In most of these methods, the primary mechanisms to remove pollutants are to

enhance settling of the particulates and to enhance infiltration into the subsurface zones. Many

container plant nurseries, where plants are grown outdoors on gravel beds lined with plastic,

construct recycling ponds to recycle the runoff from nursery beds (Lu and Sibley, 2006). The

other commonly followed BMPs to control the runoff water from container plant nurseries are

adjusting the time and rate of fertilizer application to coincide with the crop needs, using micro-

irrigation practices to minimize runoff volume, and use of controlled release fertilizers to reduce

NO3- N leaching from containers.









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Master of Science

NITROGEN DYNAMICS IN A CONSTRUCTED WETLAND RECEIVING PLANT
NURSERY RUNOFF IN SOUTHEASTERN UNITED STATES

By

Bhuvaneswari Govindarajan

May 2008
Chair: Jyotsna Sharma
Major: Interdisciplinary Ecology

Constructed wetlands are a cost effective method for on-site nutrient removal from

polluted water from nonpoint sources. Nitrogen (N) concentration and denitrification potential

were studied in a constructed wetland receiving plant nursery runoff over five months in

southeastern US. Nursery runoff, before entering the wetland, contained an average total

nitrogen concentration of 34 mg L^1. Maj ority of nitrogen was in the form of nitrate nitrogen

(NO3--N; 30 mg L^1) followed by ammonium nitrogen (NH4 -N; 2 mg L^1) and organic nitrogen

(2 mg L^)~. Total phosphorus concentration was 1 mg L^1 and dissolved organic carbon was 11

mg L^1. Mean nutrient removal efficiency of the constructed wetland was 40% for TN, 40% for

NO3--N, 59% for NH4 -N, and 16% for TP. Nitrate nitrogen removal efficiency was inversely

related to the loading rate of NO3--N; and directly related to temperature of the water column.

The gradual increase in temperature over months might have enhanced NO3- N removal by

influencing plant growth and activity of microorganisms, which is also evidenced by the gradual

decrease in the concentration of DO in the water column during the study period. The theoretical

hydraulic residence time was estimated to be 3-5 days in the model. The nutrients (N and P)

analyzed in this study exhibited nutrient gradient from the inflow to outflow indicating removal

of these nutrients within the wetland. The denitrification potential of the water column was









in a subsurface flow wetland receiving plant nursery runoff (Huett et al., 2005). The absence of

attachment sites for bacteria and/or low organic carbon availability possibly resulted in low

denitrification rates in the water, as reported by Toet et al. (2003). The distribution of

denitrifying microbes was most likely regulated by the availability of organic material, with

higher denitrification rates in the sediments than on surface in the water column. The DO

concentration in the water column of the study cell ranged from 6 mg L-1 to 12 mg L^1, which

might have affected the denitrification process, as well. Poe et al. (2003) reported that the

denitrification potential was influenced by the concentration ofNO3- -N, and the rates of

denitrification were observed to increase with an increase ofNO3- N. Such trend was not

observed in this study.

Denitrification potential of macrophyte rhizosphere soil. Denitrification potential of

macrophyte rhizosphere soil was influenced by months; however, there was no trend between

April and August 2007. The mean denitrification potential in the rhizosphere of the two species

combined was 7.6, 1.7, 0.9, and 3.0 mg N20-N kg-l hr- in May, June, July, and August 2007,

respectively. Only the values obtained in May significantly differed from other months.

Although there was a gradual increase in temperature of the water column from May (22"C) to

August (30"C), the rhizosphere denitrification potential did not show significant difference over

months. The mean temperature (26"C) of the wetland during the study period was close to the

optimum range, 20-25"C, for denitrification activity.

Results of this study showed that mean denitrification potential over five months for

Canna flaccida and Typha latifolia was 3 mg and 4 mg N20-N kg-l hr- respectively. Data

variability within replicates was observed for denitrification potential values in this study. The

variation in values obtained for replicates in this study might be due to the heterogeneity in the









sampling locations (Fig. 2-9). On the north side (the inflow side), one location was selected

within each of the three sections of the primary cell from where the samples were collected. This

sampling point was located approximately 3 m inward from the northern edge of the cell and

approximately 3 m inward from the eastern edge. Two sets of three replicate water samples were

collected at the surface and at 0.5 m depth of the water column at each of the three locations at

monthly intervals between April 2007 and August 2007. One set of samples (three replications)

was collected in 125 ml plastic bottles and analyzed for TKN, NH4 -N, NO3--N, TP, and DOC

and the second set of samples (three replicates) was collected in 25 ml vials (Fisher Scientific,

Pittsburgh, PA) and used to determine denitrification enzyme activity (DEA), to estimate

denitrification potential in the samples. The surface water samples were collected as 'grab'

samples in the undisturbed water column. To collect samples from the bottom of the water

column, the sampling bottle was turned upside down and forced through the water column by

hand. After reaching the bottom of the water column, the bottle was turned slowly to fill it and

then rapidly brought straight up to the surface.

On the south side (outflow side) of the study cell, sampling points were selected

approximately 2 m inward from the location of outflow pipe, to ensure minimum disturbance in

the water column. Owing to its depth, the surface and bottom of the water column samples at the

outflow side were collected by using a swing sampler. For the bottom of the water column

samples, the bottle tied to the swing sampler was sent upside down through the water column.

After reaching the bottom of the water column, the pole was turned slowly to fill the bottle and

then rapidly brought straight up to the surface. Two sets of three replicate water samples were

collected at the surface and bottom of the water column at each of the three locations at monthly

intervals between April 2007 and August 2007. One set of samples (three replications) was









effectiveness of three floating and six emergent aquatic plants in improving domestic wastewater

quality based on their oxygen release capacities. Plants such as H. umbellata transported oxygen

2.5 times more rapidly than E. cra~ssipes, which transports oxygen four times more rapidly than

P. stratiotes. Radial oxygen loss (ROL), which is influenced by the external oxygen demand,

was higher in Juncus e~ffitsus L. (common bulrush) (9.5 A 1 x 10- mol 02 h-l root- ) than in I.

inflexus L. (inland bulrush) (4.5 & 0.5 x 10-7 mol 02 h-l root- ) (Sorrell, 1999). Higher rates of

(6.8 mg N m-2 d- ) denitrification were obtained in the wetland sediments planted with Juncus

spp as compared to the unplanted plots (Smialek et al., 2006). This indicates the importance of

vegetation in the constructed wetlands to facilitate denitrification.

Besides oxygen, plant roots also exude organic compounds, which serve as a carbon source

for denitrifying microorganisms. The influence of temperature on plant metabolism in turn

affects the concentration of root exudates. The magnitude of organic compound release, which

enhances the NO3- N removal in constructed wetlands, ranged from 5-25% of the

photosynthetically fixed carbon (Platzer, 1996). Bachand and Horne (2000) observed that the

different vegetation types resulted in significantly different denitrification rates (T.1atifolia 565

mg N m-2 day- J. e~ffu~sus 261 mg N m-2 day l, and mixed vegetation 835 mg N m-2 day )~.

Productivity, physical structure, C:Nlitter ratio, and plant fiber content have been observed to

differ between plants. Hence they concluded that these factors affected the organic carbon

availability of plants, thereby resulting in different denitrification rates. In organic carbon

limited, free surface wetlands, it was recommended that a mixture of labile (submergent and

floating) and more recalcitrant (emergent and grasses) vegetation be planted to enhance

denitrification rates (Bachand and Horne, 2000).









mg L^1) TP concentration was lower than the TP in the bottom (3 mg L^1) of the water column.

In contrast, the concentrations of TP at the two depths were similar at outflow. The other two-

way interactions between location~month, and depth month did not show any particular trend

in the concentration of TP during the period of this study.

Dissolved Organic Carbon (DOC). The concentration of DOC was affected (p=0.0008)

by an interaction among location, depth, and month (Table 3-2). At inflow and outflow, the

surface and lower depth of the water column did not show difference in the concentration of

DOC over the Hyve months of this study. There was no trend observed for the DOC

concentrations during this study for both locations and depth within location. Of all the two-way

interactions only location* depth showed significant differences between treatment means. The

surface of the water column at inflow had higher (11 mg L^1) concentration of DOC as compared

to surface (10 mg L^1) of the water column at outflow. Similarly the DOC concentration was

higher (12 mg L^1) at inflow and lower at outflow (10 mg L 1) for bottom of the water column.

The Hyve month mean concentration of DOC was similar at two depths both at inflow and

outflow of the wetland.

DOC: NO3--N ratio. The DOC: NO3--N ratio was influenced (p=0.0009) by a three-way

interaction among location, depth, and months. At inflow, the DOC: NO3--N ratio at the lower

depth of the water column was higher as compared to the surface water column between April to

August. Similar trend of higher DOC: NO3--N ratio at the lower depth was observed at the

outflow of the study cell, except for April. However, there was no temporal pattern observed

during this period both at inflow and outflow of the study cell. All two-way interactions,

depth~month, location~month, and location~depth, significantly influenced the ratio at a=0.05

level. The mean DOC: NO3--N ratio over Hyve months at the inflow and outflow was 0.9:1 and
















storm water


ceR~ia


11m
sistle Hohi pm

+ FC amle
+Inlowsamle
+ uflwsape

Toalsmpe =2
Netscl

Fiur 2-.Lcto n ubr fwtrsmls(niatdb tr)a nlw andotlwo h td elt eemn h
nurin removal effienc oftecosrctdwtlns









way that it captures half an inch of rainfall and the excess rainfall is diverted either to the

irrigation ponds for recycling or actively pumped into the wetlands.

A surface flow wetland was constructed at the nursery in 1997 to treat runoff effluents

before releasing the excess water into a natural creek adj acent to the nursery property. This

constructed wetland system is utilized to treat stormwater and runoff from three plant production

areas (catchment area) totaling 48.5 hectares. The wetland was constructed by creating pits in a

naturally depressed area of 3.8 hectares. A 12.5 cm think layer of bark chips was placed at the

bottom of the pits to support the planting and also to serve as a source of carbon for the microbial

activity in the constructed system. The wetland was constructed as several cells such as primary,

secondary, and test cells (Fig. 2-5). The primary cells were designated as 1A and 2A; the

secondary cells were designated as IB and 2B; and the test cells were designated as 3,4,5,6, and

7. For this entire study, only one primary cell (lA) was used. There exists a difference in the

size and depth of cells, and in the distribution and abundance of macrophytes between the

primary and secondary cells. The area of the primary cells 1A and 2A is 1.17 hectares and 0.57

hectares respectively. The area of the secondary cells IB and 2B is 0.69 hectares and 0.65

hectares respectively. The primary cells are 0.75 m deep and the secondary cells are 0.3m deep.

The depth at the outflow (South) side of the primary cells is comparatively higher (1.2 m) than

the inflow (North) side, which is 0.75 m deep. Each primary and secondary cell was divided into

three sections by creating two earthen berms which extend to approximately %/th of the width

(north-south) of the cells. These earthen berms were created to maintain a linear water

movement. The difference in vegetation of the primary and secondary cells is attributed to

variation in planting selection and natural growth of plants. The macrophytes originally planted

in the deeper primary cells were Can2na flaccida Salisb. (Canna Lily), Scirpus validus Vahl.





























Fiur 25.Ovrie o te ontucedweladsa Monrovi Nusry aro A A naeria img (to rih)fpato

Monovi NrlseyinldnthtramnwelnsinCioGAB)Ashmtcoeveoftecntcedelns
(bottm lef) isalso rese ted h rosidct h ieto fwte oeettruhtewtad.1 n A
andIB nd B rpreen th prmar an seonaryclsepciey









Denitrification occurs in the anoxic water column of aquatic ecosystems. Many

researchers observed lower significant denitrification potential activity in the water column as

compared to the sediments in different aquatic environments. The denitrification rates of the

water samples in the wetland receiving sewage treatment plant effluent were observed to be 0.06

- 0.96 Clg N L-1 hr- (Toet et al., 2003). Bastviken et al. (2003) obtained higher denitrification

rates (6.5-7.5 kg N ha-l day- ) in the sediment than on surface of the water column.

Aquatic plants serve as a maj or source of organic carbon to the microorganisms and

provide large surface area (commonly referred as "biofilms") for nitrifying and denitrifying

bacteria. Plants also provide oxygen, which vary from species to species, to their root zones

(Stoottmeister et al., 2003). Diffusion of oxygen to root zones is reported to influence

biogeochemical cycles within the rhizosphere. But it was shown that the amount of oxygen

being released by the plants around the roots is limited (Brix, 1994; Wood, 1995). The limited

release of oxygen around roots ensures that anaerobic conditions will predominate unless the

organic load to the wetland is low and wetland is shallow, as the amount of oxygen decreases

with increasing depth (Ayaz and Acka, 2000).The limited release of oxygen through roots,

however, also provides the required oxygen for aerobic nitrifying microbes for the oxidation of

NH4 -N to NO3--N/nitrite nitrogen (NO2--N), which in turn provides nitrate/nitrite for the

anaerobic denitrifying microbes for denitrification (Risgaard-Petersen and Jenson, 1997). The

oxygen release rates of floating and emergent aquatic plants such as Pistia stratiotes L. (water

lettuce), Eichhornia cra~ssipes (Mart.) Solms (water hyacinth), Hydrocotyle umbellata

L.(pennywort), Pontederia spp.L. (pickerel weed), Phragmites spp. (Reed) and Typha latifolia L.

(cattail) were studied by Moorhead and Reddy (1988); the oxygen release rate in Phragmites

spp. was estimated to be from 0.02 g m-2 dayl to 12 g m-2 day- Reddy et al. (1989) studied the




































Figure 2-1. Map of the state of Georgia, United States, showing county boundaries. Grady
County, in which Monrovia nursery is located, is indicated by a callout box. Source:
http ://www.epodunk.com/cgi-bin/genInfo.php?locne=98 October 2007.






















Org~anic N
mNH4-N

mNO3-N


~~U ~~ -~ "
ct = a,,=
"I:~C~~C~


Oufo


Flo~w Control Chlannlel


Figure 3-1. Mean concentration (mg L^1) of total nitrogen (TN) and composition of TN [nitrate (NO3~ N), ammonium (NH4 -N) and
organic nitrogen concentration, mg L- ] between April to August of 2007 in the water samples obtained from the flow
control channel (n=3), inlet (n=9), and outlet (n=9) pipes of a constructed wetland receiving nursery runoff.


Inflow r









(June) to 12 mg L^1 (August), and at outflow it ranged from 10 mg L^1 (July) to 11 mg L^1 (May)

Most of the means were statistically similar, although the overall model showed significant

interaction among locations and months. The Hyve month average concentrations of DOC at FCC

and inflow of the study cell were similar. The mean concentration of DOC present in the inflow

water was 1 1 mg L^1 which was higher (p<0.001) than the concentration of DOC leaving the

system, averaging 10 mg L-1 during the study period.

TN: TP ratio. The computed TN: TP ratio was affected (p<0.0001) by an interaction

among location of sample collection and months during which samples were collected. At

inflow, the ratio was highest in the month of April (36: 1) and lowest in July (17:1i). The ratio at

outflow was also highest in April (35:1), but it was lowest in August (5:1). At FCC, the ratio

ranged from 12: 1(July) to 33:1(August). The Hyve month mean TN: TP ratio, at inflow and

outflow, was 25:1 and 18:1, respectively.

DOC: NO3--N ratio. The computed DOC: NO3--N ratio was affected (p<0.0001) by an

interaction among location of sample collection and months during which samples were

collected. At both inflow and outflow, there was gradual increase in the DOC: NO3--N ratio over

Hyve months. At inflow, the ratio gradually increased from April (0.2:1) to August (0.9: 1);

similarly the ratio at outflow also gradually increased from April (0.2:1) to August (3.3:1). At

FCC, the ratio ranged from 0.3:1(April and May) to 0.8:1(July), with values not showing any

trend over the period of the study. The Hyve month mean DOC: NO3--N ratios at inflow and

outflow were 0.5:1 and 1.3:1, respectively.

Relationship between nutrient loading and removal rates. The correlation between

loading and removal rate (g m-2 day l) of NO3- N, NH4 -N and TP are presented in Figure 3-7 to

3-9. The Pearson' s correlation coefficient (r) value of 0. 19 (p<0.21) was obtained for the









means. The two-way interaction among month~section and plant~month occurred, however most

of the means were statistically similar. During May 2007, the denitrification potential of T.

latifolia was 12 mg N20-N kg-l hr- Similarly the denitrification potential in the western section

(17 mg N20-N kg-l hr- ) was found to be significantly different from rest of the treatments during

the month of May 2007. Denitrification potential was influenced (p = 0.01) by the main effect

of month; however, there was no trend between April and August 2007. The mean

denitrification potential was 8, 2, 1, 3 mg N20-N kg-l hr- during May, June, July, and August

2007, respectively. Only the denitrification potential during May differed from other months.

The five month mean denitrification potential for Can2na flaccida and Typha latifolia were 3 mg

and 4 mg N20-N kg-l hr- respectively.









The removal efficiency of the NO3--N during the study period ranged from 12 -77% over

five months, indicated the temporal variation in effectively removing NO3- N. The highest

NO3- N removal was obtained in the month of August 2007 and the lowest removal efficiency

was obtained during April. During the same period, the concentration of NO3- N entering and

leaving the wetland gradually decreased from April to August. There was an inverse relationship

between the concentration of NO3- N at inflow and the removal efficiency of the wetland, i.e.,

when the concentration of NO3- N entering the wetland was higher, the removal efficiency was

lower during April. Similar relationship was observed between loading rates and removal

efficiency of NO3- N in the constructed wetlands. Gradual increase in removal efficiency of

NO3- N from April (12%) to August (77%) was observed, as the inflow concentration gradually

decreased from 47 mg L- to 13 mg L^1. Lin et al. (2002) observed similar decrease in the

efficiency of nitrate removal but the rate of removal increased as the nitrate loading rate was

increased. Darbi et al. (2003) also reported inverse relationship between loading rate and the

removal efficiency of NO3--N. Reduction in the removal efficiency of NO3- N from 2.46 to

1.64 kg NO3 -N m-3 dayl was obtained, as the inflow NO3--N concentration increased from 175

to700 mg L^ (Park et al., 2002). Most of the studies reported high loads of NO3--N as compared

to the load in the present study. As the loading rate of NO3--N was observed to be less than 7 g

m-2 day l, this study provides a useful insight into nutrient removal at lower loading rate. Similar

results of low NO3--N loading rates were observed (<2 g m-2 day- ) by Headly et al. (2001) and

Huett et al. (2005) in subsurface flow reed beds receiving plant nursery runoff in Australia.

Although the loading rates are comparatively low, the removal rate of NO3- N (1.75 g m-2 day l)

in this study is comparable and in many cases higher as compared to the values of other studies

(Table 4.1).










Seitzinger, S.P. 1993. Denitrification measurements in aquatic sediments: A comparison of three
methods. Biogeochemistry 23:147-167.

Shipin, O., T. Koottatep, N.T.T. Khanh and C. Polprasert. 2005. Integrated natural treatment
systems for developing communities: low-tech N-removal through the fluctuating
microbial pathways. Water Science and Technology 51:299-306.

Sirivedhin,T., and K.A. Gray 2006. Factors affecting denitrification rates in experimental
wetlands: Field and laboratory studies. Ecological Engineering 26:167-181.

Smialek, J., V. Bouchard, B. Lippmann, M. Quigley, T. Granata, J. Martin, and L. Brown. 2006.
Effect of a woody (Salix nigra) and an herbaceous (Juncus effuses) macrophyte species
on methane dynamics and denitrification. Wetlands 26:509-517.

Sorrell, B.K. 1999. Effect of external oxygen demand on radial oxygen loss by Juncus roots in
titanium citrate solutions. Plant, Cell & Environment 22: 1587-1593.

Southeastern Regional Climate Assessment report. 1999.
http:.//weather.m sfc. nasa.gov/regi onal/water~quality. html

Spieles, D.J., and W.J. Mitsch. 2000. The effects of season and hydraulic and chemical loading
on nitrate retention in constructed wetlands: a comparison of low- and high-nutrient
riverine systems. Ecological Engineering. 14:77-91.

Stottmeister,U., A. Wiesner, P. Kuschk, U. Kappelmeyer, M. Kastner, O. Bederski, R.A. Miidler
and H. Moormann. 2003. Effects of plants and microorganisms in constructed wetlands
for wastewater treatment. Biotechnology advances 22:93-117.

Tanner, C.C., and R.H. Kadlec. 2003. Oxygen flux implications of observed nitrogen removal
rates in subsurface flow wetlands. Water science and Technology 48:191-198.

Taylor M.D., S.A. White, S.L. Chandler. 2006. Nutrient mitigation for large scale nurseries. p
540-543. hz M. Taylor. (ed.) Water Management. Proceedings of Research Conference,
GA, 2006, Southern Nursery Association, Atlanta, GA.

Thomas, M.V., and D. Joyner. 2002. Total Maximum Daily Load (TMDL) Environmental
Regulations. Proceedings Conference Fort Worth, Texas, USA. 11-13 March, 2002.

Thompson, S.P., M.F. Piehler, and W.H. Paerl. 2000. Denitrification in an estuarine headwater
creek within an agricultural watershed. Journal of Environmental Quality 29:1914-1923.

Tiedje, J.M. 1982. Denitrification. P. 1011-1026 hz A.L.Page, R.H.Miller, and D.R. Keeney (ed.)
Methods of soil analysis. Part 2. (2nd ed). Agronomy Monograph 9. ASA and SSSA.
Madison, WI.

Tiedj e, J.M. 1988. Ecology of denitrification and dissimilatory nitrate reduction to ammonium,
p. 179-24 hz AJB Zehnder (ed.), Biology of anaerobic microorganisms. John and Wiley &
Sons, Inc., New York.

















Storm water
Rletention

ceHn





ceR






Naiscl

Fiur -1.Ilutrtinofte tdycelshwngte octonad ube inictd ystr) f othypln rotsmpe t sss
tede nirfcto oeta nteriopeesi fCrafacd n yh aioi










Toet, S., L.H.F.A. Huibers, R.S.P. Van Logtestijn, and J.T.A.Verhoeven. 2003. Denitrification in
the periphyton associated with plant shoots and in the sediment of a wetland system
supplied with sewage treatment plant effluent. Hydrobiologia 501:29-44.

Toet, S., R.S.P. Van Logtestijn, R. Kampf, M. Schreijer and J.T.A. Verhoeven. 2005. The effect
of hydraulic retention time on the removal of pollutants from sewage treatment plan
effluent in a surface-flow wetland system. Wetlands 25:375-391.

USEPA. 1983. Methods of chemical analysis of water and wastes. Environmental Monitoring
Support Lab., Cincinnati, OH.

USEPA. 1986. Quality criteria for water 1986 and update no. I and update no. 2 (May 1, 1987):
Washington, D.C., EPA-440/5-86-001

USEPA. 1993a. NPS. Guidance Specifying Management Measures for Sources of Nonpoint
Pollution in Coastal Waters EPA 840-B-92-002.

USEPA, 1993b. Constructed Wetlands for Wastewater Treatment and Wildlife Habitat 17 Case
Studies EPA832-R-93-005

USEPA. 1994. EPA's Polluted brochure EPA-841-F-94-005,
http://www. epa.gov/owow/np s/qa.html

USEPA. 2002a. The Successful Implementation of the Clean Water Act's Section 3 19 Nonpoint
Source Pollution Program http://www. epa.gov/nps/Section3 19III/intro.htm

USEPA.2002b. National Coastal Condition Report.
http://www. ep a.gov/owow/oc ean s/nc cr/i ndex. html

USGS. 1998. Water Quality in the Georgia-Florida Coastal Plain, Georgia and Florida, 1992-96
By Marian P. Berndt, Hilda H. Hatzell, Christy A. Crandall, Michael Turtora, John R.
Pittman, and Edward T. Oaksford USGS Circular 1151

USGS. 2004. http://pubs.usgs.gov/circ/2004/1265/pdf/ciclal6.pdf.

White, J.R., and K.R. Reddy. 1999. Influence of nitrate and phosphorus loading on denitrifying
enzyme activity in Everglades wetland soils. Soil Science Society of America journal
63:1945-1954.

Wood, A. 1995. Constructed wetlands in water pollution control: fundamentals to their
understanding. Water Science and Technology 32:21-29.

Wood, S.L., E.F. Wheeler, R.D. Berghage, and R.E. Graves 1999. Temperature effects on
wastewater nitrate removal in laboratory scale constructed wetlands. American Society of
Agricultural Engineers 42:185-190.









collected in 125 ml plastic bottles and analyzed for TKN, NH4 -N, NO3--N, TP, and DOC and

the second set of samples (three replicates) was collected in 25 ml vials and analyzed for DEA.

If the water was too turbid, suspended particles were allowed to settle by setting aside for a few

minutes without any disturbance. Both surface and bottom of the water column samples were

filtered by cheese cloth (Hermitage Inc, Camden, SC) to remove large debris, dead leaves, twigs,

barks, and plants of Wolfia spp. In addition, three replicate water samples also were collected for

denitrification potential analysis, at the point where the flow control channel (FCC) releases

water into the holding pond or into the stormwater retention basin. There were 36 samples for

nutrient analysis and 39 samples for denitrification potential analysis in total. Samples were

stored in a cooler filled with ice-packs, transported to the lab, and stored at 4 OC until further

processing.

In addition, pH, dissolved oxygen (DO), and temperature were measured every month

during the study by using a handheld YSI 556 MPS (YSI Incorporated, Yellow Springs, OH)

multi-probe system. The parameters were measured at both surface and bottom of the water

column at inflow and outflow points on the days of water sample collection.

Plant sampling (for determining the denitrification potential of macrophyte

rhizosphere soil). To assess the denitrification potential of macrophyte rhizosphere soil, two

macrophytes were selected based on their abundance and distribution in the study cell, and based

on the reported contribution of these macrophytes to nutrient removal in constructed wetlands.

The plants selected for this study were T.1atifolia and C. flaccida. Roots of three plants of each

species were collected from each of the three sections at monthly intervals from plants growing

on the north (inflow) side (Fig. 2-10). The sample plants were selected from where a group of

the same species existed to avoid contamination with other species. The selected plants were










Faeth, P. 2000. Fertile ground: Nutrient Trading's potential to cost effectively improve water
quality. World Resources Institute. Washington DC. USA

Franklin, E.C., J.D. Gregory, D.W. Hazel, and J.E. Parsons. 2000. Management of filter zones
for dispersion and treatment of agricultural runoff. Water Resources Institute, Raleigh,
NC, USA. 312/202.

Frisbee, A.E. 2007. Nitrate nitrogen dynamics in tributaries of the SantaFe river watershed, north
central Florida. MS Thesis University of Florida.

Gale, P.M., I. Devai, K.R. Reddy and D.A. Gratez. 1993. Denitrifieation potential of soils from
constructed and natural wetlands. Ecological Engineering 2: 119-130.

Gersberg, R. M., B. V. Elkins and C. R. Goldman. 1983. Nitrogen removal in artificial wetlands.
Water Research 17: 1009-1014.

Gersberg, R.M., B.V. Elkins, S.R. Lyons, and C.R. Goldman. 1986. Role of aquatic plants in
wastewater treatment by artificial wetlands. Water Research 20:363-368.

Gren, I.M. 1995. The value of investing in wetlands for nitrogen abatement. European review of
Agricultural economics 22:157-172.

Hall, J.M., E. Paterson, and K. Killham. 1998 The effect of elevated CO2 COncentration and soil
pH on the relationship between plant growth and rhizosphere denitrifieation potential.
Global Change Biology 4:209-216.

Hammer, D.A. 1992. Designing constructed wetlands systems to treat agricultural nonpoint
source pollution. Ecological Engineering 1:49-82.

Hammer, D. A., and R.L. Knight. 1994. Designing constructed wetlands for nitrogen removal.
Water science and Technology 29:15-17.

Headley, T.R., D.O. Huett, and L. Davidson. 2001. The removal of nutrients from plant nursery
irrigation runoff in subsurface horizontal flow wetlands. Water Science and Technology
44:77-84.

Hoag, J.C. 1997. A Constructed wetland system for water quality improvement of nursery
irrigation wastewater. p 132-135. In: Landis, T.D., Thompson, J.R. (ed). National
Proceedings, Forest and Conservation Nursery Associations. General Technical Report.
PNW-GTR-419. Portland. OR: US Department of Agriculture, Forest Service, Pacific
Northwest Research Station.

Hodges, A.W., and J.J. Haydu. 2002. Economic Impacts of the Florida Environmental
Horticulture Industry, 2000. Economic Information Report EIR-02-3, Food and Resource
Economics Department, University of Florida, Gainesville, FL.
http:.//hortbusine ss.ifas.ufl .edu/EIRO2-3 .pdf









is approximately 69% woodland, 17% cropland, 11% pastureland, and 3% urban (Berndt et al.,

1996). According to the USGS (1998) report on water quality in the Georgia-Florida coastal

plain during 1992-1996; "more than 20% of groundwater samples from the surficial aquifers in

the agricultural areas throughout the study area had nitrate nitrogen (NO3- N) concentrations

greater than the USEPA drinking water standard of 10 mg L^1." The concentrations of other

nutrients in groundwater including ammonia, orthophosphate and total phosphorus were low.

Nitrogen concentrations (as NO3- N) in streams did not exceed drinking water standards or

guidelines, but were higher in streams draining agricultural and mixed basins. Nearly 30% of the

water samples from the streams had dissolved phosphorus concentrations greater than the aquatic

criteria (0.1 mg L^1; USEPA, 1986) for total phosphorus. Nearly 80% of the streams in

agricultural areas had phosphorus concentrations greater than the USEPA standard to prevent

algal growth (USGS 1998). The study area, situated in the Georgia-Florida coastal plain,

overlies the Upper Floridan aquifer, which is the maj or source of drinking water for that area.

The unconfined Upper Floridan aquifer has karst features, which is vulnerable to groundwater

contamination, similar to that of sandy soils (USGS, 1998).

Water Quality and Ornamental Plant Industry in Georgia and Florida

Florida is one of America's leading agricultural states which produces a wide range of

commodities. The state has a humid subtropical climate with mean rainfall of up to 140 cm per

year, mean monthly temperature ranging from 4 to 33"C during a calendar year, and is prone to

hurricanes (Black, 1993). Florida has the largest total acreage of 'Aquods', (wet sandy soils with

an organic stained subsoil layer; www.nrcs.gov, November 2007), are susceptible to

groundwater contamination due to poor water holding capacity. Hence, pollution of surface and

groundwater resources resulting from agricultural runoff, urban stormwater runoff, and erosion









CHAPTER 2
MATERIALS AND METHODS

Site Description

The study was carried out at a constructed wetland at Monrovia Nursery, which is one of

the largest (total area over 445 hectares) commercial nurseries in the United States, located in

Cairo, Georgia. Cairo is located at 30052'40"N, 84012'32"W in Grady County, Georgia (Fig. 2-

1), adj oining Leon and Gadsden counties of Florida. The annual production of the nursery

ranges from 11 to 15 million plants (www.monrovia.com, October 2007).

Gravel beds are created outdoors on the ground and are overlaid with plastic sheets on

which plants are grown in plastic containers ranging from 5 50 L capacity. In the catchment

area for the wetlands, large (50 L) containerized ornamental trees and shrubs are grown in the

sloped nursery beds, which are lined with plastic sheets. Fertilization of the nursery plants is

done by two means; controlled release fertilizer (CRFs) mixed in the potting medium and

fertigation (liquid fertilizers applied in irrigation water). Topdressing of CRFs is also done based

on the needs of the plants. Overhead sprinkler or micro-irrigation is used approximately 3-5

times a day. While larger containers are irrigated via micro irrigation, runoff(Fig. 2-2) from

these nursery beds is still significant due to the water application rates and sloped terrain.

Runoff from the nursery beds and from stormwater is directed by lined or unlined waterways and

drained into a flow control channel (FCC) which is 500 m in length (Fig. 2-3 and 2-4). Runoff

eventually flows into a holding pond (HP) where it is held for recycling or for release into the

wetlands. The FCC controls the movement of water and allows sedimentation to occur.

Depending upon the holding capacity of the HP, the water from the FCC is either directed into

the HP or diverted offsite through the stormwater retention basin. The HP is created in such a









0.5 -y 0.536E

OA -~O~


03






0.0 0.2 0A 0.6 0.8

~~ Ammonlnwm noea wr~mgu leadi re (3 dJ)




Figure 3-8. Pearson's correlation between loading and removal rate of ammonium nitrogen (g m-2 day l) between April 2007 and
August 2007 in a constructed wetland receiving nursery runoff









N and thereby may have contributed to the increased removal efficiency of the NO3- N. Studies

have shown that macrophytes in the wetlands remove substantial amount of NO3--N directly by

uptake or indirectly by providing carbon exudates from roots, which enhance microbial

transformations (Gersberg et al., 1983; Bachand and Horne, 2000).

The NH4 -N contributed only less than 1% of the TN load into the wetland. The removal

efficiency of the NH4 -N ranged from 41% (June) to 93% (August). The concentration of NH4 -

N coming into the wetland was approximately 2 mg L^1, and the outflow concentration less than

1 mg L^1. The NH4 -N removal is likely because of plant uptake or nitrification process in the

aerobic sites of the wetlands. Immobilization by plants and microbes enhances the removal of

NH4 -N. In addition, microbe mediated nitrification process, in which NH4 -N is oxidized to

NO3~ N also contributes to NH4 -N removal in the aerobic zones of wetlands (Burger and

Jackson, 2004). The volatilization ofNH4 -N to ammonia would not be a significant process for

NH4 -N removal in this wetland as the mean pH value (6.0) was lower than optimal (> 8) for this

chemical process to occur (USEPA, 1993, Sartoris et al., 2000). The pH values in this system

did not reach that optimum level during the study period. Plant growth and the size of denitrifier

population within the rhizosphere are observed to be coupled. Hence the influence of pH on

plant growth in turn affects the denitrification potential (Hall et al., 1998). However, the

influence of pH on plant growth was not studied in this system during this study.

The concentration of TP entering and leaving the wetland cell was approximately 1 mg L^1

The removal efficiency of TP ranged from 0-32%, which varied over months. The removal of

TP (which consists orthophosphate) may be attributed to biological uptake, adsorption onto

sediment, etc. The percent TP removal efficiency over months did not follow any particular

trend, however. Similar results of variable TP net removal during the growing season were









BIOGRAPHICAL SKETCH

Bhuvaneswari Govindaraj an, shortly called as Bhuvana, was born in 1976 in India. She

spent the greater portion of her life growing up in a small town, named Vandavasi, Tamilnadu,

India. Hailing from a farming family, she was fascinated by field of agriculture. She majored in

horticulture sciences and received her Bachelor' s degree in Horticulture (1994-1998) from

Tamilnadu Agricultural University, Coimbatore, Tamilnadu, India. She completed her Master of

Science in Vegetable Crops (1999-2001), with a minor in plant breeding, from Punj ab

Agricultural University, Ludhiana, Punjab, India.

For the next four years, she worked in several organizations, as research associate in a

private biofertilizer production unit, as a lecturer in agricultural college, as a technical consultant

for an MNC, on National Horticulture Board proj ect, and as a Development Executive for an

NGO in a watershed development proj ect. While working in an NGO, she was exposed to

problems associated with environmental degradation/pollution and inclined towards the field of

environmental science. She then decided to continue her education in the field of environmental

sciences. Upon achieving the graduate research assistantship, she came to United States, to

pursue her Masters degree in Interdisciplinary Ecology at University of Florida. Bhuvana has

had an opportunity to work on a project as well as her thesis research focusing on nutrient

removal from nonpoint source pollution using constructed wetlands.

After gaining hands on experience in the US, She wish to return to her county, India,

where her knowledge and skills will be applied towards achieving cleaner, greener, and

sustainable environment.









CHAPTER 3
RESULTS

Removal efficiency of the constructed wetlands. The concentration of nutrients

entering and leaving the constructed wetland was determined to evaluate the effectiveness of the

study cell in removing nutrient contaminants over a period of five months from April to August

2007. The plant nursery runoff contained an average total nitrogen (TN) concentration of 34 mg

L- during the study period. When assessing mean TN composition (Fig. 3-1), it was observed

that the maj ority of Nitrogen (N) entering and leaving the wetland during the study period was in

the form of nitrate nitrogen (NO3--N; 30 mg L^1) followed by ammonium nitrogen (NH4 -N; 2

mg L^1) and organic nitrogen (2 mg L^)~. Total phosphorus concentration was 1 mg L^1 and

dissolved organic carbon was 10 mg L^1. The nursery runoff entering the constructed wetland

over the study period consisted of 87% of the TN in the form of NO3--N. The mean nutrient

removal efficiency of the study cell, based on load-in and load-out rates, was observed to be 40%

for TN, 40% for NO3- N, 59% for NH4 -N, and 16% for TP (Table 3-1).

Nitrate Nitrogen (NO3--N). The concentration ofNO3--N was influenced (p<0.0001) by

an interaction among location of sample collection and month during which samples were

collected (Fig. 3-2). The inflow NO3--N concentration gradually decreased from April (47 mg L-

1) to August (13 mg L^)~; similarly the outflow concentration also gradually decreased from April

(42 mg L^1) to August (3 mg L^)~. At flow control channel (FCC), NO3--N concentration did not

show any trend over time, but the absolute values were highest in April (52 mg L^1) and lowest in

July (13 mg L^)~. The mean removal efficiency ofNO3--N in the study cell was 40%, highest in

August (77%) and lowest in April (12%). The removal efficiency was inversely related to the

loading rate of NO3--N coming into the wetland; i.e., when the loading rate of NO3--N was higher

(7 g m-2 day- ; April), the removal efficiency was lower (12%) and when the loading rate was









differences in the availability of organic carbon in the respective macrocosms. The influence of

different macrophyte rhizosphere on denitrifieation rates within the constructed wetlands

receiving plant nursery runoff is not known. The goal of this study was to assess the N dynamics

in a constructed wetland receiving plant nursery runoff for on-site N removal. To achieve this

goal, following specific obj ectives and hypotheses were addressed.

Objectives and Hypotheses

Objective 1: To determine the N and P removal efficiency of a constructed wetland
receiving plant nursery runoff

Hypothesis 1: The concentration of N and P will vary over months based on the timing
of fertilizer application at the nursery, irrigation, and rainfall. The concentration of N and
P will be higher at inflow as compared to the outflow water, due to the nutrient removal
within the constructed wetlands.

Objective 2: To theoretically model the effect of hydraulic loading rate of the runoff and
of estimated plant density on the residence time, which influences N removal in a
constructed wetland.

Hypothesis 2: The theoretical residence time will increase with increasing plant density
and decreasing loading rate.

Objective 3: To assess the concentration of nutrients (N and P) and denitrification
potential within the water column at varying depths (surface and bottom) and variation
associated with the locations across the wetland (inflow and outflow) over Hyve months in
a constructed wetland receiving plant nursery runoff.

Hypothesis 3: The concentration of nutrients (N and P) will be higher at inflow as
compared to the outflow. The denitrifieation potential will be higher in the water
samples collected from the lower depths as compared to the surface of the water column
due to the anoxic conditions in the deeper parts of the water column.

Objective 4: To determine the denitrification potential in the rhizosphere soil (soil
closely adhering to the roots) of Can2na flaccida and Typha latifolia in a constructed
wetland receiving plant nursery runoff

Hypothesis 4: The denitrification potential of the rhizosphere soil (soil closely adhering
to the roots) will vary from species to species because macrophytes can vary in their
capacity to diffuse oxygen and different carbon compounds in their root zones.









shaker for an hour to ensure even distribution of C2H2 in the water samples. Denitrification

Enzyme Assay (DEA) solution was prepared by adding nitrate (as KNO3) and carbon (as

C6H1206) at the rate of 404mg L^1 KNO3 and 720 mg- L^ C6H1206, TOSpectively, along with 250

mg- L^ chloramphenicol in distilled water. Chloramphenicol was added to inhibit the production

of new denitrifying enzymes while the activity of previously existing denitrifying enzymes are

measured. Eight ml of DEA solution was added to 20 ml sample. The 7 ml of head space of gas

was collected by using a 10 ml syringe immediately after adding DEA to the samples and after

taking the pressure readings. After that, the bottles were kept in an end-to-end shaker in a dark

room at 25oC. In the month of April 2007 the headspace gas samples were collected at 30, 60

120, 180 minutes. From May 2007 and to August 2007 the headspace gas samples were

collected every 2 hours, up to 6 hours, as there was no gas production observed until 2 hours in

the month of April. At the end of each pre-selected time period, 7 ml of gas (presumably nitrous

oxide) was extracted by using a syringe and placed in 3-4ml capped, crimped, pre-evacuated

glass serum bottles. The bottles with DEA solution and the original sample were then

immediately returned to the shaker. The 7 ml sample of gas was stored at 25oC until analyzed in

a Gas Chromatograph (GC).

The concentration of Nitrous oxide (N20) was measured on a Shimadzu GC 14A

(Shimadzu Scientific, Kyoto, Japan) by using a 63Ni electron capture detector. Column

temperature was 30oC, detector temperature was 240 oC, and inj ector temperature was 120 oC.

The carrier gas was a mix of 95% Argon (Ar) and 5% methane (CH4). The gas samples were

injected into GC and the concentration ofN20 was determined. Denitrification potential was

determined by calculating the slope of the linear curve obtained when the gas concentrations

were plotted over time.





























,~ ~i~t~
J~ ;-~


Figure 2-7. Outflow pipe of the study cell










LIST OF FIGURES


Figure page

1-1 Diagram illustrating potential pathways for chemical transformations of nitrogen (N)
in a w etland. .............. ...............29....

2-1 Map of the state of Georgia, United States, showing county boundaries. .......................41

2-2 Runoff generated from the microirrigated nursery beds at Monrovia Nursery .................42

2-3 Runoff generated from the nursery beds drains into the flow control channel ................. .43

2-4 Runoff draining into the flow control channel ................. ...............44..............

2-5 Overview of the Constructed wetlands at Monrovia Nursery, Cairo, GA. ......................45

2-6 Inflow pipes into the study cell ................. ...............46........... ..

2-7 Outflow pipe of the study cell ................. ...............47........... ..

2-8 Location and number of water samples (indicated by stars) at inflow and outflow of
the study cell to determine the nutrient removal efficiency of the constructed
w wetlands. ............. ...............48.....

2-9 Locations of water samples (indicated by stars) in the study cell to assess the spatial
variation in the concentration of nutrients and denitrification potential of the water
column, between locations (inflow and outflow) and between depths (top and
bottom) within locations over five months. ............. ...............49.....

2-10 Illustration of the study cell showing the location and number (indicated by stars) of
monthly plant root samples to assess the denitrification potential in the rhizosphere
soil of Can2na flaccida and Typha latifolia .............. ...............50....

3-1 Mean concentration (mg L^1) of total nitrogen (TN) and composition of TN [nitrate
(NO3- N), ammonium (NH4 -N) and organic nitrogen concentration, mg L^1]
between April to August of 2007 in the water samples obtained from the flow control
channel (n=3), inlet (n=9), and outlet (n=9) pipes of a constructed wetland receiving
nursery runoff ................. ...............65......__ .....

3-2 Mean concentration (mg L^1) of nitrate nitrogen (NO3--N) between April 2007 and
August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the
study cell in the constructed wetland receiving nursery runoff. .............. ...................66

3-3 Mean concentration (mg L^1) of ammonium nitrogen (NH4+ N) between April 2007
and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9)
of the study cell of a constructed wetland receiving nursery runoff. ............ .................67









8- A
A b A A B

A


A

p4
r9,a Inkt








Marv June July Ananet
unntes



Figure 3-13. Physiochemical parameter, pH, as affected by location of sample collection, and months (April and August 2007) of the
constructed wetland receiving nursery runoff (n=3). Capital letter indicates statistical difference between locations within
each month (May to August); lower-case letters indicate statistical difference between months at each location (inlet and
outlet). Means followed by the same letter are statistically similar.









the water column showed significant influence on DO level, which is expected in waterlogged

conditions. The anoxic condition of the wetland was enhanced with increasing depth of the

water column. The DO concentration gradually decreased from April to August, when the

temperature gradually increased from April to August. The reduction in concentration of DO

could be attributed to increased activity of microorganisms, which enhances N removal, with the

gradual increase in temperature.

The nutrients (N and P) analyzed in this study varied in their concentration at inflow and

outflow. This variation in the concentration of nutrients may be attributed to removal of

nutrients within the study wetland. Similar nutrient gradient from the inlet to outlet points was

observed in Everglades's wetlands soil (White and Reddy, 1999) and in experimental wetlands

receiving agricultural runoff (Sirivedhin and Gray, 2006). The low concentration of NO3- N at

outflow is likely due to the removal of nitrogen by various mechanisms such as plant uptake,

microbial assimilation, and denitrifieation along the flow of water from inflow. The low

concentration of NO3- N at lower depth of the water column is likely due to the removal of

NO3- N by plant uptake and denitrifieation process in the anaerobic zones of the water column.

Another possible reason might be the lack of diffusion of NO3- N to the lower depths of the

water column, which will also limit the denitrifieation activity.

Although the concentration of NH4 -N was always higher at inflow as compared to

outflow of the wetland, there was no trend observed in the spatial distribution of NH4 -N within

the wetland cell over the period of Hyve months. The lower concentration of NH4 -N at outflow

could be attributed to the removal of NH4 -N within the study cell. However, the concentration

of NH4 -N was observed to be higher in lower depth of the water column, which is likely due to

lack of nitrifieation in the anoxic zones of the water column.









pulled out and the roots were severed and then placed in zip-lock bags, stored in a cooler filled

with ice-packs, transported to the lab, and stored at 40C. There were 18 rhizosphere samples

(three replicates of each species per section) in total.

Analytical Methods

Nutrient analyses. The day after sampling, part of all the water samples collected in 125

ml bottles was filtered through Whatman 0.45Cpm filter (Pall Corporation, Ann arbor, MI) and

collected in two 25 ml vials for analysis of NH4 -N, NO3- -N, and DOC. The remaining

unfiltered part of each sample was used for analysis of TKN and TP. After filtering, all the

filtered and unfiltered water samples (except the ones collected for analyzing denitrification

potential), were acidified to a pH of 2 with one drop of ultra pure concentrated sulfuric acid

(H2SO4) foT CVery 20-25 ml of sample. The filtered and unfiltered portions of samples were

stored in a refrigerator until further analysis. Samples were analyzed for respective nutrients

within 28 days of sampling as recommended by the USEPA.

Total Kj eldal Nitrogen was measured by digesting the unfiltered, acidified samples by the

Kj eldahl procedure (Method 3 51.2, USEPA, 1983). The water samples were digested with

sulfuric acid (H2SO4) and a copper sulfate mixture (CuSO4) to convert organic forms of nitrogen

to ammonium. The digested samples were analyzed in AQ2' automated discrete analyzer (Seal

Analytical, Mequon, WI) Method no: USEPA 111 A (Method 351.2, USEPA, 1983). Nitrate

nitrogen was analyzed calorimetrically using the cadmium (Cd) reduction method on a rapid

flow analyzer (Alpkem rapid flow Analyzer 300) or in the automated AQ2+ discrete analyzer

(Method 353.2, USEPA) or AQ2 Method no: USEPA 132 A (Method 353.2, USEPA 1983).

During April, July, and August 2007 the filtered, acidified samples were analyzed for NO3--N on

a rapid flow analyzer. In the month of May and June 2007, the discrete analyzer was used to










Burgin, A.J., and S.K. Hamilton. 2007. Have we overemphasized the role of denitrification in
aquatic systems? A review of nitrate removal pathways. Frontiers in Ecology and the
Environment 5:89-96.

Carpenter, S.R., N.F. Caraco, D.L.Correll, R.W. Howarth, A.N. Sharpley, and V.H. Smith. 1998.
Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological
Applications 83:559-568.

Coffey, S. 1997. Selected agricultural best management practices to control nitrogen in the
Neuse River Basin. North Carolina State University, Raleigh, NC, USA. Technical
Bulletin 1-50.

Cooper, P.F., and B.C. Findlater. 1990. Constructed wetlands in water pollution control.
Pergamon Press, Oxford.

Crumpton, W.G., T.M. Isenhart, and S.W Fisher 1993. Controlling agricultural runoff by use of
constructed wetlands. In: Moshiri, A. (Ed.), Constructed Wetlands for Water Quality
Improvement. Lewis Publishers, Ann Arbor, MI.

D'Angelo, E.M., and K.R. Reddy. 1999. Regulators of heterotrophic microbial potentials in
wetland soils. Soil Biology and Biochemistry 31:815-830.

Dahl, T.E. 1990 Wetlands losses in the United States, 1780' s to 1980' s. Report to the Congress.
United States Department of the Interior. Fish and Wildlife Service. Washington DC
USA.

Darbi, A., T. Viraraghavan, R. Butter and D. Corkal. 2003. Pilot-scale evaluation of select nitrate
removal technologies. Journal of Environmental Science and Health 38:1703-1715.

DeBusk, W.F. 1999. Functional role of wetlands in watersheds.SL 169 Florida Cooperative
Extension Service. IFAS. University of Florida. Gainesville. http://edis.ifas.ufl .edu.

Desimone, L.A., and B.L. Howes. 1996 Denitrification and nitrogen transport in a coastal aquifer
receiving wastewater discharge. Environmental Science and Technology 30: 1152-1162

Dortch, M. S. 1992. Literature analysis addresses the functional ability of the wetlands to
improve water quality. The wetlands research program bulletin. 2: 1-4 USACE
Waterways Experimental station, Vicksburg, MS.

Dunne, E.J., N. Culleton, G.O. Donovan, R. Harrington, and A.E. Olson. 2005. An integrated
constructed wetland to treat contaminants and nutrients from dairy farmyard dirty water.
Ecological Engineering 24:221-234.

Eriksson, P.G., and S.E.B. Weisner 1997. Nitrogen removal in a wastewater reservoir: The
importance of denitrification by epiphytic biofilms on submersed vegetation. Journal of
Environmental Quality 26:905-910.









Significant spatial distribution in total phosphorus concentration was observed within the

wetland. However, there were no temporal effects on TP concentration during the period of this

study. The TP was observed to be higher in the deeper water column as compared to the surface

of the water column, but only at the inflow side of the wetland. This could be due to the release

of adsorbed P from the sediments into immediate atmosphere, or due to lack of adsorption sites

near the inflow of the study cell. The sorption sites on the inflow side might have become

saturated because of the higher concentration of TP at inflow. The DOC concentration did not

show any differences in its distribution within the depth of the water column, suggesting its

uniform distribution within the wetland.

The denitrification potential of the water column in the study cell was either negligible or

undetectable. The maximum value of denitrification potential observed in the water column was

0.03 mg N20-N L-1 hr- Lower values of denitrification potential are likely due to a lack of

particulate material in the water column to support the microbial activity. Little research has

been conducted on denitrification potential of the water column in constructed wetlands, but

whenever it was included in wetland studies, the reported values were always relatively low.

Denitrification rates of the water samples in a constructed wetland receiving sewage treatment

plant effluent were observed to be between 0.06 0.96 Clg N m-2 d-l (Toet et al., 2003). The

water column of the study cell had low dissolved organic carbon (DOC) concentration, and this

may be limiting NO3- N removal due to non availability of energy for denitrifying microbes.

The optimum C:N (plant carbon added:NO3--N in water) ratio for the denitrification in wetlands

is reported to be 4:1 (Ingersoll and Baker, 1998) to 5:1(Baker, 1998). Only during the month of

April, the DOC: NO3- N ratio reached near 4: 1 in the wetland. The mean DOC: NO3- N ratio

was low (<1) during most of the study period. Low DOC: TN reported to limit NO3- N removal









The Pearson correlation coefficient value (r = 0.19) indicated a little/no association

between loading and removal rate of NO3- N. The strong positive association between loading

and removal rate of NH4 -N (r = 0.75) and a weak positive association between loading and

removal rate of TP (r = 0.6). The positive correlation coefficient indicates that the loading and

removal rate tend to increase or decrease together. Similar linear relationship of increasing

removal rate with increasing loading rate has been observed by Headley et al. (2001) for TN (r =

0.76).

Plant nursery runoff in this study consisted 87% of the TN in the form of NO3- N. The

dominance of NO3- N in runoff water entering the wetland could be attributed to the controlled

release (coated urea), and soluble nitrate fertilizers (Ammonium Nitrate) applied to nursery

plants (S. Chandler, Horticulture Manager Monrovia Nursery personal communication). It is not

uncommon to apply up to 200 mg L^1 nitrate nitrogen to ornamental crops during peak growing

season. The concentration of NO3- N (mg L^1) entering the wetland varied during the period of

study, i.e., it gradually decreased from April to August 2007. The higher concentration of NO3~ -

N during April was likely due to fertilization practices at the nursery during their peak growing

season (S. Chandler, Horticulture Manager Monrovia Nursery personal communication). In

addition, the limited rainfall (15-17 cm during April and May 2007) might have contributed to

increased concentration of nutrients in the runoff. Similarly, increased concentration of nutrients

in the nursery runoff was observed due to fertilization in subsurface horizontal flow reed beds in

subtropical region of Australia (Headley et al., 2001). During April and May there were

comparatively higher concentrations (44 mg L^1 and 36 mg L^1, respectively) of NO3- N

entering the wetland as compared to June, July and August. Higher concentration of NO3- N

during April (averaging 19 mg L^1) and May (averaging 18 mg L^1) were also obtained by Taylor









ACKNOWLEDGMENTS

I would like to thank my advisor Dr. Jyotsna Sharma for her guidance, wisdom, and

support, as well as my committee members, Drs. K.R.Reddy, Mark W Clark, and Kanika S

Inglett for their inputs on this proj ect. Additional thanks to Dr. Jyotsna Sharma for all her time,

especially during process of writing my thesis.

I would like to thank Dr. Jim Jawitz for his valuable inputs throughout my research. I wish

to thank Dr. Wilson for the statistical inputs. I am also thankful to Dr. Dunne for his assistance

and inputs. I thank everyone in the Wetland Biogeochemistry Laboratory for invaluable training

and inputs during lab experiments.

Finally, I would like to thank my friends and family for their love and support. A special

thanks to my husband, Subramaniam, for his patience and sacrifices through the long years of

graduate school, and for his support and help in all the endeavors I have chosen to pursue.









Denitrification potential (Rhizosphere). The stored root samples were processed as

early as possible after reaching the lab. The excessive and loosely adhering soil was removed by

shaking the root system, and also by gently removing by hand. The soil very closely adhered to

the roots was defined as 'rhizosphere soil' which was removed by washing the roots with

autoclaved double distilled water. Thirty-five ml of water was used to remove the firmly

attached soil by washing with the use of pipette. Twenty ml of the obtained soil slurry was used

for the determination of denitrification potential. Ten ml of the rhizosphere solution was used to

determine the moisture content of the soil. The moisture content was determined gravimetrically

by drying 10 ml of the rhizosphere solution at 70oC for 72 hours.

The denitrification potential of the rhizosphere soil was determined as described above.

The volume of the glass serum bottle used in this experiment was 120 ml. Eight ml of DEA

solution (720 mg- L^ glucose, 404 mg- L^ KNO3, and 500 mg- L^ chloroamphenicol) was added

to 20 ml of samples. Gas samples were collected every 30, 60 120, 180 minutes. Rest of the

procedure was as described for measuring denitrification potential in water samples.

Statistical Methods

Statistical analysis of the data was performed by using SAS 9. 1 (SAS Institute Inc, Cary,

NC). The dependent variables were subjected to analysis of variance (ANOVA) by using the

GLM procedure along with Scheffe for all pair wise comparison, which indicate not only

whether the means are different from each other, but also which means differ from which other

means. Test of statistical significance was done at a = 0.05. The physiochemical parameters of

water viz., temperature and dissolved oxygen were correlated with denitrification potential by

using CORR procedure to determine Pearson's correlation coefficient. The concentration of

NO3--N was also correlated with denitrification potential by using the same procedure.











Table 4-2. Denitrification potentials of different substrate reported by other researchers in constructed wetlands. While the units of
measure are different and make direct comparisons challenging, these data are useful for making general inferences for
wetlands dominated by different plant taxa. The acronyms used are as follows: Denitrification Enzyme activity (DEA),
Acetylene Inhibition Technique (AIT), Membrane Inlet Mass Spectrometry (MIMS), and N tracers (N 5).
Substrate Wastewater type DEA Range of DEA values Method Reference
(mg N m d )


Phragmites australis shoots

Elodea nuttallii shoots


Sewage treatment plant emfuent

Sewage treatment plant emfuent

Sewage treatment plant emfuent

Sewage treatment plant emfuent


AIT Toet et al., 2003

AIT Toet et al., 2003

AIT Toet et al., 2003

AIT Toet et al., 2003


Sediments


Water


Spartina spp., Cladium spp.
Juncus spp. sediment cores
Juncus spp. sediment

Everglades soils


Agricultural runoff

Agricultural runoff

Agricultural runoff


44.4 121 mg N m d

14.8 33.1 mg N m d '

0.5 25.5 mg N m d

0.4 3.9 pLg N m d

0.7 9.2 mg N m d

0.1-6.0 mg N m~2d

0.004 7.75 mg N kg' hr

2-11.8 mg N m d

150 900 ng N g' h ~


Mixed plants
sediments


Agricultural tile drainage

Wetland microcosms


MIMS Poe at al., 2003

AIT Thompson et al., 2000

AIT White and Reddy, 1999

AIT /N'5 Xue et al.,1999

AIT Hall at al., 1998

AIT This study


Lolium perene (Ryegrass)
Rhizosphere soil
Typha latifolia


2.51 mg N kg h-'

4.11 mg N kg h-'


Plant nursery runoff


Canna flaccida Plant nursery runoff


AIT This study









N20-N kg-l h- ) as compared to 0.51 mg N20-N kg-l h-l in the underlying soil, between 10-30

cm, of the wetlands.

The residence time of water in a wetland is critical for removal of N because it affects the

duration of the contact between polluted water and the biotic and abiotic components such as

plant roots, soil, sediments, and microbes. The longer residence time may enhance removal of N

by promoting retention and biochemical processes such as denitrifieation, sedimentation, and

plant uptake (Kadlec and Knight, 1996; Mitsch and Gosselink, 2000). Kadlec and Knight (1996)

also showed that the N removal efficiency is logarithmically related to the residence time.

Temperature is a significant abiotic factor that influences microbial activities. Wood et al.

(1999) reported that the temperature influences the biological and physical activities in the

wetlands, and denitrifieation is a temperature dependent process. The rates of denitrifieation have

been shown to increase by 1.5-2 times with each increase in 100C (Reddy and Patrick, 1984).

The denitrifieation rates were higher (9.2 mg N m-2 h-1) in summer and lower (0.7 mg N m-2 h-

1) in fall in the constructed wetlands (Poe et al., 2003); and similarly Xue et al.(1999) reported

higher denitrifieation rates in summer (11.8 mg N m-2 h-1) and lower (2.0 mg N m-2 h-1) in

winter. White and Reddy (1999) reported highest (2.69 mg N20-N kg-1 h-1) activity of

denitrifying enzymes during the summer when the temperature, hydraulic loading, and nutrient

loading were highest. Spieles and Mitsch (2000) reported the optimum range of temperature for

denitrification in wetland sediments as 20 to 25 O C. Water temperatures less than 150C or greater

than 300C have been shown to limit the rate of denitrification (Reddy and Patrick, 1984). The

most effective pH for denitrification ranges from 7.0 to 8.5, with the optimum at 7.0 (Reddy et

al., 2000).

































To every individual who nurtured my intellectual curiosity
throughout my lifetime
making this milestone possible









denitrification rates between the studies is often challenging (Seitzinger, 1993). The AIT

method is one of the most commonly used techniques for measuring denitrification rates. This

technique measures the denitrification rates under ambient anaerobic conditions, in which

acetylene (C2H2) inhibits the reduction of N20 to N. The extent of denitrification is quantified

by the progressive accumulation ofN20. This is also known as Denitrification enzyme Assay

(DEA), which is an indicator of denitrifier biomass present in the samples and serves as an

integrated measure of denitrification potential.

Factors Influencing Denitrification in Wetlands

Denitrification in wetland systems is primarily controlled by organic carbon availability

(Reddy et al., 1982; Gale et al., 1993; D'Angelo and Reddy, 1999), aeration status (dissolved

oxygen levels; Tanner and Kadlec, 2003), and NO3- N concentration (Cooper and Findlater,

1990; Gale et al., 1993; Martin and Reddy, 1997). These factors are highly influenced by

temporal (diurnal and seasonal) and spatial (horizontal and vertical) variation (White and Reddy,

1999; Kadlec and Reddy, 2001). The other factors that influence denitrification rates in wetlands

are the make-up of the vegetative and microorganismal communities (Gersberg et al., 1986;

Sirivedhin and Gray, 2006; Smialek et al., 2006), depth of the water column (Sirivedhin and

Gray, 2006), hydraulic residence time, and pH of the water and soil (Reddy and Patrick, 1984).

The maj or regulator of denitrification is the oxygen status of the wetland, as this process is

facilitated by facultative anaerobic bacteria. Denitrification is observed to take place only in low

oxygen zones of the wetlands (Knowles, 1982; Knowles, 1990; Tiedje, 1988). In anoxic

conditions, denitrification is controlled by NO3- N and carbon availability.

The importance of organic carbon as an electron donor in the denitrification process is well

documented (Eriksson and Weisner, 1997; Lin et al., 2002; Bastviken et al., 2003). While

studying the relationship between denitrification rates and dissolved organic carbon (DOC),


































O 2008 Bhuvaneswari Govindarajan





















Aug 30 Aa 6 Aa 4.4 Aa
Bottom May 22 Aa 4 Aa 6.2 Aa
June 27 Aa 5 Aa 7.2 Aa
July 27 Aa 3 Aa 5.4 Aa
Aug 30 Aa 2 Aa 3.7 Aa
Top May 25 Aa 25 Aa 5.5 Aa
June 27 Aa 15 Aa 6.5 Aa
July 27 Aa 11 Aa 5.8 Aa
Aug 29 Aa 12 Aa 6.9 Aa
Bottom May 22 Ba 15 Aa 6.4 Aa


Table 3-3. Physiochemical parameters (temperature, dissolved oxygen, pH) as affected by location of sample collection, depth within
location, and months (April and August 2007) in the constructed wetland receiving nursery runoff (n=3). Capital letters
indicate statistical difference between depths (top and bottom) at inlet and outlet over five months; lower-case letters
indicate statistical difference between locations (inlet and outlet) at top and bottom depths over five months. Means
followed by the same letter are statistically similar.


Dissolved oxygen (mg L1)
13 Aa
7 Aa
7 Aa


Location
Inlet









Outlet


Depth
Top


Month
May
June
July


pH
Aa
Aa
Aa


Temperature oC
23 Aa
28 Aa
28 Aa


June
July
Aug


25 Aa
26 Aa


10 Aa
5 Aa


6.6 Aa
5.0 Aa
6.0 Aa


Aa 4 Aa









N during April and May in this study was likely due to fertilization practices at the nursery

during their peak growing season. In addition, the limited rainfall (15-17 cm during April and

May 2007) might have contributed to increased concentration of nutrients in the runoff.

The concentration of nutrients, as expected, was higher in the influent as compared to the

outflow. The mean nutrient removal efficiency of the constructed wetland was 40% for TN, 40%

for NO3--N, 59% for NH4 -N, and 16% for TP. While comparing to other studies, the study cell

appears to be functioning moderately efficiently in removing N and P. The NO3--N removal

efficiency was inversely related to the loading rate of NO3--N in this study. The P removal might

be due to plant, algal uptake and adsorption/precipitation reactions. The negative P removal

efficiency in July indicates that at times there may be internal loading of P in the system.

Ammonium Nitrogen removal is likely due to plant uptake or nitrification process in the aerobic

zones of the wetlands. The volatilization of NH4 -N to ammonia is probably not a significant

process for NH4 -N removal in the wetland because the observed pH in the wetland was lower

that what is reported to be optimal for this process to occur. The low DOC: NO3--N ratio in the

influent water indicates that the system might be limited by carbon availability, which could

adversely affect NO3- N removal.

Hydraulic loading rate and the plant density were observed to influence the theoretical

residence time in the model. When higher values of hydraulic loading rate were used in the

model, the theoretical residence time decreased by a day for this system. The model showed that

increasing the wetland porosity from 0.75 to 0.95 and/or 1.0 would increase the residence time in

the study cell by one day.

The nutrients (N and P) analyzed in this study exhibited a gradient from the inflow to

outflow due to removal of nutrients within the wetland. The low NO3--N concentration at lower












TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES .........__.. ..... .__. ...............6....


LIST OF FIGURES .............. ...............7.....


AB S TRAC T ......_ ................. ............_........9


CHAPTER


1 INTRODUCTION ................. ...............11.......... ......


Introducti on ................... ....... ........ ...............11.......
Water Quality in the United States ................ ...............12...............
Water Quality in Southeastern United States .............. .... ....... ...........1
Water Quality and Ornamental Plant Industry in Georgia and Florida ................. ...............14
Management Strategies to Control Nonpoint Source Pollution .......................... ..............1
Importance of Natural and Constructed Wetlands for Improving Water Quality ..................18
Nitrogen Cycling in Wetlands .............. ...............20....
Denitrification in Wetlands............... ........ ... .........2
Factors Influencing Denitrification in Wetlands .............. ...............22....
The Problem Statement ................. ...............27......._.. ....

Obj ectives and Hypotheses ........._.___..... .___ ...............28....

2 MATERIALS AND METHODS .............. ...............30....


Site Description .............. ...............30....
Field M ethods .............. ...............33....

Analytical Methods............... ...............36
Statistical Methods............... ...............40


3 RE SULT S ........._.___..... .__. ...............5 1....


4 DI SCUS SSION ........._.___..... .___ ...............78....


5 SUMMARY AND CONCLUSIONS ........._._.. ....__. ...............91..


LIST OF REFERENCES ........._.___..... .___ ...............96....


BIOGRAPHICAL SKETCH ........._.___..... .__. ...............105....












18

16

14 B

12

10

S8. B C



2 -










Figure 3-12. Dissolved oxygen concentration (mg L 1) as affected by location of sample collection, depth within location, and months
(April and August 2007) in the constructed wetland receiving nursery runoff (n=3). Capital letters indicate statistical
difference within location, within depth, and within months. Means followed by the same letter are statistically similar.









The Problem Statement

Among the various management strategies developed to mitigate nonpoint source

pollution, constructed wetlands are considered as a cost effective method for on-site removal of

nutrients from nonpoint source pollution. The performance of the constructed wetlands varies

with site, characteristics of the wastewater, type and design of the wetland. Hence, a "systems

approach" which recognizes site specific conditions, is essential for the successful management

of the plant nursery runoff (Mitsch and Jorgensen 1989; NRCS, 2002; Dunne et al., 2005). Most

research has focused on use of constructed wetlands to treat wastewater from municipal water

source, animal waste, urban runoff, stormwater runoff, and agricultural runoff. Very few studies

have focused on the constructed wetlands treating nursery runoff (Headley et al., 2001; Lea-cox

et al., 2002; Huett et al., 2005). Constructed wetlands that receive nursery runoff show

considerable difference in water quality as compared with other wastewater sources. One of the

main differences is the low concentration of dissolved organic carbon (DOC) in the nursery

runoff (Huett et al., 2005), which is due to type of medium used in the containerized plants.

Nursery runoff can have seasonal and hydraulic variability based on the timing of fertilizer

application and rainfall. Due to the characteristic features of the plant nursery runoff, it is

important to understand the nutrient removal mechanisms in a constructed wetland treating

nursery runoff.

One of the methods to increase the efficiency of nutrient removal is to implement

management and design practices to increase the denitrification process in constructed wetlands.

Although denitrification is considered a maj or removal mechanism for N in constructed

wetlands, the importance of this process in wetlands receiving plant nursery runoff has not been

studied. Earlier studies (Bachand and Horne, 2000) indicated that different vegetation types

resulted in different denitrification rates in the wetlands which could be attributed to the










Patrick, W.H. Jr. 1994. From wastelands to wetlands. Journal of Environmental Quality 23:892-
896.

Platter C. 1996. Enhanced nitrogen elimination in subsurface flow artificial wetlands a multi
stage concept. International Conference of wetland systems for water pollution control.
Vienna, Austria.

Poe, A.C., M.F. Piehler, S.P. Thompson, and H.W. Paerl. 2003. Denitrification in a constructed
wetland receiving agricultural runoff. Wetlands 23:817-826.

Reddy, K. R. and E. M. D'Angelo. 1994. Soil processes regulating water quality in wetlands.p.
309-324. In: Mitsch, J.W. (ed.) Global Wetlands: Old World and New. Elsevier, NY.

Reddy, K. R., and R. D. Delaune. 2007. Biogeochemistry of Wetlands: Science and
Applications. CRC Press, Boca Raton, Fl. (in press).

Reddy, K.R., and W.H. Patrick 1984. Nitrogen transformations and loss in flooded soils and
sediments. Critical Reviews in Environmental Control 13:273-309.

Reddy, K.R., E.M. D'Angelo and T.A. DeBusk. 1989. Oxygen transport through aquatic
macrophytes: the role in wastewater treatment. Journal of Environmental Quality
19:261-267.

Reddy, K.R., P.S.C. Rao, and R.E. Jessup. 1982. The effect of carbon mineralization on
denitrification kinetics in mineral and organic soils. Soil science society of America
journal 46:62-68.

Reddy, K.R., W.H. Patrick, and C.W. Lindau. 1989. Nitrification-denitrification at the plant root
sediment interface in wetlands. Limnology and Oceanography 34: 1004-1113.

Reed, S.C., R.W. Crites and E.J. Middlebrooks, 1995 Natural Systems for Waste Management
and Treatment (2nd ed.), McGraw Hill, New York .

Reinhardt, M., B. Muller, R.Gachter, B. Wehrli. 2006. Nitrogen removal in a small constructed
wetland: An isotope mass balance approach. Environmental science and technology
40:3313-3319.

Reinhardt, M., R.Gachter, B. Wehrli, and B. Muller. 2005. Phosphorus retention in small
constructed wetlands treating agricultural drainage water. Journal of Environmental
Quality 34: 1251-1259.

Risgaard-Petersen, N., and K. Jenson. 1997. Nitrification and denitrification in the rhizosphere of
the aquatic macrophyte Lobelia dortmanna L. Limnology and Oceanography 42:529-537.

Sartoris, J.J., J.S. Thullen, L.B. Barber, D.E. Salas. 2000. Investigation of nitrogen
transformations in a southern California constructed wastewater treatment wetland.
Ecological Engineering 14:49-65.









less than 0.4m) layer of surface water over mineral or organic soils. The persistent emergent

vegetation such as Typha spp.L. (Cattail), Phragmites spp. Adans. (Reed), and Scirpus spp. L.

(Bulrush) are commonly used, along with floating and submerged aquatic vegetation, and shrubs

and trees. In subsurface wetlands, the basin (less than 0.6m) is filled with coarse substrate such

as gravel, and the water level is maintained below ground with water flowing through the gravel

and the roots of vegetation (DeBusk, 1999). Both surface flow and subsurface flow constructed

wetlands have been reported to remove N and P from plant nursery runoff(Taylor et al., 2006).

A surface flow wetland was reported to remove 79-99% of nitrogen from the plant nursery

runoff (Taylor et al., 2006).

Nitrogen Cycling in Wetlands

Constructed wetlands are recognized as a means to improve water quality through N

removal. The maj or contributors of organic and inorganic N to the constructed wetlands include

biological N fixation and point and nonpoint source runoff, respectively. Organic N undergoes

mineralization/decomposition and results in the release of inorganic N forms. The processes

such as biological nitrification, denitrification, and volatilization result in the production of

inorganic N forms (ammonia, nitrate, nitrite and N2 gaS).

Most of the N transformations in the constructed wetlands are either plant or microbe

mediated. Nitrification is an aerobic process, in which oxidation of ammonium nitrogen (NH4 -

N) to nitrate nitrogen (NO3- N) is mediated by nitrifying bacteria. In contrast, denitrification

occurs in anaerobic conditions, whereby NO3- N is reduced to N2 gaS. There is growing

evidence that other microbe mediated processes such as Dissimilatory Nitrate Reduction to

Ammonium (DNRA) and anaerobic ammonium oxidation (anammox) facilitate N

transformations in wetlands (Stottmeister et al., 2003; Shipin et al., 2005; Reddy and Delaune,

2007; Burgin and Hamilton, 2007; Paredes et al., 2007).









sampling point as affected by plants and surrounding environment, water movement, etc.

Similar data variability in the denitrification rates between the replicates was observed by

Bastviken at al. (2003).

Regardless, the denitrification values obtained in this study are relatively high when

compared to several other studies (Table 4-2). Hunt et al. (2003) obtained higher denitrification

potential of 0.21 mg N kg-l hr- in Typha dominated wetland soils, as compared to 0.52 mg N kg

1 hr- in Juncus dominated wetlands used for treatment of swine wastewater. Denitrification

potential values observed in this study also lie in the range of denitrification potential (0.004 -

7.75 mg N kg-l hr- ) obtained for everglades wetland soils (White and Reddy, 1999).

Denitrification rates in soil samples from Juncus-planted wetland plots were reported to be

higher (12.3 mg N m-2 d- ) as compared to Salix plots (2.65 mg N m-2 d- ). The denitrification

rates obtained from soil samples amended with nitrate and/or glucose, indicated that the

denitrification was limited by the availability of carbon and NO3- N in the Salix plots, whereas

only by NO3- -N in the Juncus plots (Smialek et al 2006). Denitrification potential in rhizosphere

soil of Lolium perenne L.Melinda (Rye grass) was observed to be 150 900 ng N g-l hr- and

followed the pattern of plant growth, indicating that the plant growth and the size of denitrifier

population within the rhizosphere are likely correlated (Hall et al., 1998). Ottosen et al. (1999)

observed low denitrification activity in the rhizosphere of Zostera marina and Potamoge~~~~tttt~~~~tttto n

pectinatus (1.5 to 5 Cpmol N m-2 h- ) as compared to Lobelia dortmanna and Littorella uniflora

vegetated sediments (24 and 30 Cpmol N m-2 h- ).

Statistically similar denitrif ication potential in Canna flaccida and Typha latifolia in this

present study might be due to their similar effect on the rhizosphere. Other researchers have

observed differences in denitrification potential in wetland soil and sediments planted with










LIST OF REFERENCES


Aldrich, R.A., and J.W. Bartok. 1994. Greenhouse engineering. Natural Resource Agriculture
and Engineering Service (NRAES) Publication No. 33, Ithaca, NY.

Armentano, T.V., and J.T.A. Verhoeven. 1990. Wetland processes: Global biogeochemical
cycles. In: Patten, B.C. (ed.) Wetlands and shallow continental water bodies. Vol. 1. SPB
Academic Publishing, The Hague.

Ayaz, S.C., and L. Acka. 2001. Treatment of wastewater by natural systems. Environmental
International 26:189-195.

Bachand, P.A.M., and A.J. Home. 2000. Denitrification in constructed free water surface
wetlands: Effects of vegetation and temperature. Ecological Engineering 14:17-32.

Baker, L.A. 1998. Introduction to nonpoint source pollution in the United States and prospects
for wetland use. Ecological Engineering 1:1-26.

Bastviken, S.K., P.G. Eriksson, I. Martins, J.M. Neto, L. Leonardson, and K. Tonderski. 2003.
Potential nitrification and denitrification on different surfaces in a constructed treatment
wetland. Journal of Environmental Quality 32:2414-2420.

Bayley, M.L., L. Davidson, and T.R. Headley. 2003. Nitrogen removal from domestic effluent
using subsurface flow constructed wetlands: influence of depth, hydraulic residence time
and pre-nitrification. Water Science and Technology 48:175-182.

Berghage, R.D., E.P. MacNeal, E.F. Wheeler, and W.H. Zachritz. 1999. Green water treatment
for the green industries: opportunities for biofiltration of greenhouse and nursery
irrigation water and runoff with constructed wetlands. Hortscience 34:50-54.

Bemdt, M.P., E.T. Oaksford, M.R. Darst, and R.L. Marella. 1996. Environmental setting and
factors that affect water quality in the Georgia-Florida Coastal Plain: US Geological
Survey Water-Resources Investigations.

Black, R.J. 1993. Florida Climate Data. Florida Energy Extension Service Publication EES-5,
University of Florida, Gainesville. http://edis.ifas.ufl .edu.

Blahnik, T and J Day. 2000. The effects of varied hydraulic and nutrient loading rates in water
quality and hydrologic distributions in a natural forested treatment wetland. Wetlands
20:48-61.

Brix, H. 1997. Do macrophytes play a role in constructed treatment wetlands? Water Science and
Technology 35:11-17.

Burger, M., and L.E. Jackson. 2004. Plant and microbial nitrogen use and turnover: Rapid
conversion of nitrate to ammonium in soils with roots. Plant and soil 266:289-301.









et al. (2006) during their four year (2002 2005) monitoring of influent of the constructed

wetland at Monrovia Nursery. In addition, application of fertilizers at a nearby horticultural

nursery contributed to the increase in the concentration of NO3- N (0.03 to 4.73 mg L^1) through

runoff in the tributaries of Santa Fe river watershed in north central Florida (Frisbee, 2007).

While assessing the composition of TN during the study period, it was observed that

concentration of organic N at the outflow was higher as compared to inflow of the study cell,

which might be due to internal generation of organic N in the wetland during decay of plant and

microbes. Persistence of organic N in low concentrations was always observed in the outflow

water of a constructed wetland (Kadlec and Knight, 1996; Headley et al., 2001).

Water temperature in study wetland was observed to gradually increase from May to

August. The NO3--N removal efficiency was observed to gradually increase during the same

period. Nitrate nitrogen removal might have been enhanced by the increase in temperature,

which probably resulted in vigorous plant growth and activity of the microorganisms. The mean

temperature of the wetland (over all locations, depths and months) was 260 C. The favorable

range of temperature for denitrification is between 20 and 25 O C (Spieles and Mitsch, 2000),

which is close to the range of the mean temperature of the study cell. The influence of

temperature on nitrogen removal was explained by Reddy and Patrick (1984). The rates of

denitrification have been shown to increase by 1.5-2 times with each increase in 100C. Water

temperatures less than 150C or greater than 300C have been shown to limit the rate of

denitrification (Reddy and Patrick, 1984).

The gradual increase in NO3--N removal efficiency over April to August could also be

related to the growth of the plants in the wetland. The perennial plants in the wetland started

growth in April. During the vigorous growing stages (April, May, and June), plants utilize NO3 -










significant positive correlation was observed by Reddy et al. (1982) and Gale et al. (1993) for a

wide range of wetlands. The positive linear relationship indicates there is direct relationship

between denitrification and DOC, i.e. denitrification increases with increasing DOC

concentrations. Hence the denitrification rate is related to the rate of mineralization of C and

level of bio-available C to the denitrifying microbes (Reddy et al., 1982; Gale et al., 1993).

Martin and Reddy (1997) suggested that denitrification rates are limited by NO3- N

concentrations rather than C availability and by diffusion rates of NO3- N from aerobic to

anaerobic phase in wetlands. Many researchers observed positive correlation between NO3- N

concentrations and denitrification rates (White and Reddy 1999; Sartoris et al., 2000; Gale et al.,

1993; Poe et al., 2003). Responses of denitrification rates to changes in NO3- N concentrations

between 0.7 10.5 mg L^1 were assessed (by NO3- N addition experiments) and observed

significant correlation between nitrate concentration and denitrification rates (Poe et al., 2003).

The nutrient gradient from the inlet to outlet points and along the soil depth in a

constructed wetland is reported to influence the denitrification rates (White and Reddy, 1999;

White and Reddy, 2003; Sirivedhin and Gray, 2006). Increased denitrification rates were

observed at the inflow of the wetlands where there was high NO3- N concentration as compared

to the outflow, where there was low NO3- N concentration. The denitrification potential

decreased exponentially with increasing distance, due to decreasing NO3- N concentration, from

inflow point to the outflow point (White and Reddy, 1999). Similar results of increased

denitrification potential near the inlets and decreased denitrification potential in the middle and

outlet of the constructed wetlands were observed by Sirivedhin and Gray (2006). The higher

denitrification potential was observed in the surface soil, between 0-15 cm, (averaging 14.1 mg






































Figure 2-6. Inflow pipes into the study cell












1

I~

zi


I


Top


Figure 3-11. Mean concentrations of nitrate nitrogen (mg L 1) as affected by location of sample collection and depth within location,
between April and August 2007 in the constructed wetland receiving nursery runoff (n=18). Capital letters indicate
statistical difference between location (inlet and outlet) at top and bottom of the water column; lower-case letters indicate
statistical difference between depths (top and bottom) at inlet and outlet over five months. Means followed by the same
letter are statistically similar.


IInld
IC~tld


Bottom









CHAPTER 4
DISCUSSION

Concentration of nutrients in the inflow and outflow water. The mean concentration

(over all locations and months) of nutrients in plant nursery runoff entering the constructed

wetland were 34 mg L1 total nitrogen (TN) [30 mg L^1 nitrate nitrogen (NO3- N); 2 mg L^1

ammonium nitrogen (NH4 -N); and 2 mg L^1 organic nitrogen], 1 mg L1 total phosphorus (TP),

and 10 mg L^1 Dissolved Organic Carbon (DOC). The concentration of DOC, less than 20 mg L

1, is considered low as compared to other wastewater sources (Headley et al., 2001; Huett et al.,

2005; Crumpton et al., 1993). The nutrient removal efficiency of the constructed wetland was

observed to be 40% for TN, 40% for NO3- N, 59% for NH4 -N, and 16% for TP. The

constructed wetland systems in the US are reported to provide removal efficiencies of 66-95%

for TN and TP (Dortch, 1992), 30-60% for N (Hammer and Knight, 1994). Headly et al. (2001)

have reported > 84% TN and > 65% TP removal in the subsurface horizontal flow wetlands

receiving plant nursery runoff. Huett et al. (2005) obtained > 95% removal efficiency for both

TN and TP in a subsurface flow wetlands treating plant nursery runoff. Reinhardt et al. (2005)

reported 23% TP removal efficiency in constructed wetlands treating agricultural drainage water.

While comparing the results, the study cell appears to be functioning moderately efficiently in

removing TN and TP.

Mean removal efficiency for NO3- N of the constructed wetland over five months was

40%. Nitrate nitrogen removal efficiency of treatment wetlands in the US ranges from 30-60%

(Hammer and Knight, 1994). Poe et al. (2003) obtained 53% of NO3- N removal in a

constructed wetlands treating agricultural runoff. It is possible that a longer sampling period at

the study cell may have given different results.









60-


50_ A B

a A








0 "

x..e

Fiue3-.Mancnenrto (gL) fntrt iroe NO-N btenApi 00 n Ags 00 nth lw oto
chne n3,ilt(=)adteote n9 ftesuyceli h osrce eln eevn usr uof ro
bar iniaepu tnadero.Cptllte nictssaitcldfernebtenlctoswihnec ot Myt
Augst) loe-aeltesidct ttsia ifrnebtenm nh tec oain(C ,iltadote) en
followed by th a e etr r taitcal imlr










80
A

70


60-

C A
i~50 A b
sob

40 A EFCC
fB A
ZI B
3 0 -b c [nle
A
o\20 C A~ B
c: d~ e


to


Apdl May June July Angs
Months



Figure 3-4. Mean concentration (mg L ') of total nitrogen (TN) between April 2007 and August 2007 in the flow control channel
(n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetland receiving nursery runoff. Error bars
indicate plus standard error. Capital letter indicates statistical difference between locations within each month (May to
August); lower-case letters indicate statistical difference between months at each location (FCC, inlet and outlet). Means
followed by the same letter are statistically similar.









CHAPTER 1
INTTRODUCTION

Introduction

Globally, many aquatic ecosystems have experienced an impairment of water quality due

to excess nutrient loading from both point and nonpoint sources. Point source pollution refers to

pollutants that come from a definite, single identifiable source. Point source pollution has been

reduced effectively in recent years because point source pollutants are easy to identify, monitor,

and control as compared with nonpoint source pollutants. Due to the development of new

wastewater treatment technologies and implementation of stricter regulatory controls, point

source pollution has effectively been reduced in the United States.

Nonpoint source pollution generally results from land runoff, precipitation, atmospheric

deposition, drainage, seepage, or hydrologic modification [United States Environmental

Protection Agency (USEPA), 1993a]. Nonpoint source pollutants are nutrients, metals, salts,

sediments, pathogens, and toxics that come in unidentifiable runoff from agriculture, urban and

suburban stormwater, mining, and oil and gas operations [National Research Council (NRC),

1992]. It is difficult to identify, control, and regulate the nonpoint source pollution, as it comes

from vast and diverse landscapes, many diffuse sources, and varies by time of year. The ability

of these nonpoint source pollutants to reach waterbodies is enhanced by rainfall, snowmelt, and

irrigation. The nonpoint source pollutants have harmful effects on groundwater and surface

water resources, drinking water supplies, recreation, fisheries, and wildlife (USEPA, 1994). The

maj or sources of nonpoint source pollution are identified as agriculture and urban activities,

which includes industry and transportation (Carpenter et al., 1998). The lack of effective control

measures on polluted runoff especially from agricultural farms (including plant nurseries), urban









assimilation and sediment and/or soil denitrification. Rhizosphere denitrification potential in this

study was observed to significantly contribute towards NO3- -N removal.









A TMDL specifies the maximum amount of a pollutant that a water body can receive and

still meet water quality standards, and establishes pollutant loadings among point and nonpoint

pollutant sources (www.epa.gov, October 2007). The 1999 Florida Legislature adopted

comprehensive TMDL legislation, and concluded that the development of a TMDL program will

promote improvements in water quality throughout the state through the coordinated control of

point and nonpoint sources of pollution. "The scientifically based TMDL program was found to

be necessary to fairly and equitably allocate pollution loads to both nonpoint and point sources",

(Thomas and Joyner, 2002). In Georgia, over 850 TMDLs had been approved by the end of

2002 by Georgia Environmental Protection Division and USEPA. Most of the TMDLs in

Georgia, however, were established for fecal coliform bacteria (pathogens), metals, sediments,

and dissolved oxygen (http://pubs.caes.uga.edu/caespubs/pubs/PD/B242-2.pdf, October 2007).

Importance of Natural and Constructed Wetlands for Improving Water Quality

Wetlands and forested riparian buffers are most commonly utilized management strategies

for on-site nutrient removal from polluted waters generated from agricultural practices, including

plant nurseries (Franklin et al., 2000). Among these, wetlands can play an important role in

ecosystem management and can act as a sink for nutrients. Wetlands, however, often have been

destroyed for agricultural use of land and for urban development. In the United States alone,

approximately 54% of the original wetlands have been destroyed (Patrick, 1994). By the mid-

1980's, Florida had lost approximately 46% of its estimated original wetlands (Dahl, 1990).

Georgia had lost approximately 23% of its estimated original wetlands, by the mid 1980's (Dahl,

1990). However, in the past few decades, planned use of natural wetlands for treating

wastewater treatment has been seriously studied and implemented in a controlled manner

(USEPA, 1993b). Gren (1995) reported that wetland restoration is the most cost effective

method of decreasing nonpoint source pollution.









relationship between loading and removal rate of NO3- N. The NH4 -N and TP exhibited a

linear relationship with Pearson' s correlation coefficient (r) value of 0.75 (p<0.0001) and 0.60

(p<0.0001), respectively.

Determination of theoretical hydraulic residence time. The effect of hydraulic

loading rate and plant density on hydraulic residence time was computed by utilizing a model

and the results are shown in Figure 3-10. The residence time was calculated for two different

loading rates with three different plant density conditions, which provide three different wetland

porosity values. Hydraulic loading rate was inversely related to residence time; i.e., when the

loading rate increased from 1300 to 1734 m3day l, the theoretical residence time decreased by a

day for all three wetland porosity values. As the wetland porosity value increased from 0.75 to

0.95, the theoretical residence time increased at least by one day in this model. There was no

difference in the residence time when the wetland porosity value increased from 0.95 to 1.0 at

both hydraulic loading rates.

Spatial effects on nutrients, physiochemical parameters and denitrification potential of the

water column:

Nitrate Nitrogen (NO3--N). The concentration of NO3--N was not influenced

(p=0.0812) by an interaction among the treatments (location, depth, and months), presented in

table 3-2. All two-way interactions, depth~month, location~month, and location~depth

significantly affected the concentration of NO3--N at a=0.05 level. The interaction among the

location of sample collection (inflow and outflow) and the months during which samples were

collected, showed that the concentration of NO3--N changed from inflow to outflow except in

April. In April, the inflow concentration (44 mg L^1) of NO3--N was similar to that at outflow

(42 mg L^)~, but it was significantly higher than other means. There was no trend observed










Table 3-4. Denitrification potential (mg N20-N kg-' hr- ) as affected by macrophyte rhizosphere
soil in three sections (east, middle, and west) between May 2007 and August 2007 in
the constructed wetland receiving nursery runoff. (n =3). Capital letter indicates the
statistical difference between plant species within each section and month. Means
followed by the same letter are statistically similar.
Plant species Month Section Denitrification potential
mg kg-1 hr-1
Canna flaccida May East 1.9 A
Middle 2.4 A
West 4.1 A
June East 2.0 A
Middle 2.7 A
West 3.4 A
July East 1.0 A
Middle 0.5 A
West 3.1 A
August East 2.2 A
Middle 3.9 A
West 3.0 A
Typha latifolia May East 1.5 A
Middle 6.1 A
West 29.7 A
June East 2.1 A
Middle 0.2 A
West 0.1 A
July East 0.6 A
Middle 0.3 A
West 0.1 A
August East 6.1 A
Middle 0.3 A
West 2.3 A





















































---
-
C II -
4 % i~
~;~ ~~-C ~"
I P-
,crr,~
~s
:~
c
:. -.I
.t;-..
~ --..i
1..r
s~--~


Figure 2-3. Runoff generated from the nursery beds drains into the flow control channel




















































43


t-ir~i~lCj~brc-~' ~t~


~-Jss~_ ~;r~c

--I
-- .---=-~-C--~p,
r~Z
~c-L,


r~---~--~-- '




--









Table 3-1. Mean load-in, load-out, removal rate (g m-2 day l) of nutrients and percentage nutrient removal efficiency of the
constructed wetland over five months (April and August 2007). Standard deviation values in are presented in parentheses.
Total Nitrate Ammonium Total
Nitrogen (TN) Nitrogen (NO3-) Nitrogen (NH4 ) Phosphorus (TP)

(g m-2 day- ) (g m-2 day- ) (g m-2 day- ) (g m-2 day- )


Load-in rate
Load-out rate
Removal rate

% Removal


5.07 (2.64)
3.07 (2.50)
1.99 (0.15)


4.43 (2.21)
2.68 (2.18)
1.75 (0.75)


0.32 (0.21)
0.13 (0.14)
0.19 (0.11)


0.19 (0.05)
0.16 (0.04)
0.03 (0.03)













A





B A

BB c
o C A b n~
C cd d 1ea
b outlet

dl c eB







April My June July Angut




Figure 3-5. Mean concentration (mg L 1) of total phosphorus (TP) and Dissolved organic carbon between April 2007 and August
2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetland receiving
nursery runoff. Error bars indicate plus standard error. Capital letter indicates statistical difference between locations
within each month (May to August); lower-case letters indicate statistical difference between months at each location
(FCC, inlet and outlet). Means followed by the same letter are statistically similar.










depth of the water column is likely due to utilization of NO3--N by microorganisms, during the

process of denitrification in the anoxic zones. The high NH4 -N concentration in the lower depth

of the water column is likely due to the lack of nitrification process in the anoxic lower depth of

the water column, and hence the accumulation of NH4 -N. Higher concentration of TP in the

lower depth of the water column is likely due to release of adsorbed P from the sediments, or due

to lack of adsorption sites near the inflow of the study cell. The sorption sites on the inflow side

might have become saturated because of the higher concentration of TP at inflow. Concentration

of DOC did not show any differences in distribution along the depth of the water column, and

across the locations over months, indicated uniform distribution within the wetland.

Gradual increase in the temperature of the water column from May to August 2007 might

have enhanced NO3- N removal by affecting plant growth and activity of the microorganisms.

Denitrification potential values were observed to follow the pattern of plant growth, indicating

that the plant growth and the size of denitrifier population within the rhizosphere might be

coupled. Hence, physiological parameters affecting the plant growth can also indirectly affect

denitrification. A direct relationship was observed between the NO3- N removal efficiency and

temperature of the water column during the study period. As the temperature gradually

increased from May to August, the DO level gradually decreased. This is likely due to the

enhanced activity of the microorganisms utilizing oxygen as an electron acceptor. The higher

DO level at outflow can be attributed to open water towards the deeper side of the wetland. Due

to the potential differences in the temperature over months, DO level in the water column, and

concentration ofNO3- -N at inflow and outflow, it was expected that the denitrification potential

of the water column would be different as well. However, the low (0.01-0.03 mg N20-N L-1 hr- )

or undetectable denitrification potential of the water column is possibly due to the low DOC









observed by Taylor et al. (2006) in the constructed wetland at Monrovia nursery. The few

negative removal efficiency values of TP indicate that there is likely internal loading of

phosphorus. Macrophytes, algae, and microorganisms utilize P as an essential nutrient and can

enhance temporary removal of TP. Long term TP removal, however, is limited to the P sorption

capacity of the sediments (Kadlec, 1999).

The concentration of DOC did not follow a particular trend over months. The distribution

of DOC seemed uniform through the wetland in the study cell. The nursery runoff had a low

dissolved organic carbon (DOC) concentration (a characteristic of nursery runoff). The low

dissolved organic carbon (DOC < 20 mg L 1) was observed in the subsurface flow reed bed

systems receiving plant nursery and agricultural runoff(Headley et al., 2001; Huett et al., 2005;

Crumpton et al., 1993) in Australia. The DOC: NO3- N ratio was observed to be low (<3:1)

during most of the study period.

The modeled effect of hydraulic loading rate and plant density on theoretical residence

time was noticeable. When higher values of hydraulic loading rate were used in the model, the

theoretical residence time decreased by a day for this system. The model showed that increasing

the wetland porosity from 0.75 to 0.95 and/or 1.0 would increase the residence time in the study

cell by one day. Based on the model, the theoretical residence time for this system ranged from

3 to 5 days. Headley et al. (2001) reported that the maximum N removal can be achieved even at

2 day residence time in the subsurface flow reed beds. It was estimated that a nursery of 1 ha

area would require a reed bed area of 200 m2 for a 2 day hydraulic residence time.

Spatial effects on nutrients, physiochemical parameters and denitrification potential

of the water column. The higher DO level at outflow can be attributed to the higher prevalence

of open water towards the outflow side of the wetland compared to the inflow side. The depth of










Horner et al 1994 Horner, R.R., J.J. Skupien, H.L. Eric and H. E. Shaver. 1994. Fundamentals of
Urban Runoff Management: Technical and Institutional Issues, Terrence Institute.

Howarth, R. W., G. Billen, D. Swaney, A. Townsend, N.Jaworski, K. Lajtha, J. A. Downing, R.
Elmgren, N. Caraco,T. Jordan, F. Berendse, J. Freney, V. Kudeyarov, P. Murdoch, and
Zhu Zhao-liang. 1996. Regional nitrogen budgets and riverine inputs of N and P for the
drainages to the North Atlantic Ocean: natural and human influences. Biogeochemistry
35:75-139.

Huett, D.O., S.G. Morris, G. Smith, and N. Hunt. 2005. Nitrogen and phosphorus removal from
plant nursery runoff in vegetated and unvegetated subsurface flow wetlands. Water
Research.39:3259-3272.

Ingersoll, T.L., and L.A. Baker. 1998. Nitrate removal in wetland microcosms. Water
research.32:677-684.

Jerardo, A. 2005. Floriculture and nursery crops situation and outlook yearbook. FLO-2005,
Economic Research Service / USDA. Washington, DC.

Kadlec, R.H. 1999. The limits of phosphorus removal in wetlands. Wetlands Ecology and
Management 7: 165-175, 1999.

Kadlec, R.H., and R.L. Knight 1996 Treatment Wetlands. CRC Press, Boca Raton, FL, USA.

Kadlec, R.H., and K.R. Reddy. 2001. Temperature effects in treatment wetlands. Water
Environment Research 73:543-55.

Knowles, R. 1982. Denitrification. Microbiological Reviews 46:43-70.

Knowles, R. 1990. Acetylene inhibition technique: development, advantages and potential
problems. p 151-165 In Revsbech, N.P. and Sorensen, J. (eds.) Denitrifieation in Soil and
Sediment. Plenum Press, New York, NY, USA.

Kozu, D.D.and S.K. Lieh. 1999. Assessing denitrifieation rate limiting factors in a constructed
wetland receiving landfill leachate. Water Science and technology 40:75-82.

Kristiansen, S., and M.T. Schaanning. 2002. Denitrifieation in the water column of an
intermittently anoxic fj ord Hydrobiologia 469:77-86.

Lea-cox, J.D., and D.S. Ross. 2001. Clean water policy and the rationale for developing a water
and nutrient management planning process for container nursery and greenhouse
operations. Journal of environmental Horticulture 19: 1-8.

Lea-cox, J.D., D.S. Ross, A.G. Ristvey, and J.D. Murray. 2002. Estimating nitrogen and
phosphorus Total Maximum Daily Loads for container nursery and greenhouse
production systems. p 466-471.1In Total Maximum Daily Load (TMDL) Environmental
Regulations: Proceedings Conference Fort Worth, Texas, USA. 11-13 March, 2002.









Over the recent years, it has been realized that in addition to restoring the natural wetlands,

it is useful to construct artificial wetlands for cost effective, on-site remediation of polluted water

before discharging it into the natural aquatic systems. A number of studies have provided

evidence that constructed wetland systems provide an effective means to control the nonpoint

source pollution (including nutrient contaminants), thereby improving water quality (Mitsch,

1992; Hammer, 1992; Kadlec and Knight, 1996; Hoag, 1997; Headley et al., 2001; Poe et al.,

2003; Huett et al., 2005; Yang et al., 2007). Constructed wetlands are engineered systems that

differ from natural systems but are used to treat water by mimicking the processes occurring in

natural wetlands. These are designed to provide an ideal habitat for microbial communities for

achieving nutrient transformations in the contaminated water (Kadlec and Knight, 1996).

Wetland soil is highly saturated and exists in a chemically reduced state. The oxidized forms of

nutrients undergo transformations when it come in contact with saturated and reduced soil

conditions. Biological activity in the biofilms that attach to wetland soil and plants accounts for

much of the transformations of pollutants. Nitrogen transformations such as nitrification and

denitrification in the sediments were observed to occur rapidly in the aerobic/anaerobic

microsites in the wetlands (Kadlec and Knight, 1996).

Constructed wetlands are increasingly being used for treating N rich wastewater (Gersberg

et al., 1983; Reddy and D'Angelo, 1994; Bachand and Horne, 2000; Headley et al., 2001; Poe et

al., 2003; Huett et al., 2005; Reinhardt et al., 2006; Yang et al., 2007). Such systems have been

subjected to wastewater discharges from municipal, industrial, agricultural and surface runoff,

irrigation return flows, urban stormwater discharges, and other sources of polluted water. The

constructed wetland systems can be broadly classified into two main categories; surface flow and

subsurface flow systems. The surface flow wetlands are designed to have a shallow (typically










Lin, Y.F., S.R. Jing, D.Y. Lee, and T.W. Wang. 2002. Effect of macrophytes and external carbon
sources on nitrate removal from groundwater in constructed wetlands. Environmental
pollution 119:413-420.

Lu, W., and J.L. Sibley. 2006. Modeling of water requirements for container production using
overhead irrigation. p 518-520. In M. Taylor. (ed.) Water Management. Proceedings of
Research Conference, GA, 2006, Southern Nursery Association, Atlanta, GA.

Marella, R.L., and J.L. Fanning. 1996. National water quality assessment of the Georgia-Florida
Coastal Plain study unit water withdrawals and treated wastewater discharges, 1990.
US Geological Survey WRIR.

Martin, J.F., and K.R. Reddy. 1997. Interaction and spatial distribution of wetland nitrogen
processes. Ecological modeling 105:1-21.

Mitsch, W.J. 1992. Landscape design and the role of created, restored, and natural riparian
wetlands in controlling nonpoint source pollution. Ecological Engineering 1:27-47.

Mitsch, W.J., and J.G. Gosselink 2000. Wetlands. John and Wiley & Sons, Inc., New York.

Mitsch, W.J., and S.E. Jorgensen. 1989. Ecological engineering: An introduction to
ecotechnology. John and Wiley & Sons, Inc., New York.

Moorhead, K.K., and K.R. Reddy. 1988. Oxygen transport through selected aquatic
macrophytes. Journal of Environmental Quality 17:138-142.

National Research Council (NRC). 1992. Restoration of aquatic ecosystems Washington DC;
National Academy press.

National Water Quality Inventory (NWQI). 1998. http ://www.pollutionissues. com/Na-
Ph/Nonp oi nt- Source-P olluti on. html

National Resources Conservation Service (NRCS). 2002 Constructed wetlands. Environmental
Engineering National Engineering Handbook. USDA and NRCS, Washington DC.

Ottosen, L.D.M., N. Risgaard-Petersen, and L.P. Nielsen. 1999. Direct and indirect
measurements of nitrification and denitrification in the rhizosphere of aquatic
macrophytes. Aquatic Microbial Ecology 19:81-91.

Paredes, D., P. Kuschk, F. Stange, R.A.Muller, and H. Koser., 2007. Model experiments on
improving nitrogen removal in laboratory scale subsurface constructed wetlands by
enhancing the anaerobic ammonia oxidation. Water Science and Technology 56:145-
150.

Park, J.H., H.S. Shin, I.S. Lee, and J.H. Bae. 2002. Denitrification of High NO3--N Containing
Wastewater Using Elemental Sulfur; Nitrogen Loading Rate and N20 Production.
Environmental Technology 23:53-65.









different taxa of plants. These differences were attributed to differences in the ability of plant

roots to diffuse oxygen into the rhizosphere, and also due to differences in leakage of available

organic carbon from plant roots.

Most of the studies on denitrification potential have used the sediments and bulk soil in the

rhizosphere of the plants as substrate. Results from this study show that denitrification rates

comparable to or higher than those reported from bulk soil of vegetated plots can be obtained in

rhizosphere soil closely adhered to the roots.









concentration of the inflow (6 mg L^1) was significantly (p=0.0005) lower when compared to that

of outflow (11 mg L^)~. The DO was observed to significantly decrease from surface (11 mg L^1)

to bottom (6 mg L^1) of the water column. Although the overall analysis showed significant

difference between months, only the value for May (12 mg L^1) was significantly different from

rest of the months.

pH. The pH was not affected by interactions between locations, depth and months (Table

3-3), except for interaction among location and months (Fig. 3-13), and because of the main

effect of months. At inflow, the pH in August (4) was lower than in May (6), June (7) or July

(6). In August, the inflow pH (4) was lower than outflow pH (6). The other interactive means

were similar. When compared over months, mean pH in June was higher (7) than the values

obtained for July (6) and August (5). The mean pH of this system was 6 during the period of this

study .

Denitrification potential of water. The concentration of nitrous oxide nitrogen (N20-N)

produced per liter of water per hour (mg N20-N L-1 hr- ) was calculated and it was observed that

the N20-N production was either negligible (0.01 to 0.03 mg N20-N L-1 hr- ) or undetectable, as

was the case in most of the samples. The data, thus, are neither presented graphically, nor in a

tabular form.

Denitrification potential of macrophyte rhizosphere soil. The denitrification potential

of the rhizosphere soil was influenced (p=0.0419) by an interaction among macrophyte species,

sections, and months (Table. 3-4). The significant differences for the denitrification potential

were observed between very few means, however. The rhizosphere soil of Typha latifolia

produced significantly higher denitrification potential (30 mg N20-N kg-l hr- ) in the western

section of the cell during May, which was higher than the values obtained from the rest of the

















Storm twar

comna basi









statla Hohdag pend

een







Naltiescale


Figure 2-9. Locations of water samples (indicated by stars) in the study cell to assess the spatial variation in the concentration of
nutrients and denitrification potential of the water column, between locations (inflow and outflow) and between depths (top
and bottom) within locations over five months.









1.5:1, respectively. The Hyve month mean DOC: NO3--N ratio at the surface and bottom of the

water column were 1.0:1 and 1.3: 1, respectively.

Temperature. The temperature of the water column was affected (p=0.0141) by an

interaction among locations (inflow and outflow), depth within location (surface and bottom),

and months (April to August 2007). Although the overall model showed significant difference,

the mean comparisons showed similar values for the bottom and surface of the water column

temperature both at inflow and outflow during the period of this study (Table. 3-3). For bottom

of the water column on the inflow side, the effect of months on temperature was apparent as it

gradually increased from May (22 O C) to August (30 0 C). The temperature at lower depth of the

water column at outflow was also significantly affected by month of sampling. Surface water

temperature on the inflow side was lowest during May (23 O C) and highest (30 0 C) in August.

Similar trend over time was observed at the surface of the water column on the outflow side of

the study cell. There was significant difference between bottom (22 0 C) and surface (25 O C)

water column temperature at outflow during the month of May, but the temperature at surface

and lower depth of the water column was similar during other months. The Hyve month mean

temperature (averaged over surface and lower depth values) of the wetland at inflow (27 O C) was

higher than the outflow (26 O C). The Hyve month mean temperature at the surface of the water

column (averaged over inflow and outflow locations) was higher (26 O C) than the bottom (27 O C)

of the water column.

Dissolved Oxygen. The dissolved oxygen (DO) concentration in the water column was

not affected by any interaction effect (Table 3-3) between the treatments (locations, depth and

months). The main effects of location of sample collection, depth of the water column and

months showed significant difference at a = 0.05 level (Fig. 3-12). The dissolved oxygen










Xue, Y., D.A. Kovacic, M.B. David, L.E.Gentry, R.L.Mulvaney, and C.W. Lindau. 1999. In situ
measurements of denitrification in constructed wetlands. Journal of Environmental
Quality 28:263-269.

Yang, Q., Z.H. Chen, J.G. Zhao, and B.H. Gu. 2007. Contaminant Removal of Domestic
Wastewater by Constructed Wetlands: Effects of Plant Species. Journal of Integrative
Plant Biology 49:437-446.

Yeager, T.H., R.D. Wright, D. Fare, C. Gilliam, J. Johnson, T. Bilderback, and R. H. Zondag.
1993. Six state survey of container nursery nitrate nitrogen runoff. Journal of
Environmental Horticulture 11:206-208.










3-4 Mean concentration (mg L 1) of total nitrogen (TN) between April 2007 and August
2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study
cell of a constructed wetland receiving nursery runoff ........................... ...............68

3-5 Mean concentration (mg L 1) of total phosphorus (TP) and Dissolved organic carbon
between April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9)
and the outlet (n=9) of the study cell of a constructed wetland receiving nursery
runoff ................. ...............69.................

3-6 Mean concentration (mg L^1) of dissolved organic carbon (DOC) between April 2007
and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9)
of the study cell of a constructed wetland receiving nursery runoff ............... ... .........._..70

3-7 Pearson' s correlation between loading and removal rate of nitrate nitrogen (g m-2
day- ) between April 2007 and August 2007 in a constructed wetland receiving
nursery runoff ................. ...............71.................

3-8 Pearson's correlation between loading and removal rate of ammonium nitrogen (g m
2 day- ) between April 2007 and August 2007 in a constructed wetland receiving
nursery runoff ................. ...............72.................

3-9 Pearson' s correlation between loading and removal rate of total phosphorus (g m
2day- ) between April 2007 and August 2007 in a constructed wetland receiving
nursery runoff ................. ...............73.................

3-10 Theoretical hydraulic residence time (in days) as affected by hydraulic loading rate
(m3 day l) and three estimates of wetland porosity values (0.75, 0.95 and 1.0) in the
study cell of a constructed wetland receiving plant nursery runoff ................. ................74

3-11 Mean concentrations of nitrate nitrogen (mg L 1) as affected by location of sample
collection and depth within location, between April and August 2007 in the
constructed wetland receiving nursery runoff (n=18). ............ ...... ............... 7

3-12 Dissolved oxygen concentration (mg L 1) as affected by location of sample
collection, depth within location, and months (April and August 2007) in the
constructed wetland receiving nursery runoff (n=3) ................. ................. ..........76

3-13 Physiochemical parameter, pH, as affected by location of sample collection, and
months (April and August 2007) of the constructed wetland receiving nursery runoff
(n=3) ................. ...............77.................









ornamental nursery production requires intensive irrigation and fertilization for production of

marketable plants. While many growers use controlled release fertilizers (CRFs) incorporated

into the medium, liquid fertilizers are also applied through overhead irrigation systems at several

operations. The method of fertilizer application has been shown to influence the concentration

of nutrients in the runoff. The CRFs are observed to be suitable to minimize the concentration of

nutrients in the runoff. For example, NO3- N concentration was observed to be low (0.5 to 33

mg L^1, averaging 8 mg L^1) when CRFs were used, as compared to 0.1 to 135 mg L^1 (averaging

20 mg L^1) when a combination of CRFs and liquid fertilizers was used (Yeager et al., 1993).

The liquid fertilizers are typically used because CRFs are often more expensive and do not

provide the nutrients as readily as the liquid formulations. Nitrate fertilizers dissolve readily,

they are highly mobile in the water, hence leach readily, and are often found in the runoff water.

Compared to NO3- N, phosphorus is rapidly retained as insoluble inorganic compounds and

sorbed to soil surfaces. Overhead irrigation is the primary method of applying irrigation to

ornamental plants in small containers. The irrigation application efficiency is reported to be low,

ranging from 15% to 30% (Lu and Sibley, 2006). The overhead sprinkler irrigation of 1.8 x 105

L ha -1 yr -1 (Aldrich and Bartok, 1994) is estimated to produce 1.8 9 x 104 L ha-l yr- of

wastewater as runoff (Berghage et al., 1999) accounting for 10-50% loss as runoff of the

irrigation water. The irrigation practices lead to runoff of excessive fertilizers which find their

way into nearby water resources either through percolation in case of bare ground or through

runoff in case of plastic covered nursery beds. Due to high amount of rainfall and soil conditions

in Florida and Georgia, there is significant potential for discharge of large volume of runoff with

high nutrient concentration to nearby surface and groundwater resources. Hence runoff from

plant nursery poses a threat to water quality in Florida and Georgia. The restrictions/legi stations









N~ugen in atmosphere [IA)


NI6


Figure 1-1. Diagram illustrating potential pathways for chemical transformations of nitrogen (N) in a wetland.


16ol 16



















y= 0331x
R~ = -0.72


g
E


..


X


atra t~ea stralndin rate (g aa- ~)


Figure 3-7. Pearson' s correlation between loading and removal rate of nitrate nitrogen (g m-2 day l) between April 2007 and August
2007 in a constructed wetland receiving nursery runoff










(Giant Bulrush), Pontederia cordata L. (Pickerelweed), Sagittaria latifolia Willd. (Arrowhead),

and PanicuntPPP~~~~PPP~~~PPP hentitonzon J.A Schultes. (Maiden cane). Currently the primary cells are dominated

by S. validus, Typha latifolia L. (Cattail), C. flaccida, Hydrocotyle unabellata L. (Pennywort),

Lenana minor L. (Duckweed), and Wolffia spp. Horkel ex Schleid. (Water meal). There are few

open water areas in the primary cells and several floating islands of vegetation occur near the

outflow (South) side of the cell. The macrophytes originally planted in the shallow cells were C.

flaccida, S. latifolia, P. hentitonzon, and Juncus effusus L. (Soft Rush). Currently the shallow

secondary cells are predominantly covered by T. latifolia; and the other dominant species in the

shallow cells are S. latifolia, H. unabellata, and L. minor (Duckweed).

When it is necessary to discharge water from the HP into the wetland, it is pumped through

6 cm inner diameter PVC pipes (Fig. 2-6) to the primary cells 1A and 2A, at 9 locations and 8

locations respectively. The water from each primary cell passes under a road [through 15 cm

diameter PVC pipes; Fig. 2-7] and enters the secondary cell before being released into the

grassed waterway that carries the treated water to the stilling pools. The water flow from

primary to secondary cell and from secondary cell to the grassed waterway is based on gravity,

and there is no external control to regulate the water flow. However, the water depth (and

consequently the volume) of the primary cell can be controlled by adjusting the height of the

inflow pipe of the secondary cell. The stilling pools enhance the settling of suspended sediments

before discharging the treated water into the natural creek. The amount of water to be treated is

influenced by the rate and frequency of rainfall and irrigation. Based on the amount of water

available for treatment, four different 'run-times' (number of hours per day of operation of the

pump) are utilized; 0, 12, 18, 24 hours per day. The most frequently used run-times are 18 and

24 hours per day. During very low rainfall and minimal irrigation period, the run-time of 12









CHAPTER 5
SUMMARY AND CONCLUSIONS

Constructed wetlands are cost effective and viable method for on-site removal of nonpoint

source contaminants from water before it is released into aquatic systems. These systems are

used to treat wastewater from municipal water source, animal waste, urban runoff, stormwater

runoff, and agricultural runoff. Few studies have focused on constructed wetlands treating

nursery runoff (Headley et al., 2001; Lea-cox et al., 2002; Huett et al., 2005). In general, the

performance of constructed wetlands varies and partly depends on the source and the

characteristic features of influent wastewater (Mitsch and Jorgensen, 1989; NRCS, 2000a;

Dunne et al., 2005).

When constructed wetlands are to be used for nutrient, especially N, removal, management

strategies can be optimized by understanding the mechanisms by which N is removed within the

constructed wetlands. In view of the necessity of a 'systems approach', which recognizes the

site specific conditions, nitrogen (N) dynamics in the constructed wetland receiving plant nursery

runoff was investigated to facilitate improved understanding of N transformations, factors

influencing N removal, and to optimize the performance of the system for on-site N removal.

The concentration of nutrients N and phosphorus (P) entering and leaving the constructed

wetland was monitored from April 2007 to August 2007 to determine the nutrient removal

efficiency of the study cell in the constructed wetlands receiving plant nursery runoff. The

dominant nutrient in the plant nursery runoff was N, in the form of nitrate nitrogen (NO3--N),

followed by phosphorus (P). The concentration of DOC was low (< 20 mg L^1) which is a

characteristic feature of the plant nursery runoff

It was hypothesized that the concentration of N and P would vary over months based on

the timing of the fertilizer application, irrigation, and rainfall. The higher concentration of NO3~ -









hours or even 0 hours per day is used. The 12 and 18 hours per day run-times are achieved by

running the pump for 2 hours and then turning it off for the next 2 hours, or 3- hours 'on' and 1-

hour 'off' cycles, respectively, during a 24 hour period. The pump supplies water at an average

of 2,271 liters per minute, which is distributed to all 17 inflow pipes of the primary cells. Hence

the hydraulic loading rate of the wetland depends on the number of hours the pump is

operati onal .

Field Methods

Water sampling (for determining the nutrient removal efficiency of the constructed

wetlands). Water samples were collected from the study wetland cell at monthly intervals from

April 2007 to August 2007. Three replicate water samples per section of the wetland cell were

collected by taking grab samples in 125 ml plastic bottles (Fisher Scientific Co LLC, Suwannee,

GA.) at 9 inflow points (one sample from each of the 9 pipes) and 3 outflow points (three

replicate samples from each of the 3 pipes) as shown in Figure 2-8. In addition, three replicate

samples were collected at a location where the flow control channel (FCC) releases water into

the holding pond or into the stormwater retention basin. In summary, there were 9 inflow

samples, 9 outflow samples and 3 FCC samples, totaling 21 samples for the water nutrient

analysis. Samples were stored in a cooler filled with ice-packs, transported to the lab, and stored

at 4"C until further processing. Water samples were analyzed for Total Kj eldal Nitrogen (TKN),

ammonium nitrogen (NH4'-N), nitrate nitrogen (NO3--N), total phosphorus (TP), and dissolved

organic carbon (DOC).

Water sampling (for assessing the spatial effects on concentration of nutrients,

physiochemical parameters and denitrification potential of the water column). Nutrient

concentrations, physiochemical parameters, and denitrification potential were measured at select






































Figure 2-4. Runoff draining into the flow control channel









NITROGEN DYNAMICS IN A CONSTRUCTED WETLAND RECEIVING PLANT
NURSERY RUNOFF IN SOUTHEASTERN UNITED STATES





















By

BHUVANESWARI GOVINDARAJAN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2008









relatively low (0.01 to 0.03 mg N20-N L-1 hr- ), and could be due to carbon limitation in the

wetland water column as evident by low DOC concentration and low DOC: NO3--N ratio. The

presence of few attachment sites for bacteria and lack of particulate substances to support the

microbial activity possibly resulted in low denitrification rates in the water column. The mean

denitrification potential in the rhizosphere soil of Typha latifolia (3 mg N20-N kg -1 hr- ) and

Can2naflaccida (4 mg N20-N kg -1 hr- ) over five months were statistically similar. As the

combined denitrification potential of the water column (0.72 mg N20-N kg-l day- ) and the

rhizosphere soil (96 mg N20-N kgl day- ) in this study was less than the total removal (1.75 g

m-2 day )~, it is likely that other NO3--N processes are contributing to the observed NO3--N

removal in this constructed wetland. Plant uptake and microbial denitrification in the sediments

and/or soil are considered as maj or NO3--N removal mechanisms in constructed wetlands.

Hence, it is concluded that rhizosphere denitrification is significantly contributing to NO3- -N

removal in the constructed wetland receiving plant nursery runoff.










10

A
A
A








5 B FCC

a B A mnlakt
A
S3 C c
o\ tC B
2 -B a b
A
1~ Cc A
A C
b b

Alnil My June July Angut





Figure 3-3. Mean concentration (mg L 1) of ammonium nitrogen (NH4+ N) between April 2007 and August 2007 in the flow control
channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetland receiving nursery runoff. Error
bars indicate plus standard error. Capital letter indicates statistical difference between locations within each month (May to
August); lower-case letters indicate statistical difference between months at each location (FCC, inlet and outlet). Means
followed by the same letter are statistically similar.









lower (2 g m-2 day- ; August), the removal efficiency was higher (77%). The concentration of

NO3--N at FCC significantly differed from inflow concentration during the period of study,

except for April. The mean inflow NO3--N concentration over the study period was 30 mg L^1,

which was significantly higher (p<0.0001) than the outflow concentration of NO3--N, averaging

18 mg L^1

Ammonium Nitrogen (NH4 -N). The nursery runoff that enters the constructed wetland

over the study period consisted less than 1% of the TN in the form of NH4 -N. A significant

interaction (p < 0.0001) was observed among location of sample collection and month during

which samples were collected, for the concentration of NH4 -N (Fig. 3-3). At all locations

(inflow, outflow, and FCC), the NH4 -N concentration did not show any particular trend over

months, but the mean values were always higher at inflow and lower at outflow for each month

of this study. The average removal efficiency of NH4 -N was 59% and there was no trend

observed over months. There were significant differences in the concentration of NH4 -N

between FCC and inflow concentration over months, which indicate that the NH4 -N removal in

the holding pond was noticeable. The average concentration of the NH4 -N coming into the

wetland over the study period was 2. 1 mg L^1 which was higher (p<0.0001) than the

concentration of NH4 -N leaving the system, averaging 0.9 mg L^1

Total Nitrogen (TN). The TN was computed by summing the values of Total Kj eldal

Nitrogen (TKN) and NO3--N. The trend was similar to that of NO3--N. The concentration of the

TN was influenced (p<0.0001) by an interaction among location of sample collection and month

during which samples were collected (Fig. 3-4). The inflow concentration of TN gradually

decreased from April (60 mg L^1) to August (15 mg L^)~; similarly the concentration of TN at

outflow was also gradually decreased from April (49 mg L^1) to August (4 mg L^)~. At FCC, TN










crops only utilize 40-60% of N fertilizer applied to fields, and the remainder is incorporated into

soil organic matter, volatilized, denitrified, lost as runoff, or enters groundwater (Coffey, 1997).

Water Quality in Southeastern United States

In the Southeastern Regional Climate Assessment Report (1999), it has been reported that

in many cases, water quality indices are either below recommended levels or nearly so. The

maj or sources of nonpoint source pollution, including intense agricultural practices, urban

development, coastal processes, and mining activities, impair the quality of water in the

southeastern United States. In the southeastern region, especially in Georgia and Florida, the

natural features such as geology, climate, hydrology, and soils, significantly influence the

transport of nonpoint source pollutants from land to water. The soils of the Atlantic and Gulf

coastal plains include clays, loams, and large areas of gray, sandy soils. For example, in

loam/clay soils and sandy soils, the N export, as a percentage of fertilizer inputs, from

agricultural systems to adj acent waterbodies ranges from 10 to 40% and 25 to 80%, respectively

(Howarth et al., 1996). The transport of nutrients is also influenced by the rate, season, chemical

forms and method of application, amount and timing of rainfall after application, and vegetative

cover (Carpenter et al., 1998). In Florida, well drained sandy soils [underlain by gravel and

carbonate karstt) rocks] are susceptible to groundwater contamination due to low water holding

capacity, rapid infiltration and downward movement of water and chemicals. In contrast, the

areas with poorly drained clay soils are susceptible to runoff, which leads to stream

contamination, rather than groundwater contamination. Due to the poor draining capacity of the

clay soils, excessive irrigation or rainfall quickly drains to the adj acent streams as runoff (USGS,

1998).

The southern coastal plain of the US holds significant agricultural and hydrological

importance, with wide range of soil types and crop management systems. Land use in this region









between April and August, both in inflow and outflow concentration of NO3--N, but the

concentration during each month was higher at inflow than at outflow. The interaction among

location of sample collection (inflow and outflow) and depth within location (surface and bottom

of the water column) showed that surface (24 mg L^1) of the water column at inflow had higher

mean concentration of NO3--N than at the lower depth (21 mg L^1) at inflow and the surface (18

mg L^1) of outflow (Fig. 3-1 1). On the outflow side, the mean concentration of NO3--N over Hyve

months was higher (18 mg L 1) at the surface than the lower depth of the water column (16 mg L-

).The Hyve month mean concentration of NO3--N was always higher at surface water column as

compared to lower depth of the water column.

Ammonium Nitrogen (NH4 -N). The mean concentration of NH4 -N was influenced

(p=0.0001) by an interaction among location, depth, and months (Table. 3-2). Most of the means

were similar, however. The distribution of NH4 -N within the wetland did not vary. The

concentration of NH4 -N was also influenced by an interaction among location~month, and by

depth~month. Although both depth~month, and location~month interactions yielded some

significant means, there was no clear trend over time observed in the concentration ofNH4 -N.

At inflow, the concentration of NH4 -N ranged from 3.1 mg L^1 (April) to 0.2 mg L^1 (May). At

outflow the NH4 -N concentration ranged from 0.1 mg L^1 (May) to 1.8 mg L^1 (June).

Total Phosphorus (TP). The concentration of TP was influenced (p=0.0009) by an

interaction among location of sample collection, depth within locations, and month (Table 3-2),

however most of the means were similar. The absolute TP concentration at inflow and outflow

was higher at lower depth of the water column than that of the surface of the water column

between April to August. In addition to three-way interaction, location of sample

collection~depth was observed to highly affect TP concentration. At the inflow, the surface (2




Full Text

PAGE 1

1 NITROGEN DYNAMICS IN A CONSTR UCTED WETLAND RECEIVING PLANT NURSERY RUNOFF IN SOUTHEASTERN UNITED STATES By BHUVANESWARI GOVINDARAJAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Bhuvaneswari Govindarajan

PAGE 3

3 To every individual who nurtured my intellectual curiosity throughout my lifetime making this milestone possible

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor Dr. Jyotsna Sharma for her guidance, wisdom, and support, as well as my committee members, Dr s. K.R.Reddy, Mark W Clark, and Kanika S Inglett for their inputs on this project. Additiona l thanks to Dr. Jyotsna Sharma for all her time, especially during process of writing my thesis. I would like to thank Dr. Jim Ja witz for his valuable inputs throughout my research. I wish to thank Dr. Wilson for the statistic al inputs. I am also thankful to Dr. Dunne for his assistance and inputs. I thank everyone in the Wetland Biog eochemistry Laboratory for invaluable training and inputs during lab experiments. Finally, I would like to thank my friends and fa mily for their love a nd support. A special thanks to my husband, Subramaniam, for his patie nce and sacrifices thr ough the long years of graduate school, and for his support and help in all the endeavors I have chosen to pursue.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........6LIST OF FIGURES.........................................................................................................................7ABSTRACT.....................................................................................................................................9CHAPTER 1 INTRODUCTION..................................................................................................................11Introduction................................................................................................................... ..........11Water Quality in the United States......................................................................................... 12Water Quality in Southeastern United States......................................................................... 13Water Quality and Ornamental Plant Industry in Georgia and Florida.................................. 14Management Strategies to Control Nonpoint Source Pollution............................................. 17Importance of Natural and Constructed Wetlands for Improving Water Quality.................. 18Nitrogen Cycling in Wetlands................................................................................................ 20Denitrification in Wetlands.................................................................................................... .21Factors Influencing Den itrification in Wetlands.................................................................... 22The Problem Statement...........................................................................................................27Objectives and Hypotheses.....................................................................................................282 MATERIALS AND METHODS........................................................................................... 30Site Description......................................................................................................................30Field Methods.........................................................................................................................33Analytical Methods............................................................................................................. ....36Statistical Methods............................................................................................................ ......403 RESULTS...............................................................................................................................514 DISCUSSION.........................................................................................................................785 SUMMARY AND CONCLUSIONS.....................................................................................91LIST OF REFERENCES...............................................................................................................96BIOGRAPHICAL SKETCH.......................................................................................................105

PAGE 6

6 LIST OF TABLES Table page 3-1 Mean load-in, load-out, removal rate (g m-2 day-1) of nutrients and percentage nutrient removal efficiency of the constr ucted wetland over five months (April and August 2007). ...................................................................................................................613-2 Mean concentrations of nutrients (mg L-1) as affected by location of sample collection, depth within location, and months (April and August 2007) in the constructed wetland receiving nursery runoff (n=18). ..................................................... 623-3 Physiochemical parameters (temperature, dissolved oxygen, pH) as affected by location of sample collection, depth within location, and months (April and August 2007) in the constructed wetland rece iving nursery runoff (n=3). ..................................633-4 Denitrification potential (mg N2O-N kg-1 hr-1) as affected by macrophyte rhizosphere soil in three sections (eas t, middle, and west) between May 2007 and August 2007 in the constructed wetland receiv ing nursery runoff. (n =3).................................................. 644-1 Nitrate removal rates in constructed wetlands, under different hydraulic residence periods, as reported by other researchers........................................................................... 894-2 Denitrification potentials of different substrate reported by ot her researchers in constructed wetlands. ....................................................................................................... 90

PAGE 7

7 LIST OF FIGURES Figure page 1-1 Diagram illustrating potential pathways for chemical transformations of nitrogen (N) in a wetland. .......................................................................................................................292-1 Map of the state of Georgia, Unite d States, showing county boundaries. ....................... 412-2 Runoff generated from the microirriga ted nursery beds at Monrovia Nursery................. 422-3 Runoff generated from the nursery beds drains into the flow control channel.................. 432-4 Runoff draining into th e flow control channel...................................................................442-5 Overview of the Constructed wetlan ds at Monrovia Nursery, Cairo, GA. ......................452-6 Inflow pipes into the study cell..........................................................................................462-7 Outflow pipe of the study cell............................................................................................472-8 Location and number of water samples (ind icated by stars) at in flow and outflow of the study cell to determine the nutrient removal efficiency of the constructed wetlands.............................................................................................................................482-9 Locations of water samples (indicated by stars) in the study cell to assess the spatial variation in the concentration of nutrients and denitrification potential of the water column, between locations (inflow and outflow) and between depths (top and bottom) within locations over five months........................................................................ 492-10 Illustration of the study cell showing the location and number (indicated by stars) of monthly plant root samples to assess the denitrification potential in the rhizosphere soil of Canna flaccida and Typha latifolia ........................................................................503-1 Mean concentration (mg L-1) of total nitrogen (TN) a nd composition of TN [nitrate (NO3 N), ammonium (NH4 +-N) and organic nitrogen concentration mg L-1] between April to August of 2007 in the wate r samples obtained from the flow control channel (n=3), inlet (n=9), and outlet (n=9 ) pipes of a constructed wetland receiving nursery runoff................................................................................................................. ....653-2 Mean concentration (mg L-1) of nitrate nitrogen (NO3 --N) between April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell in the constructed wetl and receiving nursery runoff......................................... 663-3 Mean concentration (mg L-1) of ammonium nitrogen (NH4+ N) between April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed we tland receiving nursery runoff. ............................... 67

PAGE 8

8 3-4 Mean concentration (mg L-1) of total nitrogen (TN) between April 2007 and August 2007 in the flow control channel (n=3), in let (n=9) and the outlet (n=9) of the study cell of a constructed wetla nd receiving nursery runoff...................................................... 683-5 Mean concentration (mg L-1) of total phosphorus (TP) a nd Dissolved organic carbon between April 2007 and August 2007 in the flow control channel (n =3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetland receiving nursery runoff..................................................................................................................................693-6 Mean concentration (mg L-1) of dissolved organic car bon (DOC) between April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetland receiving nursery runoff.................................. 703-7 Pearsons correlation between loading a nd removal rate of nitrate nitrogen (g m-2 day-1) between April 2007 and August 2007 in a constructed wetland receiving nursery runoff................................................................................................................. ....713-8 Pearsons correlation between loading and removal rate of ammonium nitrogen (g m2 day-1) between April 2007 and August 2007 in a constructed wetland receiving nursery runoff................................................................................................................. ....723-9 Pearsons correlation between loading a nd removal rate of total phosphorus (g m2day-1) between April 2007 and August 2007 in a constructed wetland receiving nursery runoff................................................................................................................. ....733-10 Theoretical hydraulic residence time (in days) as affected by hydr aulic loading rate (m3 day-1) and three estimates of wetland poros ity values (0.75, 0.95 and 1.0) in the study cell of a constructed wetland receiving plant nursery runoff................................... 743-11 Mean concentrations of nitrate nitrogen (mg L-1) as affected by location of sample collection and depth within location, between April and August 2007 in the constructed wetland receiving nursery runoff (n=18). ..................................................... 753-12 Dissolved oxygen concentration (mg L-1) as affected by location of sample collection, depth within location, and months (April and August 2007) in the constructed wetland receiving nursery runoff (n=3).......................................................... 763-13 Physiochemical parameter, pH, as aff ected by location of sample collection, and months (April and August 2007) of the constructed wetla nd receiving nursery runoff (n=3).......................................................................................................................... .........77

PAGE 9

9 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Master of Science NITROGEN DYNAMICS IN A CONSTR UCTED WETLAND RECEIVING PLANT NURSERY RUNOFF IN SOUTHEASTERN UNITED STATES By Bhuvaneswari Govindarajan May 2008 Chair: Jyotsna Sharma Major: Interdisciplinary Ecology Constructed wetlands are a cost effective method for on-site nutrient removal from polluted water from nonpoint sources. Nitrogen (N) concentration and de nitrification potential were studied in a constructed wetland receiving plant nursery runoff over five months in southeastern US. Nursery runoff, before ente ring the wetland, contai ned an average total nitrogen concentration of 34 mg L-1. Majority of nitrogen was in the form of nitrate nitrogen (NO3 --N; 30 mg L-1) followed by ammonium nitrogen (NH4 +-N; 2 mg L-1) and organic nitrogen (2 mg L-1). Total phosphorus concentration was 1 mg L-1 and dissolved organic carbon was 11 mg L-1. Mean nutrient removal efficiency of th e constructed wetland was 40% for TN, 40% for NO3 --N, 59% for NH4 +-N, and 16% for TP. Nitrate nitroge n removal efficiency was inversely related to the loading rate of NO3 --N; and directly related to temperature of the water column. The gradual increase in temperature over months might have enhanced NO3 N removal by influencing plant growth and activ ity of microorganisms, which is also evidenced by the gradual decrease in the concentration of DO in the wate r column during the study period. The theoretical hydraulic residence time was estimated to be 3-5 da ys in the model. Th e nutrients (N and P) analyzed in this study exhibited nutrient gradient from the inflow to outflow indicating removal of these nutrients within the wetland. The denitrification potential of the water column was

PAGE 10

10 relatively low (0.01 to 0.03 mg N2O-N L-1 hr-1), and could be due to carbon limitation in the wetland water column as evident by lo w DOC concentration and low DOC: NO3 --N ratio. The presence of few attachment sites for bacteria a nd lack of particulate s ubstances to support the microbial activity possibly resulted in low denitrification rates in the water column. The mean denitrification potential in the rhizosphere soil of Typha latifolia (3 mg N2O-N kg -1 hr-1) and Canna flaccida (4 mg N2O-N kg -1 hr-1) over five months were statistically similar. As the combined denitrification potential of the water column (0.72 mg N2O-N kg-1 day-1) and the rhizosphere soil (96 mg N2O-N kg-1 day-1) in this study was less than the total removal (1.75 g m-2 day-1), it is likely that other NO3 --N processes are contributing to the observed NO3 --N removal in this constructed wetland. Plant uptake and microbial denitrification in the sediments and/or soil are considered as major NO3 --N removal mechanisms in constructed wetlands. Hence, it is concluded that rh izosphere denitrification is significantly contributing to NO3 -N removal in the constructed wetland receiving plant nursery runoff.

PAGE 11

11 CHAPTER 1 INTRODUCTION Introduction Globally, many aquatic ecosystems have experien ced an impairment of water quality due to excess nutrient loading from both point and nonpoint sources. Point source pollution refers to pollutants that come from a defi nite, single identifiable source. Point source pollution has been reduced effectively in recent years because poin t source pollutants are easy to identify, monitor, and control as compared with nonpoint source po llutants. Due to the development of new wastewater treatment technologies and implementation of strict er regulatory controls, point source pollution has effectively been reduced in the United States. Nonpoint source pollution generally results fr om land runoff, precipitation, atmospheric deposition, drainage, seepage, or hydrologic modification [United States Environmental Protection Agency (USEPA), 1993a]. Nonpoint s ource pollutants are nutrie nts, metals, salts, sediments, pathogens, and toxics that come in unidentifiable runoff from agriculture, urban and suburban stormwater, mining, and oil and gas operations [National Research Council (NRC), 1992]. It is difficult to identify, control, and re gulate the nonpoint sour ce pollution, as it comes from vast and diverse landscapes, many diffuse s ources, and varies by time of year. The ability of these nonpoint source pollutants to reach wate rbodies is enhanced by rainfall, snowmelt, and irrigation. The nonpoint source pollutants have harmful effects on groundwater and surface water resources, drinking water supplies, recrea tion, fisheries, and wildlife (USEPA, 1994). The major sources of nonpoint source pollution are iden tified as agriculture and urban activities, which includes industry and transportation (Carpenter et al., 1998). The lack of effective control measures on polluted runoff especially from agricultural farms (including plant nurseries), urban

PAGE 12

12 areas and forestry operations are the primary reas on for the lack of improvement in controlling nonpoint source pollution. Water Quality in the United States At the beginning of the twenty-first cent ury, nonpoint source pollu tion stands as the primary cause of water quality impairment with in the United States. According to the USEPA (2002a), nonpoint source pollution is the main reas on that approximately 39% of surveyed rivers, 45% of lakes and 51% of estuaries are not clean enough to meet basic uses such as fishing or swimming. The report from National Coastal Co ndition (USEPA, 2002b) projected that the 70% of the US estuary conditions would worsen by 2020 due to eutrophicat ion caused by nonpoint source pollutants. The National Water Quality Assessment (NAWQA) program [United States Geological Survey (USGS), 1998] of the US assessed the quality of water resources, especially streams, river basins, groundwater, and aquifer systems in the Unites States from 1991-2001 and reported that the contamination of streams and groundwater is widespread in agricultural and urban areas due to nonpoint source pollutants. States and other jurisdictions reported in th e National Water Quality Inventory (NWQI, 1998) that agriculture and urban runoff are among the leading contributors to deteriorate water quality nationwide. The most common nonpoint source pollutants cau sing water quality impairment include nutrients [nitrogen (N) and ph osphorus (P)], pesticides, chemicals, siltation (soil particles), metals, and pa thogens (bacteria and viruses). According to Faeth (2000), agriculture is the largest source of pollution in the US degrading the quality of surface waters such as rivers and lakes, with croplands alone accounting for nearly 40 % of the N and 30% of the P pollution. A major nonpoint so urce pollutant from these activit ies is an excess of nutrients, which can occur through applications of crop fertilizers. Fo r example, in agricultural systems,

PAGE 13

13 crops only utilize 40-60% of N fer tilizer applied to fields, and the remainder is incorporated into soil organic matter, volatilized, de nitrified, lost as runoff, or enters groundwater (Coffey, 1997). Water Quality in Southeastern United States In the Southeastern Regional Climate Assessment Report (1999), it has been reported that in many cases, water quality indices are either below recommended levels or nearly so. The major sources of nonpoint source pollution, including intense agricultural practices, urban development, coastal processes, and mining activities, impair the quality of water in the southeastern United States. In the southeastern region, especially in Georgia and Florida, the natural features such as geology, climate, hydrology, and soils, significantly influence the transport of nonpoint source pollutants from land to water. The soils of the Atlantic and Gulf coastal plains include clays, loams, and larg e areas of gray, sandy soils. For example, in loam/clay soils and sandy soils, the N export, as a percentage of fertilizer inputs, from agricultural systems to adjacent waterbodies rang es from 10 to 40% and 25 to 80%, respectively (Howarth et al., 1996). The transpor t of nutrients is also influenced by the rate, season, chemical forms and method of application, amount and timing of rainfall after app lication, and vegetative cover (Carpenter et al ., 1998). In Florida, well drained sandy soils [underlain by gravel and carbonate (karst) rocks] are susceptible to groundwater cont amination due to low water holding capacity, rapid infiltration and downward movement of water and chemicals. In contrast, the areas with poorly drained clay soils are susceptible to r unoff, which leads to stream contamination, rather than groundwater contamina tion. Due to the poor draining capacity of the clay soils, excessive irrigation or rainfall quickly drains to the adjacent streams as runoff (USGS, 1998). The southern coastal plain of the US holds significant agricult ural and hydrological importance, with wide range of soil types and crop management systems. Land use in this region

PAGE 14

14 is approximately 69% woodland, 17% cropland, 11% pastureland, a nd 3% urban (Berndt et al., 1996). According to the USGS (1998) report on water quality in the Ge orgia-Florida coastal plain during 1992-1996; more than 20% of groundwater samples from the surficial aquifers in the agricultural areas throughout the study area had nitrate nitrogen (NO3 N) concentrations greater than the USEPA drinki ng water standard of 10 mg L-1. The concentrations of other nutrients in groundwater includ ing ammonia, orthophosphate and total phosphorus were low. Nitrogen concentrations (as NO3 N) in streams did not exceed drinking water standards or guidelines, but were higher in streams draining ag ricultural and mixed basi ns. Nearly 30% of the water samples from the streams had dissolved phos phorus concentrations greater than the aquatic criteria (0.1 mg L-1; USEPA, 1986) for total phosphorus. Nearly 80% of the streams in agricultural areas had phosphorus concentrations greater than the USEPA st andard to prevent algal growth (USGS 1998). The study area, situated in the Georgia-Florida coastal plain, overlies the Upper Floridan aquife r, which is the major source of drinking water for that area. The unconfined Upper Floridan aquifer has karst features, which is vuln erable to groundwater contamination, similar to that of sandy soils (USGS, 1998). Water Quality and Ornamental Plant Industry in Georgia and Florida Florida is one of America's leading agricultu ral states which produces a wide range of commodities. The state has a humid subtropical c limate with mean rainfall of up to 140 cm per year, mean monthly temperat ure ranging from 4 to 33 C during a calendar year, and is prone to hurricanes (Black, 1993). Florida has the largest total acreage of Aquods, (wet sandy soils with an organic stained subsoil layer; www.nr cs.gov, November 2007), are susceptible to groundwater contamination due to poor water hold ing capacity. Hence, pollution of surface and groundwater resources resulting from agricultural runoff, urban stormwater runoff, and erosion

PAGE 15

15 sedimentation is a significant pr oblem in Florida, where approximately one-third of landforms are wetlands, and drinking water is mostly dr awn from underground aquifers which are close to the land surface (Marella and Fanning, 1996). The generally temperate climate in Georgia is influenced by the proximity to the Atlantic Ocean and to the Gulf of Mexico. The coastal plains, underlain by sand and limestone, occupy over 60% of the state. Because of the sandy so il and flat topography in the coastal areas, infiltration is a larger concern ra ther than runoff. Nearly one third of the state is underlain by sandy/clay loam soils, which support variety of agricultural crops. The state typically receives 100-125 cm of rainfall every year (www.georgi a.org; www.coastgis.marsci.uga.edu, October 2007). Florida is the second leading producer of orna mental plants in the US. The greenhouse and nursery industry comprises 50,992 hectares of production area in Florida (Hodges and Haydu, 2002). According to the United States Department of Agriculture (USDA) estimates, the value of nursery and greenhouse crops in Florida was approximate ly $1.63 billion in 2004 (Jerardo, 2005). Georgia is among the top ten producers of floriculture cr ops, ranking tenth in the US. The value of nursery and greenhouse crops in Georgia was approximately $400 million in 2006 [National Agricultural Statistics Service ( NASS), http://www.nass.us da.gov/, accessed on October, 2007]. While greenhouse crop production generates sign ificant runoff, this problem is most noticeable in the container nursery industry wher e large numbers of containerized plants are grown outdoors. Generally, for prod uction of containerized plants outdoors, plants are grown in containers placed on gravel beds which are lined with plastic to prev ent weed growth, and irrigation is achieved through ove rhead sprinklers (USDA, 1998a; Lea-cox and Ross, 2001). The

PAGE 16

16 ornamental nursery production requires intensive irrigation and fertilization for production of marketable plants. While many growers use cont rolled release fertilizers (CRFs) incorporated into the medium, liquid fertilizers are also app lied through overhead irrigation systems at several operations. The method of fertiliz er application has been shown to influence the concentration of nutrients in the runoff. The CRFs are observed to be suitable to minimize the concentration of nutrients in the runoff. For example, NO3 N concentration was observed to be low (0.5 to 33 mg L-1, averaging 8 mg L-1) when CRFs were used, as compared to 0.1 to 135 mg L-1 (averaging 20 mg L-1) when a combination of CRFs and liquid fertilizers was used (Yeager et al., 1993). The liquid fertilizers are typically used because CRFs are often more expensive and do not provide the nutrients as readily as the liquid formulations. Nitr ate fertilizers dissolve readily, they are highly mobile in the water, hence leach readily, and are often f ound in the runoff water. Compared to NO3 N, phosphorus is rapidly retained as insoluble inorganic compounds and sorbed to soil surfaces. Overhead irrigation is the primary method of applying irrigation to ornamental plants in small containe rs. The irrigation application efficiency is reported to be low, ranging from 15% to 30% (Lu and Sibley, 2006). The overhead sprinkler irrigation of 1.8 x 105 L ha -1 yr -1 (Aldrich and Bartok, 1994) is estimated to produce 1.8 9 x 104 L ha-1 yr-1 of wastewater as runoff (Bergha ge et al., 1999) accounting for 10-50% loss as runoff of the irrigation water. The irrigation practices lead to runoff of excessive fertilizers which find their way into nearby water resources either through percolation in case of bare ground or through runoff in case of plastic covered nursery beds. Due to high amount of rainfall and soil conditions in Florida and Georgia, there is significant potential for discharg e of large volume of runoff with high nutrient concentration to nearby surface a nd groundwater resources. Hence runoff from plant nursery poses a threat to wa ter quality in Florida and Georgia. The restrictions/legislations

PAGE 17

17 such as Watershed Restoration Act and Total Maximum Daily Load (TMDLs), imposed by the environmental protection agencies against c ontamination of water resources necessitate recycling/treatment of nursery r unoff before discharging into near by waterbodies (Headley et al., 2001; Lea-cox et al., 20 02; Huett et al., 2005). Management Strategies to Control Nonpoint Source Pollution The control of nonpoint sour ce pollution centers on adoption of Best Management Practices (BMPs) to manage land and to control the release of pollutants into the atmosphere; further, establishment of threshold levels of contaminants (e.g., TMDLs) is also a strategy for limiting nonpoint source pollution (Carpenter et al., 1998). Best Management Practices can be groupe d into two categories; structural and nonstructural (Horner et al ., 1994). Non structural BMPs include practices such as preservation of natural areas and drainage systems, land us e methods, and efficient use of fertilizers and irrigation water. Structural BMPs include esta blishment of detention/ retention and recycling ponds, grass swales, filter stri ps, riparian buffer strips, fl oodplains, and wetlands. The commonly used structural BMPs are retention an d recycling ponds, riparian buffers strips, and grass swales to control nonpoint source pollution from agricultur al operations, including plant nurseries. In most of these methods, the prim ary mechanisms to remo ve pollutants are to enhance settling of the particulates and to enhanc e infiltration into the subsurface zones. Many container plant nurseries, where plants are grown outdoors on grav el beds lined with plastic, construct recycling ponds to recycle the runoff from nursery beds (Lu and Sibley, 2006). The other commonly followed BMPs to control the ru noff water from container plant nurseries are adjusting the time and rate of fertilizer applicat ion to coincide with the crop needs, using microirrigation practices to minimize runoff volume, and use of controlled release fertilizers to reduce NO3 N leaching from containers.

PAGE 18

18 A TMDL specifies the maximum amount of a pollutant that a water body can receive and still meet water quality standards, and establishes pollutant loadings among point and nonpoint pollutant sources (www.epa.gov, October 2007). The 1999 Florida Legislature adopted comprehensive TMDL legislation, and concluded that the development of a TMDL program will promote improvements in water quality throughout the state through the c oordinated control of point and nonpoint sources of pollution. The scientifically based TMDL program was found to be necessary to fairly and equitably allocate pollution loads to both nonpoint and point sources, (Thomas and Joyner, 2002). In Georgia, ove r 850 TMDLs had been approved by the end of 2002 by Georgia Environmental Protection Divisi on and USEPA. Most of the TMDLs in Georgia, however, were established for fecal co liform bacteria (pathogens), metals, sediments, and dissolved oxygen (http://pubs.caes.uga.edu/caespubs/pubs/PDF/B1242-2.pdf, October 2007). Importance of Natural and Constructed Wetlands for Improving Water Quality Wetlands and forested riparian buffers are most commonly utilized management strategies for on-site nutrient removal from polluted waters generated from agricultural practices, including plant nurseries (Franklin et al., 2000). Among these, wetlands can play an important role in ecosystem management and can act as a sink for nutrients. Wetlands, however, often have been destroyed for agricultural use of land and for urban development. In the United States alone, approximately 54% of the origin al wetlands have been destroyed (Patrick, 1994). By the mid1980's, Florida had lost approximately 46% of its estimated original wetlands (Dahl, 1990). Georgia had lost approximately 23% of its esti mated original wetlands, by the mid 1980s (Dahl, 1990). However, in the past few decades, pl anned use of natural wetlands for treating wastewater treatment has been seriously studied and implem ented in a controlled manner (USEPA, 1993b). Gren (1995) repo rted that wetland restoration is the most cost effective method of decreasing nonpoint source pollution.

PAGE 19

19 Over the recent years, it has b een realized that in addition to restoring the natural wetlands, it is useful to construct artificial wetlands for cost effective, on-site remediation of polluted water before discharging it into the natural aquatic systems. A number of studies have provided evidence that constructed wetland systems provide an effective means to control the nonpoint source pollution (including nutrient contaminants ), thereby improving water quality (Mitsch, 1992; Hammer, 1992; Kadlec and Kn ight, 1996; Hoag, 1997; Headley et al., 2001; Poe et al., 2003; Huett et al., 2005; Yang et al ., 2007). Constructed wetlands are engineered systems that differ from natural systems but are used to tr eat water by mimicking th e processes occurring in natural wetlands. These are designed to provide an ideal habitat for microbial communities for achieving nutrient transformations in the contaminated water (Kadlec and Knight, 1996). Wetland soil is highly saturated and exists in a chemically reduced state. The oxidized forms of nutrients undergo transformations when it come in contact with saturated and reduced soil conditions. Biological activity in the biofilms that attach to wetland soil and plants accounts for much of the transformations of pollutants. Nitr ogen transformations such as nitrification and denitrification in the sediments were observe d to occur rapidly in the aerobic/anaerobic microsites in the wetlands (Kadlec and Knight, 1996). Constructed wetlands are increasi ngly being used for treating N rich wastewater (Gersberg et al., 1983; Reddy and DAngelo, 1994; Bachand and Horne, 2000; Headley et al., 2001; Poe et al., 2003; Huett et al., 2005; Reinhardt et al., 20 06; Yang et al., 2007). Such systems have been subjected to wastewater discharg es from municipal, industrial, agricultural and surface runoff, irrigation return flows, urban st ormwater discharges, and other sources of polluted water. The constructed wetland systems can be broadly classifi ed into two main categories; surface flow and subsurface flow systems. The surface flow we tlands are designed to have a shallow (typically

PAGE 20

20 less than 0.4m) layer of surface water over mineral or organic soils. The persistent emergent vegetation such as Typha spp.L. (Cattail), Phragmites spp. Adans. (Reed), and Scirpus spp. L. (Bulrush) are commonly used, al ong with floating and submerged aquatic vegetation, and shrubs and trees. In subsurface wetlands, the basin (less than 0.6m) is filled with coarse substrate such as gravel, and the water level is maintained below ground with water fl owing through the gravel and the roots of vegetation (DeBusk, 1999). Both surface flow and subsurface flow constructed wetlands have been reported to remove N and P from plant nursery runoff (Taylor et al., 2006). A surface flow wetland was reported to remove 79-99% of nitrogen from the plant nursery runoff (Taylor et al., 2006). Nitrogen Cycling in Wetlands Constructed wetlands are rec ognized as a means to improve water quality through N removal. The major contributors of organic an d inorganic N to the constructed wetlands include biological N fixation and point and nonpoint s ource runoff, respectively. Organic N undergoes mineralization/decomposition and results in the re lease of inorganic N forms. The processes such as biological nitrificati on, denitrification, and volatilizati on result in the production of inorganic N forms (ammonia, nitrate, nitrite and N2 gas). Most of the N transformations in the constr ucted wetlands are eith er plant or microbe mediated. Nitrification is an aerobic process, in which oxida tion of ammonium nitrogen (NH4 +N) to nitrate nitrogen (NO3 N) is mediated by nitrifying bacteria. In contrast, denitrification occurs in anaerobic conditions, whereby NO3 N is reduced to N2 gas. There is growing evidence that other microbe mediated processe s such as Dissimilatory Nitrate Reduction to Ammonium (DNRA) and anaer obic ammonium oxidation (anammox) facilitate N transformations in wetlands (Stottmeister et al., 2003; Shipin et al., 2005; Reddy and Delaune, 2007; Burgin and Hamilton, 2007; Paredes et al., 2007).

PAGE 21

21 The temporary removal of N in wetlands occurs through plant uptake and microbial immobilization, whereas, permanent removal of N o ccurs through denitrification (Fig. 1-1). In wetlands, biological denitrificat ion is the major pathway of N removal (Reddy et al., 1989; Tanner and Kadlec, 2003; Poe et al., 2003; Toet et al., 2003; Smia lek et al., 2006). At high pH (typically higher than 8.0) volatil ization of ammonia also contribut es to N loss to the atmosphere (Reddy and Patrick, 1984). Other processes such as sedimentation, adsorption, uptake by plants and microbes lead to N storage in the wetlands (Fig. 1-1). Denitrification in Wetlands Denitrification is the stepwise reduction of nitrogenous oxides (NO3 and NO2 -) to produce gaseous N products (N2O and N2). These gases are released into the atmosphere, thus resulting in a removal of N from the wetland. It is a microbe mediated process in which facultative anaerobic bacteria ( Pseudomonas spp., Alcaligenes spp., Flavobacterium spp., Paracoccus spp., and Bacillus spp.) use nitrate (NO3 -) as the terminal electron acceptor during the oxidation of carbon (C) in the absence of oxygen (O2) (Tiedje, 1988). The low redox potential and high carbon content in the wetlands favors the process of denitrification. Gl obally, N loss due to denitrification process is estimated as 19 Tg yr-1 (Armentano and Verhoeven, 1990). The equation and the chemical transformations of denitrification pro cess are as follows 2NO3 + 10e+ 12H+ N2 + 6H2O Nitrate (NO3 -) Nitrite (NO2 -) Nitrous oxide (N2O) Nitrogen (N2) The microbe mediated denitrif ication process was observed to be the dominant mechanism for N removal in the constructed wetlands (X ue et al., 1999; Bachand and Horne 2000). Hence the quantification of de nitrification potential is critical. The commonly used methods are mass balance, Acetylene Inhi bition Technique (AIT), 15N tracers, and Membrane Inlet Mass Spectrometry (MIMS). Due to the differe nces in the methodology, the comparison of

PAGE 22

22 denitrification rates between the studies is often challenging (Seitzi nger, 1993). The AIT method is one of the most commonly used technique s for measuring denitrification rates. This technique measures the denitr ification rates under ambient anaerobic conditions, in which acetylene (C2H2) inhibits the reduction of N2O to N. The extent of de nitrification is quantified by the progressive accumulation of N2O. This is also known as Denitrification enzyme Assay (DEA), which is an indicator of denitrifier biomass present in the samples and serves as an integrated measure of de nitrification potential. Factors Influencing Denitrification in Wetlands Denitrification in wetland systems is primar ily controlled by organic carbon availability (Reddy et al., 1982; Gale et al., 1993; DAnge lo and Reddy, 1999), aeration status (dissolved oxygen levels; Tanner and Kadlec, 2003), and NO3 N concentration (Cooper and Findlater, 1990; Gale et al., 1993; Martin and Reddy, 1997). These factors are highly influenced by temporal (diurnal and seasonal) and spatial (horizontal and vert ical) variation (White and Reddy, 1999; Kadlec and Reddy, 2001). The other factors that influence denitrification rates in wetlands are the make-up of the vegetative and microorganismal communities (Gersberg et al., 1986; Sirivedhin and Gray, 2006; Smialek et al., 2006), depth of the water column (Sirivedhin and Gray, 2006), hydraulic residence time, and pH of the water and soil (Reddy and Patrick, 1984). The major regulator of denitrific ation is the oxygen status of th e wetland, as this process is facilitated by facultative anaerobic bacteria. Deni trification is observed to take place only in low oxygen zones of the wetlands (Knowles, 1982; Knowles, 1990; Tiedje, 1988). In anoxic conditions, denitrificati on is controlled by NO3 N and carbon availability. The importance of organic carbon as an electr on donor in the denitrific ation process is well documented (Eriksson and Weisner, 1997; Lin et al., 2002; Bastviken et al., 2003). While studying the relationship between denitrificati on rates and dissolved organic carbon (DOC),

PAGE 23

23 significant positive correlation was observed by Re ddy et al. (1982) and Gale et al. (1993) for a wide range of wetlands. The positive linear rela tionship indicates there is direct relationship between denitrification and DOC, i.e. denitrification increases with increasing DOC concentrations. Hence the denitrif ication rate is related to the rate of mineralization of C and level of bio-available C to the denitrifying mi crobes (Reddy et al., 1982; Gale et al., 1993). Martin and Reddy (1997) suggested that denitrification rates are limited by NO3 N concentrations rather than C ava ilability and by diffusion rates of NO3 N from aerobic to anaerobic phase in wetlands. Many researchers observed positive correlation between NO3 N concentrations and denitrification rates (White and Reddy 1999; Sartoris et al., 2000; Gale et al., 1993; Poe et al., 2003). Responses of de nitrification rates to changes in NO3 N concentrations between 0.7 10.5 mg L-1 were assessed (by NO3 N addition experiments) and observed significant correlation between nitrate concentration and denitrification rates (Poe et al., 2003). The nutrient gradient from the inlet to outlet points and along the soil depth in a constructed wetland is reported to influence th e denitrification rates (White and Reddy, 1999; White and Reddy, 2003; Sirivedhin and Gray, 2006) Increased denitrif ication rates were observed at the inflow of the wetlands where there was high NO3 N concentration as compared to the outflow, where there was low NO3 N concentration. The denitrification potential decreased exponentially w ith increasing distance, due to decreasing NO3 N concentration, from inflow point to the outflow point (White and Reddy, 1999). Similar results of increased denitrification potential near th e inlets and decreased denitrifi cation potential in the middle and outlet of the constructed wetla nds were observed by Sirivedhin and Gray (2006). The higher denitrification potential was obs erved in the surface soil, between 0-15 cm, (averaging 14.1 mg

PAGE 24

24 N2O-N kg-1 h-1) as compared to 0.51 mg N2O-N kg-1 h-1 in the underlying soil, between 10-30 cm, of the wetlands. The residence time of water in a wetland is cr itical for removal of N because it affects the duration of the contact between polluted water a nd the biotic and abiotic components such as plant roots, soil, sediments, and microbes. Th e longer residence time ma y enhance removal of N by promoting retention and biochemical processes such as denitrificati on, sedimentation, and plant uptake (Kadlec and Knight, 1996; Mitsch and Gosselink, 2000). Kadlec and Knight (1996) also showed that the N removal efficiency is logarithmically related to the residence time. Temperature is a significant abio tic factor that influences micr obial activities. Wood et al. (1999) reported that the temperature influences the biological and phys ical activities in the wetlands, and denitrification is a temperature dependent process. Th e rates of denitrification have been shown to increase by 1.5-2 times with each increase in 10oC (Reddy and Patrick, 1984). The denitrification rates were higher (9.2 mg N m-2 h-1) in su mmer and lower (0.7 mg N m-2 h1) in fall in the constructed wetlands (Poe et al., 2003); and similarly Xue et al.(1999) reported higher denitrification rates in summer (11.8 mg N m-2 h-1) and lower (2.0 mg N m-2 h-1) in winter. White and Reddy (1999) reported hi ghest (2.69 mg N2O-N kg-1 h-1) activity of denitrifying enzymes during the summer when the temperature, hydraulic loading, and nutrient loading were highest. Spieles and Mitsch (2000) reported the op timum range of temperature for denitrification in wetland sediments as 20 to 25 C. Water temperatures less than 15C or greater than 30C have been shown to limit the rate of denitrification (Reddy and Patrick, 1984). The most effective pH for denitrification ranges fr om 7.0 to 8.5, with the optimum at 7.0 (Reddy et al., 2000).

PAGE 25

25 Denitrification occurs in the anoxic water column of aquatic ecosystems. Many researchers observed lower significant denitrification potential activity in the water column as compared to the sediments in different aquatic environments. The denitr ification rates of the water samples in the wetland receiving sewage tr eatment plant effluent were observed to be 0.06 0.96 g N L-1 hr-1 (Toet et al., 2003). Bastviken et al. (2003) obtained higher denitrification rates (6.5-7.5 kg N ha-1 day-1) in the sediment than on su rface of the water column. Aquatic plants serve as a major source of organic carbon to the microorganisms and provide large surface area (commonly referred as biofilms) for nitrifying and denitrifying bacteria. Plants also provide oxygen, which vary from species to species, to their root zones (Stoottmeister et al., 2003). Diffusion of oxygen to root zone s is reported to influence biogeochemical cycles within the rhizosphere But it was shown th at the amount of oxygen being released by the plants around the roots is limited (Brix, 1994; Wood, 1995). The limited release of oxygen around roots ensures that an aerobic conditions will predominate unless the organic load to the wetland is low and wetland is shallow, as the amount of oxygen decreases with increasing depth (Ayaz and Acka, 2000).T he limited release of oxygen through roots, however, also provides the required oxygen for aer obic nitrifying microbes for the oxidation of NH4 +-N to NO3 --N/nitrite nitrogen (NO2 --N), which in turn provides nitrate/nitrite for the anaerobic denitrifying microbes for denitrifica tion (Risgaard-Petersen and Jenson, 1997). The oxygen release rates of floating and em ergent aquatic plants such as Pistia stratiotes L. (water lettuce), Eichhornia crassipes (Mart.) Solms (water hyacinth), Hydrocotyle umbellata L.(pennywort), Pontederia spp.L. (pickerel weed), Phragmites spp. (Reed) and Typha latifolia L. (cattail) were studied by Moorhead and Reddy (1988); the oxygen release rate in Phra gmites spp. was estimated to be from 0.02 g m-2 day-1 to 12 g m-2 day-1. Reddy et al. (1989) studied the

PAGE 26

26 effectiveness of three floating and six emergent aquatic plants in improvi ng domestic wastewater quality based on their oxygen rel ease capacities. Plants such as H. umbellata transported oxygen 2.5 times more rapidly than E. crassipes which transports oxygen four times more rapidly than P stratiotes Radial oxygen loss (ROL), which is in fluenced by the external oxygen demand, was higher in Juncus effusus L. (common bulrush) (9.5 1 10 7 mol O2 h-1 root-1) than in J. inflexus L. (inland bulrush) (4.5 0.5 10-7 mol O2 h-1 root-1) (Sorrell, 1999). Higher rates of (6.8 mg N m-2 d-1) denitrification were obtained in the wetland sediments planted with Juncus spp as compared to the unplanted plots (Smialek et al., 2006). This indi cates the importance of vegetation in the constructed wetlands to facilitate denitrification. Besides oxygen, plant roots also exude orga nic compounds, which serve as a carbon source for denitrifying microorganisms. The influence of temperature on plant metabolism in turn affects the concentration of root exudates. The magnitude of organi c compound release, which enhances the NO3 N removal in constructed wetlands, ranged from 5-25% of the photosynthetically fixed carbon (Pla tzer, 1996). Bachand and Horn e (2000) observed that the different vegetation types resulted in signi ficantly different deni trification rates ( T.latifolia 565 mg N m-2 day-1, J effusus 261 mg N m-2 day-1, and mixed vegetation 835 mg N m-2 day-1). Productivity, physical structure, C:Nlitter ratio, and plant fiber content have been observed to differ between plants. Hence they concluded th at these factors affect ed the organic carbon availability of plants, thereby resulting in different denitrific ation rates. In organic carbon limited, free surface wetlands, it was recommended that a mixture of labile (submergent and floating) and more recalcitrant (emergent a nd grasses) vegetation be planted to enhance denitrification rates (B achand and Horne, 2000).

PAGE 27

27 The Problem Statement Among the various management strategies developed to mitigate nonpoint source pollution, constructed wetlands are considered as a cost effectiv e method for on-site removal of nutrients from nonpoint source pollution. The perf ormance of the constructed wetlands varies with site, characteristics of th e wastewater, type and design of the wetland. Hence, a systems approach which recognizes site specific conditions is essential for the successful management of the plant nursery runoff (M itsch and Jorgensen 1989; NRCS, 2002; Dunne et al., 2005). Most research has focused on use of constructed wetl ands to treat wastewater from municipal water source, animal waste, urban runo ff, stormwater runoff, and agricu ltural runoff. Very few studies have focused on the constructed wetlands treatin g nursery runoff (Headley et al., 2001; Lea-cox et al., 2002; Huett et al., 2005). Constructed wetlands that receive nursery runoff show considerable difference in water quality as compar ed with other wastewater sources. One of the main differences is the low concentration of dissolved organic carbon (DOC) in the nursery runoff (Huett et al., 2005), which is due to type of medium used in the containerized plants. Nursery runoff can have seasonal and hydraulic variability based on the timing of fertilizer application and rainfall. Due to the characterist ic features of the plant nursery runoff, it is important to understand the nutri ent removal mechanisms in a constructed wetland treating nursery runoff. One of the methods to increase the efficien cy of nutrient removal is to implement management and design practices to increase the de nitrification process in constructed wetlands. Although denitrification is considered a major removal mechanism for N in constructed wetlands, the importance of this process in wetla nds receiving plant nursery runoff has not been studied. Earlier studies (Bacha nd and Horne, 2000) indicated th at different vegetation types resulted in different denitrification rates in the wetlands which could be attributed to the

PAGE 28

28 differences in the availability of organic carbon in the respectiv e macrocosms. The influence of different macrophyte rhizosphere on denitrification rates with in the constructed wetlands receiving plant nursery runoff is not known. The goal of this study was to assess the N dynamics in a constructed wetland receiving plant nursery runoff for onsite N removal. To achieve this goal, following specific objectives and hypotheses were addressed. Objectives and Hypotheses Objective 1: To determine the N and P removal efficiency of a constructed wetland receiving plant nursery runoff. Hypothesis 1: The concentration of N and P will vary over months based on the timing of fertilizer application at the nursery, irrigati on, and rainfall. The concentration of N and P will be higher at inflow as compared to the outflow water, due to the nutrient removal within the constructed wetlands. Objective 2: To theoretically model the effect of hydraulic loading rate of the runoff and of estimated plant density on the residen ce time, which influences N removal in a constructed wetland. Hypothesis 2: The theoretical residence time will increase with increasing plant density and decreasing loading rate. Objective 3: To assess the concentration of nutri ents (N and P) and denitrification potential within the water column at varying depths (surface and bottom) and variation associated with the locations across the wetla nd (inflow and outflow) over five months in a constructed wetland receiving plant nursery runoff. Hypothesis 3: The concentration of nutrients (N and P) will be higher at inflow as compared to the outflow. The denitrifica tion potential will be higher in the water samples collected from the lower depths as compared to the surface of the water column due to the anoxic conditions in the d eeper parts of the water column. Objective 4: To determine the denitrification potential in the rhizosphere soil (soil closely adhering to the roots) of Canna flaccida and Typha latifolia in a constructed wetland receiving plant nursery runoff. Hypothesis 4: The denitrification potential of th e rhizosphere soil (soil closely adhering to the roots) will vary from species to species because macrophytes can vary in their capacity to diffuse oxygen and different carbon compounds in th eir root zones.

PAGE 29

29 Figure 1-1. Diagram illustrating potentia l pathways for chemical transformati ons of nitrogen (N) in a wetland.

PAGE 30

30 CHAPTER 2 MATERIALS AND METHODS Site Description The study was carried out at a constructed wetl and at Monrovia Nurs ery, which is one of the largest (total area over 445 hectares) commercial nurseries in the United States, located in Cairo, Georgia. Cairo is located at 30 40 N, 84 32 W in Grady County, Georgia (Fig. 21), adjoining Leon and Gadsden counties of Flor ida. The annual production of the nursery ranges from 11 to 15 million plants (www.monrovia.co m, October 2007). Gravel beds are created outdoors on the ground and are overlaid with plastic sheets on which plants are grown in plastic containers ranging from 5 50 L capacity. In the catchment area for the wetlands, large (50 L) containerized ornamental trees and shrubs are grown in the sloped nursery beds, which are lined with plastic sheets. Fertilization of the nursery plants is done by two means; controlled release fertilizer (CRFs) mi xed in the potting medium and fertigation (liquid fertilizers a pplied in irrigation water). Topd ressing of CRFs is also done based on the needs of the plants. Overhead sprinkler or micro-irrigation is used approximately 3-5 times a day. While larger containers are irriga ted via micro irrigation, runoff (Fig. 2-2) from these nursery beds is still significant due to the water application rates and sloped terrain. Runoff from the nursery beds and from stormwater is directed by lined or unlined waterways and drained into a flow control channel (FCC) whic h is 500 m in length (Fig. 2-3 and 2-4). Runoff eventually flows into a holding pond (HP) where it is held for recy cling or for release into the wetlands. The FCC controls the movement of water and allows sedimentation to occur. Depending upon the holding capacity of the HP, the wa ter from the FCC is e ither directed into the HP or diverted offsite through the stormwater retention basin. The HP is created in such a

PAGE 31

31 way that it captures half an inch of rainfall and the excess rainfall is diverted either to the irrigation ponds for recycling or actively pumped into the wetlands. A surface flow wetland was constructed at the nursery in 1997 to treat runoff effluents before releasing the excess water into a natural creek adjacent to the nursery property. This constructed wetland system is utilized to treat stormwater and runoff from three plant production areas (catchment area) totaling 48.5 hectares. The wetland was cons tructed by creating pits in a naturally depressed area of 3.8 hectares. A 12.5 cm think layer of bark ch ips was placed at the bottom of the pits to su pport the planting and also to serve as a source of carbon for the microbial activity in the constructed system. The wetland was constructed as several cells such as primary, secondary, and test cells (Fig. 2-5). The primary cells were designated as 1A and 2A; the secondary cells were designated as 1B and 2B; and the test cel ls were designated as 3,4,5,6, and 7. For this entire study, only one primary cell (1A) was used. There exists a difference in the size and depth of cells, and in the distribu tion and abundance of m acrophytes between the primary and secondary cells. The area of the pr imary cells 1A and 2A is 1.17 hectares and 0.57 hectares respectively. The area of the seconda ry cells 1B and 2B is 0.69 hectares and 0.65 hectares respectively. The primar y cells are 0.75 m deep and the s econdary cells are 0.3m deep. The depth at the outflow (South) side of the primary cells is co mparatively higher (1.2 m) than the inflow (North) side, which is 0.75 m deep. Each primary and secondary cell was divided into three sections by creating two earthen berms which extend to approximately th of the width (north-south) of the cells. These earthen be rms were created to maintain a linear water movement. The difference in vegetation of the primary and secondary cells is attributed to variation in planting selection a nd natural growth of plants. Th e macrophytes originally planted in the deeper primary cells were Canna flaccida Salisb. (Canna Lily), Scirpus validus Vahl.

PAGE 32

32 (Giant Bulrush), Pontederia cordata L. (Pickerelweed), Sagittaria latifolia Willd. (Arrowhead), and Panicum hemitomon J.A Schultes. (Maiden cane). Currently the primary cells are dominated by S. validus Typha latifolia L. (Cattail), C. flaccida Hydrocotyle umbellata L. (Pennywort), Lemna minor L. (Duckweed), and Wolffia spp. Horkel ex Schleid. (Water meal). There are few open water areas in the primary cells and severa l floating islands of vegetation occur near the outflow (South) side of the cell. The macrophytes originally plan ted in the shallow cells were C. flaccida S. latifolia P. hemitomon, and Juncus effusus L. (Soft Rush). Currently the shallow secondary cells are predominantly covered by T. latifolia; and the other dominant species in the shallow cells are S. latifolia H. umbellata and L. minor (Duckweed). When it is necessary to discha rge water from the HP into the wetland, it is pumped through 6 cm inner diameter PVC pipes (Fig. 2-6) to th e primary cells 1A and 2A, at 9 locations and 8 locations respectively. The water from each primary cell passes under a road [through 15 cm diameter PVC pipes; Fig. 2-7] and enters the secondary cell before being released into the grassed waterway that carries the treated wate r to the stilling pools. The water flow from primary to secondary cell and from secondary cell to the grassed waterway is based on gravity, and there is no external contro l to regulate the water flow. However, the water depth (and consequently the volume) of the primary cell can be controlled by adjusting the height of the inflow pipe of the secondary cell. The stilling pools enhance the settling of suspended sediments before discharging the treated water into the natu ral creek. The amount of water to be treated is influenced by the rate and frequency of rainfa ll and irrigation. Based on the amount of water available for treatment, four different run-tim es (number of hours per day of operation of the pump) are utilized; 0, 12, 18, 24 hours per day. The most frequently used run-times are 18 and 24 hours per day. During very low rainfall and minimal irrigation period, the run-time of 12

PAGE 33

33 hours or even 0 hours per day is used. The 12 and 18 hours per day run-times are achieved by running the pump for 2 hours and then turning it off for the next 2 hours, or 3hours on and 1hour off cycles, respectively, during a 24 hour period. The pump supplies water at an average of 2,271 liters per minute, which is distributed to all 17 inflow pipes of the primary cells. Hence the hydraulic loading rate of the wetland depends on the number of hours the pump is operational. Field Methods Water sampling (for determining the nutrient removal efficiency of the constructed wetlands). Water samples were collected from the st udy wetland cell at monthly intervals from April 2007 to August 2007. Three replicate water samples per section of the wetland cell were collected by taking grab samples in 125 ml plastic bottles (Fisher Scientific Co LLC, Suwannee, GA.) at 9 inflow points (one sample from each of the 9 pipes) and 3 outflow points (three replicate samples from each of the 3 pipes) as shown in Figure 2-8. In addition, three replicate samples were collected at a location where the fl ow control channel (FCC) releases water into the holding pond or into the stor mwater retention basin. In summary, there were 9 inflow samples, 9 outflow samples and 3 FCC samples, totaling 21 samples for the water nutrient analysis. Samples were stored in a cooler filled with ice-packs, transported to the lab, and stored at 4C until further processing. Water samples were analyzed for Total Kjeldal Nitrogen (TKN), ammonium nitrogen (NH4 +-N), nitrate nitrogen (NO3 --N), total phosphorus (TP), and dissolved organic carbon (DOC). Water sampling (for assessing the spatial effects on concentration of nutrients, physiochemical parameters and denitrificat ion potential of the water column). Nutrient concentrations, physiochemical parameters, and deni trification potential were measured at select

PAGE 34

34 sampling locations (Fig. 2-9). On the north side (the inflow side), one location was selected within each of the three sections of the primary cell from where the samples were collected. This sampling point was located approximately 3 m in ward from the northern edge of the cell and approximately 3 m inward from the eastern edge. Two sets of three replicate water samples were collected at the surface and at 0.5 m depth of the water column at each of the three locations at monthly intervals between Apr il 2007 and August 2007. One set of samples (three replications) was collected in 125 ml plastic bot tles and analyzed for TKN, NH4 +-N, NO3 --N, TP, and DOC and the second set of samples (three replicates) was collected in 25 ml vials (Fisher Scientific, Pittsburgh, PA) and used to determine denitrif ication enzyme activity (DEA), to estimate denitrification potential in the samples. The surface water samples were collected as grab samples in the undisturbed water column. To co llect samples from the bottom of the water column, the sampling bottle was turned upside down and forced through the water column by hand. After reaching the bottom of the water colum n, the bottle was turned slowly to fill it and then rapidly brought straight up to the surface. On the south side (outflow side) of th e study cell, sampling points were selected approximately 2 m inward from the location of ou tflow pipe, to ensure minimum disturbance in the water column. Owing to its depth, the surface and bottom of the water column samples at the outflow side were collected by using a swing sa mpler. For the bottom of the water column samples, the bottle tied to the swing sample r was sent upside down th rough the water column. After reaching the bottom of the water column, th e pole was turned slowly to fill the bottle and then rapidly brought straight up to the surface. Two sets of three replicate water samples were collected at the surface and bottom of the water co lumn at each of the three locations at monthly intervals between April 2007 and August 2007. One set of samples (three replications) was

PAGE 35

35 collected in 125 ml plastic bot tles and analyzed for TKN, NH4 +-N, NO3 --N, TP, and DOC and the second set of samples (three replicates) was colle cted in 25 ml vials and analyzed for DEA. If the water was too turbid, susp ended particles were allowed to settle by setting aside for a few minutes without any disturbance. Both surface and bottom of the water column samples were filtered by cheese cloth (Hermitage Inc, Camden, SC ) to remove large debris, dead leaves, twigs, barks, and plants of Wolfia spp. In addition, three replicate wa ter samples also were collected for denitrification potential analysis at the point where the flow control channel (FCC) releases water into the holding pond or in to the stormwater retention ba sin. There were 36 samples for nutrient analysis and 39 samples for denitrificati on potential analysis in total. Samples were stored in a cooler filled w ith ice-packs, transported to the lab, and stored at 4 C until further processing. In addition, pH, dissolved oxygen (DO), and temperature were measured every month during the study by using a handheld YSI 556 MPS (YSI Incorporated, Yellow Springs, OH) multi-probe system. The parameters were meas ured at both surface and bottom of the water column at inflow and outflow points on th e days of water sample collection. Plant sampling (for determining the denitrification potential of macrophyte rhizosphere soil). To assess the denitrification potential of macrophyt e rhizosphere soil, two macrophytes were selected based on their abundan ce and distribution in th e study cell, and based on the reported contribution of these macrophytes to nutrient removal in constructed wetlands. The plants selected for this study were T.latifolia and C. flaccida. Roots of three plants of each species were collected from each of the three sections at monthly intervals from plants growing on the north (inflow) side (Fig. 2-10). The sa mple plants were selected from where a group of the same species existed to avoid contamination with other species. The selected plants were

PAGE 36

36 pulled out and the roots were severed and then placed in zip-lock bags, stored in a cooler filled with ice-packs, transported to the lab, and stored at 4 C. There were 18 rhizosphere samples (three replicates of each species per section) in total. Analytical Methods Nutrient analyses. The day after sampling, part of all the water samples collected in 125 ml bottles was filtered through Whatman 0.45 m filter (Pall Corporati on, Ann arbor, MI) and collected in two 25 ml vials for analysis of NH4 +-N, NO3 -N, and DOC. The remaining unfiltered part of each sample was used for anal ysis of TKN and TP. After filtering, all the filtered and unfiltered water samples (except the ones collected for analyzing denitrification potential), were acidified to a pH of 2 with one drop of ultra pure con centrated sulfuric acid (H2SO4) for every 20-25 ml of sample. The filtered and unfiltered portions of samples were stored in a refrigerator until further analysis. Samples were analyzed for respective nutrients within 28 days of sampling as recommended by the USEPA. Total Kjeldal Nitrogen was measured by dige sting the unfiltered, acidified samples by the Kjeldahl procedure (Method 351.2, USEPA, 1983). The water sa mples were digested with sulfuric acid (H2SO4) and a copper sulfate mixture (CuSO4) to convert organic forms of nitrogen to ammonium. The digested samples were analyzed in AQ2+ automated discrete analyzer (Seal Analytical, Mequon, WI) Method no: USEPA 11 1 A (Method 351.2, USEPA, 1983). Nitrate nitrogen was analyzed calorimetrically usi ng the cadmium (Cd) reduction method on a rapid flow analyzer (Alpkem rapid flow Analyzer 300) or in the automated AQ2+ discrete analyzer (Method 353.2, USEPA) or AQ2 Method no: USEPA 132 A (Method 353.2, USEPA 1983). During April, July, and August 2007 the filtered acidified samples we re analyzed for NO3 --N on a rapid flow analyzer. In the month of May and June 2007, the discrete analyzer was used to

PAGE 37

37 analyze NO3 --N. Ammonium nitrogen was analyzed calorimetrically in filtered, acidified samples on AQ2+ (Method no: USEPA 103 A (Method 350.1, USEPA, 1983). Total Nitrogen (TN) was calculated by summing up the TKN (organic nitrogen and NH4+-N), and NO3 -N. To measure TP, unfiltered, acidified samples were digested by autoclaving and analyzed further by following the AQ2 Method no: USEPA 119 A (Method 365.2, USEPA, 1983). Dissolved organic carbon was analyzed in filtered, acidi fied samples on a Shim adzu TOC 5050a (USEPA 415.1). All the analysis were performed accord ing the Quality Assurance/Quality Control requirements (a spike, repeat, continuing calibration st andard, blank, Practical Quantitation Limit (PQL) to be run for every 20 samples) set by th e University of Florida Wetland Biogeochemistry Lab. Theoretical hydraulic residence time. The hydraulic residence time is a function of the ratio of wetland volume to water inflow rate Because the study wetlands are operated for variable number of hours per day, the hydraulic loading rate (rati o of volumetric flow rate to wetland surface area/volume) under two predominantly occurring conditions, 18 and 24 hours pump operation per day, were factored into the ca lculations. The effect of vegetation (plant density) was also factored into the calculations while determining the theoretical residence time. The calculations were based on the assumption that the entire volume of water in the wetland is involved in the flow. The theoretical reside nce time was calculated by using the following formulae: = V/Q, in which = theoretical residence time in days; V = wetland water volume in m3; Q = water flow rate in m3/day. The wetland volume (V) can be ca lculated using the formula:

PAGE 38

38 V = Ah, in which V = wetland water volume in m3; = wetland porosity in m3/ m3; A = wetland area in m2; h = mean water depth in m. Wetland porosity is a fraction of total wetland volume available through which water can flow (USEPA, 2000). It is the amount of we tland water volume not occupied by plants and expressed as a decimal (NRCS, 2002). Theref ore, wetland porosity and plant density are inversely related; the lower th e wetland porosity values, higher the plant density. As it is difficult to accurately measure wetland porosity in the field, highly variable porosity values were reported in the literature. For example, Reed et al. (1995) reported values ranging from 0.65 to 0.75 for fully vegetated wetlands, and for dense to less mature wetlands, respectively; Kadlec and Knight (1996) reported average wetland porosity values ranging from 0.95 to 1.0. According to USEPA (2000), it was suggested to use the wetland porosity value of 0.65 to 0.75 for fully vegetated wetlands, while consid ering design of constructed wetlands. Denitrification potential (Water). Denitrification potential in water samples was determined by measuring the deni trification enzyme activity (DEA) within a week of sample collection (Tiedje, 1982; White and Reddy, 1999). Tw enty ml of water sample (unfiltered) was taken into either 120 ml or 160 ml glass serum bottle. The bottles were capped and crimped ensuring proper seal. The bottles were purged with nitrogen gas (N2) for approximately 5 minutes to achieve anaerobic conditions. The initia l pressure of less than 20 psi was maintained. High grade acetylene (C2H2) gas was generated by adding N2 purged water to a separate bottle filled with Calcium Carbide (CaC2) rocks, which was already ca pped, crimped and subjected to N2 purging to remove any oxygen. Nine ml of acetylene gas was injected into each sample bottle from which 9 ml headspace was removed before a dding acetylene. Samples were then kept in a

PAGE 39

39 shaker for an hour to ensure even distribution of C2H2 in the water samples. Denitrification Enzyme Assay (DEA) solution was prepared by adding nitrate (as KNO3) and carbon (as C6H12O6) at the rate of 404mg L-1 KNO3 and 720 mg L-1 C6H12O6, respectively, along with 250 mg L-1 chloramphenicol in distilled water. Chlo ramphenicol was added to inhibit the production of new denitrifying enzymes while the activity of previously ex isting denitrifying enzymes are measured. Eight ml of DEA solution was added to 20 ml sample. The 7 ml of head space of gas was collected by using a 10 ml syringe immediat ely after adding DEA to the samples and after taking the pressure readings After that, the bottles were kept in an end-to-end shaker in a dark room at 25oC. In the month of April 2007 the headspace gas samples were collected at 30, 60 120, 180 minutes. From May 2007 and to August 2007 the headspace gas samples were collected every 2 hours, up to 6 hours, as ther e was no gas production observed until 2 hours in the month of April. At the end of each pre-sele cted time period, 7 ml of gas (presumably nitrous oxide) was extracted by using a syringe and pl aced in 3-4ml capped, crimped, pre-evacuated glass serum bottles. The bottl es with DEA solution and the original sample were then immediately returned to the shaker. The 7 ml sa mple of gas was stored at 25C until analyzed in a Gas Chromatograph (GC). The concentration of Nitrous oxide (N2O) was measured on a Shimadzu GC 14A (Shimadzu Scientific, Kyoto, Japan) by using a 63Ni electron capture detector. Column temperature was 30oC, detector temperature was 240 oC, and injector temperature was 120 oC. The carrier gas was a mix of 95% Argon (Ar) and 5% methane (CH4). The gas samples were injected into GC and the concentration of N2O was determined. Denitrification potential was determined by calculating the slope of the linea r curve obtained when the gas concentrations were plotted over time.

PAGE 40

40 Denitrification potential (Rhizosphere). The stored root sample s were processed as early as possible after reaching the lab. The excessive and loosely adhering soil was removed by shaking the root system, and also by gently remo ving by hand. The soil ve ry closely adhered to the roots was defined as rhizosphere soil which was removed by washing the roots with autoclaved double distilled water. Thirty-five ml of water was used to remove the firmly attached soil by washing with the use of pipette. Twenty ml of the obtained soil slurry was used for the determination of denitrific ation potential. Ten ml of the rhizosphere solution was used to determine the moisture content of the soil. The moisture content was determined gravimetrically by drying 10 ml of the rh izosphere solution at 70oC for 72 hours. The denitrification potential of the rhizosphere soil was determin ed as described above. The volume of the glass serum bottle used in th is experiment was 120 ml. Eight ml of DEA solution (720 mg L-1 glucose, 404 mg L-1 KNO3, and 500 mg L-1 chloroamphenicol) was added to 20 ml of samples. Gas samples were collected every 30, 60 120, 180 minutes. Rest of the procedure was as described for measuring den itrification potential in water samples. Statistical Methods Statistical analysis of the data was performe d by using SAS 9.1 (SAS Institute Inc, Cary, NC). The dependent variables were subjected to analysis of variance (ANOVA) by using the GLM procedure along with Scheffe for all pair wise comparison, whic h indicate not only whether the means are different from each other, but also which means differ from which other means. Test of statistic al significance was done at = 0.05. The physiochemical parameters of water viz., temperature and dissolved oxygen were correlated with denitrification potential by using CORR procedure to determine Pearsons co rrelation coefficient. The concentration of NO3 --N was also correlated with denitrificati on potential by using the same procedure.

PAGE 41

41 Figure 2-1. Map of the state of Georgia, Un ited States, showing c ounty boundaries. Grady County, in which Monrovia nursery is locate d, is indicated by a callout box. Source: http://www.epodunk.com/cgi-bin/genInfo.php?locIndex=7968, October 2007.

PAGE 42

42 Figure 2-2. Runoff generated from the microi rrigated nursery beds at Monrovia Nursery

PAGE 43

43 Figure 2-3. Runoff generated from the nursery beds drains into the flow control channel

PAGE 44

44 Figure 2-4. Runoff draining into the flow control channel

PAGE 45

45 Figure 2-5. Overview of the Constructed we tlands at Monrovia Nursery, Cairo, GA. A) An aerial image (top right) of part of Monrovia Nursery, including the treatment wetlands, in Cairo, GA. B) A schematic overview of the constructed wetlands (bottom left) is also presented. The arrows indicate the direction of water movement th rough the wetlands. 1A and 2A, and 1B and 2B represent the primary and secondary cells, respectively.

PAGE 46

46 Figure 2-6. Inflow pipes into the study cell

PAGE 47

47 Figure 2-7. Outflow pipe of the study cell

PAGE 48

48 Figure 2-8. Location and number of water samples (indicated by st ars) at inflow and outflow of the study cell to determine the nutrient removal efficiency of the constructed wetlands.

PAGE 49

49 Figure 2-9. Locations of water samples (indicated by stars) in the study cell to asse ss the spatial variati on in the concentration of nutrients and denitrification pot ential of the water column, between locations (inflow and outflow) and between depths (top and bottom) within locations over five months.

PAGE 50

50 Figure 2-10. Illustration of the study cell showing the location and numbe r (indicated by stars) of monthly plant root samples to assess the denitrification potential in the rhizosphere soil of Canna flaccida and Typha latifolia

PAGE 51

51 CHAPTER 3 RESULTS Removal efficiency of th e constructed wetlands. The concentration of nutrients entering and leaving the constructe d wetland was determined to evaluate the effectiveness of the study cell in removing nutrient contaminants over a period of five months from April to August 2007. The plant nursery runoff contained an averag e total nitrogen (TN) concentration of 34 mg L-1during the study period. When assessing mean TN composition (Fig. 3-1), it was observed that the majority of Nitrogen (N) entering and leaving the wetland during the study period was in the form of nitrate nitrogen (NO3 --N; 30 mg L-1) followed by ammonium nitrogen (NH4 +-N; 2 mg L-1) and organic nitrogen (2 mg L-1). Total phosphorus concentration was 1 mg L-1 and dissolved organic carbon was 10 mg L-1. The nursery runoff entering the constructed wetland over the study period consisted of 87 % of the TN in the form of NO3 --N. The mean nutrient removal efficiency of the study cell, based on load -in and load-out rates, was observed to be 40% for TN, 40% for NO3 N, 59% for NH4 +-N, and 16% for TP (Table 3-1). Nitrate Nitrogen (NO3 --N). The concentration of NO3 --N was influenced (p<0.0001) by an interaction among location of sample coll ection and month during which samples were collected (Fig. 3-2). The inflow NO3 --N concentration gradually decreased from April (47 mg L1) to August (13 mg L-1); similarly the outflow concentration also gradually decreased from April (42 mg L-1) to August (3 mg L-1). At flow control channel (FCC), NO3 --N concentration did not show any trend over time, but the absolute values were highest in April (52 mg L-1) and lowest in July (13 mg L-1). The mean removal efficiency of NO3 --N in the study cell was 40%, highest in August (77%) and lowest in April (12%). The re moval efficiency was inversely related to the loading rate of NO3 --N coming into the wetland; i.e ., when the loading rate of NO3 --N was higher (7 g m-2 day-1; April), the removal efficiency was lower (12%) and when the loading rate was

PAGE 52

52 lower (2 g m-2 day-1; August), the removal efficiency was higher (77%). The concentration of NO3 --N at FCC significantly differed from infl ow concentration during the period of study, except for April. The mean inflow NO3 --N concentration over the study period was 30 mg L-1, which was significantly higher (p<0.0001) than the outflow concentration of NO3 --N, averaging 18 mg L-1. Ammonium Nitrogen (NH4 +-N). The nursery runoff that ente rs the constructed wetland over the study period consisted less than 1% of the TN in the form of NH4 +-N. A significant interaction (p < 0.0001) was observed among loca tion of sample collection and month during which samples were collected, for the concentration of NH4 +-N (Fig. 3-3). At all locations (inflow, outflow, and FCC), the NH4 +-N concentration did not show any particular trend over months, but the mean values were always higher at inflow and lower at outflow for each month of this study. The average removal efficiency of NH4 +-N was 59% and there was no trend observed over months. There were significan t differences in the concentration of NH4 +-N between FCC and inflow concentration ove r months, which indicate that the NH4 +-N removal in the holding pond was noticeable. The average concentr ation of the NH4 +-N coming into the wetland over the study period was 2.1 mg L-1 which was higher (p<0.0001) than the concentration of NH4 +-N leaving the system, averaging 0.9 mg L-1. Total Nitrogen (TN). The TN was computed by summing the values of Total Kjeldal Nitrogen (TKN) and NO3 --N. The trend was similar to that of NO3 --N. The concen tration of the TN was influenced (p<0.0001) by an interaction among location of sample collection and month during which samples were collected (Fig. 3-4). The inflow concentration of TN gradually decreased from April (60 mg L-1) to August (15 mg L-1); similarly the con centration of TN at outflow was also gradually decreased from April (49 mg L-1) to August (4 mg L-1). At FCC, TN

PAGE 53

53 concentration was highest in April (70 mg L-1) and lowest in July (15 mg L-1), but did not appear to consistently increase or decrease over this period of time. Overa ll, the average removal efficiency of TN was 40% which gradually increased from April 2007 (18%) to August 2007 (73%). It was observed that the removal efficiency was inversely related to the loading rate of TN into the wetland. The mean concentration of TN over five months at inflow (34 mg L-1) was significantly higher than the concentration of total nitr ogen at outflow (20 mg L-1) of the study cell. Total Phosphorus (TP). The concentration of TP wa s influenced (p<0.0001) by location of sample collection and month during which samples were collected. The concentration of TP at all three locations (inflow, outflow, and FCC) did not follow any particular trend between April and August 2007 (Fig 3-5). The inflow concentration of TP over months ranged from 1.7 mg L-1 (April) to 0.8 mg L-1 (August); and the con centration of TP at outflow ranged from 1.4 mg L-1 (April) to 0.8 mg L-1 (August). Although the interaction among location of sample collection and month showed significant influence, most of the means were statistically similar. The removal efficiency of TP ranged from 0-32%, which varied over months. In July 2007, the removal efficiency was negative, which indicates that there might be possible internal loading of phosphorus in the sy stem. There was no significant difference in the concentration of TP between FCC and inflow of the wetland. The mean concentration over five months, of TP coming into the wetland was 1.3 mg L-1, which was higher (p<0.0001) than the concentration of TP leaving the system, averaging 1.1 mg L-1. Dissolved Organic Carbon (DOC). The concentration of DOC was influenced (p<0.0001) by an interaction among location of sample collection and month during which samples were collected (Fig 3-6). The DOC concentration at inflow ranged from 10 mg L-1

PAGE 54

54 (June) to 12 mg L-1 (August), and at outflow it ranged from 10 mg L-1 (July) to 11 mg L-1 (May) Most of the means were statis tically similar, alth ough the overall model showed significant interaction among locations and months. The five month average concentrations of DOC at FCC and inflow of the study cell were similar. The mean concentration of DOC present in the inflow water was 11 mg L-1 which was higher (p<0.001) than the concentration of DOC leaving the system, averaging 10 mg L-1 during the study period. TN: TP ratio. The computed TN: TP ratio was a ffected (p<0.0001) by an interaction among location of sample collection and months during which samples were collected. At inflow, the ratio was highest in the month of April (36:1) and lowest in July (17:1). The ratio at outflow was also highest in Apr il (35:1), but it was lowest in A ugust (5:1). At FCC, the ratio ranged from 12:1(July) to 33:1(August). The five month mean TN: TP ratio, at inflow and outflow, was 25:1 and 18:1, respectively. DOC: NO3 --N ratio. The computed DOC: NO3 --N ratio was affected (p<0.0001) by an interaction among location of sample collect ion and months during which samples were collected. At both inflow and outflow, there was gradual increase in the DOC: NO3 --N ratio over five months. At inflow, the ratio gradually in creased from April (0.2:1) to August (0.9:1); similarly the ratio at outflow also gradually increased from April (0.2:1) to August (3.3:1). At FCC, the ratio ranged from 0.3:1(April and May) to 0.8:1(July), with va lues not showing any trend over the period of the study. The five month mean DOC: NO3 --N ratios at inflow and outflow were 0.5:1 and 1.3:1, respectively. Relationship between nutrient loading and removal rates. The correlation between loading and removal rate (g m-2 day-1) of NO3 N, NH4 +-N and TP are presented in Figure 3-7 to 3-9. The Pearsons correlation coefficient (r) value of 0.19 (p<0.21) was obtained for the

PAGE 55

55 relationship between loadi ng and removal rate of NO3 N. The NH4 +-N and TP exhibited a linear relationship with Pearsons correla tion coefficient (r) value of 0.75 (p<0.0001) and 0.60 (p<0.0001), respectively. Determination of theoretical hydraulic residence time. The effect of hydraulic loading rate and plant density on hydraulic residence time was computed by utilizing a model and the results are shown in Figure 3-10. The re sidence time was calculated for two different loading rates with three different plant density conditions, which provide three different wetland porosity values. Hydraulic loadi ng rate was inversely related to residence time; i.e., when the loading rate increa sed from 1300 to 1734 m3day-1, the theoretical residence time decreased by a day for all three wetland porosity values. As the wetland porosity value increased from 0.75 to 0.95, the theoretical residence time increased at least by one day in this model. There was no difference in the residence time when the wetl and porosity value increased from 0.95 to 1.0 at both hydraulic loading rates. Spatial effects on nutrients, phy siochemical parameters and deni trification potential of the water column: Nitrate Nitrogen (NO3 --N). The concentration of NO3 --N was not influenced (p=0.0812) by an interaction among the treatments (l ocation, depth, and months), presented in table 3-2. All two-way interactions, de pth*month, location*mont h, and location*depth significantly affected the concentration of NO3 --N at =0.05 level. The interaction among the location of sample collection (inflow and outfl ow) and the months during which samples were collected, showed that the concentration of NO3 --N changed from inflow to outflow except in April. In April, the infl ow concentration (44 mg L-1) of NO3 --N was similar to that at outflow (42 mg L-1), but it was significantly higher than othe r means. There was no trend observed

PAGE 56

56 between April and August, both in inflow and outflow concentration of NO3 --N, but the concentration during each month was higher at in flow than at outflow. The interaction among location of sample collection (inflow and outflow ) and depth within location (surface and bottom of the water column) showed that surface (24 mg L-1) of the water column at inflow had higher mean concentration of NO3 --N than at the lower depth (21 mg L-1) at inflow and the surface (18 mg L-1) of outflow (Fig. 3-11). On the outflow side, the mean concentration of NO3 --N over five months was higher (18 mg L-1) at the surface than the lower dept h of the water column (16 mg L1). The five month mean concentration of NO3 --N was always higher at surface water column as compared to lower depth of the water column. Ammonium Nitrogen (NH4 +-N). The mean concentration of NH4 +-N was influenced (p=0.0001) by an interaction among lo cation, depth, and months (Table 3-2). Most of the means were similar, however. The distribution of NH4 +-N within the wetland did not vary. The concentration of NH4 +-N was also influenced by an interaction among location*month, and by depth*month. Although both depth*month, and lo cation*month interactions yielded some significant means, there was no clear trend ov er time observed in th e concentration of NH4 +-N. At inflow, the concentration of NH4 +-N ranged from 3.1 mg L-1 (April) to 0.2 mg L-1 (May). At outflow the NH4 +-N concentration ranged from 0.1 mg L-1 (May) to 1.8 mg L-1 (June). Total Phosphorus (TP). The concentration of TP was influenced (p=0.0009) by an interaction among location of samp le collection, depth within lo cations, and month (Table 3-2), however most of the means were similar. The ab solute TP concentration at inflow and outflow was higher at lower depth of the water column than that of the surface of the water column between April to August. In addition to three-way interaction, location of sample collection*depth was observed to highly affect TP concentration. At the inflow, the surface (2

PAGE 57

57 mg L-1) TP concentration was lower than the TP in the bottom (3 mg L-1) of the water column. In contrast, the concentrations of TP at the tw o depths were similar at outflow. The other twoway interactions between locati on*month, and depth month did not show any particular trend in the concentration of TP during the period of this study. Dissolved Organic Carbon (DOC). The concentration of DOC was affected (p=0.0008) by an interaction among location, depth, and month (Table 3-2). At inflow and outflow, the surface and lower depth of the water column did not show difference in the concentration of DOC over the five months of this study. There was no trend observed for the DOC concentrations during this study for both locations and depth within locati on. Of all the two-way interactions only location* depth showed significant differences between treatment means. The surface of the water column at inflow had higher (11 mg L-1) concentration of DOC as compared to surface (10 mg L-1) of the water column at outflow. Similarly the DOC concentration was higher (12 mg L-1) at inflow and lower at outflow (10 mg L-1) for bottom of the water column. The five month mean concentration of DOC wa s similar at two depths both at inflow and outflow of the wetland. DOC: NO3 --N ratio. The DOC: NO3 --N ratio was influenced (p=0.0009) by a three-way interaction among location, depth, and months. At inflow, the DOC: NO3 --N ratio at the lower depth of the water column was higher as compared to the surface water column between April to August. Similar trend of higher DOC: NO3 --N ratio at the lower depth was observed at the outflow of the study cell, except for April. Ho wever, there was no temp oral pattern observed during this period both at inflow and outflow of the study cell. All two-way interactions, depth*month, location*month, and location*depth, significantly influenced the ratio at =0.05 level. The mean DOC: NO3 --N ratio over five months at th e inflow and outflow was 0.9:1 and

PAGE 58

58 1.5:1, respectively. The fi ve month mean DOC: NO3 --N ratio at the surface and bottom of the water column were 1.0:1 and 1.3: 1, respectively. Temperature. The temperature of the water colu mn was affected (p=0.0141) by an interaction among locations (inflow and outflow), depth within location (surface and bottom), and months (April to August 2007) Although the overall model s howed significant difference, the mean comparisons showed similar values for the bottom and surface of the water column temperature both at inflow and outflow during the period of this study (Table. 3-3). For bottom of the water column on the inflow side, the effect of months on temperature was apparent as it gradually increased from May (22 C) to August (30 C). The temperature at lower depth of the water column at outflow was al so significantly affected by month of sampling. Surface water temperature on the inflow side was lowest during May (23 C) and highest (30 C) in August. Similar trend over time was observed at the surf ace of the water column on the outflow side of the study cell. There was significant difference between bottom (22 C) and surface (25 C) water column temperature at out flow during the month of May, but the temperature at surface and lower depth of the water column was simila r during other months. The five month mean temperature (averaged over surface and lower depth values) of the wetland at inflow (27 C) was higher than the outflow (26 C). The five month mean temper ature at the surface of the water column (averaged over inflow and outflow locations) was higher (26 C) than the bottom (27 C) of the water column. Dissolved Oxygen. The dissolved oxygen (DO) concentration in the water column was not affected by any interaction e ffect (Table 3-3) between the tr eatments (locations, depth and months). The main effects of location of sample collection, depth of the water column and months showed significant difference at = 0.05 level (Fig. 3-12). The dissolved oxygen

PAGE 59

59 concentration of the inflow (6 mg L-1) was significantly (p=0.0005) lo wer when compared to that of outflow (11 mg L-1). The DO was observed to significan tly decrease from surface (11 mg L-1) to bottom (6 mg L-1) of the water column. Although the overall analysis showed significant difference between months, only the value for May (12 mg L-1) was significantly different from rest of the months. pH. The pH was not affected by interactions be tween locations, depth and months (Table 3-3), except for interaction am ong location and months (Fig. 313), and because of the main effect of months. At inflow, th e pH in August (4) was lower than in May (6), June (7) or July (6). In August, the inflow pH (4) was lower th an outflow pH (6). Th e other interactive means were similar. When compared over months, mean pH in June was higher (7) than the values obtained for July (6) and August (5). The mean pH of this system was 6 during the period of this study. Denitrification potential of water. The concentration of nitrous oxide nitrogen (N2O-N) produced per liter of water per hour (mg N2O-N L-1 hr-1) was calculated and it was observed that the N2O-N production was either neg ligible (0.01 to 0.03 mg N2O-N L-1 hr-1) or undetectable, as was the case in most of the samples. The data, thus, are neither presente d graphically, nor in a tabular form. Denitrification potential of macrophyte rhizosphere soil. The denitrification potential of the rhizosphere soil was influenced (p= 0.0419) by an interaction among macrophyte species, sections, and months (Table. 3-4) The significant di fferences for the den itrification potential were observed between very few means, however. The rhizosphere soil of Typha latifolia produced significantly higher denitr ification potential (30 mg N2O-N kg-1 hr-1) in the western section of the cell during May, which was higher th an the values obtained from the rest of the

PAGE 60

60 means. The two-way interaction among month*sect ion and plant*month occurred, however most of the means were statistically similar. During May 2007, the denitrification potential of T. latifolia was 12 mg N2O-N kg-1 hr-1. Similarly the denitrification potential in the western section (17 mg N2O-N kg-1 hr-1) was found to be significantly different from rest of the treatments during the month of May 2007. Denitrification potential was influenced (p = 0.01) by the main effect of month; however, there was no trend between April and August 2007. The mean denitrification potential was 8, 2, 1, 3 mg N2O-N kg-1 hr-1, during May, June, July, and August 2007, respectively. Only the denitr ification potential during May differed from other months. The five month mean denitrification potential for Canna flaccida and Typha latifolia were 3 mg and 4 mg N2O-N kg-1 hr-1, respectively.

PAGE 61

61Table 3-1. Mean load-in, load-out, removal rate (g m-2 day-1) of nutrients and percentage nut rient removal efficiency of the constructed wetland over five months (April and August 2007). Standard deviation valu es in are presented in parentheses. Total Nitrate Ammonium Total Nitrogen (TN) Nitrogen (NO3 -)Nitrogen (NH4 +) Phosphorus (TP) (g m-2 day-1) (g m-2 day-1) (g m-2 day-1) (g m-2 day-1) Load-in rate 5.07 (2.64) 4.43 (2.21) 0.32 (0.21) 0.19 (0.05) Load-out rate 3.07 (2.50) 2.68 (2.18) 0.13 (0.14) 0.16 (0.04) Removal rate 1.99 (0.15) 1.75 (0.75) 0.19 (0.11) 0.03 (0.03) % Removal 40 40 59 16

PAGE 62

62Table 3-2. Mean concentra tions of nutrients (mg L-1) as affected by location of sample coll ection, depth within location, and months (April and August 2007) in the constructed wetland receiving nursery r unoff (n=18). Capital lett ers indicate statistical difference between depths (top and bottom) at inlet and outl et over five months; lower-case letters indicate statistical difference between locations (inlet and outlet) at top and bottom de pths over five months; asterisks(*, **) indicate statistical difference between locations at top and bottom for every month. Mean s followed by the same letter and same number of asterisks are statistically similar. Location Depth Month NO3 --Nmg L-1 NH4 --Nmg L-1 TP mg L-1 DOCmg L-1 Inlet Top April 46A a 4.0A a 1.8 A a 11A a May 42B a 0.2BC a 2.0 A a 13B a June 8D a 2.3AB a 1.4 A a 10A a July 15C a 0.2B a 1.3 A a 11A a Aug 11CD a 1.3ABC a 0.9 A a 11A a Bottom April 42A a 2.2AB a 2.1 A a 11AB a May 36B b 0.2B a 2.4 A a 12AB a June 7D a 2.0AB a 3.5 A b 10B a July 11C a 0.9BC a 2.3 A a 11AB a Aug 7CD b 3.6A a 3.0 A b 13A a Outlet Top April 43A a 1.0A a 1.6 A a 11A a May 32B a ** 0.1A a 1.2 A a 11A a June 4C a ** 1.7A a 0.9 A a 9A a July 7C a ** 0.1A a 1.2 A a 9A a Aug 4C a ** 0.1A a 0.7 A a 10A a Bottom April 42A a 2.3A a 1.5 A a ** 10A a May 26B a ** 0.1A a 1.3 A a 11A a June 4C a 1.8A a 0.9 A a 10A a July 5C a ** 0.1A a 1.5 A a ** 9A a Aug 3C b ** 0.1A a ** 1.0 A a 10A a **

PAGE 63

63Table 3-3. Physiochemical parameters (tem perature, dissolved oxygen, pH) as affected by location of sample collection, depth w ithin location, and months (April and August 2007) in the constructed wetla nd receiving nursery runoff (n=3). Capital letters indicate statistical difference between depths (top and bottom) at inlet and ou tlet over five months; lower-case letters indicate statistical difference between lo cations (inlet and outlet) at top and bottom depths over five months. Means followed by the same letter are statistically similar. Location Depth Month Temperature C Dissolved oxygen (mg L-1) pH Inlet Top May 23A a 13A a 6.3A a June 28A a 7A a 7.2A a July 28A a 7A a 7.0A a Aug 30A a 6A a 4.4A a Bottom May 22A a 4A a 6.2A a June 27A a 5A a 7.2A a July 27A a 3A a 5.4A a Aug 30A a 2A a 3.7A a Outlet Top May 25A a 25A a 5.5A a June 27A a 15A a 6.5A a July 27A a 11A a 5.8A a Aug 29A a 12A a 6.9A a Bottom May 22B a 15A a 6.4A a June 25A a 10A a 6.6A a July 26A a 5A a 5.0A a Aug 28A a 4A a 6.0A a

PAGE 64

64 Table 3-4. Denitrifi cation potential (mg N2O-N kg-1 hr-1) as affected by macrophyte rhizosphere soil in three sections (eas t, middle, and west) between May 2007 and August 2007 in the constructed wetland receiv ing nursery runoff. (n =3). Capital letter indicates the statistical difference between plant species within each section and month. Means followed by the same letter are statistically similar. Plant species Month Section Denitrification potential mg kg-1 hr-1 Canna flaccida May East 1.9A Middle 2.4A West 4.1A June East 2.0A Middle 2.7A West 3.4A July East 1.0A Middle 0.5A West 3.1A August East 2.2A Middle 3.9A West 3.0A Typha latifolia May East 1.5A Middle 6.1A West 29.7A June East 2.1A Middle 0.2A West 0.1A July East 0.6A Middle 0.3A West 0.1A August East 6.1A Middle 0.3A West 2.3A

PAGE 65

65 Figure 3-1. Mean concentration (mg L-1) of total nitrogen (TN) and composition of TN [nitrate (NO3 N), ammonium (NH4 +-N) and organic nitrogen concentration mg L-1] between April to August of 2007 in the water samples obtained from the flow control channel (n=3), inlet (n=9), and outlet (n=9) pi pes of a constructed wetland receiving nursery runoff.

PAGE 66

66 Figure 3-2. Mean concentration (mg L-1) of nitrate nitrogen (NO3 --N) between April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell in th e constructed wetla nd receiving nursery runoff. Error bars indicate plus standard error. Capital letter indicates statistical difference between loca tions within each month (May to August); lower-case letters indica te statistical difference between months at each location (FCC, inlet and outlet). Means followed by the same letter are statistically similar.

PAGE 67

67 Figure 3-3. Mean concentration (mg L-1) of ammonium nitrogen (NH4+ N) between April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed we tland receiving nurse ry runoff. Error bars indicate plus standard error. Capital letter indicates statistical difference between loca tions within each month (May to August); lower-case letters indica te statistical difference between months at each location (FCC, inlet and outlet). Means followed by the same letter are statistically similar.

PAGE 68

68 Figure 3-4. Mean concentration (mg L-1) of total nitrogen (TN) be tween April 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetla nd receiving nursery runoff. Error bars indicate plus standard error. Capital letter indicates statis tical difference between locations within each month (May to August); lower-case letters indica te statistical difference between months at each location (FCC, inlet and outlet). Means followed by the same letter are statistically similar.

PAGE 69

69 Figure 3-5. Mean concentration (mg L-1) of total phosphorus (TP) and Dissolved organic carbon between April 2007 and August 2007 in the flow control channel (n=3), in let (n=9) and the outlet (n=9) of the study cell of a constructe d wetland receiving nursery runoff. Error bars indicate plus standard error. Capital letter indicate s statistical difference between locations within each month (May to August); lower-case letters indi cate statistical difference be tween months at each location (FCC, inlet and outlet). Means followed by the same letter are statistically similar.

PAGE 70

70 Figure 3-6. Mean concentration (mg L-1) of dissolved organic carbon (DOC) between Ap ril 2007 and August 2007 in the flow control channel (n=3), inlet (n=9) and the outlet (n=9) of the study cell of a constructed wetland receiving nursery runoff. Error bars indicate plus standard error. Capita l letter indicates statistical difference between locations within each month (May to August); lower-case letters indica te statistical difference between months at each location (FCC, inlet and outlet). Means followed by the same letter are statistically similar.

PAGE 71

71 Figure 3-7. Pearsons correlation between loading and removal rate of nitrate nitrogen (g m-2 day-1) between April 2007 and August 2007 in a constructed wetland r eceiving nursery runoff.

PAGE 72

72 Figure 3-8. Pearsons correlation between loading and removal rate of ammonium nitrogen (g m-2 day-1) between April 2007 and August 2007 in a constructed wetlan d receiving nursery runoff.

PAGE 73

73 Figure 3-9. Pearsons correlati on between loading and removal rate of total phosphorus (g m-2day-1) between April 2007 and August 2007 in a constructed wetland r eceiving nursery runoff.

PAGE 74

74 Figure 3-10. Theoretical hydraulic residence time (i n days) as affected by hydraulic loading rate (m3 day-1) and three estimates of wetland porosity values (0.75, 0.95 and 1.0) in the study cell of a constructed we tland receiving plant nursery runoff

PAGE 75

75 Figure 3-11. Mean concentrations of nitrate nitrogen (mg L-1) as affected by location of sample collection and dept h within location, between April and August 2007 in the constructed wetland receiving nursery runoff (n=18). Capital letters indicate statistical difference between location (inl et and outlet) at top and bottom of the water column; lower-case letters indicate statistical difference between depths (t op and bottom) at inlet and outlet over five months. Means followed by the same letter are statistically similar.

PAGE 76

76 Figure 3-12. Dissolved oxygen concentration (mg L-1) as affected by location of sample coll ection, depth within location, and months (April and August 2007) in the constructed wetland receiving nursery runoff (n=3). Capital letters i ndicate statistical difference within location, within depth, a nd within months. Means followed by the same letter are statistically similar.

PAGE 77

77 Figure 3-13. Physiochemical parameter, pH, as affected by location of sample colle ction, and months (April and August 2007) of the constructed wetland receiving nursery runoff (n=3). Capital letter indicates statistical difference between locations within each month (May to August); lower-case lett ers indicate statistical di fference between months at each location (inlet and outlet). Means followed by the same letter are statistically similar.

PAGE 78

78 CHAPTER 4 DISCUSSION Concentration of nutrients in the inflow and outflow water. The mean concentration (over all locations and months) of nutrients in plant nursery runoff entering the constructed wetland were 34 mg L-1 total nitrogen (TN) [30 mg L-1 nitrate nitrogen (NO3 N); 2 mg L-1 ammonium nitrogen (NH4 +-N); and 2 mg L-1 organic nitrogen], 1 mg L-1 total phosphorus (TP), and 10 mg L-1 Dissolved Organic Carbon (DOC). The concentration of DOC, less than 20 mg L1, is considered low as compared to other wastew ater sources (Headley et al., 2001; Huett et al., 2005; Crumpton et al., 1993). The nutrient removal efficiency of the constructed wetland was observed to be 40% for TN, 40% for NO3 N, 59% for NH4 +-N, and 16% for TP. The constructed wetland systems in the US are report ed to provide removal efficiencies of 66-95% for TN and TP (Dortch, 1992), 30-60% for N (Ha mmer and Knight, 1994). Headly et al. (2001) have reported > 84% TN and > 65% TP remova l in the subsurface horizontal flow wetlands receiving plant nursery runoff. Huett et al. (2005) obtained > 95% removal efficiency for both TN and TP in a subsurface flow wetlands treating plant nursery runoff. Reinhardt et al. (2005) reported 23% TP removal efficiency in constructe d wetlands treating agricultu ral drainage water. While comparing the results, the study cell appear s to be functioning moderately efficiently in removing TN and TP. Mean removal efficiency for NO3 N of the constructed wetl and over five months was 40%. Nitrate nitrogen removal efficiency of tr eatment wetlands in the US ranges from 30-60% (Hammer and Knight, 1994). Poe et al. (2003) obtained 53% of NO3 N removal in a constructed wetlands treating agricu ltural runoff. It is possible that a longer sampling period at the study cell may have give n different results.

PAGE 79

79 The removal efficiency of the NO3 --N during the study period ra nged from 12 -77% over five months, indicated the temporal variation in effectively removing NO3N. The highest NO3 N removal was obtained in the month of A ugust 2007 and the lowest removal efficiency was obtained during April. During the same period, the concentration of NO3 N entering and leaving the wetland gradually decrea sed from April to August. Th ere was an inverse relationship between the concentration of NO3 N at inflow and the removal efficiency of the wetland, i.e., when the concentration of NO3 N entering the wetland was higher, the removal efficiency was lower during April. Similar relationship was observed between loading rates and removal efficiency of NO3 N in the constructed wetl ands. Gradual increase in removal efficiency of NO3 N from April (12%) to Augus t (77%) was observed, as the in flow concentration gradually decreased from 47 mg L-1 to 13 mg L-1. Lin et al. (2002) observe d similar decrease in the efficiency of nitrate removal but the rate of removal increased as the nitrate loading rate was increased. Darbi et al. (2003) also reported inverse relationshi p between loading rate and the removal efficiency of NO3 --N. Reduction in the removal efficiency of NO3 N from 2.46 to 1.64 kg NO3 --N m-3 day-1 was obtained, as the inflow NO3 --N concentration increased from 175 to700 mg L-1 (Park et al., 2002). Most of th e studies reported high loads of NO3 --N as compared to the load in the present study. As the loading rate of NO3 --N was observed to be less than 7 g m-2 day-1, this study provides a useful insight into nutrient removal at lower loading rate. Similar results of low NO3 --N loading rates were observed (<2 g m-2 day-1) by Headly et al. (2001) and Huett et al. (2005) in subsurface flow reed beds receiving plant nursery runoff in Australia. Although the loading rates are comparat ively low, the removal rate of NO3 N (1.75 g m-2 day-1) in this study is comparable and in many cases high er as compared to the values of other studies (Table 4.1).

PAGE 80

80 The Pearson correlation coefficient value (r = 0.19) indicated a little/no association between loading and removal rate of NO3 N. The strong positive association between loading and removal rate of NH4 +-N (r = 0.75) and a weak positive association between loading and removal rate of TP (r = 0.6). Th e positive correlation coefficient indicates that the loading and removal rate tend to increase or decrease together. Similar linear relationship of increasing removal rate with increasing loading rate has b een observed by Headley et al. (2001) for TN (r = 0.76). Plant nursery runoff in this study consis ted 87% of the TN in the form of NO3 N. The dominance of NO3 N in runoff water entering the wetland c ould be attributed to the controlled release (coated urea), and soluble nitrate fert ilizers (Ammonium Nitrat e) applied to nursery plants (S. Chandler, Horticulture Manager Monr ovia Nursery personal comm unication). It is not uncommon to apply up to 200 mg L-1 nitrate nitrogen to ornament al crops during peak growing season. The concentration of NO3 N (mg L-1) entering the wetland vari ed during the period of study, i.e., it gradually decreased from April to August 2007. Th e higher concentration of NO3 N during April was likely due to fertilization prac tices at the nursery du ring their peak growing season (S. Chandler, Horticulture Manager M onrovia Nursery personal communication). In addition, the limited rainfall (15-17 cm during Ap ril and May 2007) might have contributed to increased concentration of nutrients in the runoff. Similarly, increased con centration of nutrients in the nursery runoff was observed due to fertilization in subsurface horizontal flow reed beds in subtropical region of Australia (Headley et al., 2001). Duri ng April and May there were comparatively higher concentrations (44 mg L-1 and 36 mg L-1, respectively) of NO3 N entering the wetland as compared to June, July and August. Higher concentration of NO3 N during April (averaging 19 mg L-1) and May (averaging 18 mg L-1) were also obtained by Taylor

PAGE 81

81 et al. (2006) during their four year (2002 2005) monitoring of influent of the constructed wetland at Monrovia Nursery. In addition, application of ferti lizers at a near by horticultural nursery contributed to the incr ease in the concentration of NO3 N (0.03 to 4.73 mg L-1) through runoff in the tributaries of Santa Fe river waters hed in north central Flor ida (Frisbee, 2007). While assessing the composition of TN dur ing the study period, it was observed that concentration of organic N at the outflow was hi gher as compared to inflow of the study cell, which might be due to internal generation of or ganic N in the wetland during decay of plant and microbes. Persistence of organic N in low conc entrations was always observed in the outflow water of a constructed wetland (Kadlec a nd Knight, 1996; Headley et al., 2001). Water temperature in study wetland was obser ved to gradually increase from May to August. The NO3--N removal efficiency was obs erved to gradually increase during the same period. Nitrate nitrogen removal might have been enhanced by the increase in temperature, which probably resulted in vigorous plant growth and act ivity of the microorganisms. The mean temperature of the wetland (over all locations, depths and months) was 26 C. The favorable range of temperature for denitrification is between 20 and 25 C (Spieles and Mitsch, 2000), which is close to the range of the mean temperature of the study ce ll. The influence of temperature on nitrogen removal was explained by Reddy and Patrick (1984). The rates of denitrification have been shown to increase by 1 .5-2 times with each in crease in 10oC. Water temperatures less than 15C or greater than 30C have been shown to limit the rate of denitrification (Reddy and Patrick, 1984). The gradual increase in NO3 --N removal efficiency over Ap ril to August could also be related to the growth of the plants in the wetlan d. The perennial plants in the wetland started growth in April. During the vigorous growing stages (April, May, and Ju ne), plants utilize NO3 --

PAGE 82

82 N and thereby may have contributed to th e increased removal efficiency of the NO3 N. Studies have shown that macrophytes in the we tlands remove substantial amount of NO3 --N directly by uptake or indirectly by provi ding carbon exudates from roots, which enhance microbial transformations (Gersberg et al., 1983; Bachand and Horne, 2000). The NH4 +-N contributed only less than 1% of the TN load into the wetland. The removal efficiency of the NH4 +-N ranged from 41% (June) to 93% (August). The concentration of NH4 +N coming into the wetland was approximately 2 mg L-1, and the outflow con centration less than 1 mg L-1. The NH4 +-N removal is likely because of plant uptake or nitrification process in the aerobic sites of the wetlands. Immobilization by plants and microbes enhances the removal of NH4 +-N. In addition, microbe mediated nitrification process, in which NH4 +-N is oxidized to NO3 N also contributes to NH4 +-N removal in the aerobic zones of wetlands (Burger and Jackson, 2004). The volatilization of NH4 +-N to ammonia would not be a significant process for NH4 +-N removal in this wetland as the mean pH value (6.0) was lower than optimal (> 8) for this chemical process to occur (USEPA, 1993, Sartoris et al., 2000). The pH values in this system did not reach that optimum level during the study period. Plant growth and the size of denitrifier population within the rhizosphere are observed to be coupled. Hence the influence of pH on plant growth in turn affects th e denitrification pote ntial (Hall et al., 1998 ). However, the influence of pH on plant growth was not st udied in this system during this study. The concentration of TP entering and leavi ng the wetland cell was approximately 1 mg L-1. The removal efficiency of TP ranged from 0-32%, which varied over months. The removal of TP (which consists orthophosphate) may be attr ibuted to biological uptake, adsorption onto sediment, etc. The percent TP removal effici ency over months did not follow any particular trend, however. Similar results of variable TP net removal during the growing season were

PAGE 83

83 observed by Taylor et al. (2006) in the constructed wetland at Monrovia nursery. The few negative removal efficiency values of TP indi cate that there is likely internal loading of phosphorus. Macrophytes, algae, and microorganism s utilize P as an essential nutrient and can enhance temporary removal of TP. Long term TP removal, however, is limited to the P sorption capacity of the sediments (Kadlec, 1999). The concentration of DOC did not follow a part icular trend over mont hs. The distribution of DOC seemed uniform through the wetland in the study cell. The nursery runoff had a low dissolved organic carbon (DOC) concentration (a characteristic of nursery runoff). The low dissolved organic carbon (DOC < 20 mg L-1) was observed in the subsurface flow reed bed systems receiving plant nursery and agricultural runoff (Headley et al., 20 01; Huett et al., 2005; Crumpton et al., 1993) in Australia. The DOC: NO3 N ratio was observed to be low (<3:1) during most of the study period. The modeled effect of hydraulic loading rate and plant density on theoretical residence time was noticeable. When higher values of hydraulic loading rate were us ed in the model, the theoretical residence time decreased by a day for th is system. The model showed that increasing the wetland porosity from 0.75 to 0.95 and/or 1.0 would increase the residence time in the study cell by one day. Based on the model, the theoretical residence time for this system ranged from 3 to 5 days. Headley et al. ( 2001) reported that the maximum N removal can be achieved even at 2 day residence time in the subsurface flow reed beds. It was estimated that a nursery of 1 ha area would require a re ed bed area of 200 m2 for a 2 day hydraulic residence time. Spatial effects on nutrients, physiochemical parameters and denitrification potential of the water column The higher DO level at outflow can be attributed to the higher prevalence of open water towards the outflow side of the we tland compared to the inflow side. The depth of

PAGE 84

84 the water column showed significant influence on DO level, which is expected in waterlogged conditions. The anoxic condition of the wetland was enhanced with increasing depth of the water column. The DO concentration graduall y decreased from April to August, when the temperature gradually increased from April to August. The reduction in concentration of DO could be attributed to increased activity of microorganisms, whic h enhances N removal, with the gradual increase in temperature. The nutrients (N and P) analyzed in this study varied in their concentration at inflow and outflow. This variation in the concentration of nutrients may be attributed to removal of nutrients within the study wetland. Similar nutrient gradient from the inlet to outlet points was observed in Evergladess wetlands soil (White and Reddy, 1999) and in experimental wetlands receiving agricultural runoff (Sirivedhin and Gray, 2006). The low concentration of NO3 N at outflow is likely due to the removal of nitroge n by various mechanisms such as plant uptake, microbial assimilation, and denitrification along the flow of water from inflow. The low concentration of NO3 N at lower depth of the water colu mn is likely due to the removal of NO3 N by plant uptake and denitr ification process in the anaerobi c zones of the water column. Another possible reason might be the lack of diffusion of NO3 N to the lower depths of the water column, which will also limit the denitrification activity. Although the concentration of NH4 +-N was always higher at inflow as compared to outflow of the wetland, there was no trend observed in the spatia l distribution of NH4 +-N within the wetland cell over the pe riod of five months. The lower concentration of NH4 +-N at outflow could be attributed to the removal of NH4 +-N within the study cell. However, the concentration of NH4 +-N was observed to be higher in lower depth of the water column, which is likely due to lack of nitrification in the anoxi c zones of the water column.

PAGE 85

85 Significant spatial distribution in total phosphorus concentrat ion was observed within the wetland. However, there were no temporal effect s on TP concentration during the period of this study. The TP was observed to be higher in the deeper water column as compared to the surface of the water column, but only at th e inflow side of the wetland. This could be due to the release of adsorbed P from the sediments into immediate atmosphere, or due to lack of adsorption sites near the inflow of the study cell. The sorpti on sites on the inflow side might have become saturated because of the higher co ncentration of TP at inflow. The DOC concentration did not show any differences in its distribution within the depth of the water column, suggesting its uniform distribution within the wetland. The denitrification pote ntial of the water colu mn in the study cell was either negligible or undetectable. The maximum value of denitrificati on potential observed in the water column was 0.03 mg N2O-N L-1 hr-1. Lower values of denitrification pot ential are likely due to a lack of particulate material in the water column to support the microbial activity. Little research has been conducted on denitrification potential of the water column in constructed wetlands, but whenever it was included in wetland studies, the reported values were always relatively low. Denitrification rates of the water samples in a constructed wetland receiving sewage treatment plant effluent were observed to be between 0.06 0.96 g N m-2 d-1 (Toet et al., 2003). The water column of the study cell had low dissolved organic carbon (DOC) concentration, and this may be limiting NO3 N removal due to non availability of energy for denitrifying microbes. The optimum C:N (plant carbon added:NO3 --N in water) ratio for the denitrification in wetlands is reported to be 4:1 (Ingersol l and Baker, 1998) to 5:1(Baker, 1998). Only during the month of April, the DOC: NO3 N ratio reached near 4:1 in the wetland. The mean DOC: NO3 N ratio was low (<1) during most of the study period. Low DOC: TN reported to limit NO3 N removal

PAGE 86

86 in a subsurface flow wetland receiving plant nu rsery runoff (Huett et al., 2005). The absence of attachment sites for bacteria and/or low organic car bon availability possibly resulted in low denitrification rates in the wa ter, as reported by Toet et al (2003). The distribution of denitrifying microbes was most lik ely regulated by the availabili ty of organic material, with higher denitrification rates in the sediment s than on surface in the water column. The DO concentration in the water column of the study cell ranged from 6 mg L-1 to 12 mg L-1, which might have affected the denitrification process, as well. Poe et al. (2003) reported that the denitrification potential was infl uenced by the concentration of NO3 -N, and the rates of denitrification were observed to increase with an increase of NO3 N. Such trend was not observed in this study. Denitrification potential of macrophyte rhizosphere soil. Denitrification potential of macrophyte rhizosphere soil was influenced by months; however, there was no trend between April and August 2007. The mean denitrification pot ential in the rhizosphe re of the two species combined was 7.6, 1.7, 0.9, and 3.0 mg N2O-N kg-1 hr-1, in May, June, July, and August 2007, respectively. Only the values obtained in May significantly differed from other months. Although there was a gradual incr ease in temperature of the water column from May (22 C) to August (30 C), the rhizosphere denitrific ation potential did not show significant difference over months. The mean temperature (26 C) of the wetland during the study period was close to the optimum range, 20-25 C, for denitrification activity. Results of this study showed that mean deni trification potential over five months for Canna flaccida and Typha latifolia was 3 mg and 4 mg N2O-N kg-1 hr-1, respectively. Data variability within replicates was observed for denitrification potential values in this study. The variation in values obtained for replicates in this study might be due to the heterogeneity in the

PAGE 87

87 sampling point as affected by plants and surrounding environmen t, water movement, etc. Similar data variability in the denitrificati on rates between the repl icates was observed by Bastviken at al. (2003). Regardless, the denitrification values obtai ned in this study are relatively high when compared to several other studies (Table 4-2). Hunt et al. (2003) obtai ned higher denitrification potential of 0.21 mg N kg-1 hr-1 in Typha dominated wetland soils, as compared to 0.52 mg N kg1 hr-1 in Juncus dominated wetlands used for treatment of swine wastewater. Denitrification potential values observed in this study also lie in the range of denitrif ication potential (0.004 7.75 mg N kg-1 hr-1) obtained for everglades wetla nd soils (White and Reddy, 1999). Denitrification rates in soil samples from Juncus -planted wetland plots were reported to be higher (12.3 mg N m-2 d-1) as compared to Salix plots (2.65 mg N m-2 d-1). The denitrification rates obtained from soil samples amended with nitrate and/or glucose, indicated that the denitrification was limited by th e availability of carbon and NO3 N in the Salix plots, whereas only by NO3 -N in the Juncus plots (Smialek et al 2006). Denitr ification potential in rhizosphere soil of Lolium perenne L.Melinda (Rye grass) was observed to be 150 900 ng N g-1 hr-1, and followed the pattern of plant grow th, indicating that the plant grow th and the size of denitrifier population within the rhizosphere are likely correlated (Hall et al ., 1998). Ottosen et al. (1999) observed low denitrification activity in the rhizosphere of Zostera marina and Potamogeton pectinatus (1.5 to 5 mol N m-2 h-1) as compared to Lobelia dortmanna and Littorella uniflora vegetated sediments (24 and 30 mol N m-2 h-1). Statistically similar denitrification potential in Canna flaccida and Typha latifolia in this present study might be due to their similar effect on the rhizosphere. Other researchers have observed differences in denitrification potenti al in wetland soil and sediments planted with

PAGE 88

88 different taxa of plants. These differences were attributed to di fferences in the ability of plant roots to diffuse oxygen into the rhizosphere, and al so due to differences in leakage of available organic carbon from plant roots. Most of the studies on denitrific ation potential have used the sediments and bulk soil in the rhizosphere of the plants as substrate. Results from this study show th at denitrification rates comparable to or higher than those reported from bulk soil of vegetated pl ots can be obtained in rhizosphere soil closely a dhered to the roots.

PAGE 89

89Table 4-1. Nitrate removal rates in cons tructed wetlands, under different hydraulic re sidence periods, as reported by other re searchers. System Wastewater Hydraulic residence time (days) Nitrate removal rates (g m-2 day-1) Flow through wetland microcosms Constructed free water surface wetlands Constructed wetlands Constructed free water surface wetlands Nitrate contaminated water Landfill leachates Agricultural tile drainage River flow 2.4 10-12 7 1-10 2.8-5 0.63 0.291.51 0.56 Ingersoll and Baker, 1998 Kozu and Lieh, 1999 Xue et al., 1999 Bachand and Horne, 2000 Natural forested treatment wetland Constructed wetlands Municipal wastewater Groundwater 0.9 1.1 4.2 0.10 0.94 Blahnik and Day, 2000 Lin et al., 2002 Subsurface flow constructed wetlands Domestic effluent 10.5 0.61 Bayley et al., 2003 Surface flow constructed wetlands Nurs ery runoff 3-5 1.75 This study

PAGE 90

90Table 4-2. Denitrification potentials of di fferent substrate reported by other research ers in constructed wetlands. While the units of measure are different and make direct co mparisons challenging, these data are useful for maki ng general inferences for wetlands dominated by different plant taxa. The acronyms used are as follows: Denitrifica tion Enzyme activity (DEA), Acetylene Inhibition Technique (AIT), Membra ne Inlet Mass Spectrometry (MIMS), and N15 tracers (N15). Substrate Wastewater type DEA (mg N m-2 d-1) Range of DEA values Method Reference Phragmites australis shoots Sewage treatment plant effluent 44.4 121 mg N m-2 d-1 AIT Toet et al., 2003 Elodea nuttallii shoots Sewage treatment plant effluent 14.8 33.1 mg N m-2 d-1 AIT Toet et al., 2003 Sediments Sewage treatment plant effluent 0.5 25.5 mg N m-2 d-1 AIT Toet et al., 2003 Water Sewage treatment plant effluent 0.4 3.9 g N m-2 d-1 AIT Toet et al., 2003 Spartina spp., Cladium spp. Juncus spp. sediment cores Agricultural runoff 2.5 0.7 9.2 mg N m-2 d-1 MIMS Poe at al., 2003 Juncus spp. sediment Agricultural runoff 2.3 0.1-6.0 mg N m-2 d-1 AIT Thompson et al., 2000 Everglades soils Agricultural runoff 0.004 7.75 mg N kg-1 hr-1 AIT White and Reddy, 1999 Mixed plants sediments Agricultural tile drainage 2-11.8 mg N m-2 d-1 AIT /N15 Xue et al.,1999 Lolium perene (Ryegrass) Rhizosphere soil Wetland microcosms 150 900 ng N g-1 hr-1 AIT Hall at al., 1998 Typha latifolia Canna flaccida Plant nursery runoff Plant nursery runoff 2.51 mg N kg-1 hr-1 4.11 mg N kg-1 hr-1 AIT AIT This study This study

PAGE 91

91 CHAPTER 5 SUMMARY AND CONCLUSIONS Constructed wetlands are cost effective and viable method for on-site removal of nonpoint source contaminants from water before it is rele ased into aquatic systems. These systems are used to treat wastewater from municipal water source, animal waste, urban runoff, stormwater runoff, and agricultural runoff. Few studies have focused on constr ucted wetlands treating nursery runoff (Headley et al., 2001 ; Lea-cox et al., 2002; Huett et al., 2005). In general, the performance of constructed wetlands varies and partly depends on the source and the characteristic features of influent wastew ater (Mitsch and Jorgensen, 1989; NRCS, 2000a; Dunne et al., 2005). When constructed wetlands are to be used for nutrient, especially N, removal, management strategies can be optimized by understanding the mechanisms by which N is removed within the constructed wetlands. In view of the necessity of a systems approach, which recognizes the site specific conditions, nitrogen (N) dynamics in the constructed wetland receiving plant nursery runoff was investigated to f acilitate improved understanding of N transformations, factors influencing N removal, and to optimize the perfor mance of the system for on-site N removal. The concentration of nutrients N and phosphorus (P) entering and leav ing the constructed wetland was monitored from April 2007 to A ugust 2007 to determine the nutrient removal efficiency of the study cell in the constructed wetlands receiving plant nursery runoff. The dominant nutrient in the plan t nursery runoff was N, in the form of nitrate nitrogen (NO3 --N), followed by phosphorus (P). The concentration of DOC was low (< 20 mg L-1) which is a characteristic feature of the plant nursery runoff. It was hypothesized that the concentration of N and P would vary over months based on the timing of the fertilizer application, irrigati on, and rainfall. The hi gher concentration of NO3

PAGE 92

92 N during April and May in this study was likely due to fertilization practices at the nursery during their peak growing seas on. In addition, the limited rain fall (15-17 cm during April and May 2007) might have contributed to increased concentration of nutrients in the runoff. The concentration of nutrients, as expected, was higher in the influent as compared to the outflow. The mean nutrient removal efficiency of the constructed wetland was 40% for TN, 40% for NO3 --N, 59% for NH4 +-N, and 16% for TP. While comparing to other studi es, the study cell appears to be functioning moderately efficiently in removing N and P. The NO3 --N removal efficiency was inversely relate d to the loading rate of NO3 --N in this study. The P removal might be due to plant, algal uptake and adsorption/precipitation reactions. The negative P removal efficiency in July indicates that at times ther e may be internal loading of P in the system. Ammonium Nitrogen removal is like ly due to plant uptake or nitrif ication process in the aerobic zones of the wetlands. The volatilization of NH4 +-N to ammonia is probably not a significant process for NH4 +-N removal in the wetland because the observed pH in the wetland was lower that what is reported to be optimal for this process to occur. The low DOC: NO3 --N ratio in the influent water indicates that the system mi ght be limited by carbon availability, which could adversely affect NO3 N removal. Hydraulic loading rate and the plant density were observed to influence the theoretical residence time in the model. When higher valu es of hydraulic loading rate were used in the model, the theoretical residence time decreased by a day for this system. The model showed that increasing the wetland porosity from 0.75 to 0.95 and/ or 1.0 would increase th e residence time in the study cell by one day. The nutrients (N and P) analyzed in this st udy exhibited a gradient from the inflow to outflow due to removal of nutrients within the wetland. The low NO3 --N concentration at lower

PAGE 93

93 depth of the water column is likely due to utilization of NO3 --N by microorganisms, during the process of denitrification in the anoxic zones. The high NH4 +-N concentration in the lower depth of the water column is likely due to the lack of nitrification process in the anoxic lower depth of the water column, and hence the accumulation of NH4 +-N. Higher concentra tion of TP in the lower depth of the water column is likely due to release of adsorbed P from the sediments, or due to lack of adsorption site s near the inflow of the study cell. The sorption sites on the inflow side might have become saturated because of the higher concentration of TP at inflow. Concentration of DOC did not show any differences in distri bution along the depth of the water column, and across the locations over mont hs, indicated uniform distri bution within the wetland. Gradual increase in the temper ature of the water column from May to August 2007 might have enhanced NO3 N removal by affecting plant growth an d activity of the microorganisms. Denitrification potential values were observed to follow the patte rn of plant growth, indicating that the plant growth and the size of denitrif ier population within the rhizosphere might be coupled. Hence, physiological parameters affecting the plant growth can al so indirectly affect denitrification. A direct relati onship was observed between the NO3 N removal efficiency and temperature of the water column during the study period. As the temperature gradually increased from May to August, the DO level gra dually decreased. This is likely due to the enhanced activity of the microorganisms utilizin g oxygen as an electron acceptor. The higher DO level at outflow can be attrib uted to open water towards the d eeper side of the wetland. Due to the potential differences in the temperature over months, DO level in the water column, and concentration of NO3 -N at inflow and outflow, it was expected that the denitrification potential of the water column would be different as well. However, the low (0.01-0.03 mg N2O-N L-1 hr-1) or undetectable denitrification potential of the wate r column is possibly due to the low DOC

PAGE 94

94 concentration/low DOC: NO3 -N, which limits denitrification. The presence of few attachment sites for bacteria possibly resulted in low denitrification rates in the water column. Mean rhizosphere denitrification potential over five months for Canna flaccida and Typha latifolia was 3 mg and 4 mg N2O-N kg-1 hr-1, respectively. The mean de nitrification potential of two species was statistically sim ilar. Regardless, the denitrifica tion values obtained in this study are relatively high when compared to several othe r study studies. Hunt et al. (2003) obtained higher denitrification po tential of 0.21 mg N kg-1 hr-1 in Typha dominated wetland soils, as compared to 0.52 mg N kg-1 hr-1 in Juncus dominated wetlands used for treatment of swine wastewater. Rhizosphere denitrif ication potential values of Canna flaccida and Typha latifolia in this study were comparatively higher than the rh izosphere denitrificati on potential values of other plant species Lolium perenne Zostera marina, Potamogeton pectinatus, Lobelia dortmanna and Littorella uniflora Results from this study show that denitrification rates comparable to or higher than those reported from bulk soil of vegetated pl ots can be obtained in rhizosphere soil closely a dhered to the roots. Nitrate nitrogen removal efficiency of the constructed wetland receiving plant nursery runoff was observed be 40% over five month pe riod, with an average re moval rate of 1.75 g m-2 day-1. The mean loading rate of NO3 -N was 4.43 g m-2 day-1. Compared to mean removal rate of NO3 -N (1.75 g m-2 day-1), the amount of NO3 -N removed through coupled (0.1 g N2O-N kg1 day-1) denitrification potential [mean denitrifica tion potential of water column (0.72 mg N2O-N kg-1 day-1) plus mean rhizosphere denitr ification potential (96 mg N2O-N kg-1 day-1)] was lower. Therefore, it is likely that the rest of NO3 -N would be removed by other pathways contributing to removal of NO3 -N. The major NO3 -N removal pathways are plant and microbial

PAGE 95

95 assimilation and sediment and/or so il denitrification. Rhizosphere denitrification pot ential in this study was observed to signifi cantly contribute towards NO3 -N removal.

PAGE 96

96 LIST OF REFERENCES Aldrich, R.A., and J.W. Bartok. 1994. Greenhouse engineering. Natural Resource Agriculture and Engineering Service (NRAES) Publication No. 33, Ithaca, NY. Armentano, T.V., and J.T.A. Verhoeven. 1990. Wetland processes: Global biogeochemical cycles. In : Patten, B.C. (ed.) Wetlands and shallow continental water bodies. Vol. 1. SPB Academic Publishing, The Hague. Ayaz, S.C., and L. Acka. 2001. Treatment of wa stewater by natural systems. Environmental International 26:189-195. Bachand, P.A.M., and A.J. Horne. 2000. Denitr ification in construc ted free water surface wetlands: Effects of vegetation and temp erature. Ecological Engineering 14:17-32. Baker, L.A. 1998. Introduction to nonpoint source pollution in the United States and prospects for wetland use. Ecological Engineering 1:1-26. Bastviken, S.K., P.G. Eriksson, I. Martins, J. M. Neto, L. Leonardson, and K. Tonderski. 2003. Potential nitrification and denitrification on different surfaces in a constructed treatment wetland. Journal of Environmental Quality 32:2414-2420. Bayley, M.L., L. Davidson, and T.R. Headley. 20 03. Nitrogen removal from domestic effluent using subsurface flow constructed wetlands: influence of depth, hydraulic residence time and pre-nitrification. Water Sc ience and Technology 48:175-182. Berghage, R.D., E.P. MacNeal, E.F. Wheeler, and W.H. Zachritz. 1999. Green water treatment for the green industries: opportunities fo r biofiltration of greenhouse and nursery irrigation water and runoff with constructed wetlands. Hortscience 34:50-54. Berndt, M.P., E.T. Oaksford, M.R. Darst, a nd R.L. Marella. 1996. Environmental setting and factors that affect water quality in the Ge orgia-Florida Coastal Plain: US Geological Survey Water-Resources Investigations. Black, R.J. 1993. Florida Climate Data. Florida Energy Extension Service Publication EES-5, University of Florida, Gaines ville. http://e dis.ifas.ufl.edu. Blahnik, T and J Day. 2000. The effects of varied hydraulic and nutrient loading rates in water quality and hydrologic distributions in a na tural forested treatment wetland. Wetlands 20:48-61. Brix, H. 1997. Do macrophytes play a role in co nstructed treatment wetlands? Water Science and Technology 35:11-17. Burger, M., and L.E. Jackson. 2004. Plant and mi crobial nitrogen use and turnover: Rapid conversion of nitrate to ammonium in so ils with roots. Pl ant and soil 266:289-301.

PAGE 97

97 Burgin, A.J., and S.K. Hamilton. 2007. Have we ove remphasized the role of denitrification in aquatic systems? A review of nitrate rem oval pathways. Frontiers in Ecology and the Environment 5:89-96. Carpenter, S.R., N.F. Caraco, D.L.Correll, R. W. Howarth, A.N. Sharpley, and V.H. Smith. 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological Applications 83:559. Coffey, S. 1997. Selected agricu ltural best management practices to control nitrogen in the Neuse River Basin. North Carolina State Un iversity, Raleigh, NC, USA. Technical Bulletin 1-50. Cooper, P.F., and B.C. Findlater. 1990. Constr ucted wetlands in water pollution control. Pergamon Press, Oxford. Crumpton, W.G., T.M. Isenhart, and S.W Fisher 1993. Controlling agricultural runoff by use of constructed wetlands. In : Moshiri, A. (Ed.), Constructed Wetlands for Water Quality Improvement. Lewis Publishers, Ann Arbor, MI. DAngelo, E.M., and K.R. Reddy. 1999. Regulators of heterotrophic microbial potentials in wetland soils. Soil Biology and Biochemistry 31:815-830. Dahl, T.E. 1990 Wetlands losses in the United St ates, 1780`s to 1980`s. Report to the Congress. United States Department of the Interior. Fish and Wildlife Service. Washington DC USA. Darbi, A., T. Viraraghavan, R. Butter and D. Co rkal. 2003. Pilot-scale evaluation of select nitrate removal technologies. Journal of Envir onmental Science and Health 38:1703-1715. DeBusk, W.F. 1999. Functional role of wetlands in watersheds.SL 169 Florida Cooperative Extension Service. IFAS. University of Fl orida. Gainesville. http://edis.ifas.ufl.edu. Desimone, L.A., and B.L. Howes. 1996 Denitrification and nitrogen transport in a coastal aquifer receiving wastewater discharge. Environmental Science and Technology 30:1152-1162 Dortch, M.S. 1992. Literature analysis addresses the functional ability of the wetlands to improve water quality. The wetlands rese arch program bulletin. 2: 1-4 USACE Waterways Experimental station, Vicksburg, MS. Dunne, E.J., N. Culleton, G.O. Donovan, R. Harrington, and A.E. Olson. 2005. An integrated constructed wetland to treat contaminants and nutrients from dairy farmyard dirty water. Ecological Engineering 24:221-234. Eriksson, P.G., and S.E.B. Weis ner 1997. Nitrogen removal in a wastewater reservoir: The importance of denitrification by epiphytic bi ofilms on submersed vegetation. Journal of Environmental Quality 26:905-910.

PAGE 98

98 Faeth, P. 2000. Fertile ground: Nutrient Tradings potential to cost effectively improve water quality. World Resources In stitute. Washington DC. USA Franklin, E.C., J.D. Gregory, D.W. Hazel, and J.E. Parsons. 2000. Manage ment of filter zones for dispersion and treatment of agricultural runoff. Water Resources Institute, Raleigh, NC, USA. 312/202. Frisbee, A.E. 2007. Nitrate nitrogen dynamics in tributaries of the SantaF e river watershed, north central Florida. MS Thesis University of Florida. Gale, P.M., I. Devai, K.R. Reddy and D.A. Grat ez. 1993. Denitrification pot ential of soils from constructed and natural wetlands. Ecological Engineering 2:119-130. Gersberg, R. M., B. V. Elkins and C. R. Gold man. 1983. Nitrogen removal in artificial wetlands. Water Research 17:1009. Gersberg, R.M., B.V. Elkins, S.R. Lyons, and C. R. Goldman. 1986. Role of aquatic plants in wastewater treatment by artificial wetlands. Water Research 20:363. Gren, I.M. 1995. The value of investing in wetland s for nitrogen abatement. European review of Agricultural economics 22:157-172. Hall, J.M., E. Paterson, and K. Killham. 1998 The effect of elevated CO2 concentration and soil pH on the relationship between plant growth and rhizosphere denitr ification potential. Global Change Biology 4:209-216. Hammer, D.A. 1992. Designing constructed wetlands systems to treat agricultural nonpoint source pollution. Ecological Engineering 1:49-82. Hammer, D. A., and R.L. Knight. 1994. Designing constructed wetlands for nitrogen removal. Water science and Technology 29:15-17. Headley, T.R., D.O. Huett, and L. Davidson. 2001. The removal of nutrients from plant nursery irrigation runoff in subsurface horizontal fl ow wetlands. Water Science and Technology 44:77-84. Hoag, J.C. 1997. A Constructed wetland system for water quality improvement of nursery irrigation wastewater. p 132-135. In : Landis, T.D., Thompson, J.R. (ed). National Proceedings, Forest and Conservation Nurser y Associations. General Technical Report. PNW-GTR-419. Portland. OR: US Department of Agriculture, Forest Service, Pacific Northwest Research Station. Hodges, A.W., and J.J. Haydu. 2002. Economic Impacts of the Florida Environmental Horticulture Industry, 2000. Economic Inform ation Report EIR-02-3, Food and Resource Economics Department, University of Florida, Gainesville, FL. http://hortbusiness.ifas.ufl.edu/EIR02-3.pdf

PAGE 99

99 Horner et al 1994 Horner, R.R., J.J. Skupien, H.L. Eric and H. E. Shaver. 1994. Fundamentals of Urban Runoff Management: Technical and Institutional Issues, Terrence Institute. Howarth, R. W., G. Billen, D. Swaney, A. Townse nd, N.Jaworski, K. Lajtha, J. A. Downing, R. Elmgren, N. Caraco,T. Jordan, F. Berendse, J. Freney, V. Kudeyarov, P. Murdoch, and Zhu Zhao-liang. 1996. Regional nitrogen budgets and riverine inputs of N and P for the drainages to the North Atlantic Ocean: natu ral and human influences. Biogeochemistry 35:75. Huett, D.O., S.G. Morris, G. Smith, and N. Hunt. 2005. Nitrogen and phosphorus removal from plant nursery runoff in vegetated and unve getated subsurface flow wetlands. Water Research.39:3259-3272. Ingersoll, T.L., and L.A. Baker. 1998. Nitrate removal in wetland microcosms. Water research.32: 677. Jerardo, A. 2005. Floriculture and nursery crops situation and outlook yearbook. FLO-2005, Economic Research Service / USDA. Washington, DC. Kadlec, R.H. 1999. The limits of phosphorus removal in wetlands. Wetlands Ecology and Management 7:165,1999. Kadlec, R.H., and R.L. Knight 1996 Treatment Wetlands. CRC Press, Boca Raton, FL, USA. Kadlec, R.H., and K.R. Reddy. 2001. Temperatur e effects in treatment wetlands. Water Environment Research 73:543-55. Knowles, R. 1982. Denitrification. Microbiological Reviews 46:43-70. Knowles, R. 1990. Acetylene inhibition technique: development, advantages and potential problems. p 151-165 In Revsbech, N.P. and Sorensen, J. (e ds.) Denitrifica tion in Soil and Sediment. Plenum Press, New York, NY, USA. Kozu, D.D.and S.K. Lieh. 1999. Assessing denitrification rate limiting fact ors in a constructed wetland receiving landfill leachate. Water Science and technology 40:75-82. Kristiansen, S., and M.T. Schaanning. 2002. Den itrification in the wa ter column of an intermittently anoxic fjord Hydrobiologia 469:77-86. Lea-cox, J.D., and D.S. Ross. 2001. Clean water po licy and the rationale for developing a water and nutrient management planning proce ss for container nursery and greenhouse operations. Journal of environm ental Horticulture 19:1-8. Lea-cox, J.D., D.S. Ross, A.G. Ristvey, a nd J.D. Murray. 2002. Estimating nitrogen and phosphorus Total Maximum Daily Loads for container nursery and greenhouse production systems. p 466-471. In Total Maximum Daily Load (TMDL) Environmental Regulations: Proceedings Conference Fort Worth, Texas, USA. 11-13 March, 2002.

PAGE 100

100 Lin, Y.F., S.R. Jing, D.Y. Lee, and T.W. Wang. 2002. Effect of macrophytes and external carbon sources on nitrate removal from groundwater in constructed wetlands. Environmental pollution 119:413. Lu, W., and J.L. Sibley. 2006. Modeling of water requirements for container production using overhead irrigation. p 518-520. In M. Taylor. (ed.) Water Management. Proceedings of Research Conference, GA, 2006, Southern Nursery Association, Atlanta, GA. Marella, R.L., and J.L. Fanning. 1996. National wa ter quality assessment of the GeorgiaFlorida Coastal Plain study unit wate r withdrawals and treated wastewater discharges, 1990. US Geological Survey WRIR. Martin, J.F., and K.R. Reddy. 1997. Interaction and spatial distribution of wetland nitrogen processes. Ecological modeling 105:1-21. Mitsch, W.J. 1992. Landscape design and the role of created, restored, and natural riparian wetlands in controlling nonpoint source pol lution. Ecological Engineering 1:27-47. Mitsch, W.J., and J.G. Gosselink 2000. Wetlands John and Wiley & Sons, Inc., New York. Mitsch, W.J., and S.E. Jorgensen.1989. Ecol ogical engineering: An introduction to ecotechnology. John and Wiley & Sons, Inc., New York. Moorhead, K.K., and K.R. Reddy. 1988. Oxyge n transport through selected aquatic macrophytes. Journal of Envi ronmental Quality 17:138-142. National Research Council (NRC). 1992. Restora tion of aquatic ecosystems Washington DC; National Academy press. National Water Quality Inventory (NWQI). 1998. http://www.pollutionissues.com/NaPh/Nonpoint-Source-Pollution.htm l National Resources Conservation Service (NRCS). 2002 Constructed wetlands. Environmental Engineering National Engineering Handbook. USDA and NRCS, Washington DC. Ottosen, L.D.M., N. Risgaard-Petersen, and L.P. Nielsen. 1999. Di rect and indirect measurements of nitrification and denitr ification in the rhiz osphere of aquatic macrophytes. Aquatic Microbial Ecology 19:81-91. Paredes, D., P. Kuschk, F. Stange, R.A.Mulle r, and H. Koser., 2007. Model experiments on improving nitrogen removal in laboratory scale subsurface constructed wetlands by enhancing the anaerobic ammonia oxida tion. Water Science and Technology 56:145 150. Park, J.H., H.S. Shin, I.S. Lee, and J.H. Bae. 2002. Denitrification of High NO3 --N Containing Wastewater Using Elemental Sulf ur; Nitrogen Loading Rate and N2O Production. Environmental Technology 23:53-65.

PAGE 101

101 Patrick, W.H. Jr. 1994. From wastelands to wetlands. Journal of Environmental Quality 23 : 892 896. Platzer C. 1996. Enhanced nitrogen elimination in subsurface flow artific ial wetlands a multi stage concept. International Conference of wetland systems for water pollution control. Vienna, Austria. Poe, A.C., M.F. Piehler, S.P. Thompson, and H. W. Paerl. 2003. Denitrification in a constructed wetland receiving agricultural runoff. Wetlands 23:817-826. Reddy, K. R. and E. M. DAngelo.1994. Soil processes regulating water qua lity in wetlands.p. 309-324. In : Mitsch, J.W. (ed.) Global Wetlands: Old World and New. Elsevier, NY. Reddy, K. R., and R. D. Delaune. 2007. Biogeochemistry of Wetlands: Science and Applications. CRC Press, Boca Raton, Fl. (in press). Reddy, K.R., and W.H. Patrick 1984. Nitrogen transformations and loss in flooded soils and sediments. Critical Reviews in Environmental Control 13:273-309. Reddy, K.R., E.M. DAngelo and T.A. De Busk. 1989. Oxygen transport through aquatic macrophytes: the role in wastewater treat ment. Journal of Environmental Quality 19: 261. Reddy, K.R., P.S.C. Rao, and R.E. Jessup. 1982. The effect of carbon mineralization on denitrification kinetics in mineral and orga nic soils. Soil scien ce society of America journal 46:62-68. Reddy, K.R., W.H. Patrick, and C.W. Lindau. 1989. Nitrification-de nitrification at the plant root sediment interface in wetlands. Limnology and Oceanography 34:1004. Reed, S.C., R.W. Crites and E. J. Middlebrooks, 1995 Natural Systems for Waste Management and Treatment (2nd ed.), McGraw Hill, New York Reinhardt, M., B. Muller, R.Gachter, B. Wehr li. 2006. Nitrogen removal in a small constructed wetland: An isotope mass balance appro ach. Environmental science and technology 40:3313-3319. Reinhardt, M., R.Gachter, B. Wehrli, and B. Muller. 2005. Phosphorus retention in small constructed wetlands treating ag ricultural drainage water. Journal of Environmental Quality 34:1251. Risgaard-Petersen, N., and K. Jenson. 1997. Nitrificat ion and denitrification in the rhizosphere of the aquatic macrophyte Lobelia dortmanna L. Limnology and Oceanography 42:529-537. Sartoris, J.J., J.S. Thullen, L.B. Barber, D.E. Salas. 2000. Investigation of nitrogen transformations in a southern California constructed wastewater treatment wetland. Ecological Engineering 14:49-65.

PAGE 102

102 Seitzinger, S.P. 1993. Denitrification measurements in aquatic sediments: A comparison of three methods. Biogeochemistry 23:147-167. Shipin, O., T. Koottatep, N.T.T. Khanh and C. Polprasert. 2005. Integrated natural treatment systems for developing communities: lowtech N-removal through the fluctuating microbial pathways. Water Science and Technology 51:299. Sirivedhin,T., and K.A. Gray 2006. Factors affec ting denitrification ra tes in experimental wetlands: Field and labora tory studies. Ecological Engineering 26:167-181. Smialek, J., V. Bouchard, B. Lippmann, M. Quig ley, T. Granata, J. Martin, and L. Brown. 2006. Effect of a woody (Salix nigra ) and an herbaceous ( Juncus effuses ) macrophyte species on methane dynamics and denitrification. Wetlands 26:509. Sorrell, B.K. 1999. Effect of external oxygen demand on radial oxygen loss by Juncus roots in titanium citrate solutions. Plant, Cell & Environment 22: 1587. Southeastern Regional Climate Assessment report. 1999. http://weather.msfc.nasa.gov/ regional/water_quality.htm l Spieles, D.J., and W.J. Mitsch. 2000. The effect s of season and hydraulic and chemical loading on nitrate retention in cons tructed wetlands: a comparison of lowand high-nutrient riverine systems. Ecol ogical Engineering. 14: 77. Stottmeister,U., A. Wiesner, P. Kuschk, U. Kappel meyer, M. Kstner, O. Bederski, R.A. Mller and H. Moormann. 2003. Effects of plants a nd microorganisms in constructed wetlands for wastewater treatment. Biotechnology advances 22: 93. Tanner, C.C., and R.H. Kadlec. 2003. Oxygen fl ux implications of observed nitrogen removal rates in subsurface flow wetlands. Water science and Technology 48:191-198. Taylor M.D., S.A. White, S.L. Chandler. 2006. Nutr ient mitigation for large scale nurseries. p 540-543. In M. Taylor. (ed.) Water Management. Proceedings of Research Conference, GA, 2006, Southern Nursery Association, Atlanta, GA. Thomas, M.V., and D. Joyner. 2002. Total Ma ximum Daily Load (TMDL) Environmental Regulations. Proceedings Conference Fort Worth, Texas, USA. 11-13 March, 2002. Thompson, S.P., M.F. Piehler, and W.H. Paerl. 2000. Denitrification in an estuarine headwater creek within an agricultura l watershed. Journal of Environmental Quality 29:1914-1923. Tiedje, J.M. 1982. Denitrification. P. 1011-1026 In A.L.Page, R.H.Miller, and D.R. Keeney (ed.) Methods of soil analysis. Part 2. (2nd ed). Agronomy Monograph 9. ASA and SSSA. Madison, WI. Tiedje, J.M. 1988. Ecology of den itrification and dissimilatory n itrate reduction to ammonium, p.179-24 In AJB Zehnder (ed.), Biology of anaerobic microorganisms. John and Wiley & Sons, Inc., New York.

PAGE 103

103 Toet, S., L.H.F.A. Huibers, R.S.P. Van Logtes tijn, and J.T.A.Verhoeven. 2003. Denitrification in the periphyton associated with plant shoots and in the sediment of a wetland system supplied with sewage treatment pl ant effluent. Hydr obiologia 501:29-44. Toet, S., R.S.P. Van Logtestijn, R. Kampf, M. Schreijer and J.T.A. Verhoeven. 2005. The effect of hydraulic retention time on the removal of pollutants from sewage treatment plan effluent in a surface-flow wetland system. Wetlands 25:375391. USEPA. 1983. Methods of chemical analysis of water and wastes. Environmental Monitoring Support Lab., Cincinnati, OH. USEPA. 1986. Quality criteria for water 1986 and update no. 1 and update no. 2 (May 1, 1987): Washington, D.C., EPA-440/5-86-001 USEPA.1993a. NPS. Guidance Specifying Manage ment Measures for Sources of Nonpoint Pollution in Coastal Waters EPA 840-B-92-002. USEPA, 1993b. Constructed Wetlands for Wastewat er Treatment and Wildlife Habitat 17 Case Studies EPA832-R-93-005 USEPA. 1994. EPA's Polluted brochure EPA-841-F-94-005, http://www.epa.gov/owow/nps/qa.html USEPA. 2002a. The Successful Im plementation of the Clean Water Act' s Section 319 Nonpoint Source Pollution Program http://www.epa.gov/nps/S ection319III /intro.htm USEPA.2002b. National Coastal Condition Report. http://www.epa.gov/owow/oceans/nccr/index.html USGS. 1998. Water Quality in the Georgia-Florida Coastal Plain, Georgia and Florida, 1992-96 By Marian P. Berndt, Hilda H. Hatzell, Chri sty A. Crandall, Michael Turtora, John R. Pittman, and Edward T. Oaksford USGS Circular 1151 USGS. 2004. http://pubs.usgs.gov/circ/ 2004/1265/pdf/circular1265.pdf. White, J.R., and K.R. Re ddy. 1999. Influence of nitrate and phosphorus loading on denitrifying enzyme activity in Everglades wetland soils Soil Science Society of America journal 63:1945-1954. Wood, A. 1995. Constructed wetlands in wate r pollution control: fundamentals to their understanding. Water Science and Technology 32:21-29. Wood, S.L., E.F. Wheeler, R.D. Berghage, a nd R.E. Graves 1999. Temperature effects on wastewater nitrate removal in laboratory scale constructed wetlands. American Society of Agricultural Engineers 42:185-190.

PAGE 104

104 Xue, Y., D.A. Kovacic, M.B. David, L.E.Gentry, R.L.Mulvaney, and C.W. Lindau. 1999. In situ measurements of denitrification in constructed wetlands. Journal of Environmental Quality 28:263-269. Yang, Q., Z.H. Chen, J.G. Zhao, and B.H. Gu. 2007. Contaminant Removal of Domestic Wastewater by Constructed Wetlands: Effects of Plant Species. Journal of Integrative Plant Biology 49:437 446. Yeager, T.H., R.D. Wright, D. Fare, C. Gilliam, J. Johnson, T. Bilderback, and R. H. Zondag. 1993. Six state survey of container nurser y nitrate nitrogen runoff. Journal of Environmental Horticulture 11:206.

PAGE 105

105 BIOGRAPHICAL SKETCH Bhuvaneswari Govindarajan, shortly called as Bhuvana, was born in 1976 in India. She spent the greater porti on of her life growing up in a sma ll town, named Vandavasi, Tamilnadu, India. Hailing from a farming family, she was fasc inated by field of agriculture. She majored in horticulture sciences and rece ived her Bachelors degree in Horticulture (1994-1998) from Tamilnadu Agricultural University, Coimbatore, Ta milnadu, India. She completed her Master of Science in Vegetable Crops (1999-2001), with a minor in plant breeding, from Punjab Agricultural University, L udhiana, Punjab, India. For the next four years, she worked in several organizations, as re search associate in a private biofertilizer produc tion unit, as a lecturer in agricultural college, as a technical consultant for an MNC, on National Horticulture Board pr oject, and as a Development Executive for an NGO in a watershed development project. While working in an NGO, she was exposed to problems associated with environmental degrada tion/pollution and inclined towards the field of environmental science. She then decided to cont inue her education in th e field of environmental sciences. Upon achieving the graduate research assistantship, she came to United States, to pursue her Masters degree in Inte rdisciplinary Ecology at Univers ity of Florida. Bhuvana has had an opportunity to work on a project as we ll as her thesis resear ch focusing on nutrient removal from nonpoint source polluti on using constructed wetlands. After gaining hands on experien ce in the US, She wish to re turn to her county, India, where her knowledge and skills will be applied towards achieving cleaner, greener, and sustainable environment.