<%BANNER%>

Freshwater Clams as a Tertiary Treatment for Phosphorus in Agricultural Wastewater

Permanent Link: http://ufdc.ufl.edu/UFE0022679/00001

Material Information

Title: Freshwater Clams as a Tertiary Treatment for Phosphorus in Agricultural Wastewater
Physical Description: 1 online resource (163 p.)
Language: english
Creator: Riley, Lance
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: agriculture, aquaculture, biofiltration, clam, dairy, nutrients, phosphorus, raceway, remediation, wastewater
Fisheries and Aquatic Sciences -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The objective of this study was to determine the potential of using a recirculating raceway system to remove phosphorus-containing material from agricultural wastewater streams. The focus of the research was on the biological and physical characteristics of Corbicula populations and monitoring of various water quality parameters within the system, with special emphasis on phosphorus dynamics. A prototype raceway system was designed and constructed at the University of Florida Dairy Research Unit at Hague, Florida to test the adaptability and phosphorus removal capacity of the clams in wastewater treatment. The ability of freshwater clams to capture, sequester and retain phosphorus-containing material from varying amounts of fertilizer additions was demonstrated in this study. Clam biomass contained an average phosphorus concentration of 0.299 mg P/g of whole clam DW (SE = 0.005), similar to other bivalves. Tagged clams recaptured alive over the course of the study showed growth rates of up to 0.117 mm/day in shell length (0.0024 g clam DW/day), yielding phosphorus removal rates up to 0.0079 mg P/individual/day. Overall, raceway clam populations were subject to high mortality and were unable to demonstrate significantly long-term removal of total phosphorus, dissolved phosphorus or chlorophyll a from overlying source water. High temperatures and possible impacts from amphipod infestations may have affected clam populations. Even though some clams in this study did survive and grow, use of Corbicula culture for phosphorus treatment in Florida agriculture operations may require creative solutions to temperature and parasite problems. Despite these issues, the raceway-based recirculation system design demonstrated in this study provided a dependable, easy to construct and reusable platform for testing aquaculture potential of a variety of organisms in wastewater treatment conditions at large scale. The ultimate goal of this study was to provide an effective biological remediation mechanism for removal of phosphorus from dairy waste streams; however, toxicity of dairy effluent, even at high dilutions, may prohibit application of clam-based aquaculture systems without additional treatment mechanisms.
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.
Statement of Responsibility: by Lance Riley.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Phlips, Edward J.
Local: Co-adviser: Wilkie, Ann C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022679:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022679/00001

Material Information

Title: Freshwater Clams as a Tertiary Treatment for Phosphorus in Agricultural Wastewater
Physical Description: 1 online resource (163 p.)
Language: english
Creator: Riley, Lance
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: agriculture, aquaculture, biofiltration, clam, dairy, nutrients, phosphorus, raceway, remediation, wastewater
Fisheries and Aquatic Sciences -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The objective of this study was to determine the potential of using a recirculating raceway system to remove phosphorus-containing material from agricultural wastewater streams. The focus of the research was on the biological and physical characteristics of Corbicula populations and monitoring of various water quality parameters within the system, with special emphasis on phosphorus dynamics. A prototype raceway system was designed and constructed at the University of Florida Dairy Research Unit at Hague, Florida to test the adaptability and phosphorus removal capacity of the clams in wastewater treatment. The ability of freshwater clams to capture, sequester and retain phosphorus-containing material from varying amounts of fertilizer additions was demonstrated in this study. Clam biomass contained an average phosphorus concentration of 0.299 mg P/g of whole clam DW (SE = 0.005), similar to other bivalves. Tagged clams recaptured alive over the course of the study showed growth rates of up to 0.117 mm/day in shell length (0.0024 g clam DW/day), yielding phosphorus removal rates up to 0.0079 mg P/individual/day. Overall, raceway clam populations were subject to high mortality and were unable to demonstrate significantly long-term removal of total phosphorus, dissolved phosphorus or chlorophyll a from overlying source water. High temperatures and possible impacts from amphipod infestations may have affected clam populations. Even though some clams in this study did survive and grow, use of Corbicula culture for phosphorus treatment in Florida agriculture operations may require creative solutions to temperature and parasite problems. Despite these issues, the raceway-based recirculation system design demonstrated in this study provided a dependable, easy to construct and reusable platform for testing aquaculture potential of a variety of organisms in wastewater treatment conditions at large scale. The ultimate goal of this study was to provide an effective biological remediation mechanism for removal of phosphorus from dairy waste streams; however, toxicity of dairy effluent, even at high dilutions, may prohibit application of clam-based aquaculture systems without additional treatment mechanisms.
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.
Statement of Responsibility: by Lance Riley.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Phlips, Edward J.
Local: Co-adviser: Wilkie, Ann C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022679:00001


This item has the following downloads:


Full Text





FRESHWATER CLAMS AS A TREATMENT MECHANISM FOR PHOSPHORUS IN
AGRICULTURAL WASTEWATER
















By

LANCE W RILEY


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2008


































2008 Lance W Riley



























To my parents, Captain Roy "Luke" Riley (United States Navy, Retired) and Linda C. Riley,

Thank you for all of your love, support and confidence,
I couldn't have done this without you both, I love you









ACKNOWLEDGMENTS

I would like to thank my committee chair, Dr. Edward Phlips, the co-chair, Dr. Ann Wilkie,

and the rest of my committee (Dr. Tom Crisman, Dr. Roger Nordstedt, Dr. Shirley Baker and Dr.

Patrick Baker) for their help and guidance. Special thanks goes out to Ivan Mish, Jon Mish, the

students and administration at Alee Academy (Umatilla, FL), all of the other volunteers for

helping with clam stocking, raceway monitoring and amphipod interaction investigation and Bill

Lindberg. Special thanks also to Dr. Phil Barkley, Dr. Kelly Foote and all of the staff at the UF

Student Health Center and Shands Neurosurgery Department, look at the bionhick man go!

Most of all, I want to give a very special thank you to my parents, all of my family and friends

that made this possible. Funding for this research was provided by the following entities:

* United States Department of Agriculture (USDA-CSREES Special Research Grant-
Freshwater clams as tertiary treatment for agriculture wastewater. E. Phlips, S. Baker, P.
Lazur. 2001-2004. $80,000)

* United States Department of Agriculture (USDA-CSREES Special Research Grant-
Integrating clams into a dairy wastewater treatment train. E. Phlips and P. Baker. 2003-
2004. $79,824)

* University of Florida Department of Fisheries and Aquatic Sciences (Project Facility
Construction Grant. R. Riley 2002. $30,000)









TABLE OF CONTENTS


A C K N O W L E D G M E N T S ..............................................................................................................4

LIST O F TA B LE S ............................................................................. .......................... 8

LIST OF FIGURES ........................................................ ........................... 10

ABSTRACT 12

CHAPTER

1 IN T R O D U C T IO N ............................................................................................ ... ... .. 14

2 RACEWAY-BASED RECIRCULATING WASTEWATER TREATMENT SYSTEM
DESIGN AND CON STRU CTION ..................................... ........................ ................ 20

Introduction ...................................................... .................. 20
R a c ew ay D e sig n .....................................................................................................................2 5
Blountstown Facility.............. ... ................... ........... ............................... 29
H agu e F facility ....................................................................................................... ....... .. 32
System D design Sum m ary .........................................................................................35

3 ADAPTABILITY OF Corbicula TO TREATMENT RACEWAYS.................39

In tro d u c tio n ............................................................................................................................. 3 9
M e th o d s .................. ............................ .................................................... .................... 4 1
R acew ay-based Treatm ent System ............................................................. ................ 42
S o u rc e p o n d s ............................................................................................................ 4 2
R a c ew ay s ............................................................................................................... .. 4 4
W after an aly sis ................................................... .. ................ ............... 4 5
Clam Population Dynamics In The Raceway Environment.......................................46
Stocking clam racew ays ................................................................... ................ 47
Clam racew ay population sam pling .................................................... ................ 49
T ag g e d c lam s ............................................................................................................ 5 1
C lam su rv iv al ........................................................................................................... 5 3
B iom ass ch an g es ...................................................................................................... 53
R production and recruitm ent............................................................. ................ 57
H health .................................................................................................... ....... .. 5 7
R esults.................................................................................. ...... ......................... 60
Raceway System Environm ental Param eters............................................. ................ 60
Clam Population Dynamics In Treatment Raceways.................................................66
S u rv iv a l ................................................................................................................. ... 6 6
G row th ................ ....................... ............................................................ ........ .. 6 8
R production and recruitm ent............................................................. ................ 77
H health ............................................................................................ 78
Am phipod infestation .... ................................................................................ 79









D iscu ssio n ................................................................................. ............. .................. 8 1
A m m onia C concerns ............... .. .................. .................... ...................... ..................82
Temperature ............................................ .................................. 84
F o o d A v a ila b ility .............................................................................................................8 6
D dissolved O oxygen .................................................................................................... 88
M multiple Stressors ............................................................................................. . 88
P arasites an d P red atio n ....................................................................................................8 9
Reproductive Success ......................................... ............... 93
C lam Stock A ssessm ent Issues........................................ ........................ ................ 94
G general C onclu sion ............... .. .................. .................... ............. ....... ........... ......96

4 PHOSPHORUS REMOVAL AND SEQUESTRATION IN CLAM RACEWAYS.............98

In tro d u c tio n ............................................................................................................................. 9 8
M e th o d s .................. ............................. ................................................... ..................... 1 0 2
Raceway-based Treatment System ...... ............. ............ ..................... 102
Source w ater ponds .............................. ........................................... 103
R acew ays .............................................................................. ............................... 105
W ater quality m monitoring ................. .......................................................... 105
R acew ay clam populations....................................................... ... ............... 107
P Removal From Source Water By Clam Raceways ........................ ...................108
Racew ay through-flow trials ......................................................... 108
R acew ay w ater recirculation ........................................... .... ........... ............... 112
Sequestration of phosphorus by clams in treatment raceways............................ 114
Results ....................... ............... ..... ................ .................. 116
Raceway Environmental Conditions ...... ........... .......... ..................... 116
R acew ay clam populations.................................... ...................... ...............1...... 18
Phosphorus Uptake In Clam Raceways...... .... .... ..................... 118
Raceway through-flow input/output measurements.................... .................. 118
Racew ay recirculation m easurem ents ............................................... ............... 120
Sequestration Of Phosphorus By Clams In Treatment Raceways...............................124
Phosphorus allocation in clam biomass ............... .......................125
Treatment raceway clam population phosphorus....................... .................. 127
D discussion ................... ...... ...... .... ... ... .......... ..................... 128
Distribution of Phosphorus Taken Up By Clams ..................................... ............... 128
Estim ates of Phosphorus U ptake Rates ................... ....................... .... .............. 131
Comparison Of Phosphorus Removal By Clam Raceways And Other Systems ........133
Problems With Measuring Short-term Phosphorus Uptake ............... ...................136
Dairy Application Demands And Issues ....... ... .......................... 139
Sustainability .............. ....................................................................... . ......143

5 SUMM ARY .................................................... .................. 147

R acew ay Function and A ttributes...................................... ........................ ................ 147
Adaptability of Clams to Raceway Conditions .......... .........................147
P-removal Capacity ............... ................... .. .......... ............................. 148
Future Applications ................................ .. .......... .............. ...............150


6









L IST O F R E F E R E N C E S ....................................................... ................................................ 153

B IO G R A PH IC A L SK E T C H .................................................... ............................................. 163









LIST OF TABLES


Table page

2-1 Raceway dimensions and capacities as tested, available capacities adjusted for
stan dp ip e p resen ce ............................................................................................................. 2 7

3-1 Source pond and raceway numerical designations for the treatment systems at the
H ag u e site ......................................................................................................... ........ .. 4 2

3-2 Raceway (RW) stocking and population sampling schedule for the low, medium and
high nutrient addition treatm ent system s ...................................................... ................ 46

3-3 Number of clams stocked in each raceway estimated using the volumetric method .........66

3-4 Number of live clams found alive at each sampling interval estimated using the
sp atial tech n iq u e ................................................................................................................ 6 7

3-5 Actual number of live clams stocked in each raceway and at the end of the study ...........68

3-6 Mean shell lengths measured from clams in each raceway (RW) both at stocking and
at each sampling interval .......................... .......... ........................ 70

3-7 Shell growth rates for tagged clams in each nutrient addition treatment for each
season al tim e interv al ......................................................................................................... 72

3-8 Shell size information on clams sampled for tissue biomass analysis from each
nutrient addition treatm ent racew ay system ................................................. ................ 72

3-9 Mean and range of ash content values for meat, shell and total clam tissues pooled
for all clam s sam pled ........................................................................................... 73

3-10 Results of the meat, shell and total clam tissue dry weight (DW) to shell length
correlation an aly sis ............................................................................................................ 7 3

3-11 Dry weight (DW) biomass vs length regression relationships, significance
differences and variability for whole clam, shell and meat tissues from each nutrient
ad edition treatm ent .............................................................................................................. 74

3-12 Mean, standard error (SE) and range of condition indices values (CI(WT) and CI(VOL))
calculated for the medium nutrient addition treatment at Interval 1 compared to
values calculated at all other treatment/interval combination.......................................78

3-13 Individual shell length and biomass dry weight (DW) growth rates reported for
Corbicula and other bivalves occupying different fresh and saline environments............ 81

4-1 Source pond and raceway numerical designations for the treatment systems at the
H agu e site..................................................................................................... . ......... 10 3









4-2 Raceway source water input flow rates for the period of July 1 to August 24, 2002...... 109

4-4 Monthly mean input total phosphorus (TP) concentrations in raceways and standard
error (SE), at time 0 in the recirculation trials for each nutrient addition treatment
p on d group p s.................................................................................................... ........ .. 12 0

4-5 Raceway system monthly mean input total dissolved phosphorus (TDP) at time zero
in the recirculation trials for each nutrient addition treatment...................................121

4-6 Total dissolved phosphorus (TDP) removal rates calculated from TDP slopes in the
recirculation trials for each nutrient addition treatment system during April and May
2 003 ............................................................................................. ......... 122

4-7 Raceway total dissolved phosphorus (TDP) values at time 0 for covered raceways in
the low and high nutrient addition treatments............... .........................122

4-8 Raceway chlorophyll a (chl a) values at time 0 for raceways in the low and high
nutrient addition treatments during April and May 2003 ..................... ...................123

4-9 Raceway chlorophyll a (chl a) values at time 0 for covered raceways in the low and
high nutrient addition treatm ents .................. ........................................................ 124

4-10 Mean shell length, clam wet weight (WW), meat and shell tissue dry weights (DW)
and condition index (CI) values for the sample population of clams used to determine
clam biom ass phosphorus content ......................................................... 125

4-11 Mean and range of ash content values for meat, shell and total clam tissues pooled
for all clam s sam pled ....................................................................................................... 12 5

4-12 Summary statistics for phosphorus concentrations [P] found in meat, shell and clam
tissue types pooled for all clams sampled...... ........ ..................... 126

4-13 Amounts of phosphorus (P) contained in meat, shell and clam tissues along with
percentages of total clam phosphorus allocated to meat and shell tissues for
in d iv id u al clam s .............................................................................................................. 12 6

4-14 Comparison of meat and shell phosphorus concentrations [P] in dry weight (DW)
biomass of Corbicula versus other fresh and saltwater clams...................................130

4-15 Estimated annual phosphorus removal in various biological treatment systems
applied to different effluent types using systems of varying design and scale............. 134









LIST OF FIGURES
Figure page

2-2 Blountstown system plumbing diagram .................. .................................................. 31

2-3 H ague system plum bing diagram ........................................ ....................... ................ 34

3-1 Linear regression relationship of glass sphere volume to glass sphere weight used to
estimate clam shell cavity volume for the volume-based condition index calculation......58

3-2 Air temperature readings at the Dairy Research Unit in Hague, FL over the study
p e rio d ....................................................................................................... ........ . ....... 6 1

3-3 Input water temperatures in the low, medium and high nutrient addition treatments .......61

3-4 Raceway dissolved oxygen (DO) readings in the low, medium and high nutrient
ad edition treatm ents..................................................... ................................................ 62

3-5 Raceway pH in the low, medium and high nutrient addition treatment systems ............62

3-6 Total phosphorus (TP) in the low, medium and high nutrient addition treatments...........64

3-7 Total dissolved phosphorus (TDP) in the low, medium and high nutrient addition
treatm ents ........................................................................................................ ....... .. 64

3-8 Total nitrogen (TN) in the low, medium and high nnutrient addition treatment source
w after ........................................................................................................ 6 5

3-9 Chlorophyll a (chl a) in the low, medium and high nutrient addition treatment source
p on d s .............................................................................................. ........ 6 5

3-10 Number of live clams in each nutrient addition treatment............................................69

3-11 Cumulative number of dead found on the substrate surface in the low, medium and
high nutrient addition treatm ents ........................................ ....................... ................ 69

3-12 Changes in shell lengths of tagged clams captured alive in each nutrient addition
treatm en t .......................................................................................................... ........ .. 7 1

3-12 Regression relationships for shell length vs actual and predicted (Table 3-11) whole
clam dry weight (DW) values for each nutrient addition treatment ................................75

3-13 Regression relationships for shell length vs actual and predicted (Table 3-11) shell
dry weight (DW) values for each nutrient addition treatment......................................75

3-14 Regression relationships for shell length vs actual and predicted (Table 3-11) meat
dry w eight (D W ) values for clam s....................................... ...................... ................ 76









3-15 Estimated clam dry weight (DW) biomass over time in the low medium and high
nutrient addition treatm ents ...............................................................................................77

4-1 Raceway input total phosphorus (TP) in the low nutrient addition treatment during
the through-flow trials for July and August of 2002...... ....................................... 119

4-2 Frequency distribution of changes in total phosphorus (TP) from the input to the
output in the low nutrient addition treatment raceways from July through August
2 002 ............................................................................................. ......... 12 0

4-3 Amount of phosphorus (P) sequestered in clam biomass for the low, medium and
high nutrient addition treatments over the study period ....................... ...................127









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

FRESHWATER CLAMS AS A TREATMENT MECHANISM FOR PHOSPHORUS IN
AGRICULTURAL WASTEWATER

By

Lance W Riley

August 2008

Chair: Edward J. Phlips
Co-chair: Ann Wilkie
Major: Fisheries and Aquatic Sciences

The objective of this study was to determine the potential of using a recirculating raceway

system to remove phosphorus-containing material from agricultural wastewater streams. The

focus of the research was on the biological and physical characteristics of Corbicula populations

and monitoring of various water quality parameters within the system, with special emphasis on

phosphorus dynamics. A prototype raceway system was designed and constructed at the

University of Florida Dairy Research Unit at Hague, Florida to test the adaptability and

phosphorus removal capacity of the clams in wastewater treatment.

The ability of freshwater clams to capture, sequester and retain phosphorus-containing

material from varying amounts of fertilizer additions was demonstrated in this study. Clam

biomass contained an average phosphorus concentration of 0.299 mg P/g of whole clam DW (SE

= 0.005), similar to other bivalves. Tagged clams recaptured alive over the course of the study

showed growth rates of up to 0.117 mm/day in shell length (0.0024 g clam DW/day), yielding

phosphorus removal rates up to 0.0079 mg P/individual/day. Overall, raceway clam populations

were subject to high mortality and were unable to demonstrate significantly long-term removal

of total phosphorus, dissolved phosphorus or chlorophyll a from overlying source water. High









temperatures and possible impacts from amphipod infestations may have affected clam

populations.

Even though some clams in this study did survive and grow, use of Corbicula culture for

phosphorus treatment in Florida agriculture operations may require creative solutions to

temperature and parasite problems. Despite these issues, the raceway-based recirculation system

design demonstrated in this study provided a dependable, easy to construct and reusable platform

for testing aquaculture potential of a variety of organisms in wastewater treatment conditions at

large scale. The ultimate goal of this study was to provide an effective biological remediation

mechanism for removal of phosphorus from dairy waste streams; however, toxicity of dairy

effluent, even at high dilutions, may prohibit application of clam-based aquaculture systems

without additional treatment mechanisms.









CHAPTER 1
INTRODUCTION

Phosphorus (P) produced in waste management of agricultural operations can result in

degradation of water quality in surface and groundwater flows. An addition of nutrients to

aquatic systems by human activities has resulted in the growth of nuisance macrophytes, alga

bacteria and periphyton in natural and man-made systems (Sharpley 1994, Johnson et al. 2004,

Dao et al. 2006, Wetzel 2001, Phlips et al. 2002, Timmons et al. 2002). This unwanted growth

can also cause contamination of drinking water supplies, degradation of aquatic habitat for

desirable species, fouling of engineered water systems, limitation of navigation and other

negative impacts to commercial and recreational activities (Sharpley 1994, Johnson et al.. 2004,

Dao et al. 2006). In warmer climates such as Florida, aquatic plant growth is magnified due to

the nearly year-round growing season (Scinto and Reddy 2003, DeBusk et al. 2004). The focus

of the current study was application of a clam-based biofiltration approach to remove of

particulate phosphorus from agriculture wastewater streams, with special emphasis on dairy

systems.

Excess nutrients can enter surface waters from both point and non-point sources associated

with concentrated farming activities (Sharpley et al. 1994, Knight et al. 2000, Johnson et al.

2004, Dao et al. 2006). Manure and wastes produced in dairy operations are much different than

other land-based agriculture operations since much of the waste products generated have high

water content due to milk parlor operations and bedding waste handling (NRCS 1999, Wilkie

2003). High liquid fractions in dairy wastes are encouraged by the removal of solids through

settling and through microbial decomposition by anaerobic digestion (Wilkie 2003).

Management practices that target nutrients such as phosphorus from animal manure include

oxidation ponds, facultative lagoons and storage ponds in conjunction with land application,









constructed wetlands and composting (NRCS 2006). In order to manage Florida dairy

wastewaters, liquids are stored in short-term retention ponds prior to land application that

supplies water and fertilizer to forage crops (Wilkie 2003).

There are severe limitations to the technologies currently available to reduce, effectively

and economically, nutrient levels in agricultural wastewater streams. Some nutrients are bound

in the terrestrial environment, but the remainder end up in surface water bodies (Sharpley et al.

1994) and can also enter groundwater (Johnson et al. 2004). Phosphorus is of special concern

because it is the primary limiting factor for growth of algae and other plants in most freshwater

systems (Wetzel 2001), and it cannot be removed by volatilization, as in the case of other

important nutrients, such as nitrogen and carbon (Wetzel 2001, Timmons 2002). Dairy wastes

subjected to anaerobic digestion can pose some special problems for aquatic environments

because of their high liquid content (Wilkie et al. 2004) and high dissolved phosphorus content,

making them readily available for uptake by aquatic macrophytes and algae (Sooknah and

Wilkie 2004).

In response to the special demands of nutrient removal from wastewater streams, recent

research has focused on integration of aquaculture systems. Use of aquaculture for wastewater

treatment is designed to convert phosphorus into a solid form that can be harvested as a

potentially useful commodity. In order to be successful, these culture systems must be capable

of promoting growth and reproduction of the organisms that are the end product of the process.

However, organisms capable of desirable water treatment functions, such as filtration of

particulate phosphorus-containing matter, in the natural environment may not conform to

aquaculture conditions. Therefore, the experimental system design should provide as many









natural habitat conditions for the target organisms as possible, including hydrology, food

resources, substrate and temperature.

Unlike production aquaculture, organisms in wastewater treatment systems are selected for

their ability to manipulate water quality parameters within source water and not necessarily for

their specific value as a consumer commodity. Development of experimental and eventually

commercial wastewater treatment systems is therefore driven more by public opinion on

environmental issues through government legislation than by consumer product demand and

profit normally associated with traditional aquaculture species. However, like traditional

aquaculture systems, wastewater designs must also be cost effective, not too land intensive and

have low water demands to be economically and environmentally feasible.

Various aquatic organisms have been evaluated for potential applications in the treatment

of agricultural wastewater, including dairy. Plant-based systems have received the most

attention; however, some animal based systems have also been proposed. Aquatic macrophyte

systems remove phosphorus from dairy effluents in small and large-scale systems (Reddy and

Smith 1987, Sooknah and Wilkie 2004, Lansing and Martin 2006, Wood et al. 2007).

Periphyton-based systems similar to the Algal Turf Scrubber (ATS) technology developed by

Adey and Hachney (1989) have been demonstrated to remove phosphorus from dairy

wastewaters at experimental scales (Pizarro et al. 2002) (Mulbry and Wilkie 2001). Algae

suspended in large outdoor tanks containing dairy wastewater have also demonstrated

phosphorus treatment potential (Sooknah and Wilkie 2004), as have phytoplankton in laboratory

flasks (Lincoln et al. 1993, Lincoln et al. 1996).

Freshwater, pond-scale studies of aquaculture-based wastewater treatment have focused

mostly on fish (Greer and Ziebell 1974, Dempster et al. 1995, Van Rijn 1996, Drapcho and









Brune 2000, Prein 2002, Azim et al. 2003, Ghaly 2005, Sindilariu 2007) and to a lesser extent on

filter-feeding bivalves (Busch 1974, Buttner and Heidinger 1980, Buttner 1986). Use of bivalves

in polyculture with other organisms has also been demonstrated as a treatment mechanism in

freshwater aquaculture operations in fish polyculture ponds (Buttner and Heidinger 1980,

Buttner 1986, Soto and Mena 1999). Use of bivalves for wastewater treatment is more

pronounced in the mariculture industry, where more elaborate systems have been used in

conjunction with phytoplankton and/or seaweed to remove nutrients generated from finfish and

shrimp culture (Shpigel 1993, Shpigel and Neori 1996, Lefebvre et al. 2000, Jones et al. 2001,

Mazzola and Sara 2001).

Freshwater clams were proposed and for use nutrient reduction by Greer and Ziebell (1972)

and Stanley (1974) who made calculations based on available literature; however, no large-scale

systems have been tested. Little is known about phosphorus sequestration by these organisms or

their potential for large-scale culture using dairy wastewater. Freshwater clams may provide an

ideal P-removal vector for dairy wastewater because of their ability to remove and sequester

phosphorus from the overlying water column (Fuji 1979). In a typical clam-based system,

dissolved phosphorus from wastewater is converted to a particulate form through a

phytoplankton intermediary and is coupled with other wastewater particulates to feed clam

populations (Greer and Ziebell 1972, Stanley 1974). Filter feeding allows the clam to pump

overlying water through its siphon and into the mantle cavity where particulates are then

removed by the gills and converted to biomass (McMahon and Bogan 2001). Sequestered

phosphorus in clam biomass and sediment depositions can be periodically removed at harvest

intervals to remove phosphorus permanently from the treatment system.









One of the clam species that has been the focus of past efforts in treatment systems is the

well known invasive, Corbicula. This organism is a recent invader to North America as

described in McMahon and Bogan (2001), and is a significant contributor to fouling in power

plants and industrial raw water systems Williams and McMahon (1986). The success of this

organism as a biofouling agent is due to its high reproductive fecundity stemming from self-

fertilization no complex life cycle needing water-born gametes and intermediate hosts, unlike

freshwater mussels that possess all of these traits (McMahon and Bogan 2001, McMahon 2002).

A statewide distribution of Corbicula has been noted in most Florida waterways (Blalock and

Herod 1999), so the clam can be considered as a naturalized species instead of a potential

invasive species. Its high reproductive potential makes the clam an ideal candidate for

propagation under aquaculture conditions, since it should repopulate rapidly following harvest of

only a few individuals.

Corbicula can be found throughout most of North American are all expected to be very

similar due to reproductive characteristics such as hermaphrodism and self-fertilization that

result in the production of exact copies of the parent lineage as examined in McMahon and

Bogan (2001). Exact taxonomic determination of clams in the genus Corbicula is subject to

intense debate and ontogenetic variation can lead to improper usage of different species

designations such asfluminea and japonica, that are commonly used in the literature to describe

most Corbicula clams found in freshwater environments. For this reason, clams in this study are

only referred to as Corbicula.

The goal of this study was to analyze the phosphorus removal potential of an engineered

raceway system containing populations of the freshwater clam Corbicula. The potential for

using freshwater clams as a mechanism for phosphorus removal was based on its ability to









remove and sequester phosphorus through active biofiltration. This study had the following

objectives:

* Design, construct and operate a large-scale raceway-based treatment system for examining
the performance of freshwater clams as a P removal mechanism

* Determine the adaptability of the freshwater clam, Corbicula, to wastewater treatment
conditions

* Determine the ability of clam raceways to remove and sequester phosphorus containing
material from agricultural waste streams


In this study, a raceway-based recirculating system was developed in order to study

phosphorus removal rates using clam populations under simulated wastewater conditions in

raceway systems. Design and construction of systems that can evaluate the performance of

organisms in wastewater aquaculture at a commercial level are critical for developing nutrient

management strategies for future applications. Systems must account for problems not only with

the aquaculture practices, but with the conditions unique to dairy wastewater effluent that may be

remedied by dilution (Sooknah and Wilkie 2004). The following questions were posed for

investigation in this study:

* What are the growth rates of clams and survival and recruitment rates of clam populations
exposed to different concentrations of nutrients?

* Does the physiological condition of raceway clam vary over time and does it correlate with
nutrient addition, environmental parameters or mortality events?

* Are clam raceway systems able to capture, sequester and retain P-containing material from
varying concentrations of dairy wastewater effluent?

* How do seasonality, clam population dynamics, temperature, algal density and P
availability affect the removal of P-containing material by clam raceway systems?

* How is P sequestered and allocated by clams into soft tissue and shell biomass?

* Is this technology suitable for use as a mechanism for P-removal by agricultural operations
in Florida?









CHAPTER 2
RACEWAY-BASED RECIRCULATING WASTEWATER TREATMENT SYSTEM DESIGN
AND CONSTRUCTION

Introduction

Most biologically-based wastewater treatment systems involve use of ponds, tanks or

raceways as steps in the removal of solids, nutrients and contaminants (Buttner 1986, Shpigel

1993, MacMillan et al. 1994, Shpigel et al. 1997, Jara-Jara et al. 1997, Jones and Preston 1999,

Jones et al. 2002, Sooknah and Wilkie 2004). The focus of this design effort was removal of

nutrients using filter-feeding bivalves as the active agent in the final stage of a process beginning

with conversion of soluble nutrients into particulate forms via production of plankton. The

design had to meet several key criteria in terms of both experimental and operational demands.

From an experimental standpoint, the system had to incorporate the ability to deal with multiple

treatment groups in a replicated manner. Operationally, the system had to be of sufficient size to

provide a reasonable measure of potential success in real life applications.

Many elements of traditional aquaculture system designs were incorporated into the design

process of the raceway-based treatment systems used in this study. Aquaculture system

hydrology flows either flow-through or recirculating water flow regimes, depending upon the

extent of water reuse and residence time (Van Rijn 1996). Recirculating systems offer the

distinct advantages of lower water consumption and confinement of wastes, reducing potential

harm to the natural aquatic environment (Timmons et al. 2002). These attributes make

recirculting systems ideal for study of wastewater treatment mechanisms because they do not

involve discharge into the environment.

Raceway-based aquaculture systems have been used to cultivate a variety of aquatic

organisms including fish, bivalves, algae and plants (Shpigel and Neori 1996, Adey et al. 1993).

The most common large-scale raceways are usually associated with the production of finfish,









such as salmonids (Timmons et al. 2002). These structures typically range in size from 3-5.5 m

in width, 24-46 m in length and 0.8-1.1 m deep (Timmons et al. 2002). This larger scale limits

construction materials most often to concrete, plastic, or earthen structures with plastic liners

(Sindilariu 2007, Van Rijn 1996). Reinforced fiberglass panels have also been used in finfish

culture to construct raceways using a modular design as an alternative to concrete (Vantaram

2004). Raceways used in the commercial rearing of Quahog clams (Merceneria merceneria)

employ long, sand-bottomed, flow-through plastic troughs to raise juveniles prior to placement in

estuarine farm sites for grow out (Lorio and Malone 1995). In addition to widespread

commercial applications, raceways have also been used in a variety of experimental aquaculture

systems targeting organisms associated with biofiltration or bioaccumulation, such as bivalves,

benthic microalgae and macrophytes (Shpigel 1993, Craggs et al. 1996).

The majority of experimental raceway designs using bivalves are involved in wastewater

remediation. In these systems, commercially valuable bivalve species are commonly cultured as

a secondary commodity on the effluent of primary culture organisms such as finfish and shellfish

(Buttner 1986, Shpigel 1993, MacMillan et al. 1994, Shpigel et al. 1997, Jara-Jara et al. 1997,

Jones and Preston 1999, Jones et al. 2002, Zhou et al. 2006). Examples of raceways for the

culture of bivalve species include: 14.4 L fiberglass tanks (Shpigel et al. 1997), 34 L plastic

tanks (Jones and Preston 1999), 340 L plastic tanks (Huchette et al. 2003), 1500L concrete tanks

(Jones et al. 2002), 1500 L fiberglass tanks (MacMillan et al. 1994), 2240 L fiberglass tanks

(Jara-Jara et al. 1997), 2 m3 V-bottom fiberglass tanks (Shpigel 1993), and 15,000 m3 concrete

tanks (Zhou et al. 2006).

A variety of raceway designs have also used algae or higher plants as the active treatment

agent. Possibly the most notable vegetative raceway system is the Algal Turf Scrubber, which









consists of an artificial stream used to culture periphytic algae. Periphyton is grown on plastic

mesh screens placed in shallow rectangular flumes. Source water is supplied using a pulse-flow

regime down the length of the raceway (Adey et al. 1993). This design has been adapted for use

at various size scales from small-scale laboratory systems (Mulbry and Wilkie 2001, Pizzarro et

al. 2002, Wilkie and Mulbry 2002, Kebebe-Westhead 2003), to 1021 m2 raceways consisting of

landfill liners between concrete sidewalls used for tertiary treatment of municipal wastewater

(Craggs et al. 1996).

Drawing on many elements of the aforementioned aquaculture technologies, a two stage

treatment system was designed for this study. The first stage involved growth of phytoplankton

in ponds supplemented with either nutrients from a dairy wastewater stream or inorganic

fertilizer. The ponds served as the source of water for a series of recirculating raceways. The

freshwater clam was used as the primary agent for nutrient removal from the source water,

through filtration of plankton and conversion into harvestable biomass.

The raceway design was chosen for this application since it mimics the small stream

environment widely occupied by Corbicula in North Florida (Blalock and Herod 1999). Like a

stream system, water is constantly supplied to the raceways, and the channel-like shape (width to

length ratio= 1 : 7.5 in this study) induces a plug-flow hydrology that has little back mixing.

This hydrology is maintained by recirculating water between the source ponds and raceways in

this system, while replenishing food particles and dissolved oxygen and removing waste

products. A coarse sand substrate was chosen for this application since it is an intermediate

aggregate size preferred by Corbicula in small stream environments (Blalock and Herod 1999,

Schmidlin and Baur 2007).









The raceway design can be constructed of a variety of materials and is scalable, adaptable

to a variety of hydrologic regimes and can be easily manipulated for sampling, cleaning and

maintenance. Raceways are also versatile, in that the structure can be used to culture a variety of

aquatic organisms, including other bivalves, fish, algae and plants using a variety of substrates.

The design used in this study is easy to disassemble and transport for reuse at different effluent

source locations, can be assembled in remote locations and can be used in short-term evaluations

of source waters without leaving a significant footprint. The modular components used in this

design can be prefabricated and assembled quickly on site without intensive labor requirements

needed to construct other large-scale systems.

Integrating Corbicula into wastewater streams via phytoplankton production has been

suggested by Stanley (1974); however no studies have been performed to evaluate the viability

of using large-scale engineered systems that emulate features that may be applicable in a full-

sized wastewater treatment system. The study of Corbicula biofiltration potential has been

limited to small- and medium-scale applications such as laboratory-based bench scale flow

chambers (Lauritsen 1985), aerated 37 L aquaria with sand substrate (Beaver et al. 1991, Brock

2000), 150 L aquaria (Greer and Ziebell 1972) and 515 L rectangular fiberglass tanks containing

mesh trays (Haines 1977). Cultivation in larger systems has focused on shallow ponds used to

produce monoculture Corbicula as a food crop in Taiwan (Phelps 1994). Use of Corbicula in a

pond-based polyculture scenario has been studied in catfish rearing ponds using benthic

sediments and suspended cages as substrates (Buttner and Heidinger 1980, Buttner 1986). The

clams have also been cultured in cages suspended within power plant discharge canals (Mattice

1977). The aquaculture system described in this study is meant to be an intermediate size scale









between the smaller experimental systems and the much larger pond and canal systems used in

previous Corbicula studies.

Investigations using experimental large-scale engineered systems are an important step in

developing commercial size treatment systems, especially when targeting an organism like

Corbicula that has not been traditionally cultured at such a scale. Predicting the adaptability of

these organisms to large-scale culture scenarios cannot be accomplished sufficiently from small-

scale experiments since the behavior of the organism in these systems may not coincide with

observations in larger systems. The systems designed for this study provide an opportunity for

replication of treatment groups without sacrificing the structural elements of real-world treatment

systems. These systems were designed using the following basic considerations necessary for

the implementation of any experimental, biological-based water treatment system:

* Must be capable of maintaining environmental conditions necessary to promote survival,
growth and reproduction of the target organism such as hydrology, substrate, food
resources, waste removal and aeration

* System design intended for scientific manipulation must be able to conform to desired
experimental treatments and replication for statistical analysis

* System must be scalable to provide an adequate surface area for the desired outcome in a
real-world treatment application

* System design and operation must be applicable to various land topographies and source
water body layouts found at different site locations

* Design must maximize energy efficiency by using gravity flow to reduce pumping
requirements

To test the efficacy of the basic pond-raceway design, two recirculating systems were

designed and constructed at different locations in northern Florida. The first system was

completed in October 2002 at the Sam Mitchell Aquaculture Demonstration Facility in

Blountstown, Florida and consisted of nine raceways supplied by two source ponds. This system

was operated from November 2002 until January 2003 when the entire facility in Blountstown









was closed permanently due to university budget cuts. Parts of this system were excavated,

dismantled and transported to the Dairy Research Unit in Hague, FL where a second system was

constructed and operated from June 2003 to October 2004. The Hague facility consisted of 3

separate systems, each with 2 ponds and 3 raceways. Multiple ponds were used in both locations

to provide an alternative to sustain phytoplankton populations in case of an unfavorable pond

condition, while multiple raceways were used for statistical rigors. Differences in the

topography and source pond design for each location required the use of different water delivery

system configurations; however, the individual raceway design remained the same for both

locations.

Multiple raceways were assembled at the Blountstown and Hague facilities for this study

using different source pond layouts and water delivery configurations. The water delivery

configurations used for these two systems were: 1) Source water was gravity-fed through the

raceways and pumped back to the source pond and 2) Source water was pumped to the raceways

and gravity-fed back to the source pond. The gravity-fed source water option was applied to the

Blountstown system, while the Hague system incorporated the pump-fed source water

configuration due to the elevations of the available areas for raceway construction in relation to

the source ponds.

Raceway Design

The central components of the raceway system were closed-ended rectangular tanks.

Raceways were formed from a 1.0 m wide by 7.4 m long channel, assembled from 10 individual

panels constructed with pressure-treated wood framing and plywood backing. These frame

sections were joined together using galvanized lag screws and a framing board (3.8 cm (1.5")

thick x 8.9 cm (3.5") wide) was placed across the width of the raceway at the bottom of each










framing section joint to help maintain the rectangular shape. The basic layout and components

of the raceways used in this study are illustrated in Figure 2-1.


Water Water Input
Distribution

meMetaer







Boards
PVC LinerLines
_'- Joint Bolts



Diverter Rubber
Figure 2-1. Components and design of individual Connector







raceway and attached using treated wood furring strips fastened to the outputside of the framing.
O" Output ..tandpipe
Standpipe System A
Feldspar, Incorporated in Edgar, Florida, was used as a substrate. Some slack was left in the
liners around the raceway sides in order to Raceway (x.5)ansion and contraction with changes in

Washer Liner and Valve
Output Lines


Figure 2-1. Components and design of individual raceways.


A custom made, black 20-mil thick ABS-PVC liner was then placfilled inside the assembled

raceway and attached using treated wood furring strips fastened to the outside of the framing.

The substrate was added before the furring strips were attached to minimize stretching as a result

of sand settling. Coarse grade SiO2 filtration sand (0.6-1.0 mm particle size), available from

Feldspar, Incorporated in Edgar, Florida, was used as a substrate. Some slack was left in the

liners around the raceway sides in order to allow for expansion and contraction with changes in

temperature. A 0.31 m thick layer of fill dirt was placed around each raceway frame before the

substrate was added to help support the raceway frames. Raceways were filled with sand to a









depth of 0.20 m so the substrate surface would be level with the raceway drain valves located on

each output standpipe elbow (Figure 2-1). Fill dirt placed around each of the raceways was used

in conjunction with the interior sand substrate in order to stabilize the sides framing and prevent

buckling. The finished raceway dimensions and capacities are listed below in Table 2-1.

Table 2-1. Raceway dimensions and capacities as tested, available capacities adjusted for
standpipe presence.
Specification Parameter Value
Inner Dimensions (empty) Width 1.0 m
Length 7.4 m
Height 0.6 m
Available Volume 4200 L (4.2 m3)
Substrate Capacity Depth 0.2 m
Available Substrate Surface Area 7.2 m2
Volume 1432 L (1.4 m3)
Water Capacity Depth 0.2 m
Available Volume 1432 L (1.4 m3)

The supply plumbing to each raceway consisted of a 10.2 cm (4") inside diameter (ID)

poly-vinyl-chloride (PVC) line stemming from the 15.2 cm (6") ID main supply line from each

source pond. Supply lines had a brass gate valve that allowed selection of the desired source

pond. After the valve, supply lines are reduced to 5.1 cm (2") ID, and the two were joined into a

single raceway input line equipped with a brass gate valve to regulate input flow and a Pitot-tube

type flow meter to help balance input water flow between the raceways in the system. Source

water was fed into the raceway through a slotted water distribution bar that dissipated the energy

of the falling water over the width of the raceway, initiating a laminar-type plug-flow

hydrological pattern. Threaded caps were used at the ends of the distribution bars to allow easy

cleaning of slotted portions to prevent blockage from biofouling.

Water exited the raceways via a standpipe system that acted as a type of weir structure to

govern water column height and channel output water to the appropriate source pond. The









standpipes were made from 15.2 cm (6") ID PVC 90 elbows with threaded male adapters on the

output ends. The threaded ends passed through holes cut in the back plywood panels of the

raceways and the liners. The threaded female adapters on the outside of the panel were tightened

to hold the standpipe system in place, and a bulkhead fitting was formed over the liner where

output lines passed through by using an aluminum washer on the inside of the raceway to prevent

leakage around the pipes.

For this study, raceway water depth was maintained at 0.20 m; however, depth could be

adjusted by extending the standpipe height using additional segments of pipe. Raceway output

water was routed to the desired source pond by placing a 20 cm (8") long removable diverter

pipe with a rubber "no-hub" connector over one of the output standpipes to divert water to the

appropriate source pond return plumbing. Small, 1.9 cm (3/4") ID, drain valves were added to

each of the 900 standpipe elbows level with the substrate surface inside the raceway to allow

removal of overlying water for periodic substrate sampling or observation, and they could be

used to drain the raceway for batch-fed experiments.

The raceways had a maximum input flow rate of 303 Liters per minute (LPM) (80 gallons

per minute (GPM)) due to the gravity flow capacity of the 15.2 cm (6") diameter output

standpipe. This flow rate limited theoretical retention time to no less than 6.3 minutes at the 0.20

m raceway water depth used in this study. Raceway flow rate was maintained at 227 LPM (60

GPM) in this study yielding a retention time of 9.5 minutes.

The one-meter width was chosen since it is approximately the limiting distance for

accessing the entire bottom area by hand from the sides of the raceways. Raceway length was

estimated from target stocking amounts of 7,000 to 10,000 adult clams per raceway at population

densities similar to the high densities ( > 1,000 clams/m2) sometimes found in naturally









occurring Corbicula populations (McMahon and Bogan 2001). Ultimately, the exact length and

height dimensions of the raceways were determined according to the dimensions of the standard

size for the plywood used in the raceway side framing, to expedite construction. In this case, the

7.4 m length is a result of using 3 lengths of a standard sheet of plywood, while the 0.6 m

raceway side height equals half of the width of a standard sheet of plywood.

The low width to length ratio of the raceway design makes the bottom area more physically

accessible than circular tanks of the same volume or surface area. When widths of Im or less are

used, bottom area can be manipulated easily by hand provided both sides of the raceway are

accessible. By scaling width and length, the amount of culture area can be expanded without

losing the plug-flow hydrology. Multiple raceways can also be employed to increase the scale of

the culture area as well as to conform to the statistical demands of experimental research.

Blountstown Facility

The raceway system constructed at the freshwater fish aquaculture farm in Blountstown,

FL site (3035.5' North, 8502.6' West) consisted of 2 source ponds supplying a group of 9

raceways arranged side-by-side. Raceways were positioned at the same elevation as the source

pond bottoms with ponds located to the north and west. Source ponds had an approximate area

of 0.20 ha (0.5 acre), and water depth was maintained at 1.5 m, yielding an estimated volume of

3084 m3 (108,900 ft3). A 15.2 cm (6") ID PVC supply line was installed through each pond

berm using a concrete anti-seep collar. Source water from the ponds was gravity fed to the

raceways, and the flow rate was regulated using a brass gate valve installed on each supply line.

A gravity fed inflow was chosen to minimize operational and equipment costs associated with

pumping water both to and from the raceways. Flow rate of the incoming water was determined

using a 15.2 cm (6") ID turbine-type, in-line flow meter for each raceway supply line.









Maximum system flow rate was limited to 950 LPM at 1.5 m source pond depth by the gravity-

fed design.

The design of the plumbing for the entire raceway set is illustrated in Figure 2-2. Supply

lines from the ponds were connected to a 15.2 cm (6") ID PVC manifold for each. Each

manifold was plumbed with a single 10.2 cm (4") ID PVC feed line for each raceway with a 10.2

cm (4") ID brass gate valve installed on each raceway feed line to select for the desired source

pond inflow. Raceway feed lines from the manifolds were then reduced to 5.1 cm (2") ID PVC

lines and joined together to form the raceway input plumbing.

The output plumbing from each raceway consisted of two separate 15.2 cm (6") ID PVC

output lines that joined corresponding manifolds made of 20 cm (8") ID PVC. Each manifold

emptied into a separate 4,542 L (1,200 gallon) concrete sump tank buried underground. Vertical

vent pipes were installed on the ends of the manifolds to prevent a suction effect in the

standpipes for the raceway furthest from the sump tanks. Each sump tank was equipped with a

120-volt, 5-horsepower centrifugal pump rated at 787 LPM (208 GPM) that was cycled using a

float switch. The pump suction lines drew water from the bottom of the sumps using a 10.2 cm

(4") ID PVC line equipped with a PVC foot valve to prevent the need for priming of the pumps

at the startup of each run cycle. A 10.2 cm (4") ID PVC outflow line leading to each pond was

plumbed from each pump and an array of 10.2 cm (4") ID brass gate valves was used to divert

water to the desired pond. This plumbing system was designed so that both source ponds could

be used simultaneously by supplying several raceways without any mixing of the two water

bodies.















0
0
\ ~E
~I.


Figure 2-2. Blountstown system plumbing diagram.

Figure 2-2. Blountstown system plumbing diagram.


--0























o -









Hague Facility

The raceway system constructed in Hague, FL (29048.0' North, 82o25.1' West) was made

of three independent test systems, each consisting of two source ponds and three raceways.

Source ponds had an approximate area of 0.05 hectare, and depths were maintained at 1.9 m,

yielding an estimated volume of 970 m3. Each pond set was fitted with aeration supplied by a

1.5-horsepower continuous-duty centripetal blower, and a 5.08 cm (2") ID PVC main line

reduced to a polyethylene line extending to the center of the ponds. A 1.9 cm (34") ID brass ball

valve was installed at each pond to balance the airflow to weighted 15.2 cm (6") long, stone

diffusers on the pond bottoms.

In each triplicate test system, raceways were positioned near the banks of the source ponds

at the north end of each pond set. Source water was supplied by continuously pumping from the

south end of the ponds to the raceways where it exited through a standpipe and was returned to

the north end of the pond by gravity-feed. A diagram of the plumbing used for each raceway

system at the Hague facility is shown in Figure 2-3. A single 120-volt, 5 HP centrifugal pump

rated at 787 LPM (208 GPM) was located on a concrete pad between the ponds on the south end

of each treatment pond pair. The suction side of each pump was plumbed using 7.6 cm (3") ID

PVC pipe with a PVC foot valve at the pond end to prevent loss of prime. A strainer made from

plastic 64 mm (1/") mesh screen was installed over the intake to prevent the passing of large

particles that might have been harmful to the pump. A pair of brass 7.6 cm (3") ID gate valves

was used to isolate the desired supply pond. The pump intakes were suspended from steel-

framed piers to one meter above the bottom of the pond bottom. A union was also added to each

suction line at the pier to enable removal of the submersed portion for regular cleaning of the

strainer screen.









Water output from each pump was delivered to the raceways using a 10.2 cm (4") ID PVC

line and regulated through a set of overpressure relief valves located near the raceways that sent

the overpressure water back to the source pond through a 10.2 cm (4") ID PVC return line. The

relief system was necessary because the continuous-duty pump configuration introduces water at

a constant flow. Therefore, in order to reduce the amount of flow to the raceways without

cycling the pump, some water volume must be relieved from the pump output line. Each pump

was also fitted with a 75-pounds per inch2-rated pressure relief valve as an emergency feature in

the event of a line blockage. After the overpressure water was relieved, the source water passed

through a 10.2 cm (4") ID turbine flow meter before entering a 15.2 cm (6") ID PVC raceway

supply manifold similar to the one used in the Blountstown system.

The manifold is reduced to a 10.2 cm (4") ID PVC fitting at each raceway and further

reduced to a 5.08 cm (2") ID PVC raceway feed line that empties into the distribution bar. Each

raceway feed line flow was measured using a Pitot-tube flow meter located between the valve

and the spreader bar. These flow meters, along with the valves at each raceway, were used to

regulate the flow balance between individual raceways, while the flow meter and relief valves

before the manifold regulated available water input to the raceway set.

As in the Blountstown system, the standpipe plumbing from each raceway consisted of 2

separate 15.2 cm (6") ID PVC output lines that joined to manifolds made of 20 cm (8") ID PVC

corresponding to the source pond receiving the outflow. Vertical vent pipes were installed on the

ends of the manifolds to prevent a suction effect in the standpipes for the raceway fartherest from

the outflow, as used in the Blountstown system.


































"C"
-o


O
\\










F e .H
,










F 2p
a)t
0~ I ^









1- C- \
.a E vs











Figur 2-3. Hagu syte plmbn digrm









System Design Summary

The Blountstown facility source ponds were three times larger than the Hague ponds;

however, only two ponds were available as opposed to the six at the Hague facility. The Hague

facility allowed for three different pond treatments to be performed simultaneously since the

raceways were distributed into three separate pond/raceway systems each consisting of three

raceways and two source ponds. This layout allowed for an alternate source pond in the event of

a catastrophic event such as phytoplankton population crash or water quality issues. The

Blountstown system could only support two simultaneous pond treatments and only one if an

alternate pond was to be incorporated. Using nine raceways per pond treatment at the

Blountstown system increased the ability to test different raceway conditions with more

statistical rigor than the three raceways per pond treatment. Consequently, the Hague system had

a 1 : 224 raceway set volume to pond volume ratio, similar to the 1 : 236 ratio found in each

Blountstown pond/raceway set when all nine raceways were being fed from a single source

pond.

The gravity inflow/pump outflow configuration constructed at Blountstown may be the

more energy efficient choice; however, in the event of pump failure, the source pond will

continue to flow, flooding the raceway site and evacuating the source pond. In order to prevent

this catastrophic failure, the only options are large self-actuating valves for the main input lines

or an emergency power generator, both of which are expensive propositions. A pneumatic or

mechanical self-actuating valve triggered by a battery-back up float switch circuit from the sump

tanks is the ideal mechanism since pump failure may occur without power outage.

Another drawback of the gravity-fed raceway inflow is that flow rate is dependent on water

depth in the source ponds. In the Blountstown system, the input flow rate to the raceway varied

with changes in pond depth. As a result of the main input flow variation, the individual raceway









flow rates would become unequal, requiring constant adjustment of the individual raceway input

valves and pond refill rate to maintain consistent flow regime. This constant monitoring was not

needed at the Hague facility since the pump input delivered a constant, inflow volume balanced

among the raceway set. Consequently, the maximum input flow rate available to the raceway set

was limited by the source pond depth and the input flow line diameter in the gravity-fed design,

whereas a larger capacity pump can be installed to increase flow in the pump-fed design to

achieve the individual raceway maximum of 227 LPM.

The raceway-based system design used in this study provided important elements of

flexibility, which allows for a broad range of applications. Integration of raceways into various

source and receiving water system configurations can be accomplished by modifying the water

delivery component to adapt to the layout of the water bodies, area topography and raceway

hydrological demands. The most energy efficient design is the gravity inflow/gravity outflow

configuration that eliminates costs and energy consumption associated with multiple pumps;

however, this can only be applied in a flow-through system with the proper raceway elevations.

In most circumstances, at least one pump will be required to feed and/or evacuate the raceway

systems. Other components, such as settling ponds, could also be integrated into the source

water stream to remove excess particulates (Krom et al. 1995, Shpigel and Fridman 1990 Van

Rijn 1996).

Changing input flow rate or adjusting the water depth by modifying standpipe height or

substrate volume allows for manipulation of raceway hydrology. Changes in these settings could

be used to adjust water retention time or linear velocity in constantly flowing systems. Water

retention time is often shortened in large commercial raceway systems either to maintain high

dissolved oxygen concentrations or to avoid ammonia build-up (Timmons et al. 2002).









Retention time can be markedly increased to increase exposure of the water for treatment

purposes by using a batch-fed regime where raceway inflow is stopped and raceway water is

circulated or agitated to mix and re-aerate it. In this study, retention time was increased to six

hours using a batch-fed configuration that incorporated a 1/3 horsepower submersible pump to

circulate and aerate raceway water. Another concern with long retention times is heat exchange,

since temperature increases in shallow raceways are due in part to radiant heat exposure from the

sun as well as evaporative heating/cooling.

The raceway structures in this study incorporated a modular design allowing for transport

and reuse at multiple sites. Portability makes the raceways ideal for application in the scientific

arena since: 1) experiments are often short term, 2) systems must be cleared from the site at the

end of the study and 3) sites may be located in remote areas. Another modular design has been

applied to reinforced fiberglass panels by Vantaram (2004) as a lightweight and less permanent

alternative to concrete. The Vantaram (2004) system requires no bracing around the sides of the

raceway to maintain the structural integrity such as in metal, fiberglass or wood-framed raceways

that need buttresses or dirt to maintain the desired shape. The raceway design used in this study

required soil backfill for structural support. This made each raceway fully accessible from all

sides, allowing for easy and comfortable physical manipulation, cleaning and sampling, an

advantage in the experimental systems. The advantage of this design over metal, fiberglass and

the Vantaram-type panels is that the materials can be readily purchased locally and assembled

without molding structures, curing time, special tools and health concerns over solvents and dust,

all of which impact budgetary demands.

The raceway-based systems constructed in this study were chosen and developed as a low-

cost, less permanent and easy to assemble alternative to other raceways constructed from









concrete, fiberglass or plastic. The raceway system designs employed here are versatile enough

to be applied to other organisms targeted for large-scale water treatment/biofiltration studies in

both fresh and saltwater conditions, a variety of locations and effluent sources.









CHAPTER 3
ADAPTABILITY OF Corbicula TO TREATMENT RACEWAYS

Introduction

The potential for using the freshwater clam Corbicula in dairy wastewater treatment

ultimately depends on the ability of the organism to adapt successfully to that environment.

Clam-based treatment raceways are essentially production aquaculture systems designed to

function using phytoplankton biomass grown on wastewater as the main food source. Growth

and harvest of clam tissue biomass represent the accumulation and removal of wastewater

derived nutrients. As in traditional aquaculture systems, growth, recruitment and health of the

clams in the treatment raceway population are important elements in assessment of the potential

success of large-scale systems.

Studies based on the production of filter-feeding organisms as a mechanism for wastewater

treatment has been applied mainly in the mariculture industry, where commercially desirable

species are cultured using finfish and shellfish farm effluents in both natural and engineered

systems (Shpigel and Blaylock 1991, Jakob et al. 1993, Shpigel et al. 1997, Lin et al. 2001). No

large-scale commercial markets for freshwater filter feeders currently exist. The freshwater clam

Corbicula may be an ideal candidate for such a wastewater treatment system in freshwater due to

its high filtration and growth rates under eutrophic conditions (Greer and Ziebell 1972, Mattice

1977, Buttner 1986, Beaver et al. 1991, Brock 2000). Corbicula is also known for its ability to

form high-density populations (Gardner 1976) and maintain high rates of reproduction (Rodgers

et al. 1977, McMahon and Williams 1986, McMahon and Bogan 2001). In freshwater

aquaculture applications, Corbicula utilizes uneaten feed and feces, in addition to phytoplankton,

produced from the polyculture of other organisms (Buttner 1986).









Sustained aquaculture of freshwater Corbicula for human consumption has been reported

in open pond-based polyculture systems (Villadolid and Del Rosario 1930, Miller and McClure

1931, Ingram 1965, Chen 1976). These open systems utilize constant exchange of water from

natural bodies that can contain juveniles and are therefore not dependent upon reproduction from

individuals within the system. Little is known about large-scale Corbicula aquaculture practices

using engineered wastewater treatment systems.

Other freshwater species of bivalves have also been considered for wastewater treatment

including Lampsilus clairbornensis (Swingle 1966), Diplodon chilensis (Soto and Mena 1999)

and Eliptio complanata (Stuart et al. 2001); however, Corbicula may be better suited for

aquaculture since it does not require an intermediate fish host for reproduction and can self-

fertilize (Kraemer et al. 1986, McMahon and Bogan 2001). This reproductive advantage over

freshwater mussels indicates that Corbicula populations should in theory be self-sustaining and

should not require re-stocking from external sources or polyculture with proper fish

intermediates to balance recruitment with biomass removed by mortality and harvest.

An application of Corbicula aquaculture to large-scale treatment of agriculture effluents

using phytoplankton as an intermediary has been proposed by Greer and Zeibell (1972), Stanley

(1974) and Olszewski et al. (1977). Greer and Ziebell (1972) used Corbicula in short-term

aquarium-based filtration experiments (16 days) to evaluate the clam's treatment potential for

waters enriched with inorganic nitrogen and phosphorus. However, the short duration and small

scale of this study may not accurately reflect the organism's ability to provide treatment over

longer time periods and in a larger system needed for real-life applications. Long-term studies

utilizing Corbicula for wastewater aquaculture have targeted freshwater finfish effluent (Habel

1970, Busch 1974, Buttner and Heidinger 1980, Buttner 1986) and municipal wastewater









(Haines 1977), but not agricultural wastewaters. In addition to nutrient reduction, Corbicula-

based systems decreases turbidity (Habel 1970, Busch 1974, Haines 1977, Buttner 1986) and

increase dissolved oxygen at dawn (Buttner 1986). The extent that these treatment actions can

be performed is primarily a function of the amount of biomass present and is strongly influenced

by population dynamics.

In this study, a large-scale raceway-based treatment system was used to examine the

adaptability of Corbicula to the raceway conditions that may be encountered in northern Florida

agriculture operations. It was hypothesized that Corbicula populations introduced to treatment

raceways would survive, grow and reproduce under wastewater conditions. Changes in clam

population were compared to raceway water quality, nutrient concentrations and phytoplankton

biomass parameters to assess the impact of environmental conditions and food availability on

population dynamics. Assessment of population parameters along with monitoring of

environmental conditions was used to determine the potential for the successful application of

Corbicula in the large-scale treatment of agricultural effluents in northern Florida.

Methods

Growth of clams in the treatment raceways was assessed as a function of changes in shell

size, tissue biomass and health, while survival and recruitment of juvenile clams was used to

assess changes in the numbers of individuals in the raceways over time. Tagging studies are

more precise, but they introduce additional handling stress resulting from tag application.

Population sampling limits handling stress but is imprecise and requires an independent

confirmation of growth from tagging studies. The combination of these two methods was

intended to provide a measure of quality assurance/quality control.









Raceway-based Treatment System

The previously described (Chapter 2) raceway-based recirculating system constructed at

the University of Florida Dairy Research Unit in Hague, Florida was used to test the adaptability

of Corbicula populations to large scale culture under exposure to simulated agricultural

wastewater conditions. Three independent pond/raceway systems were constructed in June 2002

to compare source waters with low, medium and high levels of nutrients. Each pond/raceway

system consisted of two earthen source water ponds and three wood-framed, PVC-lined

raceways. Table 3-1 shows the numerical designations for the ponds and raceways in each

nutrient addition treatment group.

Table 3-1. Source pond and raceway numerical designations for the treatment systems at
the Hague site.
Nutrient addition treatment Source pond Raceways
Low 1,2 1 3
Medium 3,4 4- 6
High 5, 6 7 9

Source ponds

Source ponds had an approximate area of 0.05 hectares (ha), with depths of approximately

2 m and volumes approximately 1000 m3. Ponds were enriched with a blend of nitrogen and

phosphorus fertilizer or anaerobically digested dairy farm wastewater, to simulate possible water

conditions associated with tertiary wastewater treatment. The low nutrient addition treatment

received no external nutrient addition. A 5 % and 10 % addition of anaerobically digested dairy

farm effluent was added to Pond 3 from the medium nutrient group and Pond 5 from the high

nutrient addition treatment group, respectively. For effluent physical and chemical

characteristics see Wilkie et al. (2004). Effluent was pumped from the digester to the source

ponds and metered through a 2.54 cm (1") turbine-type flowmeter.









Pond 4 in the medium nutrient addition treatment was dosed with 1.1 kg of triple super

phosphate (9Ca(H2PO4)2 N-P-K = 0-45-0) and 6.8 kg ammonium nitrate (NH4NO3, N-P-K =

15-0-0) resulting in a total addition of 0.23 kg of phosphorus and 1.02 kg of nitrogen per pond in

October 2002, one month prior to clam stocking. Pond 6 in the high nutrient addition treatment

was dosed with 2.2 kg of triple super phosphate (0.46 kg P) and 13.6 kg of ammonium nitrate

(2.04 kg N) in January 2002, also one month before introduction of clam populations.

Nutrient additions were designed to enhance phytoplankton biomass, which was the

putative source of particulate nutrition for the clams. Fertilizer loading levels were targeted at

increasing phosphorus and nitrogen levels by 0.23 mg/L TP and 1020 tg/L TN in the medium

nutrient treatment source pond and 0.46 mg/L TP and 2040 tg/L TN in the high nutrient

treatment source pond. Fertlizer was introduced to ponds by placing it into a burlap bag

suspended in the water column by a 0.5 m x 0.5 m floating frame constructed from 5.08 cm (2")

ID PVC pipe.

Neither of the ponds supplied with dairy effluent were exposed to the clam raceways due to

excessively high ammonia levels (2.0 mg/L or greater as NH3-N), which represented a direct

threat to the health of the clams. NH3-N levels in the effluent ponds and input and output water

from each operating raceway were monitored monthly using Aquacheck brand Ammonia

Nitrogen (NH3-N) Test Strips (Hach Incorporated, Colorado, USA) commonly used for

aquaculture and aquarium applications. Levels of NH3-N in the raceways supplied by ponds 4

and 6 never reached the 0.25 mg/L (as NH3-N) minimum value of the test kit.

Source ponds were circulated through the raceways for 10 to 20 days prior to clam

addition. Supply ponds were aerated at night throughout the study to help with mixing,

maintenance of nighttime dissolved oxygen and evaporative cooling. Floating macrophytes were









cleared by hand from each source pond several times at the start of the experiment. In addition,

six juvenile triploid grass carp ranging between 15 and 20 cm in length were stocked in each

pond in June 2002 to help reduce vegetation in ponds. No fish mortality was evident in ponds

not exposed to dairy wastewater effluent. In contrast, ponds treated with wastewater all

experienced 100 % fish mortality in less than one week after effluent addition. Fish mortality

events corresponded with high ammonia levels ( > 0.25 mg/L NH3-N).

Raceways

Raceways were 1 m in width by 7.4 m in length, with an available substrate surface area of

7.2 m2, after subtracting standpipe area. Water depths were maintained at 0.2 m, yielding a

raceway water capacity of 1.4 m3 each. Source water inflow to the raceways was maintained at

227 liters per minute (LPM) (60 GPM) during normal operating conditions. The flow rates

yielded retention times of approximately 9.5 minutes, with a linear velocity of 1.17 m/min. Flow

rates were calculated assuming near laminar flow through the raceway structure. Raceways were

filled to 0.2 m depth with a coarse grade SiO2 filtration sand (0.6-1.0 mm particle size), that was

purchased from Feldspar Incorporated in Edgar, Florida.

Aquatic plants such as Chara sp. and filamentous algae (mainly S'pii /griv sp.) growing on

the raceway substrate and liners were removed by hand from each raceway at least weekly to

reduce fouling from increased water retention time. The control of plants, especially Chara, on

the substrate made accurate estimation of clam biodeposit sedimentation impossible. This was

due to the constant resuspension of sediment deposits associated with disturbance from removal

of the vegetation that was often rooted below the sediment surface. Filamentous algae tended to

utilize the sides of the raceways where PVC liners were submersed.









Water analysis

Source water ponds and raceways were monitored once per week at 6:00 and 18:00 for

temperature, dissolved oxygen, pH, chlorophyll a, total nitrogen and total phosphorus.

Measurements in the ponds were taken at the end of sampling piers near the intake pipe leading

to the raceways. Monitoring intervals were changed to monthly in June 2003, after 75 % or

more of the clams were assumed dead from numbers of cumulative dead clams found on the

sediment surface. Temperature and dissolved oxygen was measured using a YSI model DO550,

and pH was measured using a Fisher model AP63 meter. Water temperature, dissolved oxygen

and pH readings were compared using a paired t-test (Microsoft Excel) at the raceway input

and output, as well as between nutrient addition treatments.

Water samples were collected from the sampling piers in each pond for nutrients, using a

pole sampler designed especially for this experiment. The sampler used a plunger-type

mechanism to collect a 1 L water sample from in front of the intake pipe. When the unit was

lowered to the desired depth, the plunger was actuated by the operator via a spring-loaded handle

at the opposite end of the pole. After the sample was collected, the plunger was released, sealing

a 1 L plastic (Nalgene Incorporated, USA) bottle and raised for retrieval. The sample bottle was

unscrewed from the sampler and capped for transport to the laboratory.

Water samples collected from the source ponds were analyzed at the laboratory for

phosphorus, nitrogen and phytoplankton biomass in terms of chlorophyll a. Total phosphorus

(TP) and total dissolved phosphorus (TDP) were determined using the potassium persulfate

digestion method (APHA 1998) with a Hitachi spectrophotometer. TDP determination involved

pre-filtering through a 0.7 ptm glass fiber filter. Total nitrogen was determined using potassium

persulfate digestion method (APHA 1998) with colorimeteric analysis performed using a Bran-

Luebbe auto analyzer. Phytoplankton biomass was estimated using chlorophyll a (chl a),









measured by filtering 250 mL of water onto a 0.7 [m glass fiber filter, followed by an ethanol

extraction (Sartory and Grobbelaar 1984) and spectrophotometric determination (APHA 1998)

using a Hitachi spectrophotometer. Microscope observations of phytoplankton species

composition were obtained periodically to describe dominant organisms with help from Mary

Cichra at the University of Florida Fisheries and Aquatic Sciences Department. Data obtained

from these phytoplankton species observations were not assessed quantitatively.

Clam Population Dynamics In The Raceway Environment

The adaptability of clams to the raceway environment was determined by assessing

survival, recruitment, growth and health. Clams were sampled at the time of stocking and at

specified intervals over a period of 440 days to evaluate population density, shell size and tissue

biomass. A tagging study was employed to validate survival and growth determined from

monitoring of raceway populations. A three-month interval was chosen for sampling time

duration der to reduce handling stress, while retaining the ability to assess changes in clams on a

seasonal basis. The raceway stocking and sampling time schedule is shown in Table 3-2.

Table 3-2. Raceway (RW) stocking and population sampling schedule for the low,
medium and high nutrient addition treatment systems. Time interval 0
corresponds to the time of stocking, and treatments were not stocked during
the same season due to the time required to obtain the large numbers of clams
needed.
Time interval
Year Season Low nutrient Med. nutrient High nutrient
RW 1-3 RW 4-6 RW 7-9
2002 Summer 0
2002 Fall 1 0
2003 Winter 2 1 0
2003 Spring 3 2 1
2003 Summer 4 3 2

Shell length was used as the primary indicator of clam size, since it can be measured

quickly, is not subject to the variability exhibited by soft tissue, and is stable over time. Shell









length was defined as the greatest distance anterior to posterior measured perpendicular to the

hinge line using a caliper measured to the nearest 0.01 mm. Shell length may be the best

measurable size variable compared to height and width in Corbicula because it is the largest size

variable, making it less sensitive to measurement error. Shell length also encompasses areas of

the shell that are less susceptible to erosion. Other studies of Corbicula have used shell length as

a descriptor of size, including Mattice and Wright (1986) and McMahon and Williams (1986).

In order to confirm that length provides the most dependable measurement of clam size, an

allometric analysis was performed on a sample of 500 clams obtained from the Santa Fe River

(2951.1' North, 82 37.9' West) in March 2002. Each clam was measured for shell length,

width and height. A regression analysis using SAS (PROC REG) (SAS Institute@, Cary, NC)

yielded shell length as the measurement with the highest R-square value (r2 = 0.96) compared to

the width and height (r2 > 0.94), thus making this size variable the most consistent over the size

range used in this study (shell lengths 9.4 mm to 28.4 mm). Length may also be the preferred

variable because shell erosion was apparent in the umbo region in larger clams, thereby affecting

height and width measurement values. Measurement error was determined by repeating the

length measurements three times on twenty randomly selected clams from the allometric

analysis. The maximum variance for the mean shell length was + 0.1 mm.

Stocking clam raceways

Clams for stocking the raceways were obtained from populations in three different natural

water bodies under permit number FNC-04-022 issued by the Florida Fish and Wildlife

Commission. Clams for the low nutrient group (Raceways 1-3) were collected in June 2002

from a 0.5 km stretch of the Santa Fe River near the State Road 49-bridge in Gilchrist County,

Florida (2954.2' North, 82 52.0' West). By November 2002, the rising water level of the river









made further clam excavation impossible; therefore, animals for Raceways 4-9 were collected

from lakes located in Lake County, Florida that had accessible populations of clams. Clams for

the medium nutrient group (Raceways 4-6) were collected from the southwest shore of Lake

George (2912.2' North, 81o35.7' West) in November 2002. Clams for the high nutrient group

(Raceways 7-9) were obtained from the west shore of Lake Dalhousie (28o54.0' North, 81o36.8'

West) in February 2003. Possible adaptability issues with ontogenetic differences in the

populations obtained from different locations were not addressed due to the difficulties inherent

in locating and obtaining such high numbers of clams from systems in a timely manner.

All three collection sites had coarse sand sediments similar to the substrate used in the

raceways. Clams were excavated by shoveling bottom material into weighted baskets made from

plastic mesh with 0.635 cm2 (14 in2) perforations. Clams were also excavated by hand using

trowels or a commercial clam rake modified for the small size of the Corbicula by affixing

similar plastic mesh on the inside of the collection basket. Periodic excavation of bottom

sediment using a 0.25 m2 PVC sampling quadrat was used to determine population densities for

the clams in their natural habitat. Densities ranged from 48 and 864 clams/m2 with a mean of

272 clams/m2 (standard error (SE) = 23, n = 47 observations) for all locations combined.

After excavation, clams were enumerated and divided into mesh bags. The clams were

then placed into coolers packed with wet newspapers and kept out of direct sunlight to help

minimize heat stress and dessication. They were transported directly to the aquaculture facility

in Hague, FL and scattered evenly throughout each raceway. Stocking of each raceway took up

to 15 days involving 2 to 6 people working per day. Raceways 1-3 were stocked from June 17

to 28, 2002, Raceways 4-6 from November 4 to 14, 2002, and Raceways 7-9 from February 11

to 26, 2003.









Stocking densities were estimated using a volumetric method of enumeration. The method

employed a 0.5 L plastic container that was used to transfer the clams collected in the field to the

mesh shipping bags. Ten mesh bags of clams were added to each raceway. Each mesh bag

contained approximately 1,000 clams. This method was chosen as opposed to counting each

individual or bulk weighing in order to minimize handling stress, time and equipment needed to

enumerate the large numbers of clams needed for this study.

Shell length measurements were taken on 270 clams per raceway, selected at random from

mesh basg just before stocking. A sample of clams was obtained from each mesh bag by

scooping a sample from the middle of the bag using the 0.5 L plastic container mentioned above.

A total of 27 clams were selected for the biomass analysis per raceway by keeping the tenth clam

out of every 270 clams sampled for shell length determination.

Clam raceway population sampling

Sample sites within the raceways were defined using a submersible quadrat grid and

sampling sleeve. The quadrat grid consisted of an aluminum frame divided inside into 10 cm x

10 cm squares using 2 mm thick nylon line. In order to define 100 cm2 surface areas more

accurately for excavation, a 10 cm x 10 cm ID square metal tube, 20 cm in length was used as a

sampling sleeve. The sleeve was inserted into the raceway sediment, and clams were extracted

to a depth of approximately 10 cm.

The entire grid measured 1.2 m x 0.9 m. Grids were deployed at the input, middle and

output of each raceway to cover the length of each raceway. The 10 cm wide area was not

sampled to reduce the risk of puncturing the raceway liner. Areas around the raceway edges,

under the input spreader bars, around the outflow standpipes were not sampled. Placement of

grid in all three sections allowed for 85 % coverage of the raceway bottom, providing a total of









567 possible quadrat locations per raceway. The grid rested on stainless steel pegs buried in the

raceway sediment to ensure repeatability in placement.

The number of sampling quadrats needed per event was determined in December 2001

using a power analysis with equations from Sokal and Rolhf (1995) and performed using

Microsoft Excel on a population of 500 clams introduced to the Blountstown raceway system.

Clams were distributed over a 1 m2 area of active raceway substrate and left for 5 days. A 100

cm2 PVC square quadrat was used to divide sample areas into 10 cm x 10 cm increments.

Quadrats were excavated along a lm transect in the middle of the raceway, and the number of

live clams recorded. The power analysis on the clam density data yielded a sample size of at

least 19 quadrats to achieve a 95 % confidence interval for clam density. A sample number of 27

quadrats per raceway were chosen since the sample grid consisted of 9 sample coordinates over

the width of the raceway and 3 grid placements per raceway.

Clam sampling followed a stratified random design without repetition. One length division

was chosen at random for each width division within the grid to stratify the sample areas over the

width of the raceway. Stratification of raceway sampling allowed for spatial analysis of data

over both the length and width of the raceways. Quadrat positions were repeated within each

grid placement at the input, middle and output regions of the raceways.

Live clams in each quadrat were counted, measured for shell size and returned to the same

location in the raceway. Clams used in the biomass analysis were selected from clams included

in the density analysis by retaining the third clam excavated from each sample quadrat. In the

event that there were fewer than 3 clams, the first clam excavated was kept for biomass analysis.

This procedure yielded a maximum possible sample size of 27 clams per raceway at each time

interval. This sample size was chosen to minimize the impact on raceway populations due to the









destructive nature of biomass determination. Individuals for biomass determination were placed

in numbered plastic bags and frozen until analysis.

In order to estimate raceway population densities using the spatial technique, the clam

densities obtained during each sampling event were used to calculate the average number of

clams per 100 cm2 quadrat sampled (n = 27 quadrat samples per raceway). Average quadrat

densities were then converted to the number of clams per m2 and multiplied by 7.2 m2 of

available substrate area per raceway to obtain the estimated number of clams per raceway.

Tagged clams

Clams used in the tagging study were obtained from a random sampling of individuals at

the time of stocking. The clams were marked with EZ-Code brand wire markers (Thomas &

Betts Incorporated, USA) which are self-adhesive numerical tags applied to the shells that

minimized the handling stress to the animals. This type of tag was chosen because it caused less

damage and reduced the risk of injury when compared to engraving (Mattice and Wright 1986,

McMahon and Williams 1986, Lemarie et al. 1995) or insertion of passive integrated transponder

tags into the shell cavity (Kurth et al. 2007). The pre-applied adhesive expedited the tagging

process by reducing drying time and increasing tag readability of the liquid adhesive traditionally

used for affixing numerical tags (Lemarie et al. 1995), brass washers (Toll et al. 2003), coded

wires (Layzer and Heinricher 2004) or monofilament tethers that anchor the animals to the

substrate (Foe and Knight 1986).

This type of tag was chosen because of their small size, readability, adhesive strength and

low cost compared to traditional numerical bee tags used in bivalve research. These vinyl decals

are generally used in electrical applications and have 4 mm-tall black numbers with a white

background. Tags were trimmed to 5 mm x 5 mm squares before application to towel dried

shells. Tagged clam shell lengths were measured as described earlier and placed into the









raceways using the sampling grid described above. Using the randomly generated coordinates

that were sampled over the study, one tagged clam was placed in each location using at least 33

locations at the input, middle and output sections of each raceway.

In order to test the durability of the tags, a sample of 20 live clams was obtained from the

Santa Fe River (2951.1' North, 82 37.9' West) in December 2001. Tags were affixed and

clams placed in a 1 L plastic (Nalgene Incorporated, USA) bottle along with 200 mL of coarse

sand and 200 mL of water. The mixture was capped and shaken vigorously by hand for 15

minutes, after which the clams were removed and rinsed with water for inspection. Only one

clam lost its tag and was removed while the remaining numbered clams were placed back in the

bottle. Then the clams were again shaken for fifteen minutes and removed for inspection. Tags

on six of the clams had come off during this treatment, and the shells of all of the clams

exhibited chipping around the margins. All tags remained legible after both treatments. The

conditions that these clams were exposed to are certainly harsher than the raceway environment

because of the lack of significant water movement to produce the same tumbling effect but

neglects dissolution and bacterial decay that may also account for loss of adhesive strength with

longer exposure times in aquatic systems.

A total of 108 clams were randomly selected from each raceway at the time of stocking for

the application of tags. Specimens were tagged and distributed in a stratified random fashion by

using the sampling grid to place 36 clams in the input, middle and output sections of each

raceway. Shell length, clam number and initial quadrat coordinates were recorded for each

tagged clam at stocking and at each 3-month sampling interval when found alive. Tagged clams

that were later found dead were not used even though growth may have been evident by

comparing measurements of final shell size to that at stocking.









Clam survival

Clam survival was assessed by counting live individuals in the raceways at stocking and at

designated sampling intervals. Total densities were determined at the end of the study by

counting the live and dead clams remaining in the raceway, along with cumulative counts of

dead clams removed from each raceway over the course of the study. Raceway population

densities at stocking were estimated using the volumetric technique and at each sampling interval

using the spatial technique, as described above. Clam stocking estimates were later verified

using information on cumulative counts of dead clams removed over the course of the study,

along with counts of dead and live clams at the end of the study.

In the case of visible clam mortality events, dead clam shells were removed and counted.

Counts of dead clams removed from the substrate surface were used as an overall indication of

population mortality. No attempts were made to evaluate the dead shells buried in the substrate

after the events because the removal of dead clam shells caused rususpension of sediment

deposits.

Biomass changes

Changes in clam biomass were assessed using tissue weight and length data taken from

clams obtained at stocking and at each designated sampling interval. Tissue weight to shell

length relationships were developed from a subset of clams collected at stocking and sampling

intervals. Sample clams were frozen prior to dry weight (DW) analysis (Copar and Yess 1996).

Freezing provided an alternative to live shucking, since gaping of frozen clams occurs naturally

when the clam is removed from the freezer and placed at room temperature for about 15 minutes.

Shucking can also chip shell valve edges resulting in measurement error in shell-dependent

variables.









Clam dry weights were determined for meat and shell of freshly thawed clams. Soft tissues

and shell material were separated and placed into individual dried and pre-weighed aluminum

drying dishes. Whole clam, soft tissue and shell wet-weights were recorded and placed in a

drying oven at 80 C for a period of 24 hours. This period of time was used due to the results of

preliminary tests to establish a drying time needed to attain constant weight in clams sampled

from the Santa Fe River (2951.1' North, 82 37.9' West) in June 2002. After 24 hours, the dried

tissue samples were removed from the oven and allowed to cool to room temperature in a

dessicator. Meat and shell tissue dry weights were recorded to the nearest 0.001 g and tissues

stored in a dessicator before ashing.

Dried shell and meat tissues of selected individuals were ashed to obtain ash weight, from

which ash free dry weight (AFDW) was calculated. Dried meat tissue was collected from the

drying dishes and placed into ceramic crucibles. Shell was prepared for ashing using a stainless

steel grinding device powered by a rotary hammer. The device consisted of a 7.62 cm (3") tall x

2.54 cm (1") ID chamber made by welding a segment of pipe onto a 7.62 cm2 (3")2 x 0.635 cm

(1/4") thick plate. A plunger made from a stainless steel billet was machined on one end to accept

a standard 3/8" square drive adapter attached to a Hilti T52 (Hilti Incorporated, Germany) rotary

hammer. Each specimen was placed in the chamber and the piston lowered down on top of the

shell. The rotary hammer was engaged for 10 seconds or less, long enough to pulverize the shell.

The resulting powder was collected in pre-weighed ceramic crucibles, dried for 24 hours at 80 C

and weighed to the nearest 0.001 g. Shell biomass recovered from the grinding device averaged

94.9 % (SE = 0.3, n = 238) of the original shell DW. Tissue samples were then placed in a

muffle furnace at 550 C for 6 hours and cooled to room temperature in a dessicator prior to









weighing. The ash values were used to calculate percentage ash composition and AFDW

(Wetzel 2001).

Dry weight values were primarily used to assess clam biomass since there was less

variability and a larger sample size (n = 453) than for AFDW (n = 238). However, AFDW/DW

relationships did provide insight into organic content and biomass allocation. The coefficient of

variation was very similar for DW (0.54) and wet weight (0.52) for the animals in this study (n =

453). Wet weight measurements taken from whole frozen clams at the time of analysis were also

used to help describe biomass allocation.

Shell, meat and whole clam DW measurements were performed on the low nutrient group

at stocking, interval 1 and interval 2, the medium nutrient raceways at stocking and at interval 1,

and on the high nutrient raceways at stocking. The AFDW of shell, meat and whole clams were

taken from the low nutrient raceways at stocking and at interval 1 and the medium nutrient

raceways at stocking only. The DW and AFDW analysis were discontinued after these time

periods due to the establishment of strong linear regression relationships (r2 > 0.90) between

biomass and shell.

Tissue biomass allocation was used to evaluate biomass distribution in shell and meat, as

well as to examine biomass. Biomass allocation was also used to examine water and ash content

of shell and meat tissue, for comparison with other clam studies. Clam wet and dry weights of

shell and meat were used to calculate percent shell tissue and percent water content for the whole

clam. In order to understand variability in clam tissue biomass allocation, an ANCOVA (SAS

PROC MIXED procedure, SAS Institute@, Cary, NC) was performed using both the percent

shell tissue and the percent water content as the response variables and nutrient level as the factor

with covariates shell length, time interval, nutrient level and season. The least squared means









(LSMEANS) procedure was applied to percent shell tissue and the percent water content

(Microsoft Excel) for pair-wise comparisons. Pair-wise comparisons of the means were

performed using Tukey's method to control the experiment-wise error rate. Data from individual

raceways within each nutrient addition treatment were pooled in this analysis since no blocking

effect was found in either analysis.

The mean percentage of shell tissue and water content was determined for each nutrient

addition treatment. Mean ash content was determined for clam, shell and meat tissues for each

nutrient addition treatment. The coefficient of variation was calculated for the DW, and AFDW

values determined for whole clam biomass to indicate the least variable biomass parameter.

Coefficient of variation was calculated by dividing the standard deviation by the mean for each

variable in the sample population.

A relationship between tissue DW and shell length was used to provide a means of

estimating biomass using measurements of shell size. Tissue DW and shell length were

transformed using natural logarithm (In) to best fit the polynomial regression calculated by the

SAS PROC REG procedure (SAS Institute@, Cary, NC). An ANCOVA (SAS PROC MIXED

procedure, SAS Institute@, Cary, NC) was performed on the DW and shell length data using the

natural logarithm (In) of the values for the clam, shell and meat tissue DW. Tissue DW values

were used as the response variables versus shell length, while nutrient level was the factor with

covariates time and season. Non-significant effects were removed from the ANCOVA model.

The tissue DW and shell length values were also analyzed using a correlation procedure (SAS

PROC CORR, SAS Institute@, Cary, NC).

Clam biomass was estimated using shell length to DW relationships developed in the

regression analyses. Raceway clam biomass was calculated using the regression equations to









convert shell measurements to tissue biomass. Changes in clam biomass over time were plotted

for each nutrient addition treatment at each sampling interval to assess biomass production and

clam growth.

Reproduction and recruitment

Clam populations were assessed for reproduction and recruitment by determining whether

juveniles were present in the system at any time. Adult clams were used to stock the raceways,

and therefore, any clams found with a shell length less than the smallest clam measured at

stocking should indicate successful reproduction and recruitment. In this study, juvenile clams

were defined as having a shell length less than 5 mm. Source ponds were also drained at the end

of the study to check for the presence of clams that may have been released as juveniles from the

raceway populations but did not successfully recruit to the raceway. Water samples from the

source ponds used for phytoplankton analysis were also inspected for juveniles suspended in the

water column.

Health

Changes in clam health were tested to evaluate the physiological condition of clams in the

raceway environment over time. The goal was to relate changes in condition to changes in

raceway environmental parameters. Other studies have used percent meat content to assess

population health (Haines 1977). A similar approach to the condition indices was used in this

study, which also relies on the amount of meat and shell tissues present in the clams.

Condition index (CI) was estimated using gravimetric and volumetric indices, based on dry

meat: dry shell weight and dry meat: shell cavity volume, respectively. Both indices are

defined by the ratio of a sensitive numerator- tissue dry weight, to relatively insensitive

denominators-shell weight and shell cavity volume. The resulting values were compared for









each index and used to describe changes in raceway clam populations. CI is a measure of the

nutritive status of the clam (Rainer and Mann 1992).

Clams were collected at stocking and during the seasonal time intervals described in the

biomass sampling section. Meat DW and shell DW, along with shell cavity volume values

obtained from biomass sampling, were used to calculate the gravimetric (CI(WT)) and volumetric

(CI(VOL)) indices. The following relationships, after Rainer and Mann (1992), were used to

estimate CI values:

* CI(WT) = (dry meat weight (g) x 100 / dry shell weight (g))

* CI(VOL) = (dry meat weight (g) x 100 / shell cavity volume (mL))




2.5
Sy = 0.1404x
2
R2 = 0.9887
0 1.5



0.5 -*

0
0 2 4 6 8 10 12 14 16 18
Bead Volume (ml)

Figure 3-1. Linear regression relationship of glass sphere volume to glass sphere weight used to
estimate clam shell cavity volume for the volume-based condition index calculation.


Shell cavity volume was calculated by filling the empty shell valves with 100 jtm diameter

glass beads and weighing the beads. This method was selected since the small size of Corbicula

makes it very difficult to assess volume accurately using the water displacement method (Rainer

and Mann 1992). In order to obtain a weight to volume conversion factor, a volumetric analysis









was performed by weighing glass beads. Three samples of beads were weighed at each 1 mL

increment up to 18 mL (n = 54), and a linear regression relationship was developed between

bead volume and weight (Figure 3-1) using the PROC REG procedure (SAS Institute@, Cary,

NC). The regression equation from this analysis yielded the following conversion formula:

* Clam Shell Valve Cavity Volume (mL) = Bead Weight (g) / 0.1414

A dry weight and shell volume analysis was then performed on a sample of 20 clams

obtained from the Santa Fe River (2951.1' North, 82 o37.9' West) to investigate possible

differences in volume between the two valves of each clam. Three bead weight samples were

taken from the two shell valves of each clam, and a paired, two-sample t-test for means was

performed on the measurements using Microsoft Excel. The two-sided p-value was not

significant (p = 1), therefore, no difference in shell cavity volume was found between the 2 shell

valves in each animal. Therefore, by modifying the above conversion factor, the following

equation was used to calculate the shell cavity volume for each clam:

* Clam Shell Cavity Volume (mL) = 2 x (Bead Weight (g) / 0.1414)

Two different comparative analyses were performed on CI(WT) VS CI(VOL) values to

determine the best index for assessing clam health. A linear regression analysis was performed

on each index versus length and CI(WT) VS CI(VOL) using the SAS PROC REG procedure, (SAS

Institute@, Cary, NC), and the coefficient of variation was calculated for both CI(WT) and CI(VOL).

Variation was similar for the two indices; although, CI(WT) had the lower coefficient value.

An ANOVA (SAS PROC MIXED procedure, SAS Institute@, Cary, NC) was then

performed using the condition values for both indices as the response variable and shell length as

the factor to determine the effect of clam size on condition for each index. Shell length was

subsequently removed from the model, and an ANCOVA was performed with the CI(WT) and









CI(VOL) indices as response variables, nutrient addition treatment as the factor with covariates

time interval, raceways within each nutrient level and season. The least squared means

(LSMEANS) procedure was applied to CI(WT) and CI(VOL) indices (Microsoft Excel). Pair-wise

comparisons of the means were performed using Tukey's method to control the experiment-wise

error rate.

Results

Raceway System Environmental Parameters

Air temperatures at the racewaty site at the Dairy Research Unit in Hague, FL ranged from

2 to 41 C (Figure 3-2). Recorded air temperatures varied diurnally as much as 15 C. Raceway

water temperatures ranged from 10.1 to 32.6 C (Figure 3-3) and displayed a similar seasonal

pattern as air temperature. However, water temperatures only differed diurnally by a maximum

of 2 C. Water temperatures reached 30 C or greater just after the beginning of the experiment

in July 2002 through October 2002 and from May through the end of the study in August 2003.

Raceway dissolved oxygen (DO) averaged 8.79 mg/L (SE = 0.07) and ranged from 6.10 to

12.03 mg/L (Figure 3-4). DO was higher in the afternoons by up to 4.26 mg/L with greater

diurnal differences in the warmer months. Higher DO values were observed during the

November 2002 to April 2002 period corresponding to lower air and water temperatures.

Raceway pH averaged 7.76 (SE = 0.02) and ranged from 6.87 to 8.81. Diurnal fluctuations

in pH ranged from -0.76 to 0.89. Raceway pH was significantly higher (p > 0.05) in the low and

high nutrient addition treatments (Figure 3-5); however, water temperature and DO values did

not differ significantly between the low, medium and high nutrient addition treatments. No

significant differences (p > 0.05) were detected at the input and output of each raceway or

between raceways in each nutrient addition treatment for the water temperature, dissolved

oxygen and pH parameters.










45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0


0.0
05/14/02


0g


08/17/02


11/20/02


02/23/03 05/29/03


09/01/03


Date

Figure 3-2. Air temperature readings at the Dairy Research Unit in Hague, FL over the study
period. Air temperatures varied both seasonally and diurnally with afternoon values
exceeding 40 C in the summer of 2002.


35.0

30.0

25.0

20.0

15.0

10.0

5.0
05/


14/02


08/17/02


11/20/02 02/23/03
Date


05/29/03


09/01/03


Figure 3-3. Input water temperatures in the low, medium and high nutrient addition treatments.
Water temperatures fluctuated both diurnally and seasonally, often exceeding 30 C in
the afternoons and did not differ between treatments.


*
* ** ^ *

*

*

*~ s
'*. .'


ftm! ,.J




Low
|A Med
o High










14.00

12.00

10.00

8.00

6.00


4.00 -
Low
2.00 A Med
o High
0.00 -
05/14/02 08/17/02 11/20/02 02/23/03 05/29/03 09/01/03

Date

Figure 3-4. Raceway dissolved oxygen (DO) readings in the low, medium and high nutrient
addition treatments. Values did not differ between treatments.


9.00

8.50

8.00 ,





0 Low
6.50 A Med
0 High
6.00

05/14/02 08/17/02 11/20/02 02/23/03 05/29/03 09/01/03
Date

Figure 3-5. Raceway pH in the low, medium and high nutrient addition treatment systems.
Raceway pH values fluctuated both seasonally and diurnally and tended to be
elevated in the low and high nutrient addition treatments.









Increases in phosphorus and nitrogen levels in the source ponds did not correspond to

fertilizer addition or clam mortality events. Source pond total phosphorus (TP) ranged between

0.061 and 0.211 mg/L in the low nutrient addition treatment, 0.047 and 0.471 mg/L in the

medium nutrient addition treatment and 0.043 and 0.386 mg/L in the high nutrient addition

treatment (Figure 3-6). Source pond total dissolved phosphorus (TDP) ranged from 0.002 to

0.091 mg/L in the low nutrient addition treatment, 0.007 to 0.223 mg/L in the medium nutrient

addition treatment and 0.028 to 0.300 mg/L in the high nutrient addition treatment (Figure 3-7).

A major portion of the source pond TP was made up of the dissolved form as TDP in all of the

nutrient addition treatments as TDP followed a similar pattern as TP with sharp increases in the

spring 2003.

Source pond total nitrogen (TN) ranged from 0.177 to 9.034 utg/L in the low nutrient

addition treatment, 0.328 to 12.069 [tg/L in the medium nutrient addition treatment and 1.066 to

2.786 utg/L in the high nutrient addition treatment (Figure 3-8). Total N values fluctuated in the

low and medium nutrient addition treatments and only slightly increased in the high nutrient

addition treatment over the experimental period.

Chlorophyll a (chl a) ranged from 3.218 to 27.511 mg/m3 in the low nutrient addition

treatment, 4.505 to 26.397 mg/m3 in the medium and 19.789 to 147.299 mg/m3 in the high

(Figure 3-9). Chlorophyll a in all treatments displayed an increase after February of 2003 with

peaks from April to August of 2003. Phytoplankton communities in the source ponds were

dominated by diatoms in the low nutrient addition treatment, diatoms and cyanophytes in the

medium and chlorophytes in the high. Diatoms were present in all ponds throughout the study

period.









0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00


05/14/02 08/17/02 11/20/02 02/23/03
Date


05/29/03 09/01/03


Figure 3-6. Total phosphorus (TP) in the low, medium and high nutrient addition treatments. TP
increased during the summer of 2003 in the medium and high treatments while values
fluctuated with no apparent seasonality in the low treatment.


0.50 -
0.45 ----
Low
0.40 A Med
0.35 High
0.30 o
0.25 -
0.20 -
0.15 -a A -A
0.10 o
mmmmm
0.05 -, o
0.00 51/ 08/170 A/ 0 A 1 0m 0
05/14/02 08/17/02 11/20/02 02/23/03 05/29/03 09/01/03


Date


Figure 3-7.


Total dissolved phosphorus (TDP) in the low, medium and high nutrient addition
treatments. Total phosphorus was comprised largely of TDP therefore patterns were
similar to trends in TP.


A
Low
A Med
-- High

AA

A


A A^ XwA










12.00

10.00

8.00

6.00

4.00

2.00

0.00


05/14/02


* Low
A Med
o High Aa .







SA A 0
p. *0 00
o<>0U


08/17/02 11/20/02 02/23/03 05/29/03 09/01/03


Date


Figure 3-8. Total nitrogen (TN) in the low, medium and high nnutrient addition treatment source
water. TN fluctuated in the low and medium treatments and steadily increased in the
high nutrient addition treatment.


0
05/14/02


08/17/02 11/20/02 02/23/03 05/29/03 09/01/03

Date


Figure 3-9. Chlorophyll a (chl a) in the low, medium and high nutrient addition treatment source
ponds. All treatments show an increase in chl a beginning in Febuary 2003.









Clam Population Dynamics In Treatment Raceways

Survival

Clam stocking densities were estimated at between 6,336 to 10,686 individuals per raceway

using the volumetric method (Table 3-3), which was near the target stocking density of 10,000

clams per raceway. The number of live clams in each raceway declined at each sampling

interval for all raceways in each of the nutrient addition treatment systems (Table 3-4). All nine

raceways had decreasing densities and low numbers of surviving live clams at the end of the

study period (Table 3-5). The actual number of clams stocked was reduced by 99 % by the end

of the study in all raceways; however, a small number of clams survived (Table 3-5).


Table 3-3. Number of clams stocked in each raceway estimated using the volumetric method.
The estimated number of clams per raceway (RW) and standard errors (SE) were
calculated from the mean number of clams measured per transfer cup and the
number of full cups added to each raceway. The number of clams sacrificed for
biomass analysis was subtracted to yield the estimated clams/raceway.
RW Cups/RW Mean SE n Estimated SE
clams/cup clams/RW
1 52 206 4 10 10,686 208
2 51 207 6 10 10,547 306
3 51 190 4 10 9,670 204
4 50 192 3 10 9,575 150
5 50 196 7 10 9,790 350
6 50 196 4 10 9,775 200
7 99 64 6 10 6,336 594
8 69 102 12 10 7,004 828
9 109 78 13 10 8,448 1417

The actual number of clams stocked in each raceway was overestimated using the

volumetric method of estimation at the time of stocking. Therefore, the values given in Table 3-

3 were not used in any further analysis. The final live clam densities for each raceway estimated

by the spatial method (Table 3-4) were more similar to the actual number of live clams found in

each raceway at the end of the study (Table 3-5).









Table 3-4. Number of live clams found alive at each sampling interval estimated using the
spatial technique. Clams per raceway (RW) and standard errors (SE) were
calculated using the average number of clams found per 100 cm2 quadrat sampled,
(n = 27 samples per raceway) converted to the number of clams per m2,which was
multiplied by the 7.2 m2 of available substrate area per raceway.
Summer Fall Winter Spring
2002 2002 2002 2003
RW Clams/ SE Clams/ SE Clams/ SE Clams/ SE
RW RW RW RW
1 4000 306 560 150 427 110 320 117
2 4400 340 907 179 1013 259 880 208
3 4267 295 827 166 613 251 80 44
4 n/a n/a 4282 387 1973 292 400 129
5 n/a n/a 4128 286 2747 356 640 155
6 n/a n/a 4377 563 2827 354 400 111
7 n/a n/a n/a n/a 2476 294 27 27
8 n/a n/a n/a n/a 3225 369 133 67
9 n/a n/a n/a n/a 4665 424 107 74

There were two indications that the spatial technique may have also overestimated the

actual number of clams per raceway at each time interval: 1) raceway densities estimated at the

end of the study were much higher than the actual number of live clams found in each raceway at

the end of the study and 2) the increase in clam density in raceway 9 from stocking (Table 3-5)

from the first sampling interval (Table 3-4) could not be substantiated since 123 dead clams had

been removed from the raceway during that time period and no indications of reproduction were

observed. The number of live clams in each nutrient addition treatment (Figure 3-10) was

calculated using the actual number of clams stocked (Table 3-5), the actual number of live clams

found at the end of the study (Table 3-5) and the number of clams at each sampling interval in

between (Table 3-4).

No attempt at biomass harvest was made due to the high losses of clams over the study

period. Clams were only removed from the raceways as needed for periodic mortality counts and

at seasonal sampling intervals for biomass determination. There were no indications of predation

by wildlife as evidenced by the lack of obvious disturbance to the raceway substrate and lack of









tracks that would have been left by mammalian predators around the raceway site. Large birds

such as cattle egrets often congregated on the sides of the raceways; however, predation of clams

in the raceway by avian predators was not directly observed.

Table 3-5. Actual number of live clams stocked in each raceway and at the end of the study.
Actual number of clams stocked was determined from cumulative counts of live
and dead clams over time as well as live and dead clams present at the end of the
study. The number of clams removed represents losses due to mortality and
destructive sampling but not predation.
Raceway Actual Clams removed Live clams present
clams stocked at finish
1 7083 7047 65
2 9251 9210 47
3 6794 6771 23
4 6356 6301 55
5 4663 4601 62
6 4478 4434 38
7 4880 4871 9
8 5340 5323 17
9 6338 6319 19

Clam survival could not be accurately verified using periodic counts of dead shells

removed from the substrate surface of each raceway because dead clams were also found buried

in the substrate, making them inaccessible for enumeration until the entire raceway could be

excavated at the end of the experiment. The cumulative number of dead clams collected on the

substrate surface after large mortality events (Figure 3-11) reflects trends in reductions of

raceway live clams (Figure 3-10). Dead clams removed over the course of the study and at the

end of the study accounted for more than 90 % of the total number of clams stocked.

Growth

There were no clear patterns in shell size distributions at stocking or at each sampling interval;

therefore, a cohort-based analysis of changes in shell size within the raceway population could

not be performed. Instead, the average shell length of the individuals found over time in each

raceway is given (Table 3-6). No consistent long-term trends in shell size data were observed










over time in raceways 1-6. All raceways showed some increase in mean shell length at the first

sampling interval, and clams in raceways 7-9 showed a continued increase in mean shell length

at Interval 2; however, any trends present must be viewed with caution due to the decreasing

sample size and high mortality rates.


25,000

20,000 L
\ -A-A-
15,000 -- High -

:: 10,000

5,000

0
5/14/2002 8/17/2002 11/20/2002 2/23/2003 5 29 2003 9/1 200(3
Date

Figure 3-10. Number of live clams in each nutrient addition treatment. No increases in live
clam density were found in any of the nutrient addition treatments.

25,000
-U-Low
20,000 --Med
-0-High
S15,000 -
"-d

10,000

S5,000

0
05/14/02 08/17/02 11/20/02 02/23/03 05 29 03 09/01/03
Date


Figure 3-11. Cumulative number of dead found on the substrate surface in the low, medium and
high nutrient addition treatments. Mortality was greater during the warmer periods in
all three nutrient addition treatments.











Table 3-6. Mean shell lengths measured from clams in each raceway (RW) both at
stocking and at each sampling interval. Mean shell length displayed a constant
increase in the high nutrient addition treatment raceways; however, decreases
in sample sizes (n) over time in each raceway prevented positive identification
of any discernable trends in shell size distributions and mean shell length over


time.
RW Shell
Length
1 Mean
SE
Range
n
2 Mean
SE
Range
n
3 Mean
SE
Range
n
4 Mean
SE
Range
n
5 Mean
SE
Range
n
6 Mean
SE
Range
n
7 Mean Length
SE
Range
n
8 Mean
SE
Range
n
9 Mean
SE
Range
n


Time Interval
Stocking Interval 1
18.6 20.3
0.2 0.2
10.9-30.6 16.6-30.0
270 150
18.0 20.4
0.2 0.2
8.0-27.1 8.9-28.7
270 165
18.0 20.8
0.2 0.2
8.5-28.5 13.2-24.1
270 160
19.2 20.0
0.2 0.2
9.21-28.9 10.6-29.0
270 200
19.5 19.8
0.2 0.2
12.1-27.4 12.8-33.5
270 206
19.1 18.9
0.2 0.2
11.1-26.5 11.7-31.3
270 203
22.2 25.8
0.5 0.6
10.2-34.1 16.4-34.3
270 130
24.9 25.0
0.4 0.4
10.3-38.8 17.3-33.3
270 168
22.7 26.5
0.4 0.4
10.3-34.0 5.8-41.3
270 222


Interval 3
19.5
0.6
14.4-23.3
16
16.4
0.5
11.0-23.6
38
17.4
0.4
15.0-22.4
23
22.6
0.8
13.9-25.6
15
20.0
1.0
9.7-24.8
24
20.3
1.2
10.2-25.1
15


Interval 4
21.2
0.5
18.7-23.3
12
20.1
0.4
11-24.1
35
19.6
0.3
19.1-20.0
3


Interval 2
20.1
0.5
13.2-24.1
21
15.9
0.7
10.2-24.6
34
17.5
0.8
8.8-24.4
31
22.5
0.2
16.3-29.2
74
22.6
0.3
10.0-28.0
102
22.0
0.3
8.4-30.0
105
31.4
0
31.4
1
29.2
1.5
25.0-33.0
5
27.0
1.1
23.7-28.7
4









Of the 972 total tagged clams introduced to the raceways, only 25 were recaptured alive

over the course of the study. All of the recaptured live tagged clams showed increased shell

length (Figure 3-12). The largest increases in shell length were observed in clams recovered

from the high nutrient group, and clams in the medium nutrient addition treatment appeared to

have a larger increase in shell growth during warmer months (Table 3-7). The range in shell

length for the individuals sampled for biomass analysis was 10.2 to 33.7 mm (Table 3-8). The

high nutrient addition treatment contained individuals with the largest shell size (31-33.7 mm)

compared to the low and medium nutrient addition treatment.


35.0



30.0



25.0



200


15.0


10.0 .!...
05/14/02 08/17/02 11/20/02 02/23/03 05/29/03 09/01/03

Date
Figure 3-12. Changes in shell lengths of tagged clams captured alive in each nutrient addition
treatment. Growth was observed in all live tagged clams recovered with the largest
increases found in the high and medium nutrient groups.









Table 3-7. Shell growth rates for tagged clams in each nutrient addition treatment for each
seasonal time interval. Growth rates were calculated from initial and recovered
shell lengths of live, tagged clams. Shell growth rates were similar for live,
tagged clams found at any time in the low nutrient addition treatment, while
growth rates during the spring were much higher than the winter for the medium
nutrient addition treatment.
Nutrient Season Mean SE Growth rate Initial shell n
treatment growth (mm/day) range size range
rate (mm/day) (mm)
(mm/day)

Low All 0.0119 0.0036 0.0009-0.0363 13.3-27.7 10
Med Winter 2002 0.0031 0.0008 0-0.0056 17.5-20.5 7
Med Spring 2003 0.0556 0.0158 0.0258-0.0856 17.0-23.9 4
High Spring 2003 0.0830 0.0168 0.0427-0.1173 12.2-28.6 4

Table 3-8. Shell size information on clams sampled for tissue biomass analysis from each
nutrient addition treatment raceway system. The largest individuals were found
in the high nutrient addition treatment.
Nutrient Average length SE Shell size range n
treatment (mm) (mm) (mm)
Low 19.9 0.2 10.2 -30.4 210
Med 20.3 0.2 13 -28.4 162
High 24.8 0.7 12.6 -33.7 81

The mean water content of whole clams was 31.5 % (SE = 0.4) with a range of 25.0 45.1

%, n = 453 observations. Water content did not vary significantly (p > 0.05) with shell length,

between raceways, over time or between nutrient addition treatments based on the ANOVA

analysis. The wide range of water content values observed prevented the use of tissue wet

weights to assess biomass even though clam wet weight values were highly correlated with the

whole clam DW values (corr coeff = 0.90, n = 453 observations).

Ash content of the meat tissue was much lower than the shell tissue, and shell ash made up

a greater portion of the total clam ash than meat tissue (Table 3-9). Ash content did not vary

significantly (p > 0.05) with shell length, between raceways, over time or between nutrient

addition treatments based on the ANOVA analysis.









Table 3-9. Mean and range of ash content values for meat, shell and total clam tissues pooled
for all clams sampled. Shell tissue had higher ash content and made up a greater
portion of the total clam ash than meat tissue.
Tissue Mean ash SE Ash content range n
(%) (%) (%)
Clam 94.71 0.06 90.45 97.39 238
Shell 97.48 0.02 96.06 98.23 238
Meat 13.74 0.55 2.50- 17.00 238

Overall, clams sampled in this study allocated a mean of 95.9 % (SE = 0.1) of total clam

DW biomass as shell tissue with a range of 88.4 99.7 %, n = 453. Meat biomass accounted for

the other 4.1 % of the dry weight on average (SE = 0.1) with a range of 0.3 to 11.6 %, n = 453.

The amount ofDW biomass allocated to shell tissue did not vary significantly (p > 0.05) with

shell length, between individual raceways within each nutrient addition treatment, over time in

each nutrient level or between nutrient addition treatments. Dry weight values for the meat, shell

and whole clam varied significantly (p < 0.05, n = 453) with shell length. Correlation analysis

showed a strong relationship between shell and whole clam tissue biomass and shell length

(Table 3-10).

Table 3-10. Results of the meat, shell and total clam tissue dry weight (DW) to shell length
correlation analysis. Shell and whole clam tissue biomass had a stronger
correlation to length than to meat tissue
Length Meat DW Shell DW Clam DW
Length 1 x x x
Meat DW 0.626 1 x x
Shell DW 0.966 0.599 1 x
Clam DW 0.967 0.626 0.999 1

No statistically significant differences (p > 0.05) in the length-weight regression

relationships were found between raceways, over time or seasonality in each nutrient addition

treatment using the ANOVA. However, significant differences (p > 0.05) were found in each

nutrient level for all meat, shell and whole clam tissue. Tissue DW and shell length values were









then pooled for the clams in each nutrient addition treatment, and length was removed in the

ANOVA. Regression relationships between shell length and DW are presented in Table 3-11 for

each tissue type and each nutrient addition treatment. Intercepts and slopes of the regression

lines for In (clam DW) and In (shell DW) for the high nutrient group were significantly different

(p < 0.05) than the low and medium nutrient levels. This relationship did not differ significantly

(p > 0.05) between the low and medium nutrient addition treatments; however, high nutrient

addition treatment did differ significantly (p = 0.03) from the low and medium levels.

Table 3-11. Dry weight (DW) biomass vs length regression relationships, significance
differences and variability for whole clam, shell and meat tissues from each nutrient
addition treatment. This relationship did not differ significantly between the low
and medium nutrient addition treatments The relationship for the high nutrient
addition treatment was significantly different from both the low and medium
nutrient addition treatments.
Nut. Regression Equations
Tissue level denotes significant difference (p < 0.05) r2 n
Low In (clam DW) = -3.6884 + 0.2893 (length) -0.0038 lengthh) 0.9394 243
Clam Med In (clam DW) = -3.6723 + 0.2873 (length) -0.0038 lengthh) 0.9614 161
High In (clam DW) = -4.9450* + 0.3949*(length) -0.0059*(length2) 0.9738 81
Low In (shell DW) = -3.8920 + 0.3042 (length) -0.0041 lengthh) 0.9421 243
Shell Med In (shell DW) = -3.9270 + 0.3056 (length) -0.0042 lengthh) 0.9642 161
High In (shell DW) = -5.0245* + 0.3990*(length) -0.0060*(length2) 0.9718 81
Low In (meat DW) = -4.0291 + 0.0190 (length) + 0.0021 lengthh) 0.2903 243
Meat Med In (meat DW) = -2.8659 0.0716 (length) + 0.0042 lengthh) 0.2447 161
High In (meat DW) = -6.7690* 0.2476 (length) -0.0029 lengthh) 0.8895 81

Shell length values were plotted against the actual and predicted DW values derived from

the regression equations (Table 3-11). These relationships are plotted for the whole clam, shell

and meat tissues in Figure 3-12, Figure 3-13 and Figure 3-14, respectively. The mean percentage

shell tissue calculated using the regression equations for the predicted whole clam biomass

values was similar to the tissue DW and wet weight (WW) values.











6

25

4

3
2


0
8 13 18 23 28 33

Shell length (mm)


Figure 3-12. Regression relationships for shell length vs actual and predicted (Table 3-11) whole
clam dry weight (DW) values for each nutrient addition treatment. This relationship
for the high treatment was significantly different than both low and medium nutrient
addition treatments.


7

6

5

4



2

1

0


13 18 23 28 33


Shell length (mm)


Figure 3-13. Regression relationships for shell length vs actual and predicted (Table 3-11) shell
dry weight (DW) values for each nutrient addition treatment. The relationship was
significantly different in the high nutrient addition treatment than both the low and
medium nutrient addition treatments.










0.35

m Actual m
0.30 Low Predicted
A Med Predicted
0.25 High Predicted -

0.20 m

S0.15

0.10 -

0.05 -M N

0.00 -
8 13 18 23 28 33
Shell length (mm)


Figure 3-14. Regression relationships for shell length vs actual and predicted (Table 3-11) meat
dry weight (DW) values for clams. Relationships for the high nutrient addition
treatment were significantly different than the low and medium nutrient addition
treatments.

These regression relationships can only be applied to clams in the 10.2 mm to 33.7 mm

shell length range for the high nutrient addition treatment and the 10.2 30.4 mm range for the

medium and low treatments that were used to calculate them (Table 3-8). Interpolation of the

equations (Table 3-11) for clams beyond 34 mm upper limit results in reduced or logarithmically

increased tissue dry weight estimates due to the shape of the extended regression line.

Biomass at stocking and at proceeding time intervals was estimated using length vs weight

relationships, similar to Joy and McCoy (1975). Calculations of biomass were made for each

nutrient addition treatment (Figure 3-15) using average shell length (Table 3-6). Shell length to

dry weight regression equations (Table 3-11) and live clam densities (Figure 3-10) were similar

to those found by Cataldo et al. (2001) who were not able to discern population cohorts. It was









acceptable to apply the regression equations to calculate dry weight biomass from the average

shell lengths at each interval since shell length values fell within the ranges required by each

nutrient addition treatment level (Table 3-8). Estimated population biomass decreased markedly

from stocking to the end of the study for each nutrient addition treatment.


70.000 -

60,000 -

50,000 -

40,000 -

30.000

20,000 -

10.000 -

0
05/14 02


08/1702 11 20 02 02 23 03

Date


0 1 219 0 3 019 01 03


Figure 3-15. Estimated clam dry weight (DW) biomass over time in the low medium and high
nutrient addition treatments. Increases in estimated DW biomass may not be
indicative of tissue production due to the variability in the clam density
measurements.

Reproduction and recruitment

No evidence of successful reproduction and recruitment was found in any of the nutrient

addition treatments, as indicated by declines in clam populations (Figure 3-10), the lack of

individuals less than 7 mm in shell length (Table 3-6), and no evidence of juveniles in any of the

water samples used in the phytoplankton speciation survey.









Health

Linear regression analysis yielded a correlation coefficient of 0.85 between the condition

indices based on CI(WT) and CI(VOL), indicating that they were very similar. The CI(WT) index had

a slightly lower coefficient of variation (0.53) than CI(VOL) (0.56) (n = 452). Shell length was not

a significant factor (p > 0.05) in the values of either index, so it was removed from the analysis.

Results for the ANOVA for both indices were similar, showing significantly higher

condition values for both indices (p = 0.01) in the medium nutrient level, but only at sampling

Interval 1. CI(WT) and CI(VOL) ranged from 0.302 tol3.210 and 1.435 to 15.831 per individual,

respectively (Table 3-12). Condition indices calculated for clams in the medium nutrient

addition treatment at interval 1 tended to be higher than all other nutrient addition

treatment/interval combinations. This trend was probably due to slightly higher meat dry weight

values in the medium nutrient addition treatment at interval 1. Consequently, meat DW values

were not significantly different at interval 1 compared to the other intervals within the medium

nutrient addition treatment; therefore, the CI is apparently highly susceptible to differences in

meat tissue quantity that may not be apparent using the regression analysis.


Table 3-12. Mean, standard error (SE) and range of condition indices values (CI(wT) and
CI(VOL)) calculated for the medium nutrient addition treatment at Interval 1
compared to values calculated at all other treatment/interval combination.
Values for both indices tended to be lower in all other treatment/interval
combination.
Condition index values
Population Index Mean SE Range n
Medium nutrient CI(WT) 7.758 0.238 2.798-13.1280 81
Interval 1 CI(voL) 8.931 0.262 3.406-14.9269 81
All other CI(wT) 3.601 0.080 0.302-10.033 372
combinations CI(voL) 4.581 0.131 1.435-15.831 372









Amphipod infestation

Amphipods (Hyallela azteca) began appearing in dead clam soft tissue and empty shell

valves removed from the raceways on July 26, 2002 (38 days after introduction of clams). A

sample of five live clams on the substrate surface was randomly removed for inspection on

August 9, 2002 and yielded one clam with an amphipod. The clams were immediately shucked

to reveal six additional live amphipods enclosed in the shell cavity of one clam.

A notched clam specimen was removed from the sediment surface. The clam was shucked,

treated with a 90 % ethanol solution and examined under a dissecting scope (O1x 30x) fitted

with a digital camera to reveal nine amphipods. Amphipods were identified and photographed

by Gary Warren at the Florida Fish and Wildlife Conservation Commission in Gainesville,

Florida. Amphipods were ruled out as a direct cause of the chipping due to their lack of hard

mouthparts capable of damaging clam shell material (Covich and Thorpe 2001).

In an attempt to quantify the extent of the amphipod infestation in the raceway clam

populations, collections from the following three categories of clams were made: fresh dead

clams, live clams on top of the sediment and live clams buried in the sediment. Fresh dead clams

are defined as having gaped valves with soft tissue intact. Six live clams were removed from the

sediment surface on August 9, 2002, and a sample of 10 live clams was excavated from the

substrate on August 18, 2002. All clams were collected at random from the same raceway.

Amphipods contained in fresh dead clams were counted immediately after removal from the

raceway, and any shell notching was noted. Live clams were placed in Ziploc bags, spun tight,

sealed and wrapped with a rubber band to prevent clams from opening during transport to the

laboratory for shucking and amphipod count. All clams sampled had a shell length greater than

20 mm.









The observations showed that 100 % of the fresh dead clams sampled contained live

amphipods. Amphipod abundances in these clams averaged 12 individuals per clam and ranged

from 2 to 23 individuals per clam, and 36 % of the dead shell valves sampled had notches. All

live clams taken from the substrate surface were notched, and only 50 % of these individuals

contained amphipods. Amphipod abundances within the shell cavities in these live, notched

clams ranged from 4 to 9 individuals per clam. None of the live clams found beneath the

raceway substrate surface had notches, and only one of these clams contained amphipods within

the shell cavity. This live clam, however, had the highest number of amphipods recorded per

live clam sampled, nineteen.

After the initial observations on amphipod invasion and shell chipping in raceway clams

were made, a small scale experimental system was constructed in an indoor laboratory at Alee

Academy in Eustis, FL to test the repeatability of this phenomenon in smaller-scale aquaculture

environments. A total of 9 27-L glass aquaria were stocked with between 10, 20 or 30 clams

(200 to 600 clams/m2) and 50, 75 or 100 amphipods (1075 to 2150 amphipods/m2) collected by

hand from clam raceways and transported in large aerated coolers to the laboratory in Eustis,

Florida. Lights were maintained on a 12 hour-on/12 hour-off cycle and water temperature was

regulated at 23-25 C, using submersible heaters. Sand for the aquarium substrate was obtained

from the same retailer as the Hague, FL facility. Aquaria were fed with 1 gram per week per

aquarium of dried, powdered 8pit/ "gri sp. algae obtained from the Hague raceways. No

indications of shell chipping, penetration of live clam shell cavities or high clam mortality were

recorded. The lack of clam/amphipod interaction in these small-scale systems indicates that the

Hague observations could be an isolated incident or that the negative interactions are more likely

to occur at larger scale or higher air temperatures.









Discussion

Use of the freshwater clam, Corbicula, as the primary active agent in wastewater treatment

raceway systems yielded mixed results. Significant growth and phosphorus sequestration was

observed in tagged clams, which survived through the study period. Growth in tagged clams

appeared to be substantial during warmer periods in the medium and high nutrient addition

treatment systems, with growth rates ranging from 0.043 to 0.117 mm/day (Table 3-7). These

rates are similar to those reported for other freshwater bivalves (Table 3-13).

Table 3-13. Individual shell length and biomass dry weight (DW) growth rates reported
for Corbicula and other bivalves occupying different fresh and saline
environments. Some values were not reported by the reference author and are
noted as (n/a).
Organism Shell Length Biomass Environment Type Reference
Growth Growth
(mm/day) (g DW/day)
Corbicula 0.117 0.0024 Agriculture nutrient This Study
treatment raceways
Corbicula 0.058 0.0023 Laboratory at Foe & Knight
optimum temp (1986)
Corbicula 0.180 n/a Natural river McMahon &
Williams (1986)
Corbicula 0.085 n/a Power plant Mattice &
effluent canals Wright (1986)
Corbicula n/a 0.0077 Natural estuary Fuji (1979)
japonica
Dreissena 0.080 n/a Natural lakes and McMahon &
polymorpha rivers Bogan (2001)
Dreissena 0.095 0.0150 Great Lakes Bitterman et al.
polymorpha (1994)
Elliptio 0.001 n/a Natural lakes Anthony et al.
complanata (2001)
Venerupis 0.189 0.0189 Mariculture effluent Jara-Jara et al.
pullastra treatment raceways (2000)
Tapes n/a 0.0422 Mariculture effluent Shpigel &
semidecussatus treatment ponds Fridman (1990)
Ruditapes 0.333 0.0083 Mariculture effluent Jara-Jara et al.
decussatus treatment raceways (1997)
Ruditapes 0.377 0.0730 Natural estuary Nizzoli et al.
philippinarum mariculture (2006)









Despite the positive results with tagged clams, the long-term performance of the broader

clam population in the raceways was poor, and dairy wastewater effluent could not tested due to

the high ammonia levels present in the effluent addition ponds throughout the study. The latter

problem highlights an important issue in the use of an animal-based system for wastewater

treatment: ammonia toxicity. From a broader perspective, the high rates of mortality of clams in

the raceway systems focus attention on a range of environmental and design issues that must be

dealt with in future research efforts, such as temperature, food availability, dissolved oxygen, the

effect of multiple environmental stressors, parasitism and predation, depressed reproduction in

captivity and issues with enumeration of raceway clam biomass.

Ammonia Concerns

Use of Corbicula or other freshwater bivalves in large-scale raceway-based systems as a

phosphorus treatment mechanism for dairy wastewater may be limited foremost by ammonia

toxicity concerns that could skew normal water usage and land requirements. Ammonia in the

toxic, unionized form ammonium (NH4 ), is a serious concern in the aquatic environment.

Levels of ammonium of 0.2 mM as ammonium-nitrogen (NH4 -N) negatively affect other

bivalves and finfish (Patrick et al. 1968, Epifanio and Sma 1975). Amounts of the toxic form are

directly proportional to pH and temperature and increases in ammonia-nitrogen (NH3-N) are

generally used to gauge ammonia toxicity.

Ammonia levels harmful to Corbicula and other freshwater mussels have been suggested

by Cherry et al. (2005) to be 0.11 to 0.8 mg/L total ammonia nitrogen (NH3-N). Levels of 0.041

to 0.158 mg/L NH3-N are harmful to bivalves in marine aquaculture systems and may even be

harmful at concentrations as low as 0.006 mg/L NH3-N (Harris et al. 1998). The presence of

ammonia concentrations in excess of 0.25 mg/L NH3-N in the source ponds receiving









anaerobically digested dairy effluent in this study indicate that even effluent addition rates as low

as 5 % to 10 % of source pond volume are not acceptable for production of Corbicula.

High ammonia levels in commercial aquaculture systems decimate stocks if not managed

correctly. This is especially true in recirculating systems where the lack of water exchange

necessitates removal of ammonia to achieve safe levels (Timmons et al. 2002). Keeping

ammonium within a tolerable range will be critical to using Corbicula as a wastewater treatment

mechanism (Haines 1977). Additional treatments would need to be added to the dairy

wastewater stream to manage ammonium. Coupling vegetation with Corbicula into polyculture

systems has also been proposed for nutrient uptake in agricultural waste streams (Stanley 1970,

Greer and Ziebell 1974, Mattice 1977). This type of system may provide an additional ammonia

management strategy in future applications; however, no large-scale experiments have been

developed to demonstrate the applicability of such technologies.

Anaerobically digested dairy effluents may also present additional ammonia concerns

compared to non-digested effluents, since digester systems convert nitrogen bound in solids to

ammonia by microbial degradation, which adds to the ammonia already present in the barn and

milking operation wastewaters (Wilkie et al. 2004). Potential problems with introduction of

harmful levels of ammonia were circumvented in this study by simulating algal concentration

under dairy wastewater conditions using nitrogen and phosphorous fertilizer applied at different

loading rates.

Among the raceways supplied by pond water of different nutrient concentrations, all were

subject to high rates of clam mortality over time, particularly in summer. It is possible to form

several hypotheses about the environmental factors responsible for these losses other than

ammonia, including high temperature, limited food availability, low dissolved oxygen and









parasite interactions. It is useful to explore each of these considerations in an effort to find

possible solutions and future directions for research and development.

Temperature

Temperature probably had the most impact on growth of Corbicula populations in the clam

raceways. Tagged clams recaptured alive in this study showed growth in water temperatures

between 10 and 30 C with higher growth rates occurring at temperatures above 18 C. Growth

rate in Corbicula increases with temperature as Mattice (1977) observed a maximal growth rate

of clams in power plant canals at 24 C. Other studies using lab experiments by Mattice (1977)

and Foe and Knight (1986) suggest that temperatures above 25 C cause a decrease in growth.

Buttner and Heidinger (1980) observed the highest growth rates in the temperature range of 11.2

to 24.7 C and minimal growth at 4.2 C. Conversely, exposure to temperatures exceeding 30 C

coincided with increased population mortality in this study.

McMahon and Williams (1986) suggest an upper lethal limit of 36 C for naturally

occurring populations of Corbicula. Habel (1970) reports 98 % mortality when temperature

exceeds 35 C, and Busch (1974) observed no growth and high mortality when temperatures

exceed 32 C in polyculture ponds shared with catfish. Buttner (1986) suggests that

temperatures above 30 C be avoided when culturing Corbicula, which is supported by the

results of this study since water temperatures greater than 30 C were observed through much of

the study (Figure 3-2), and losses of biomass (Figure 3-11) were consistent with exposure to

water temperature above 30 C.

Site selection based on temperature may be the key to successful application of Corbicula

to wastewater treatment. High temperatures have been implicated in limiting success of other

Corbicula-based experimental wastewater treatment systems in arid and tropical areas prone to

high temperatures such as Arizona (Greer and Ziebell 1972) and the US Virgin Islands









(Hainesl977). Pond-based experiments examining use of Corbicula for aquaculture effluent

treatment in temperate Illinois have also reported problems with high temperatures (Habel 1970,

Busch 1974, Buttner 1986). Implementation of Corbicula-based treatment systems may be

restricted to locations where water temperatures do not exceed 30 C as suggested by Greer and

Ziebell (1972) and may require geographical locations even farther north or in traditionally

colder climates than Florida or even Illinois.

Corbicula-based systems may be better suited to application in colder areas because there

are no real effective techniques for cooling water in systems of this scale or larger. Source ponds

at the Hague site were aerated at night to induce evaporative cooling, but water temperatures still

reached 30 C and above in the afternoon throughout the spring, summer and fall. Industrial-

sized water chillers are not a practical solution due to equipment and energy costs that would

exceed the value of the treatment potential. Shading of open water areas may help reduce heat

buildup, but it also lowers light availability for phytoplankton productivity. Shading of small

raceways with plywood coverings has been used by Haines (1977) to prevent nuisance

vegetation growth in Corbicula systems and greenhouse netting is also commonly used in

commercial aquaculture and agriculture operations, but the effects that covering of raceways

may have on clam populations through changes in temperature, oxygen exchange and disruption

of possible circadian rhythms are not known. Heat exchange with cold water drawn deep from

aquifers or surface water reservoirs may provide solutions, however, the energy required for

pumping and lack of specific locations with available cold water sources may prohibit

implementation.

Application of Corbicula to systems in colder climates may result in limited biomass

production at temperatures below 4.2 C (Buttner and Heidinger 1980) but cold-related









mortalities as well (McMahon and Williams 1986). Ice cover can depress dissolved oxygen

below tolerable levels in northern areas (McMahon and Bogan 2001). Temperatures in this study

did not fall into the low temperature ranges (< 5 C) identified by McMahon and Williams

(1986) as lethal; however lower clam growth rates were evident during the winter months.

Waste heat generated from power plants has been proposed by Mattice (1977) for

encouraging biomass production in Corbicula aquaculture in colder climates, however, this

resource may not be applicable to dairy operations, since locations of power plants would

determine available sites for clam raceways. Land use requirements for coupling power plants

and dairy farms may be the ultimate barrier to waste heat utilization of this type due to the

potential contamination by agriculture practices in the vicinity of surface water reservoirs needed

for power plant cooling towers. The need for transfer of wastewater over long distances in the

volumes produced by typical dairy farms (average milking herd of 359 cows = wastewater

production of 502 m3/day at Dairy Research Unit in Hague, FL (Wilkie et al. 2004), decreases

the attractiveness of this type of waste heat usage at any appreciable scale in dairy wastewater

treatment.

Food Availability

Food limitation may also have been a problem in the survival of clams in this study.

Phytoplankton availability has been implicated as a limiting factor for Corbicula biomass

production in natural systems (Foe and Knight 1986, Mattice and Wright 1986), and low particle

concentrations have been linked to poor performance in Corbicula-based wastewater treatment

systems (Haines 1977). Chlorophyll a levels remained below the 100 to 300 [tg/L range reported

by Cohen et al. (1986) to sustain growth and survival in natural systems. Source pond

phytoplankton production was probably limited to factors other than phosphorus availability.









As expected, increased in source pond water phosphorus in the high nutrient addition

treatment system corresponded with an increase in phytoplankton biomass; however,

phytoplankton biomass remained low in the medium nutrient addition treatment system (Figure

4-3). Utilization of phosphorus by phytoplankton is the first stage in this clam raceway system;

however, performance of this phase was poor since there was a large portion of dissolved

phosphorus in all of the systems indicating that phytoplankton production was probably not

limited by phosphorus. In order to identify possible limiting factors, study of nutrient dynamics

beyond nitrogen and phosphorus may be needed in large-scale systems when utilizing

phytoplankton incubation ponds for clam raceway source water.

Phosphorus and nitrogen levels in source pond water after the one-month incubation period

following fertilizer additions were similar to pre-addition levels in both fertilizer addition

treatments. Source pond total phosphorus concentrations of 0.23 and 0.46 mg P/L were expected

in the medium and high nutrient addition treatments, respectively, if the fertlilizer P would have

been retained in the water column. These total phosphorus levels were not achieved until

February 2003 after the clam raceways had been operational for at least one month. Initial

investigation of phytoplankton productivity with wastewater addition, as demonstrated in clam-

based treatment system scaling by Borges et al. (2005), should be employed to enhance large-

scale clam raceway food availability prior to clam exposure to identify and solve potential

problems.

Although Corbicula can utilize a wide variety of particles as a food (Silverman et al. 1995,

McMahon and Bogan 2001), the potential for its use in phosphorus treatment depends on the

extent that phosphorus is transformed into a particulate form for clam uptake. Phytoplankton

species management could potentially impact clam food resources since food value, and algal









phosphorus content can vary with speciation and environmental conditions (Wetzel 2001). This

may be ideal in theory, but it may not be practical at a large scale due to the susceptibility of the

outdoor systems to influx of local genetic material. Maintaining desired phytoplankton

composition in large-scale wastewater treatment through inoculation may not be effective in

governing algal speciation in outdoor, larger-scale systems as suggested by Greer and Ziebell

(1972) and Haines (1977). Management of phytoplankton nutrients in large-scale clam raceway

systems may also require additional engineering solutions such as pond liners to minimize losses

of nutrients due to the interaction with sediments and groundwater.

Dissolved Oxygen

Low dissolved oxygen levels associated with receiving waters downstream from municipal

wastewater treatment plant discharges have also been associated with high mortality and low

growth rates in Corbicula (Belanger 1991). In this study, dissolved oxygen never fell below 6.0

mg/L, which is well over the 3.0 mg/L minimum suggested by Buttner (1986) for Corbicula

under aquaculture wastewater conditions. Maintenance of adequate dissolved oxygen may be

attributed to the design of the raceway input spreader bar that was installed to help volatilize

ammonia and aerate incoming raceway water.

Multiple Stressors

Temperature alone may not have been responsible for the poor survival of clams in

raceways since other chemical and physical stressors in the environment can decrease the

threshold temperature for survival. For example, Buttner (1986) attributed Corbicula biomass

losses to temperatures up to 33 C, along with depressed oxygen levels (< 40 % saturation) at

temperatures above 25 C. Haines (1977) attributes lack of growth, high mortalities and

temperature stress, in combination with the possible presence of ammonium, at 30-32 C in

municipal wastewater conditions. Other studies have attributed no growth along with substantial









stocked biomass loses (Habel 1970, Haines 1977), to a combination of high temperatures and

low dissolved oxygen (Buttner and Heidinger 1980, Buttner 1986). Interactions of these

concerns, along with possible amphipod interactions observed in this study, may also inhibit

implementation of Corbicula-based systems in dairy wastewater treatment.

Parasites and Predation

Potential problems with amphipods that infested clam raceways at the Hague, Florida

facility were unforeseen and may have contributed to the limited adaptability of freshwater clams

to the large-scale raceway aquaculture systems. The previously undescribed phenomenon of

Corbicula infestation by the amphipod Hyalella azteca may have contributed to population

declines, decreased filtration and lack of growth. This phenomenon may have been the result of

the substrate choice and other environmental parameters unique to the raceway systems that

could be encountered in other engineered systems.

Negative impacts on Corbicula may be due to direct parasitic interactions or tissue damage

from movement and occupation within the shell cavity; or possibly its incurrence of stress from

constant valve closure stimulated by amphipod movements over the raceway substrate.

Repetitive valve closure as a result of the clam's defensive response to outside stimuli, such as

the constant activity of an amphipod infestation, may deplete energy reserves and limit its ability

to respire, expel waste products and obtain food (Gainey 1978). Shell chipping appeared to be

related directly to the ability of the amphipod to invade clam shell cavities. There was, however,

evidence that amphipod invasion of shell cavities can also occur in the absence of a chipped shell

area; therefore, amphipod invasion needed to be verified by shucking live, freshly collected

organisms.

The chipping of shell material was probably related to the amphipod infestation even

though the breech in shell tissue may not have been the only entrance mode. Chipping was









probably caused as a result of repetitive valve closure and the inability of the clams to expel sand

grains as the valves were closing. Shell margin chipping was not evident in the systems exposed

to fertilizer addition, which may have caused a difference in substrate characteristics with the

accrual of soft sediments as a result of settleable material such as phytoplankton and clam wastes

over the sand substrate. This change in sediment structure may have helped to deter shell

chipping and the phenomenon of inter-cavity invasion by amphipods, since live tissue samples

from these raceways did not contain amphipods. Clams undergoing physiological stress, such as

the mortality events in this study indicated, may be more susceptible to shell cavity invasion by

amphipods, even without the occurrence of shell chipping. This idea is supported by the low

number of live clams without evidence of shell chipping that were found containing amphipods.

Engineered systems are still in their experimental infancy and may be subject to unforeseen

parasitism and predation by organisms that are not normally pests in the natural environment as

systems achieve larger and larger scale. Negative interactions between invertebrates and

bivalves are more common in the mariculture industry, where organisms are maintained in high

densities using either closed engineered (Dame 1996, Lin et al. 2001) or open-water natural

systems (Hickman 1992, Haines et al. 1994, Dame 1996). Application of bivalve-based biofilter

technology in the freshwater environment may be susceptible to similar problems that are much

less understood at this time due to the lack of large-scale culture experience beyond the marine

environment.

Parasitic and/or amensal interactions have been described between naturally occurring and

cultured open-water marine bivalves by invertebrates such as pea crabs and other crustaceans

(Hickman 1992, Haines et al. 1994, Dame 1996, Mercado-Silva 2005). In estuarine populations

of Corbicula, attachment of barnacles has led to declines in clam health and increased mortality









(Foe and Knight 1986). Lin et al. (2001) reported mortality due to parasitic interactions between

Pyramidellid snails and the giant clam Tridacna derasa in an engineered mariculture effluent

treatment system. Oysters (Crassostrea spp.) in both natural and culture systems are susceptible

to negative effects of shell-boring polycheates (Wargo and Ford 1993, Debrosse and Allen Jr

1994). Other microscopic parasites have also have a detrimental effect on oysters, including the

endoparasite known as MSX (Wargo and Ford 1993), however, no effects from organisms of this

nature have been documented in Corbicula.

Observations and negative effects of inter-shell cavity parasites in naturally occurring

freshwater clams and mussels have been reported in a variety of systems. Parasitism of the

freshwater clam Pisidium amnicum in Finish lakes by tremetodes (Bunodera luciopercae,

Palaeorchis crassus and Phyllodistomum elongatum) has been reported by Holopainen et al.

(1997). Parasitism of freshwater unionid mussels by unionicolid mites in the United States has

been reported by Edwards and Dimock, Jr. (1988) and Fisher et al. (2000). Freshwater mussels

inhabiting Chilean salmon farms can be preyed upon by freshwater crabs and shrimp when the

valves of the mussels are open (Soto and Mena 1999), but these predators did not inhabit the

inside of the shell valves as did the amphipods in this study. Corbicula populations inhabiting

natural systems in the United States may be impacted by parasitism of soft tissue by the

oligochaete Chaetogaster limnaei, as described by Eng (1976). Corbicula stocked in fish

polyculture ponds have been subject to predation by terrestrial mammals and fish (Chen 1976,

Buttner and Heidinger 1980), but no reports of amphipods negatively affecting Corbicula

populations have been made.

This is likely the first observed report of the effects of Hyalella azteca presence as a

possibly negative aspect in Corbicula populations as well as infiltration of the mantle cavity area









by amphipods. Understanding the Corbicula and Hyallela interaction observed here will require

further study in order to accurately describe, understand and document the extent of the

interaction and the effect that it may have on the animal's health and survival. The extent that

amphipods may affect freshwater clam and mussel populations under culture conditions can be

magnified by substrate choice as well as other environmental stressors such as temperature that

can pre-stress populations prior to amphipod infestation. Potential problems with amphipods

identified here would need to be solved either by polyculture with other organisms as suggested

by Soto and Mena (1999) and Lin et al. (2001) or pesticides tolerable to clam populations. Even

with fish polyculture, invertebrate pests may still be problematic in freshwater bivalve-based

treatment systems as indicated by Soto and Mena (1999). Incidental parasites and other pests

would need to be managed in any future freshwater bivalve-based wastewater treatment system,

which is difficult since organisms such as Hyalella azteca that occur with the clam in the natural

environment may have an undescribed impact on the target organisms, especially under large-

scale monoculture conditions.

Fish/clam polyculture may also help eliminate potential problems with amphipods in clam

raceways through predation, provided that the fish do not impact clam populations as well.

Other problems from nuisance organisms encountered in the raceways, such as biofouling from

plant growth, may also be solved through polyculture of clams with other organisms such as

herbivorous snails that have been used to control biofouling by macroalgae in clam-based

mariculture systems (Lin et al. 2001).

In natural and engineered systems, the presence of larger organisms may affect survival

and growth of Corbicula; for example, predation of Corbicula in culture ponds by muskrats has

been reported by Buttner and Heidinger (1980), but no indications of such disturbances were









apparent at the Hague site. Chen (1976) reported predation of Corbicula by fish in large-scale

polyculture systems. In estuarine populations of Corbicula, attachment of barnacles has led to

declines in clam health and increased mortality (Foe and Knight 1986). Parasitism of freshwater

bivalves mussels by unionicolid mites in natural water bodies has been reported by Edwards and

Dimock, Jr. (1988) and Fisher et al. (2000), but no negative relationships between freshwater

bivalves and amphipods has been reported in the literature.

Reproductive Success

An important part of Corbicula biomass production is the clam's ability to reproduce and

recruit new individuals to the raceways. The lack of juveniles (defined as having a shell length

less than 7 mm) observed in the raceways and water samples in this study indicates that

reproduction and subsequent recruitment did not take place. The 5 mm shell length threshold

was chosen, as opposed to the 10 mm size used by Buttner (1986), since the presence of clams 7

to 10 mm were recorded during stocking and significant growth in the raceways was

questionable.

Clams were expected to repopulate the raceways readily as a result of Corbicula-specific

attributes such as high fecundity in monoecious individuals, brooding of larvae with no need for

intermediate hosts, rapid growth (Byrne et al. 2000) and their ability to recruit new individuals in

a wide variety of habitats (Sickel 1986). The lack of adapted predators has also been attributed

to recolonization success in naturally occurring populations of Corbicula (Sickel 1986). This is

advantageous for raceway-based monocultures in closed systems that can be engineered to keep

larger predators out, such as fish, raccoons and birds that may prey on smaller individuals with

softer shells (Buttner 1986). Smaller omnivorous predators such as flatworms or amphipods are

more difficult to control with mechanical methods and may prevent clam recruitment in

raceways even if successful reproduction and larvae expulsion into the water column takes place.









The lack of reproduction or recruitment in clam raceways is odd since clams readily

repopulate even obscure engineered systems such as power plant reactor plumbing (Hakenkamp

and Palmer 1999). These systems are subject to influx of larvae from natural population

reproduction where clam raceways are closed to such influx of genetic material. Clams have

also been reported to reproduce successfully at pond-scale in polyculture with fish for treatment

of aquaculture wastewater (Buttner 1986). Even with a lack of influx of natural genetic material,

high temperatures and the presence of potential clam predators (catfish), culture ponds exhibited

clam reproduction and recruitment (Buttner 1986)

As McMahon and Bogan (2001) pointed out, successful Corbicula reproduction may not be

achievable in engineered culture or experimental systems especially at smaller scales. The lack

of evidence supporting the idea that successful incidents of clam reproduction took place in the

raceways supports this theory; however, the clam raceway systems in this study were presumed

to be of large enough scale so reproduction is not inhibited. Environmental stressors such as

high temperatures, ammonia levels, amphipod infestation and food limitation that may have been

responsible for growth limitation may have also reduced the clam's reproductive capacity.

Clam Stock Assessment Issues

Another problem with comparison and implementation of Corbicula raceways at large

scale may be that population sampling can prevent an accurate assessment of phosphorus

sequestration via raceway clam biomass and settled waste products. In this study, live clam

density values estimated from seasonal spatial sampling differed greatly from clam density

estimations using actual live counts at the last sampling event of the study period. Clam biomass

losses due to mortality were not verified in this study since dead clams were not measured.

Measurement error in the raceway population spatial sampling technique probably resulted in an

over estimation of live clam density even though 3.8 % of each raceway was sampled during









each event. In the raceway design used in this study, there is an open, uniform area that

maximizes conditions for growth with far less spatial variability in biomass accrual and

hydrologic conditions than expected in natural systems. Any attempt at large-scale aquaculture

of Corbicula should pay particular attention to clam population monitoring and disturbance of

culture substrate due to the variability encountered with the methodology used in this study.

Methods for assessing clam populations at stocking also need to be addressed in order to

estimate population growth and biomass phosphorus sequestration accurately.

Clam size can have a strong negative relationship on Corbicula growth rate (Buttner and

Hiedinger 1980, Foe and Knight 1986, McMahon and Williams 1986). This relationship is

similar to the von Bertalanfy growth model used to describe differences in growth rate with age

in clam of the saltwater clam Mya arenia (Brousseau 1979). However, age is difficult to

estimate in Corbicula due to the absence of growth rings related to annual shell deposition as

found in saltwater clams, such as Mya arenia (Brousseau 1979). No conclusions could be made

on size/growth relationships since no discernable trends were found in population shell length

distributions, and low recovery of live tagged individuals did not allow for statistical analysis of

growth rate trends.

Counts of dead clams on the substrate surface may be the best indication of mortality and

actual clam biomass than the volumetric and spatial estimations used in this study. The

cumulative number of dead clams collected over time and excavated at the end of the study

accounted for up to 90 % of the clams stocked in the raceways. Stocking densities were grossly

over-estimated by the volumetric method and other techniques should be considered for future

clam stocking estimation. Raceway clam densities estimated by the spatial technique may have

also overestimated the actual live clam biomass compared to the cumulative number of dead









clams found at each sampling time. Since there was no indication of reproduction by clams, it is

assumed that the number of dead clams is the best reflection of actual live density remaining in

the raceway.

General Conclusion

Overall clam population biomass phosphorus sequestration did not occur due to high

mortality, even though significant growth rates were observed for tagged clams. Timing of

noticeable mortality events indicated that high seasonal temperature was the major factor

limiting the ability of clam raceway populations to adapt to treatment raceways. Water

temperatures in the range of 28 to 30 C and above have been implicated as the limiting factor to

success of Corbicula applications in most other waste treatment studies and in natural

populations (Greer and Ziebell 1972, Haines 1977, Mattice 1977, Buttner and Heidinger 1980,

Buttner 1986). Major population declines took place when water temperatures reached this

level, regardless of the level of nitrogen and phosphorus addition and chlorophyll a in source

pond water.

Bivalve-based treatment of dairy-derived wastewater phosphorus would require, at the very

least; scaling, study and implementation of additional treatment technologies in order to reduce

high levels of nitrogenous wastes common in dairy operations. Other than obvious issues with

implementing mechanisms untested at large scale, Corbicula-based treatment of dairy or any

other agriculture-based wastewater stream will need to manage predation and parasitism from

unforeseen organisms, as well as environmental parameters, before applications at any

appreciable scale could possibly take place.

Other environmental factors present in the raceways, including stress from the potential

problems due to infestation by amphipods (Hyalella azteca), may have affected phosphorus

removal and sequestration potential as well. Consequently, interactions reported between









Corbicula and Hyalella in this study are the first to recognize amphipods as having a potential

predatory or parasitic role in clam population dynamics. The negative interactions observed in

this study between amphipods and clams are typical of aquaculture and mariculture systems and

will need to be managed in future systems. Possible management strategies borrowed from

mariculture systems include hand removal, which is not applicable to the small size and numbers

of amphipods, top-down predation by snails or other invertebrates, polyculture with higher order

predators such as fish, or development of pesticides for this application.









CHAPTER 4
PHOSPHORUS REMOVAL AND SEQUESTRATION IN CLAM RACEWAYS

Introduction

For large-scale, clam-based raceway systems to be successful in removing and retaining

phosphorus from wastewater, they must operate within the physiological tolerances of the animal

to ensure survival, growth and reproduction. In addition to concerns over adaptability of clams

to a particular raceway environment, the system must demonstrate an ability to remove

phosphorus at rates compatible with the needs of wastewater treatment systems. Naturally

occurring populations of filter-feeding bivalves, such as Corbicula, can remove phosphorus from

the water column and sequester it into shell and meat biomass (Fuji 1979). Wastewater nutrient

management in agriculture, aquaculture, municipal and surface water sources by freshwater

bivalves, utilizing a phytoplankton intermediary, has been proposed by Stanley (1974), Haines

(1977), Mattice (1977) and Greer and Ziebell (1979).

In addition to phosphorus removal (Soto and Mena 2004), freshwater bivalves have been

used to lower turbidity (Habel 1970, Busch 1974, Haines 1977, Buttner 1986), nitrogenous waste

levels (Buttner 1986), particulate protein concentrations (Haines 1977) and seston biomass

(Mattice 1977). Actual investigations into wastewater treatment potential of bivalve-based

systems has been largely limited to the mariculture industry, where species that are commercially

desirable as a food commodity have been targeted for profitable production (Shpigel and

Fridman 1990, Jones and Iwana 1991, Shpigel and Blaylock 1991, Jakobworks et al. 1993,

Shpigel et al. 1993, Shpigel and Neori 1996, Shpigel et al. 1997, Jones and Preston 1999, Neori

et al. 2000, Jones et al. 2001). These studies focused on management of water clarity

parameters, nitrogenous compounds and removal of wastewater-generated suspended solids

including phytoplankton intermediaries, as opposed to phosphorus removal. In mariculture and









aquaculture waste systems, undigested feed and feces from a primary species, such as fish or

shrimp, has also been used as food for bivalves (Jakobworks et al. 1993, Jones et al. 2001).

Some bivalve-based treatment systems in the mariculture industry utilize multiple stages of

biofiltration including filter-feeding organisms and harvested macrophytes (seaweed) (Shpigel

and Neori 1996, Neori et al. 2000, Jones et al. 2001, Kinne et al. 2001).

Adaptation of mariculture technologies to freshwater waste streams from agriculture,

aquaculture and municipal wastewater treatment facilities requires use of freshwater organisms,

such as Corbicula, that are far less desirable commercially than other bivalves. These organisms

are potentially well-suited for wastewater treatment applications because of their high filtration,

growth and reproductive abilities that can contribute to phosphorus uptake and sequestration.

Other freshwater bivalves, including the mussel Diplodon chilensis (Soto and Mena 1999), to

reduce phosphorus in small aquarium-type systems in wastewater streams generated by

freshwater aquaculture; however, freshwater mussel species are typically sexually dimorphic and

require a fish as an intermediate host for larval development. These traits can limit the

geographical distribution and biomass production potential of mussels. Corbicula, on the other

hand, can rapidly repopulate an area as a result of a number of life history traits including high

fecundity, hermaphroditic sexuality, self-fertilization and marsupial incubation of larvae

(McMahon 1983, McMahon and Bogan 2001). Corbicula are also able to reproduce over longer

time spans due to their bivoltine (occurs twice annually) reproductive effort periodicity, as

opposed to most mussel species, which tend to be univoltine (occurs only once each year)

(McMahon 1983, McMahon and Bogan 2001). High population densities achievable by

Corbicula, exceeding 1000 individuals/m2 (McMahon 1983), can minimize the space needed for

treatment, compared to freshwater mussels.









Utilizing Corbicula for treatment of phosphorus from agricultural wastewater effluents has

been proposed by Greer and Ziebell (1972), Stanley (1974) and Mattice (1977). Similar to

bivalve-based treatment systems used in the mariculture industry (Shpigel 1993, Shpigel and

Neori 1996, Lefebvre et al. 2000 and Mazzola and Sara 2001), this approach uses a

phytoplankton intermediary to convert dissolved phosphorus into particulate matter that can be

used for food. Wastewater effluent is used to generate phytoplankton biomass, which in turn

provides clams with an unlimited food source. Sparcity of food resources has been indicated as a

limitation to Corbicula growth in natural systems (Foe and Knight 1985), however, wastewater

effluent addition decreases the likelihood that phosphorus or nitrogen becomes a limiting factor

in phytoplankton production. It is necessary to recognize that source water supply redundancy

must be designed into any engineered system to handle catastrophic events such as

phytoplankton population crashes or harmful algae blooms, which may occur under normal

operation. Even though dairy effluent addition can support algal growth (Wilkie and Mulbry

2002), phytoplankton production can be limited by a range of factors commonly found in aquatic

systems such as micronutrient limitation, seasonality of light availability and temperature.

Corbicula utilizes a wide variety of phosphorus-containing particulates. Greer and Ziebell

(1972) observed that Corbicula is not only able to remove phosphorus added to source waters

from municipal wastewater effluent through consumption of phytoplankton biomass, but can also

remove dissolved forms of phosphorus when they are converted to colloidal hydroxyl-apatite at

high pH levels. This phenomenon appears to require elevated dissolved phosphorus

concentration (5 to 15 mg/L P043-), along with a pH > 9, which can be achieved through addition

of lime and the diurnal reduction in CO2 due to phytoplankton metabolism (Greer and Ziebell

1972). Although sequestration of colloidal phosphorus has been observed only in the laboratory,









achieving removal of colloidal phosphorus from wastewater streams by Corbicula adds to the

organism's appeal as a phosphorus removal mechanism.

In the Fuji (1979) model for Corbicula in a natural population, phosphorus consumed by

clams is sequestered within the system either by sedimentation of feces and pseudofeces or

ingestion and incorporation into new clam tissue and shell. Filtration by Corbicula is

characterized by high rates of phytoplankton clearance, up to 500 mL/day/clam in

hypereutrophic lake water (Beaver et al. 1991). Like growth rate, filtration rate is primarily

dependent upon temperature and size (McMahon and Williams 1986), but can also be governed

by particle concentrations (Haines 1977). Fuji (1979) suggested that, of the total amount of

phosphorus consumed annually by filtration, 62 % is ingested, while the rest is rejected and

deposited as pseudofeces. The particulate phosphorus ingested from the water column can be

converted directly into soluble form as metabolic waste product excretions by the clams

(Lauritsen and Mozley 1989). The rest of the total amount of phosphorus ingested is exported

from the system through mortality (13 %), ejection of gametes (9 %), excretion (8 %) and

juveniles that were exported from the system by water currents and therefore not recruited into

the population (1 %) (Fuji 1979).

Phosphorus (P) content (mg P/individual) of clam biomass increases with body weight over

time in size-based age groups (Fuji 1979). Phosphorus content for clams in each age group also

varies with the time of the year due to ejection of gametes. Fuji (1979) estimate phosphorus

content of clams between 0.011 g and 3.093 g total clam dry weight to be from 0.008 mg P to

1.500 mg P per individual, yielding biomass concentrations in the range of 0.506 mg P/g to 0.800

mg P/g of total clam dry weight biomass. Using phosphorus content of clam biomass derived

from tissue analysis, Fuji (1979) estimated that clam populations can sequester 130 mg P/m2 in









one year as biomass growth (without accounting for mortality) in a population originally

containing 65 mg P/m2. Data from Fuji (1979) showed amounts of phosphorus in meat tissue

(8.293 22.667 mg P/g meat dry weight) was up to 14 times higher than in shell tissue biomass.

In this study, a large-scale raceway-based wastewater treatment system was used to assess

the phosphorus removal potential of Corbicula populations. The system was based on

phytoplankton as an intermediary to convert dissolved nutrients into particulate form (Greer and

Ziebell 1972, Stanley 1974 and Haines 1977, Mattice 1977). Source ponds that fed the raceways

in this system acted as incubation ponds for phytoplankton. The potential for using a clam-based

raceway system was assessed by the two most important mechanisms used to determine

phosphorus treatment capacity. First, direct removal of phosphorus from the overlying water

within the raceway was measured in terms of change in phosphorus concentration in the water

column. Second, the capacity of the raceways to retain and sequester phosphorus as clam

biomass was determined by changes in clam growth over time. Corbicula populations were

exposed to high, medium and low wastewater nutrient addition treatments to test the ability of

the clam systems to remove and sequester phosphorus over time in systems of different nutrient

load.

Methods

Raceway-based Treatment System

A raceway-based recirculating water treatment system constructed at the University of

Florida Dairy Research Unit in Hague, FL (See Chapter 2) was used to test the P-removal and

sequestration potential of Corbicula. Three independent pond/raceway systems were constructed

in June 2002 to compare source waters with low, medium and high levels of nutrients. Each

nutrient addition treatment group utilized two earthen source water ponds in conjunction with









three wood-framed, plastic-lined raceways. Table 4-1 shows the numerical designations for the

ponds and raceways in each nutrient addition treatment group.

Table 4-1. Source pond and raceway numerical designations for the treatment systems
at the Hague site.
Nutrient addition treatment Source Pond Raceways
Low 1,2 1 3
Medium 3,4 4- 6
High 5, 6 7 9

Source water ponds

Source ponds had an approximate area of 0.05 hectares (ha), with depths of approximately

2 m and volumes approximately 1000 m3. Ponds were enriched with a blend of nitrogen and

phosphorus fertilizer or anaerobically digested dairy farm wastewater, to simulate possible water

conditions associated with tertiary wastewater treatment. The low nutrient addition treatment

received no external nutrient addition. A 5 % and 10 % addition of anaerobically digested dairy

farm effluent was added to Pond 3 from the medium nutrient group and Pond 5 from the high

nutrient addition treatment group, respectively. For effluent physical and chemical

characteristics at the University of Florida Dairy Research Unit, see Wilkie et al. (2004).

Effluent was pumped from the digester to the source ponds and metered through a 2.54 cm (1")

turbine-type flowmeter.

Pond 4 in the medium nutrient addition treatment was dosed with 1.1 kg of triple super

phosphate (9Ca(H2PO4)2, N-P-K = 0-45-0) and 6.8 kg ammonium nitrate (NH4NO3, N-P-K = 15-

0-0) resulting in a total addition of 0.23 kg of phosphorus and 1.02 kg of nitrogen per pond in

October 2002, one month prior to clam stocking. Pond 6 in the high nutrient addition treatment

was dosed with 2.2 kg of triple super phosphate (0.46 kg P) and 13.6 kg of ammonium nitrate

(2.04 kg N) in January 2002, also one month before introduction of clam populations.









Nutrient additions were designed to enhance phytoplankton biomass, which was the

putative source of particulate nutrition for the clams. Fertilizer loading levels were targeted at

increasing phosphorus and nitrogen levels by 0.23 mg/L TP and 1.02 mg/L TN in the medium

nutrient treatment source pond and 0.46 mg/L TP and 2.00 mg/L TN in the high nutrient

treatment source pond. Fertlizer was introduced to ponds by placing it into a burlap bag that was

suspended in the water column from a floating frame (0.5 m x 0.5 m) constructed from 5.08 cm

(2") ID PVC pipe.


Neither of the ponds supplied with dairy effluent were exposed to the clam raceways due to

excessively high ammonia levels (2.0 mg/L or greater, as NH3-N), which represented a direct

threat to the health of the clams. Ammonia (NH3-N) levels in the effluent ponds and input and

output water from each operating raceway were monitored monthly using Aquacheck brand

Ammonia Nitrogen (NH3-N) Test Strips (Hach Incorporated, Colorado, USA) advertised for use

in aquaculture and aquarium applications. Levels of NH3-N in the operating raceways never

reached the 0.25 mg/L (as NH3-N) minimum value of the test kit.

Source ponds were circulated through the raceways for 10 to 20 days prior to clam

addition. Supply ponds were aerated at night throughout the study to help with mixing,

maintenance of nighttime dissolved oxygen and evaporative cooling. Floating macrophytes were

cleared by hand from each source pond several times at the start of the experiment. In addition,

six juvenile triploid grass carp ranging between 15 and 20 cm in length were stocked in each

pond in June 2002 to help reduce vegetation in ponds. No fish mortality was evident in ponds

not exposed to dairy wastewater effluent. In contrast, ponds treated with wastewater all

experienced 100 % fish mortality in less than one week after effluent addition. Fish mortality

events corresponded with high ammonia levels (> 0.25 mg/L as NH3-N).









Raceways

Raceways were 1 m in width by 7.4 m in length with an available substrate surface area of

7.2 m2, after subtracting standpipe area. Water depths were maintained at 0.2 m, yielding a

raceway water capacity of 1.4 m3 each. Source water inflow to the raceways was maintained at

227 liters per minute (LPM) (60 GPM) during normal operating conditions. The flow rates

yielded retention times of approximately 9.5 minutes, with a linear velocity of 1.17 m/min.

Turnover time for each supply water pond was approximately 24 hours. Flow rates were

calculated assuming near laminar flow through the raceway structure. Raceways were filled to

0.2 m depth with a coarse grade SiO2 filtration sand (0.6-1.0 mm particle size), that was

purchased from Feldspar Incorporated in Edgar, Florida.

Nuisance aquatic plants such as Chara sp. and filamentous algae (mainly Spirogyra sp.)

growing on the raceway substrate and liners were removed by hand from each raceway at least

weekly to reduce fouling. The control of nuisance plants, especially Chara, on the substrate

made accurate estimation of clam biodeposit sedimentation impossible. This is due the constant

resuspension of sediment deposits associated with disturbance from removal of the vegetation

that was often rooted below the sediment surface. Filamentous algae tended to utilize the sides

of the raceways where PVC liners were submersed.

Water quality monitoring

Source water ponds and raceways were monitored weekly in the morning and evening for

temperature, dissolved oxygen, pH, chlorophyll a, total nitrogen and total phosphorus.

Measurements in the ponds were taken at the end of sampling piers near the intake pipe leading

to the raceways. Monitoring was carried out monthly after 75 % or more of the clams were

assumed dead. Temperature and dissolved oxygen was measured using a YSI model DO550, and

pH was measured using a Fisher model AP63 meter by submerging the probe tips to mid-water









column depth at the raceway input and output. Water temperature, dissolved oxygen and pH

readings were compared using a paired t-test (Microsoft Excel) at the raceway input and

output, as well as between source ponds in the nutrient addition treatments.

Water samples were collected from the sampling piers in each pond for nutrients using a

pole sampler designed especially for this experiment. The sampler used a plunger-type

mechanism to collect a 1 L water sample from in front of the intake pipe. When the unit was

lowered to the desired depth, the plunger was actuated by the operator via a spring-loaded handle

at the opposite end of the pole. After the sample was collected, the plunger was released, sealing

a 1 L plastic bottle (Nalgene Incorporated, USA) bottle and raised for retrieval. The sample

bottle was unscrewed from the sampler and capped for transport to the laboratory.

Water samples collected from the source ponds were analyzed at the laboratory for

phosphorus, nitrogen and phytoplankton biomass in terms of chlorophyll a. Total phosphorus

(TP) and total dissolved phosphorus (TDP) were determined using the potassium persulfate

digestion method (APHA 1998) with a Hitachi (Japan) spectrophotometer. TDP determination

involved pre-filtering through a 0.7 ptm glass fiber filter. Total nitrogen was determined using

potassium persulfate digestion method (APHA 1998) with colorimeteric analysis performed

using a Bran-Luebbe (Germany) auto analyzer. Phytoplankton biomass was estimated using

chlorophyll a (chl a). Samples were measured by filtering 250 mL of water onto a 0.7 ptm glass

fiber filter, followed by an ethanol extraction (Sartory and Grobbelaar 1984) and

spectrophotometric determination (APHA 1998) using a Hitachi (Japan) spectrophotometer.

Microscope observations of phytoplankton species composition were obtained periodically to

describe dominant organisms with help from Mary Cichra at the University of Florida Fisheries









and Aquatic Sciences Department. Data obtained from these phytoplankton species observations

were not assessed quantitatively.

Raceway clam populations

Clams for stocking the raceways were obtained from populations in three different natural

water bodies under permit number FNC-04-022 issued by the Florida Fish and Wildlife

Commission. Clams for the low nutrient group (Raceways 1-3) were collected in June 2002

from a 0.5 km stretch of the Santa Fe River near the State Road 49-bridge in Gilchrist County,

Florida (2954.2' North, 82 52.0' West). By November 2002, the rising water level of the river

made further clam excavation impossible. Therefore, animals for Raceways 4-9 were collected

from lakes located in Lake County, Florida that had accessible populations of clams. Clams for

the medium nutrient group (Raceways 4-6) were collected from the southwest shore of Lake

George (2912.2' North, 81o35.7' West) in November 2002. Clams for the high nutrient group

(Raceways 7-9) were obtained from the west shore of Lake Dalhousie (2854.0' North, 81o36.8'

West) in February 2003.

All three collection sites had coarse sand sediments similar to the substrate used in the

raceways. Clams were excavated by shoveling bottom material into weighted baskets made from

0.635 cm (1/4") plastic mesh. Clams were also excavated by hand using trowels or a commercial

clam rake modified for the small size of the Corbicula by affixing 1" plastic mesh on the inside

of the collection basket. Periodic excavation of bottom sediment using a 0.25 m2 PVC sampling

quadrat was used to determine population densities for the clams in their natural habitat.

Densities ranged from 48 and 864 clams/m2 with a mean of 272 clams/m2 standardd error (SE) =

23, n = 47 observations) for all locations combined.

After excavation, clams were enumerated and divided into mesh bags. The clams were

then placed into coolers packed with wet newspapers and kept out of direct sunlight to help









minimize heat stress and dessication. They were transported directly to the aquaculture facility

in Hague, FL and scattered evenly throughout each raceway. Stocking of each raceway took up

to 15 days involving 2 to 6 people working per day. Raceways 1-3 were stocked from June 17

to 28, 2002, Raceways 4-6 from November 4 to 14, 2002, and Raceways 7-9 from February 11

to 26, 2003. Ten mesh bags of clams were added to each raceway. Each mesh bag contained

approximately 1,000 clams. This method was chosen as opposed to counting each individual or

bulk weighing in order to minimize handling stress, time and equipment needed to enumerate the

large numbers of clams needed for this study.

Clam populations in the raceways were monitored for density, size and biomass at

approximately 3-month intervals following stocking, as described in Chapter 3. A tag and

recapture study was also employed to verify clam growth (Chapter 3). Live clam density, shell

size and growth data from clam population surveys and tag and recapture data were used to

assess clam biomass, phosphorus removal and changes in clam phosphorus sequestration.

P Removal From Source Water By Clam Raceways

The ability of the clam raceway system to remove phosphorus and P-containing material

from overlying source water was estimated from three components: phosphorus removal from

water within a single raceway pass (through-flow): phosphorus removal from water recirculated

within the raceways and phosphorus removal from water in the source ponds through

sequestration of phosphorus in clam biomass over time.

Raceway through-flow trials

P removal from water using raceway through-flow measurements was only assessed in the

low nutrient raceways beginning after stocking in July 2002. Raceway flows were maintained at

227 LPM (60 GPM) from the time of stocking until August 2002, when input flow rates were

changed to increase raceway water retention time (Table 4-2). The low flow condition of 151









LPM (40 GPM) and a higher flow condition of 227 LPM yielded linear water velocities of 0.78

m/min and 1.17 m/min and water retention times in the raceway of 9.5 minutes and 6.3 minutes,

respectively.

Table 4-2. Raceway source water input flow rates for the period of July 1 to August 24,
2002. Flow rates were lowered in August to increase water retention time
for raceway water phosphorus removal trials using through-flow conditions
Raceway Raceway Inflow Rate (LPM)
# 7/1-7/31 8/5-8/8 8/12-8/15 8/19-8/22
1 227 227 227 151
2 227 227 189 151
3 227 227 151 151

Samples were taken from the input and output of each raceway to estimate the changes in

TP concentrations resulting from each pass of water over the clam population in the raceways.

The values were used to estimate instantaneous phosphorus removal rates from each raceway.

Environmental parameters including water temperature, DO and pH were also sampled at the

raceway input and output at 6:00, 12:00 and 18:00 hours.

The initial experiment was designed to provide baseline data on amount and variability of

phosphorus removal from raceways on a short-term basis. The sample technique involved taking

a water sample from the raceway input and output at each sample time of day. In this phase,

samples were captured from falling water under the input water distribution bar and from within

the output standpipe of each raceway. One liter of sample water was collected at each input and

output using plastic bottles (Nalgene Incorporated, USA). Samples were taken from each

raceway daily at 6:00, 12:00 and 18:00 for five, 3-day periods from July 1-31, 2002 (high flow)

and three, 3-day periods from August 5-24, 2002 (variable flow).

The differences between TP and TDP concentrations at the input and output end of the

raceways were converted to area-specific removal or addition values by multiplying by 200

L/m2. Removal rates were then calculated using the retention time estimated for each flow rate.









Negative values indicate an uptake of P, while positive values indicate a phosphorus addition

from the raceways to the water column. Rates larger than 120 mg P /m2/hr or smaller than -120

mg P/m2/hr were considered outliers and removed from the analysis to normalize the data.

The July and August through-flow TP removal data were analyzed using an ANOVA (SAS

PROC MIXED procedure, SAS Institute@, Cary, NC). The model was a randomized complete

block design (Cox 1996) with raceway as the block, where time of day and sampling period (the

5 different 3-day trials) were the factors with covariates inflow DO and water temperature.

Another ANOVA (SAS PROC MIXED procedure, SAS Institute@, Cary, NC) was performed on

the August data in order to determine the significance of flow rate. This analysis consisted of a

randomized complete block design with three factors, flow rate, time of day and raceway

number, and no covariates.

Observations from Raceway 2 from August 12-15 were excluded from the analysis due to a

low sample size, since it was the only instance where input flow rate was not 151 or 227 LPM.

Another ANOVA (PROC MIXED, SAS Institute@, Cary, NC) was performed using a

randomized complete block design with month as the block to determine if the P-removal values

were significantly different for each month. A pair-wise comparison of the means was

performed using Tukey's method to control the experiment-wise error rate in all ANOVAs.

Uneaqual variances for each time period were assumed in the ANOVA due to the significantly

low variability (x2, p = 0.033) in the 6:00 samples for the month of July. Sample variability was

assumed equal over both flow conditions for the flow rate comparison analysis (x2, p = 0.27);

however, unequal variance was assumed in ANOVA models due to significant differences in

variability between each month (x2, p < 0.001) and sampling technique (x2, p = 0.002).









In another set of phosphorus removal tests, auto-samplers were employed to increase the

sampling frequency over each 24-hour test period. This methodology allowed for a larger

sample size spread over the entire day and night. A pair of Teledyne-Isco, Inc. auto-samplers

was used to collect water samples from the input and output of raceways 1 through 3, in August

2002. Samplers were programmed to capture water at 2-hour time intervals over a 24-hour

period.

Samples were collected at two-hour time intervals beginning at 6:00 and ending at 6:00 the

following day, over two, four-day time periods (August 5-8 and August 21-24). Each of the

three raceways was only sampled for one 24-hour interval during each 4-day time period.

During the first sampling interval, raceway flows remained at 227 LPM; flows were changed to

151 LPM for the second interval. DO, water temp, air temp and pH were sampled at 6:00, 12:00

and 18:00.

An ANOVA (PROC MIXED, SAS Institute@, Cary, NC) was performed on the August

auto-sampler data using a randomized complete block design with two factors, flow rate and

time of day, and raceway number as the block. Pair wise comparisons of means were performed

using Tukey's method to control the experiment-wise error rate.

Phosphorus removal rates obtained from August samples and auto-sampler techniques were

compared to determine any differences in the two sampling methods. An ANOVA (PROC

MIXED, SAS Institute@, Cary, NC) was performed on the 6:00, 12:00 and 18:00 data using a

randomized complete block design with two factors, flow rate and time of day and sampling

method as the block. A pair-wise comparison of the means was performed using Tukey's method

to control the experiment-wise error rate. Only data collected at 6:00, 12:00 and 18:00 sampling









periods were used in the analysis, since there were far more 24-hour values. Samples taken at

189 LPM flow rate were also withheld from the comparisons.

Raceway water recirculation

To provide additional information on phosphorus uptake rates, a second method was

applied to extend the probability of exposure to the clams. This involved recirculation of water

within the raceway to extend the total exposure period. Phosphorus removal was measured in

terms of total phosphorus (TP), total dissolved phosphorus (TDP) and chlorophyll a (chl a).

A submersible pump placed within a plastic collar was used to simulate standpipe water

flow characteristics. This unit was placed just before the output standpipes, and water was piped

to the input water distribution bar at about 189 LPM within the normal operation range of 151 to

227 LPM. Raceway recirculation was started 10 minutes prior to water sampling. Periphyton

and dead clams were removed prior to recirculation. Water samples for nutrient analysis were

taken hourly at the raceway input using an ISCO auto sampler. The auto sampler intake line was

submerged approximately 3 cm from the water surface. Water temperature, DO and pH

measurements were taken at the raceway input at the beginning and end of each recirculation

time period.

Raceways 1-3 and 7-9 were utilized in this experiment. Raceways 4-6 were excluded from

the study due to high clam mortality. Clam populations were stable in raceways 1-3 and 7-9

although 7-9 had higher densities of clams. Raceway recirculation tests were conducted during

two separate experimental periods, April 11, 2003 to June 10, 2003 and June 20, 2003 to June

27, 2003. Tests were carried out for 6-hour intervals in the morning from 6:00 to 12:00 hours

and again in the evening from 18:00 to 24:00 hours.

To investigate the impact that particle deposition may have on phosphorus removal, a

control group was set up by covering the bottom of the raceways with weighted plastic covers to









eliminate interaction between the clams and the overlying water column. The control tests were

only run for 3 hours so clams would not be impacted.

P removal potential was estimated from regression relationships using TP, TDP and chl a

values in normally operating and covered raceways. Slope values formed by regression lines

were used to calculate removal rates according to the following formulae:

* TP and TDP:
Regression slope (mg/L/hr) 200 (L/m2) = removal rate (mg/m2/hr)

* Chl a:
Regression slope (mg/m3/hr) 0.20 (m3/m2) = removal rate (mg/m2/hr)

Six-hour recirculation TP, TDP and chl a removal values from the uncovered raceways

were used to perform an ANOVA (PROC MIXED, SAS Institute@, Cary, NC) using a

randomized complete block design, with slope of the line formed by TP, TDP and chl a values

over each 6-hour time period as the block, and month (April/May) and nutrient enrichment

(high/low) as the factors to determine if removal values were greater than zero. Slopes from

each am/pm time period were pooled since raceways within each treatment were not factored

into the ANOVA, due to the lack of significant differences (p > 0.05) found between am and pm

trials or raceways within each treatment. A pair-wise comparison of the means was performed

using Tukey's method to control experiment-wise error rate.

TP, TDP and chl a values from the three-hour recirculation experiments in the covered

raceways were used to perform a similar ANOVA (PROC MIXED, SAS Institute@, Cary, NC)

using a randomized complete block design with slope of the line formed by TP measurements

over each 3-hour time period as the block and nutrient enrichment (high/low) as the factors.

Slope values from all raceways in each triplicate for each sample interval were pooled for each









nutrient addition treatment as in the uncovered raceways. A pair-wise comparison of means was

also performed on the covered data using Tukey's method to control experiment-wise error rate.

An ANCOVA for repeated measures (PROC REPEATED, SAS Institute@, Cary, NC) with

a covariance structure AR-1 model was then utilized to determine any covariance between the

TP, TDP and chl a slopes for the low and high nutrient levels. This procedure was chosen due to

differences in variance between the nutrient addition treatments and lack of a significant

difference between the am and pm time periods in each nutrient addition treatment from the

uncovered analysis found in the ANOVA; therefore, time can be treated as a continuous variable.

Sequestration of phosphorus by clams in treatment raceways

Another method used to estimate phosphorus removal from the treatment water was

determination of increases in clam biomass over time, in combination with analysis of the

phosphorus content of clam tissue. Clams were randomly selected from raceways (as described

in Chapter 3) for the dry weight (DW) and ash-free dry weight (AFDW) analysis. Tests were

performed using the low and medium nutrient addition treatment raceways in June and

November 2002. A total of 26 clams were sampled per raceway, yielding a sample size of 79

clams per nutrient addition treatment group.

Shell length, wet weight, meat and shell DW and AFDW values were measured for each

clam using methods described in Chapter 3. Phosphorus content of the meat and shell tissues in

the selected clams was determined using an ignition and hydrochloric digestion method

(Andersen 1976). In this method, tissue material was dried, finely ground and ignited at 550 C

until a constant weight is achieved. The ash was then dissolved by adding IN HCI to achieve a

pH < 2, heated and diluted with DI water to neutralize the solution prior to phosphorus

determination using inductively-coupled plasma spectrometer (Copar and Yess 1996).









Meat tissue samples were prepared by dissolving the ashed material in 1 mL of IN HC1,

heating and diluting to 50 mL. For shell material, up to 1 g DW of shell ash material was taken

from each clam, dissolved in 3 mL of IN HC1, heated and diluted to 50 mL. Phosphorus

analysis was carried out on the sample solutions using an inductively coupled plasma

spectrometer and was performed by the Analytical Research Laboratory of the Soil Science

Department at the University of Florida. Phosphorus concentrations in the sample solutions were

then converted to mg P per g of DW for the meat, shell and total clam tissues using the ash DW

and tissue DW values obtained prior to digestion with the following formulae:

Meat = [P] in sample solution x _50 mL dilution volume_ x sampled meat ash DW
1000 sampled meat ash DW

Shell = [P] in sample solution x 50 mL dilution volume x crushed shell ash DW
1000 sampled shell ash DW crushed shell DW

Clam = (meat mg P/g meat DW x meat DW) + (shell mg P/g shell DW x shell DW)
(total meat DW + total shell DW)

Meat, shell and clam phosphorus concentrations were then compared to shell length values

using two different regression analyses (SAS PROC MIXED, SAS PROC GLM), where meat,

shell and clam phosphorus concentrations were the response variables with length and length2 as

factors to obtain the best-fit regression line. A polynomial regression was chosen since both

length and length2 factors were significant (p < 0.05). Both analyses yielded no significant (p >

0.05) relationships between any of the tissue phosphorus concentrations and shell length. An

ANOVA (SAS PROC MIXED procedure, SAS Institute@, Cary, NC) was then performed using

the meat, shell and clam tissue phosphorus concentrations as response variables and nutrient

addition treatment as the factor with covariates, 3-month time interval, individual raceways

within each nutrient level and season in each analysis. The least squared means (LSMEANS)

procedure was applied to meat, shell and clam tissue phosphorus concentrations (Microsoft









Excel) for pair-wise comparisons. Pair wise comparisons of the means were performed using

Tukey's method to control the experiment-wise error rate.

Data from individual raceways within each nutrient addition treatment were pooled in the

analysis since no blocking effect was found. No relationship between meat, shell and clam

phosphorus concentrations (mg P per g dry weight) and shell length could be established using

the regression analysis; therefore, shell length as a factor was also removed from the ANOVA.

A group of 12 outliers, defined as phosphorus concentration values higher than the range of 3

standard deviations away from the mean, were removed from the analysis.

Sequestration of phosphorus by raceway clam populations was estimated using population

DW biomass estimates for each nutrient addition treatment regime both at stocking and each

sampling interval. Estimated clam biomass at each time interval was multiplied by the average

phosphorus concentration of the clams, yielding the amount of phosphorus allocated in the total

raceway clam biomass. Clam biomass phosphorus values were applied to raceway population

and tagged clam results to assess removal of phosphorus by clams.

Results

Raceway Environmental Conditions

Air temperatures at the racewaty site at the Dairy Research Unit in Hague, FL ranged from

2 to 41 C (Figure 3-2). Recorded air temperatures varied diurnally as much as 15 C. Raceway

water temperatures ranged from 10.1 to 32.6 C (Figure 3-3) and displayed a similar seasonal

pattern as air temperature. However, water temperatures only differed diurnally by a maximum

of 2 C. Water temperatures reached 30 C or greater just after the beginning of the experiment

in July 2002 through October 2002 and from May through the end of the study in August 2003.

Raceway dissolved oxygen (DO) averaged 8.79 mg/L (SE = 0.07) and ranged from 6.10 to

12.03 mg/L (Figure 3-4). DO was higher in the afternoons by up to 4.26 mg/L with greater









diurnal differences in the warmer months. Higher DO values were observed during the

November 2002 to April 2002 period corresponding to lower air and water temperatures.

Raceway pH averaged 7.76 (SE = 0.02) and ranged from 6.87 to 8.81. Diurnal fluctuations

in pH ranged from -0.76 to 0.89. Raceway pH was significantly higher (p > 0.05) in the low and

high nutrient addition treatments (Figure 3-5); however, water temperature and DO values did

not differ significantly between the low, medium and high nutrient addition treatments. No

significant differences (p > 0.05) were detected at the input and output of each raceway or

between raceways in each nutrient addition treatment for the water temperature, dissolved

oxygen and pH parameters.

Increases in phosphorus and nitrogen levels in the source ponds did not correspond to

fertilizer addition or clam mortality events. Source pond total phosphorus (TP) ranged between

0.061 and 0.211 mg/L in the low nutrient addition treatment, 0.047 and 0.471 mg/L in the

medium nutrient addition treatment and 0.043 and 0.386 mg/L in the high nutrient addition

treatment (Figure 3-6). Source pond total dissolved phosphorus (TDP) ranged from 0.002 to

0.091 mg/L in the low nutrient addition treatment, 0.007 to 0.223 mg/L in the medium nutrient

addition treatment and 0.028 to 0.300 mg/L in the high nutrient addition treatment (Figure 3-7).

A major portion of the source pond TP was made up of the dissolved form as TDP in all of the

nutrient addition treatments as TDP followed a similar pattern as TP with sharp increases in the

spring 2003.

Source pond total nitrogen (TN) ranged from 0.177 to 9.034 utg/L in the low nutrient

addition treatment, 0.328 to 12.069 [tg/L in the medium nutrient addition treatment and 1.066 to

2.786 utg/L in the high nutrient addition treatment (Figure 3-8). TN values fluctuated in the low









and medium nutrient addition treatments and only slightly increased in the high nutrient addition

treatment over the experimental period.

Chlorophyll a (chl a) ranged from 3.218 to 27.511 mg/m3 in the low nutrient addition

treatment, 4.505 to 26.397 mg/m3 in the medium and 19.789 to 147.299 mg/m3 in the high

(Figure 3-9). Chl a in all treatments displayed an increase after February of 2003 with peaks

from April to August of 2003. Phytoplankton communities in the source ponds were dominated

by diatoms in the low nutrient addition treatment, diatoms and cyanophytes in the medium and

chlorophytes in the high. Diatoms were present in all ponds throughout the study period.

Raceway clam populations

Raceway live clam biomass, based on population shell length and survivorship data

decreased over time in all the nutrient addition treatment raceways (Figure 3-13). Populations

were subject to large mortality events during experimental periods (Figure 3-10), as described in

Chapter 3.

Phosphorus Uptake In Clam Raceways

Raceway through-flow input/output measurements

Temporal patterns in low nutrient addition raceway TP concentrations during the through-

flow trials from July 1-August 21, 2002 (Figure 4-1), were similar to patterns of TP

concentrations in the source ponds (Figure 3-6). Raceway input TP values varied up to 0.07

mg/L per day with no apparent diurnal trends. Input TP was slightly higher in August 2002 than

in July 2002, but did not differ significantly (p < 0.05) between the three triplicate raceways.

The results of the ANOVA showed that neither phosphorus removal (negative input/output

delta TP) nor addition (positive input/output delta) were significantly different than zero (p >

0.05) for the different times of day, flow rates, months or sampling techniques. By contrast, time

of day did have a significant effect (p = 0.02) on the magnitude of the rate of change in TP









concentrations in July trials. The ANOVA showed that the 6:00 sampling time had an estimated

phosphorus removal significantly (p = 0.001) higher than the 12:00 and 18:00 hours of day;

however, this was probably due to the lower variability for the 6:00 sampling time.

0.25 -

00 i
2 *1i
,ti : "I !:'
0.10 l ..


0.05
0.05 -.. ..


06/23/02 07/08/02 07/23/02 08/07/02 08/22/02
Date

Figure 4-1. Raceway input total phosphorus (TP) in the low nutrient addition treatment during
the through-flow trials for July and August of 2002.


Time of day did not significantly affect changes in TP between the input to the output

regardless of flow rate, month, input TP or sampling technique. Input temperature, DO, and

sampling date also had no effect (p > 0.05) on the change in TP in any of the trials. Change in

TP observations did not significantly differ (p < 0.05) between raceways or between grab

sampling and ISCO auto-sampler techniques. Due to the lack of significant differences in

experimental variables, changes in TP (ATP) were pooled for all through-flow trials.

Raceway ATP values ranged between -0.585 and 0.119 mg/L, yielding a removal rate

range of- 116.206 mg P/m2/hr to an addition rate of 99.035 mg P/m2/hr (Figure 4-2). The 25th

percentile for removal had ATP values ranging between 116.206 and -11.427 mg P/m2/hr (n =

255). ATP values appeared to be normally distributed in all trials with a slight kurtosis;

however, the ATP frequency data is not kurtotic enough (Figure 4-2) to be considered









unreasonable for assuming normality for the ANOVA analyses. The results of these trials

determined that a raceway water retention time longer than 10 minutes is needed to evaluate

significant removal of TP by the clam raceways.

60
50 -

40
0 i

20 -
10 -
0 7-,
-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120
ATP (mg P/m2/hr)

Figure 4-2. Frequency distribution of changes in total phosphorus (TP) from the input to the
output in the low nutrient addition treatment raceways from July through August
2002.

Raceway recirculation measurements

Raceway input TP values at time 0 of each recirculation trials decreased significantly (p <

0.001) from April to May 2003 in the both high and low nutrient addition treatments (Table 4-4).

Input TP was significantly higher (p < 0.001) in the high nutrient addition treatment compared to

the low (Table 4-4). Initial TP values measured at time 0 of the recirculation trials (Table 4-4)

were higher than source pond TP values (Figure 4-1), in both the low and high nutrient addition

systems.

Table 4-4. Monthly mean input total phosphorus (TP) concentrations in raceways and
standard error (SE), at time 0 in the recirculation trials for each nutrient
addition treatment pond groups.
Month Nutrient Time 0 mean SE n
treatment TP (mg/L) (mg/L)
April High 1.229 0.024 6
April Low 0.853 0.044 6
May High 0.245 0.013 6
May Low 0.067 0.030 6










Regardless of TP level at time zero, slopes of the regression lines calculated for TP values

calculated for the April-May trials did not indicate any significant decrease or addition of

phosphorus by the raceways. The ANOVA analysis showed that none of the mean TP slopes

calculated for the six-hour recirculation period were significantly different from zero in any of

the treatments during the two-month period. TP slopes calculated for the covered raceways did

not differ significantly (p > 0.05) from zero as well.

Raceway TDP values at time 0 of each recirculation trials decreased significantly (p < 0.01)

from April to May 2003 in both the high and low nutrient addition treatment pond groups (Table

4-5). Input TDP was significantly higher (p < 0.01) in the high nutrient addition treatment pond

group compared to the low (Table 4-5). Initial TDP values measured at time 0 of the

recirculation trials (Table 4-5) were higher than source pond TDP values (Figure 4-2), in both the

low and high nutrient addition systems.

Table 4-5. Raceway system monthly mean input total dissolved phosphorus (TDP) at time
zero in the recirculation trials for each nutrient addition treatment. TDP at time
zero decreased in low and high nutrient addition treatments from April to May
2003.
Month Nutrient TDP mean SE n
treatment (mg/L) (mg/L)
April High 1.119 0.024 6
April Low 0.631 0.014 6
May High 0.044 0.003 6
May Low 0.045 0.003 6

The analysis showed a significant difference (p = 0.05) in TDP over time for recirculation

trials. Significant TDP decrease was observed in both the high and low nutrient addition

treatment systems during April 2003, while TDP increased during the 6-hour recirculation trials

conducted in May 2003 (Table 4-6).









Table 4-6. Total dissolved phosphorus (TDP) removal rates calculated from TDP slopes
in the recirculation trials for each nutrient addition treatment system during
April and May 2003. TDP removal was observed during April but not in
May for both treatments.
Month Nutrient TDP SE TDP TDP n
treatment intercept (mg/L) slope removal
(mg/L) (mg/L/hr) (mg/m2/hr)
April High 1.227 0.146 -0.172 -22.566 6
April Low 0.648 0.066 -0.001 -0.131 6
May High 0.033 0.009 0.002 0.262 6
May Low 0.044 0.009 0.0001 0.123 6

Slopes of the regression lines formed by TDP values were not significantly different from

zero (p > 0.05) in any of the covered raceway trials in the low and high nutrient addition

treatments. TDP values at time 0 of the covered raceway trials were significantly (p < 0.01)

higher in the high nutrient system compared to the low (Table 4-7); however, TDP values at time

zero in the covered raceways (Table 4-7) were much higher than source pond TDP values

(Figure 4-3).

Table 4-7. Raceway total dissolved phosphorus (TDP) values at time 0 for covered
raceways in the low and high nutrient addition treatments. Raceway TDP at
time 0 was much higher than source pond TDP values in Figure 4-3.

Month Nutrient Mean TDP at SE n
treatment Time 0 (mg/L) (mg/L)

April High 0.615 0.008 3
May Low 0.554 0.008 3

Raceway chl a values at time 0 of each recirculation trials decreased significantly (p <

0.01) from April to May of 2003 in the low nutrient addition treatment pond group but not the

high (Table 4-8). In April, chl a in the low nutrient addition treatment pond group was

significantly (p < 0.05) higher than the high nutrient addition treatment group (Table 4-8). Input

chl a was significantly higher (p < 0.01) in the high nutrient addition treatment compared to the









low in May. Chl a values in the low nutrient addition treatment pond raceways at time zero

during the April 2003 recirculation trials (Table 4-8) were higher than the source pond chl a

values during the same month (Figure 4-8). Chl a values in the high nutrient addition treatment

raceways at time 0 (Table 4-8) were lower than source pond values in both April and May 2003

(Figure 4-3).

Table 4-8. Raceway chlorophyll a (chl a) values at time 0 for raceways in the low and
high nutrient addition treatments during April and May 2003. Chl a at time 0
was much higher in the low nutrient raceways during April.

Month Nutrient Chl a mean SE n
treatment (mg/m3) (mg/m3)
April High 76.682 2.440 6
April Low 112.452 2.785 6
May High 72.050 5.075 6
May Low 19.598 1.095 6

The ANOVA performed on the 6-hour recirculation trial data showed a significant (p =

0.02) reduction in chl a over time, and there were no significant interactions of time with any

other variabe. This means that the chl a slopes were the same in each experimental treatment

groups even though chl a values at time zero were significantly (p < 0.01) different between

nutrient addition treatments during each month. A mean chl a removal rate of -0.190 mg/m2/hr

was estimated from the mean slope (-1.450 mg/m3/hr) formed for chl a values over the 6-hour

recirculation time for nutrient addition treatments and months by the ANOVA.

Chl a values in the covered raceway trials involving the low nutrient group were much

higher at time zero (Table 4-9) than in the source ponds (Figure 4-3). Conversely, source pond

chl a values (Figure 4-3) were higher than values at time zero (Table 4-9) in the high nutrient

addition treatment. No chl a slopes were significantly different from zero (p < 0.05) in any of









the covered raceway treatments; therefore, it is assumed that removal of chl a in the raceways

was a result of exposure to clam populations.

Table 4-9. Raceway chlorophyll a (chl a) values at time 0 for covered raceways in the
low and high nutrient addition treatments. Chl a at time 0 was much higher
in the low nutrient raceways.
Month Nutrient Chl a mean SE n
treatment (mg/m3) (mg/m3)
April High 54.291 2.831 3
May Low 108.280 2.831 3

Significant reductions in raceway water chl a were detected after 6 hours of water

recirculation; however, these removal values are may not represent removal under normal

operation. TDP removal was only significant in the high treatment over the April 2003 period

and was estimated at nearly 22 mg P/m2/hr. Otherwise, TDP removal rates were no different

than 0, indicating that this high mean removal rate based on 3 trials may have been coincidental.

A chlorophyll a removal rate of 0.190 mg chl a /m2/hr was estimated from recirculation trials in

both low and high nutrient addition treatments since no difference in chl a removal rates could be

found between them. Some difference in removal rates was expected since both live clam

biomass and initial chl a were considerably higher in the high nutrient addition treatment

compared to the low during the recirculation trials.

Sequestration Of Phosphorus By Clams In Treatment Raceways

Shell length and DW biomass characteristics of the clam population sampled for this

analysis are shown in Table 4-10. Ash content and AFDW values for the sampled clam

population are given in Table 4-11. Shell contained the majority of the total clam ash indicating

a larger inorganic content as opposed to meat tissue.









Table 4-10. Mean shell length, clam wet weight (WW), meat and shell tissue dry weights
(DW) and condition index (CI) values for the sample population of clams used to
determine clam biomass phosphorus content. Much of the clam DW biomass
was contained in shell tissue.
Parameter Shell Clam Meat Shell Clam CI(wt) CI(vol)
length WW DW DW DW
(mm) (g) (g) (g) (g)
Mean 20.5 3.056 0.064 1.887 1.951 3.42 4.64
SE 0.2 0.079 0.002 0.047 0.050 0.07 0.16
Range 14.9- 1.200- 0.010- 0.731- 0.773- 1.00- 1.14-
30.4 9.280 0.220 5.900 6.041 0.22 15.83
n 228 228 228 228 228 228 228

Table 4-11. Mean and range of ash content values for meat, shell and total clam tissues
pooled for all clams sampled. Shell tissue had higher ash content and made
up a greater portion of the total clam ash than meat tissue.
Tissue Mean ash SE Ash content range n
(%) (%) (%)
Clam 94.71 0.06 90.45 97.39 228
Shell 97.48 0.02 96.06 98.23 228
Meat 13.74 0.55 2.50- 17.00 228

Phosphorus allocation in clam biomass

No significant differences in phosphorus concentrations for the meat, shell and clam tissues

could be found between the 2 nutrient addition treatments at stocking or between time intervals

for the low treatment using the ANOVA. No discernable relationships in phosphorus

concentrations for each tissue type (meat, shell and whole clam) were found between the tested

variables (shell length, meat DW, shell DW, clam DW, condition indices, meat AFDW, shell

AFDW, clam AFDW) in the ANOVA correlation. Values for tissue phosphorus concentrations

were then pooled for all clams used in the experiment to obtain the meat, shell and clam tissue

phosphorus concentration data (Table 4-12).

The values in Table 4-13 indicate that clams of shell lengths between 13.9 and 30.4 mm

contain from 0.143 to 1.411 mg P per individual. They also indicate that the majority of the total









phosphorus is allocated into meat tissue even though shell material constitutes the majority of the

total clam DW (Table 4-10).

Table 4-12. Summary statistics for phosphorus concentrations [P] found in meat, shell
and clam tissue types pooled for all clams sampled. Phosphorus was found
in much higher concentrations in the meat tissue than shell.
Tissue type Mean SE Range n
tissue [P] tissue [P] tissue [P]
(mg P/ g DW) (mg P/ g DW) (mg P/ g DW)
Meat 7.657 0.103 2.650- 12.853 228
Shell 0.053 0.001 0.022 0.136 228
Clam 0.299 0.005 0.141 0.606 228

Table 4-13. Amounts of phosphorus (P) contained in meat, shell and clam tissues along
with percentages of total clam phosphorus allocated to meat and shell tissues
for individual clams. The majority of the total clam phosphorus is a result of
meat phosphorus and not shell phosphorus.
Parameter Amount of P Amount of P Amount of P % of Total % of Total
contained in contained in contained in clam P clam P
meat (mg) shell (mg) clam (mg) allocated allocated
to meat to shell
Mean 0.480 0.095 0.575 82.2 17.8
SE 0.014 0.003 0.016 0.4 0.4
Range 0.087 -1.137 0.042 -0.347 0.143 -1.411 50.7-93.1 6.9-49.3
n 228 228 228 228 228

No statistically significant conclusions related to seasonality and clam phosphorus

concentration could be assessed due to the drastically decreasing sample sizes at each sampling

interval after stocking as a result of the high overall population mortality that took place in all of

the raceways. Shell phosphorus values may have been slightly higher than expected since

internal cavity area was not cleaned of residual pallail fluid and meat from shucking.

Phosphorus detected from the inner clam shell was not accounted for in the analysis and is

considered minimal, therefore no further investigation is needed










Treatment raceway clam population phosphorus

No significant reductions in net phosphorus were documented in the raceways, probably due

to high overall clam mortality. The amount of phosphorus contained in the total raceway live

clam populations was calculated using the mean clam tissue phosphorus concentration values in

Table 4-12 and the clam population DW biomass estimation in Figure 3-13, as shown in Figure

4-3. High clam population mortality prevented accurate assessment of raceway clam population

phosphorus on a temporal scale.


20,000 -
18,000 -- Low -
16,000 -- + Med -
14,000 -0-High
12,000 -
S10,000 -
S8,000-
6,000 -
4,000 -
2,000
0
05/14/02 08/17/02 11/20/02 02/23/03 05/29/03 09/01/03
Date
Figure 4-3. Amount of phosphorus (P) sequestered in clam biomass for the low, medium and
high nutrient addition treatments over the study period. Raceway clam phosphorus
decreased with clam mortality and biomass loss, and no significant accumulation of
phosphorus occurred over any of seasonal time periods.

In the high nutrient addition treatment pond raceways from February to May 2003 (Table

3-7), clams measuring 12.2 to 28.6 mm in shell length grew from 0.042 to 0.118

mm/individual/day (Figure 3-11). This equates to a phosphorus sequestration rate of 0.0022 to

0.0079 mg P/individual/day calculated from the clam DW to shell length relationship for the

high nutrient addition treatment pond group (Table 3-11) and the clam biomass phosphorus

concentration of 0.299 mg P/g DW (SE = 0.005) (Table 4-12). Individual tagged clams in the









medium nutrient addition treatment were able to sequester from 0.0015 to 0.0071 mg

P/individual/day during February to May 2003 time period. Due to low individual clam growth

rates (0.026 to 0.086 mm/day) during winter (November 2002 to February 2003) (Table 3-7),

phosphorus sequestration potential was limited to less than 0.0004 mg P/individual/day.

Discussion

The goal of this phase of the clam treatment raceway study was to evaluate phosphorus

removal potential and limitations. The three major components of the evaluation included, 1)

distribution of phosphorus taken up by clams, 2) rate of phosphrous removal by the clam

raceways and 3) sustainability of phosphorus removal. The results of the study revealed some

promising best-case scenarios for the system in terms of phosphorus removal, but also a number

of major challenges in terms of sustainability.

Distribution of Phosphorus Taken Up By Clams

Clams allocate phosphorus in both meat and shell biomass. Mean clam biomass

phosphorus concentration determined from this study was 0.299 mg P/g of clam DW (SE =

0.005, range 0.141 to 0.606, n = 228). Phosphorus concentration did not differ with clam size or

between populations of stocked clams from the Santa Fe River and Lake George. Clam biomass

phosphorus was allocated in greater amounts to clam meat tissue, as opposed to shell, as

suggested by Fuji (1979). Meat tissue had much higher average concentrations of phosphorus

(7.657 mg P/ g meat DW, SE = 0.103) than shell (0.053 mg P/g shell DW, SE = 0.001), and as a

result, more clam phosphorus was sequestered in meat than shell biomass.

On an individual clam basis, clams contained 0.143 1.411 mg P/individual based on clam

phosphorus concentration and dry weight values. This is slightly higher than values reported by

Fuji (1979) for Corbicula in an estuarine lagoon (0.008- 1.500 mg P/individual). Clam meat

phosphorus ranged from 0.087 to 1.137 mg P/individual in this study, similar to the range of









0.007 to 1.433 mg P/individual reported by Fuji (1979). Shell phosphorus ranged from 0.042 to

0.347 mg P/individual in this study, slightly higher compared to values ranging from 0.001 to

0.067 mg P/individual shell found by Fuji (1979). The relationship between shell size and

individual clam phosphorus was suggested to be exponential by Fuji (1979). In this study,

individual clam phosphorus to clam size followed a curvilinear regression relationship developed

in Chapter 3 for clam biomass to clam size. Decreasing slope with increasing size in clams

larger than 28 mm in shell length was most likely due to shell erosion evident in the larger clam

specimens from Lake Dalhousie, FL as explained in Chapter 3. Other studies of biomass

phosphorus in Corbicula (Fuji 1979) and saltwater Manilla clams, Ruditapes (Nizzoli et al.

2006) have suggested no effect of clam size or sample location on biomass phosphorus

concentration; however, these studies did find that biomass phosphorus concentration differs

seasonally. It was suggested that increases in clam phosphorus concentrations during warmer

months are related to increases in phytoplankton availability and reproductive activity, which did

not occur in the clam raceways.

Oysters contain around 1.067 mg P/g DW (Newell 2004), which is higher than observed

for Corbicula in this study (Table 4-14). Saltwater Manilla clams cultured in an open estuary

(Nizzoli et al. 2006) had similar meat phosphorus concentration as clams in this study; however,

phosphorus concentrations in clam shell material in this study (0.053 mg P/g shell DW) were on

the low end of the range observed for Manilla clams. Phosphorus concentrations reported by

Yamamuro et al. (2000) for the shells of other saltwater clams (Musculista, Ruditapes and

Anadara), are also higher (0.101 to 0.880 mg P/g shell DW) than phosphorus concentrations

reported here. The large difference between the values reported by Yamamuro et al. (2000) and









other studies may be due to chemical absorption from the overlying water column, since

Yamamuro's values were based on analysis of dead shell material.


Table 4-14. Comparison of meat and shell phosphorus concentrations [P] in dry weight
(DW) biomass of Corbicula versus other fresh and saltwater clams.
Organism Meat [P] Shell [P] Total [P] Total P per Reference
(mg P/g of (mg P/g of (mg P/g of individual
meat DW) shell DW) total DW) (mg)
Corbicula 2.650- 0.022- 0.141- 0.606 0.143- 1.411 This study
(freshwater) 12.853 0.136
Corbicula 8.293- 0.024- 0.506- 0.800 0.008- 1.500 Fuji (1979)
(estuarine) 22.667 0.056
Manilla 3.000 > 0.150 0.373 n/a Nizzoli et al.
clams (2006)
Hardshell clams 1.100 + n/a n/a n/a Capar &Yess
0.250 (SE) (1996)
Softshell clams 1.400 + n/a n/a n/a Capar & Yess
0.500 (SE) (1996)
Eastern oysters 1.100 + n/a n/a n/a Capar & Yess
0.300 (SE) (1996)
Pacific oysters 2.000 + n/a n/a n/a Capar & Yess
0.550 (SE) (1996)
Saltwater clams n/a 0.101 n/a n/a Yamamuro et
0.880 al. (2000)

Even though meat contains higher amounts of P, it is more mobile after death than shell

material and can more easily return to the overlying water column or substrate (Fuji 1979,

Nizzoli et al. 2006). Shell phosphorus, therefore, may be a better long-term sequestration

product if phosphorus is in fact incorporated into the organic portions of the shell material and

not just loosely bound to shell surfaces. Elements bound in clam shells may require additional

steps to degrade since the material is so tightly bound. The higher levels of phosphorus

sequestered into shell biomass in this study may, in part, have been due to meat biomass and

fluid retained by the shell during processing (McMahon and Bogan 2001).









Estimates of Phosphorus Uptake Rates

Estimated clam biomass phosphorus sequestration rates from the tagged clam study ranged

from 0.0022 to 0.0079 mg P/individual/day. Estimated annual phosphorus sequestration

potential of 0.803 to 2.884 mg P/individual/yr calculated from daily rates may be overestimated

due to lower winter growth of clams. Therefore, an individual clam sequestration rate between

0.201 to 0.721 mg P/individual/yr may be a more accurate estimation of clam phosphorus

sequestration, since it is based on longer-term growth estimates. Using the phosphorus

sequestration rate for clam population biomass based on tagged clam growth, a theoretical clam

population of 16,560 individuals per 21.6 m2 of raceway area in northern Florida should be able

to sequester on average 354 mg P/m2/yr, with a range of 154 to 553 mg P/m2/yr, assuming a 3-

month cessation in growth during winter.

Clam raceway phosphorus sequestration would be expected to increase beyond rates

estimated for tagged clam growth if successful reproduction and recruitment occur and clams are

not limited by environmental conditions in the raceway. High reproductive capacity of the

animals in large natural systems suggests, that under ideal conditions, raceways should be able to

reach population densities of over 2000 clams/m2 as reported for natural systems by Gardner et

al. (1976), Eng (1979), Sickel (1986) and McMahon and Bogan (2001). This is roughly 3 times

the stocked density of 766 clams/m2.

Theoretically, the ideal raceway population would be expected to follow similar

phosphorus dynamics as presented in the Fuji (1979) model, who suggested that an estuarine

Corbicula population of 65 mg P/m2 can sequester clam biomass phosphorus at an annual rate of

about 130 mg P/m2/yr as growth and recruitment when not limited by food availability. The

stocked clam densities in this study represent a total phosphorus level of 14,000 mg P (648 mg

P/m2), translating into 28,000 mg P (1296 mg P/m2) of harvested phosphorus annually using the









Fuji (1979) model. This value is much higher than the tagged clam phosphorus removal rates

given in this study, since it does not account for mortality, which would ideally be minimized by

strategic harvest of clam biomass that would also maintain younger clams capable of higher

growth rates. The Fuji (1979) model lends itself well to application in an ideal clam raceway

calculation since food resources did not appear to be limited and phosphorus sequestration was

influenced by seasonality as it is in the natural environment.

Adding to the phosphorus removal potential, phosphorus accumulated in raceways from

clam waste would be periodically removed with raceway sediment, thereby minimizing losses to

the surrounding environment. The stocked clam population densities used in this study would be

expected to produce nearly 4000 mg P/m2/yr in biosolids as feces and pseudofeces according to

the Fuji (1979) model. Ideally, a Corbicula-based raceway could be designed and operated to

maximize removal of clam wastes and other solids settling on the substrate along with growth of

clam biomass. A similar approach in bivalve-based mariculture waste treatment systems has

been employed by Shpigel et al. (1993), Shpigel et al. (1996) and Neori et al. (2000) in which

settling ponds are used to increase treatment potential. Settling of organic constituents may lead

to other problems such as dissolved oxygen demand and ammonia production due to decay

(Dame 1996). Dissolved phosphorus loss from accumulated sediments has also been observed in

freshwater bivalve-based treatment systems (Soto and Mena 1999). Harvest and substrate

removal/replacement times would need to be optimized to minimize these stressors.

An annual phosphorus uptake rate of at at least 5300 mg P/m2/yr is expected for the

Corbicula -based raceway system in this study under ideal conditions and solids management.

The theoretical clam-based raceway system would also be expected to lose phosphorus at a rate

of 1166 mg P/m2/yr as excreted wastes based on the Fuji (1979) model, that may be more









difficult to capture and are often lost in natural systems with outflowing water. Based on these

theoretical values, clam raceway P removal potential would be expected to be approximately

4100 mg P/m2/yr if solids capture simply by management of settling through optimizing raceway

hydrology or the addition of a settling stage following clam raceways.

Comparison Of Phosphorus Removal By Clam Raceways And Other Systems

Estimates for phosphorus sequestration by clams at the stocking rates used in this study are

154 to 553 mg P/m2/yr based on tagged clam growth rates. The clam phosphorus sequestration

values in this study were lower than potential values of a theoretical cultured population of

Corbicula, estimated by Mattice (1977) to be between 1200 and 1400 mg P/m2/yr (Table 4-15).

The latter range is based on annual clam wet weight biomass growth rates of naturally occurring

populations and potential mariculture density assumptions by Mattice (1977), converted to

phosphorus removal capacity using average % water and phosphorus content values. However,

the biomass phosphorus accumulation range estimated by Mattice (1977) may be unrealistically

high due to population biomass density assumptions of over 10,000 clams/m2, which may not be

possible even under ideal culture conditions.

In another study, estimated clam biomass phosphorus sequestration for Corbicula exposed

to municipal wastewater by Greer and Ziebell (1974) was calculated using biomass phosphorus

content values, yielding 877 mg P/m2/yr (Table 4-15). However, the small scale of the

experimental system used by Greer and Ziebell (1974), short time duration of the study and lack

of data given on the actual surface area in the experimental culture, casts some doubt on the

value of these estimates for large scale systems. Estuarine populations of Corbicula have been

estimated to sequester phosphorus at 134 mg P/m2/yr (Fuji 1979), while other phosphorus

sequestration values calculated by Fuji (1979) for Corbicula populations from other studies









range from 13 to 250 mg P/m2/yr compared to some other freshwater bivalves that ranged from

37 to 77 mg P/m2/yr.

Rates reported by Nizzoli et al. (2006) indicate that phosphorus removed from an estuary

through harvest of Manilla clams can range from 1300 to 2600 mg P/m2/yr using 3-month

growth intervals between harvests (Table 4-15). Phosphorus sequestration may be possible using

raceway-based Manilla clam culture in saltwater systems and could be increased significantly by

capture of waste materials lost to tidal outflow and sediment deposition in natural systems.

Phosphorus sequestration rates for bivalve culture systems are much less than the 98,000 mg

P/m2/yr rate given by Dame (1996) for an intertidal oyster reef community containing many

different organisms; therefore, this value is not indicative of bivalve phosphorus removal

potential. It is difficult to compare phosphorus sequestration in Corbicula to mariculture effluent

phosphorus treatment systems using other bivalves since treatment potential in these systems is

usually assessed on the basis of nitrogen and phytoplankton removal.


Table 4-15. Estimated annual phosphorus removal in various biological treatment
systems applied to different effluent types using systems of varying design
and scale. Estimated phosphorus removal by the Corbicula-based system in
this study is similar to other harvested systems, while harvested aquatic plant-
based systems are capable of much higher removal rates than animal-based
systems.* denotes phosphorus removal rate based on periodic harvest of
system
Effluent type Target organism P removal System type Reference authors)
(mg/m2/yr) and scale
Agriculture Corbicula 154 to 553 Raceway Measurements from
N and P tagged clam this study
fertilizer growth
Agriculture Corbicula 4100 *Large-scale a This study, ideal
N and P raceways, conditions
fertilizer theoretical
Nutrient Corbicula 877 Theoretical Greer and Ziebell
enriched water aquaculture (1974)
Nutrient Corbicula 1200-1400 Theoretical Mattice (1977)
enriched water aquaculture









Table 4-15 Continued.
Estuary, natural Brackish water 17-60 Natural tidal Fuji (1979),
clams, Corbicula estuary Yamamuro et al.
and Musculista (2000)
Estuary Manilla clam, 1300-2600 *Natural tidal Nizzoli et al. (2006)
mariculture Ruditapes as biomass estuary
Estuary, natural Oyster reef 98,000 Natural tidal Dame et al. (1989) in
ecosystem estuary Dame (1996)
Tertiary-treated Freshwater 194 to *Large-scale b Hallock and Ziebell
municipal finfish, Tilapia 5840 pond culture (1970), b Greer and
wastewater and catfish biomass & Ziebell (1974),
wastes b Bunting (2007)
Agriculture Periphyton 36,550 to *Large-scale aAdey (1993)
runoff-enriched 47,450 algal turf
surface water scrubber
Tertiary-treated Periphyton 91,250 to *Large-scale aCraggs et al. (1996)
municipal 266,450 algal turf
wastewater scrubber
Diluted dairy Periphyton 29,200 to *Small-scale c Wilkie and
wastewater 120,000 algal turf Mulbry, "',',
scrubber Pizzaro et al. (2002)
Diluted dairy Assorted aquatic 58,000 *Med.-scale a Sooknah and Wilkie (2004)
wastewater plants and algae circular tanks
Aquaculture Lettuce grown 365,000 *Med.-scale a Adler et al. (2003)
wastewater hydroponically conveyor belt
Agriculture Periphyton 320 Large-scale a DeBusk et al. (2004)
runoff-enriched dominated, wetland-type
surface water mixed vegetation raceways
Dairy effluent Corn silage 6900 *Land d Wilkie and
forage crops application Mulbry (2002)
Dairy effluent Grassland forage 18,800 to *Land Johnson et al. (2004)
crops 59,000 application
Surface water Natural wetland 146 to General Dodds (2003)
systems vegetation 803,000 Wetlands
a P removal based on annual P removal values for target organisms from seasonal-based
experiments at an appreciable scale
bP removal estimated using fish P content from Greer and Ziebell (1974) and annual fish
production rates from Hallock and Ziebell (1970) and Greer and Ziebell (1974)
c Lower P removal value from Wilkie and Mulbry (2002), higher value from Pizzaro et al.
(2002), includes Kebede-Westhead et al. (2003) P removal estimate under the same conditions
d P removal estimated by reference author using annual values from other studies

Even under ideal culture conditions, estimated removal of phosphorus by Corbicula and

other bivalves pales compared to those determined for aquatic periphyton and macrophyte-









dominated systems (Table 4-15). Harvested systems appear to be capable of higher phosphorus

removal rates, especially in plant-based treatment systems. Natural wetlands show the largest

phosphorus removal values; however, they also have the largest range, with values as low as 146

mg P/m2/yr being reported by Dodds (2003) (Table 4-15). Wetlands are subject to capacity

limitations after several years whereas harvested systems are theoretically more sustainable.

Periodic harvesting increases phosphorus removal potential of the organisms by minimizing

losses due to mortaility, maintaining high growth rates and increasing the longevity of the

system.

Vegetative systems may be better suited for management of phosphorus in a farm-scale

system due to their ability to withstand a wider variety of environmental stressors and decreased

water use compared to clam raceways as seen in this study. Vegetative systems have been added

to bivalve-based systems in the mariculture industry to provide treatment following exposure to

bivalve populations (Shpigel et al. 1993, Shpigel et al. 1996, Neori et al. 2000). Periphyton may

also be used to remove phosphorus following exposure to Corbicula-based systems since

harvested types of these systems have been successful in rapidly sequestering phosphorus from

agricultural applications especially at low water phosphorus concentrations (Scinto and Reddy

2003). Raceway systems used in this study may be applicable in vegetative-based biofilters as

well.

Problems With Measuring Short-term Phosphorus Uptake

There were no indications that the raceway systems were able to provide a noticeable

reduction in water phosphorus or chlorophyll a during the normal through-flow operation at a

range of 151 to 227 L/minute. Reduction of phosphorus by the clam raceways was not observed

even after being converted to a temporary recirculation mode to increase retention time to 6

hours from 9.5 minutes and 6.3 minutes. High temperature, low dissolved oxygen and ammonia









toxicity concerns prevented longer retention times by recirculation in an attempt to show

phosphorus reduction especially during daylight hours. There was no significant reduction of

total phosphorus levels in the covered raceways, indicating that raceway water phosphorus

reduction may not be affected by clams or settling in the raceway systems.

The most plausible explanation for the lack of large-scale phosphorus uptake may be the

obvious environmental and physiological stress that clams were undergoing over the entire study

period including during the through-flow and recirculation trials. High clam mortality was

evident throughout the study, especially in warmer months when water phosphorus removal

trials were taking place even though condition indicies did not indicate any decreases in clam

health over time. Haines (1977) attributes similar mortality and decreased water treatment

capacity in Corbicula cultured on municipal wastewater to high temperature and possibly

ammonia. Also, potential stress from amphipod infestation may have severely affected the

raceway clam population's ability to filter feed and, therefore, remove phosphorus-containing

material as described in Chapter 3.

Similarly, Kinne et al. (2001) was unable to show that TP was lowered significantly in a

medium-scale raceway system using oysters to treat shrimp farm effluent, possibly because of

soluble phosphorus excretion meeting phosphorus uptake or ammonia concerns. Conversion of

particulate phosphorus to dissolved phosphorus by excretion of soluble phosphorus by Corbicula

waste products as described by Lauritsen and Mozley (1989) may not explain the lack of

phosphorus removal in this system. The extent of phosphorus conversion within the raceways is

unknown since significant TDP addition was never detected during any recirculation trial, and

TP removal was no different in raceways containing substantial amounts of clams compared to

those with depleted stocks. Clam density is expected to limit treatment capacity for raceway









clam populations. However, raceway systems in the high nutrient addition treatment contained

approximately 460 clams/m2 (10,000 clams per raceway) more than the low nutrient system, but

still did not provide a noticeable phosphorus removal.

Chlorophyll a uptake values in the recirculation trials may have been induced by

disturbance of substrate materials related to switching raceway hydrology from through-flow to

recirculation for experimental purposes. Disturbance of raceway substrate did occur in the

recirculation trials as indicated by the higher TP, TDP and chl a values as opposed to source

pond levels at the same time. Haines (1977) attributed increased treatment potential of

Corbicula-based systems to increased particle concentration. Therefore, any disturbance and

increase in particle availability may result in removal by clam populations as well as settling.

The apparent significant chlorophyll a removal seen in both high and low nutrient addition

treatments may have been due to such a disturbance. However, removal rates were probably not

due to settling, since covered raceways failed to provide any definitive indication of chlorophyll

a uptake. Given the large differences in clam population density between the high and low

nutrient addition treatments during the recirculation trials, actual removal of chlorophyll a due to

clam filtration is also doubtful. Potential problems encountered with the raceway recirculation

modification suggests that phosphorus uptake was probably not achievable under normal

through-flow operating conditions by the clam raceways, as expected.

Problems such as high temperature, low dissolved oxygen and dangerous levels of

ammonia may result with the six-hour retention times, as seen in this study. Raceway system

hydrology can also be modified to accommodate full or partial raceway water recirculation as

demonstrated by the additional pump placement in this study to extend retention time. Water

retention time would need to be managed closely with real-time monitoring of temperature,









ammonia and dissolved oxygen to avoid levels that may negatively affect clam populations.

System water retention time greater than 10 minutes and less than 6 hours is recommended to

improve treatment potential, while minimizing environmental stress. More conclusive testing of

TP, TDP and chl a uptake by clam populations at large scale is needed before reliable short-term

removal rate estimates and system operating tolerances can be assessed.

Dairy Application Demands And Issues

Issues with water consumption may limit widespread use of this technology since digested

dairy wastewater effluent addition less than 5 % (by volume) was needed in clam raceway source

water ponds to overcome ammonia concerns. Sooknah and Wilkie (2004) have demonstrated

that a 1 : 1 dilution of anaerobically digested wastewater enhances biological uptake of

phosphorus and nitrogen in aquatic plant-based systems, substantially decreasing amounts of

ammonia in effluent water. Application of these kinds of vegetative systems may provide pre-

treatment of phosphorus and harmful nitrogenous compounds prior to clam raceway addition to

increase feasibility at a large scale. Coupling of vegetative and clam-based systems in this way

would lower water demand, while decreasing ammonia and utilizing less treatment surface area

than required by clam systems alone.

Application of clam raceway technology in a real-world scenario is perhaps best analyzed

on an individual farm basis. The University of Florida Dairy Research Unit (DRU) in Hague,

Florida is described by Wilkie et al. (2004) as having an average milking herd of 359 cows, a

wastewater production of 502 m3/day and a daily freshwater water usage of 52.25 m3/day. A

large scale, fixed film anaerobic digester is in place at the DRU and is not expected to change the

amounts of P-loading to wastewater handling systems that are restricted primarily to a secondary

settling lagoon until land application. Normal digester outflow would be similar to values









reported by Sooknah and Wilkie (2004) for water quality in a system accepting digester effluent

from the DRU, including 24 mg/L TP and 136 mg/L total ammonia nitrogen (TAN).

Phosphorus and ammonia levels in the macrophyte system output are expected to be 0.24 to

6.0 mg/L TP with a significant portion of phosphorus as dissolved phosphorus and 0.29 to 3.53

mg/L TAN (Sooknah and Wilkie 2004). Insertion of air-stripping or biofilm filtration systems

similar to ones used in the aquaculture industry (Timmons et al. 2001) or bacteria-based systems

under assessment for dairy wastewater, as indicated by Sooknah and Wilkie (2004), could

possibly be inserted following macrophyte treatment to lower TAN levels. Another possible

solution to the ammonia problem may be another vegetative treatment phase. However, these

biological-type systems may not be as dependable as mechanical ones used in commercial

aquaculture due to seasonality and temperature dependence of biological systems. Application

of aquaculture technology should help to decrease TAN and increase DO levels with little or no

excess water demand except for increased evaporation with increased treatment surface area.

The continuous growth of nuisance plants evident during raceway operation may indicate

the need for a vegetative system to be incorporated into the clam system design to get a

noticeable decrease in raceway water phosphorus. Fuji (1979) suggested that clam deposition of

organic materials provides a food source for plant growth in detrital systems; therefore, harvested

plant growth in polyculture with Corbicula may provide a way to increase phosphorus

sequestration potential of the raceway system. Use of vegetative-type filters along with bivalves

has been demonstrated in marine mariculture wastewater treatment systems to remove dissolved

phosphorus excreted from bivalves (Jones et al. 2001, Borges et al. 2005) and has been suggested

for Corbicula culture in agriculture wastewater (Stanley 1974); but no applications of this

technology coupling have been demonstrated using dairy effluent. Nuisance plant growth did









not appear to be inhibited by the high water temperatures encountered in this study, unlike clam

populations, adding to the appeal of plant-based systems.

Using the 1:1 dilution and phosphorus removal potentials suggested by Sooknah and

Wilkie (2004) for an aquatic macrophyte-based system at the DRU, I would expect freshwater

usage to increase from 502 m3/day, to 554 m3/day just for the macrophyte system. This would

increase wastewater volume to 105 m3/day and require an active system capable of sequestering

655 mg TAN/m2/day (239,075 mg TAN/m2/yr). This value is based on TAN removal of 99.6 %

from diluted wastewater containing 68 mg/L TAN and over a 31-day period using a 0.5 m x 0.36

m raceway at a water depth of 0.3 m (Sooknah and Wilkie 2004). The resulting macrophyte

system sized for TAN removal at the DRU would need 68 m2 of treatment area per day of

retention. With the suggested hydraulic retention time of 31 days to remove incoming TAN,

raceway treatment area would need to be at least 2116 m2 to meet the TAN treatment needs for

Corbicula systems using anaerobically digested wastewater at the DRU.

Removal of ammonia would be the primary goal of vegetative treatment prior to clam

raceway introduction; however, macrophyte systems have been suggested for treatment of

phosphorus as well (Sooknah and Wilkie 2004). The macrophyte system phosphorus removal

potential would be 58,000 mg P/m2/yr removing 96.5 % of the incoming phosphorus under ideal

conditions as indicated by Sooknah and Wilkie (2004). Assuming that ammonia levels could be

maintained without removal of phosphorus or further dilution, phosphorus loading from the

macrophyte system is expected to be up to 230 kg P/yr for the 105 m3/day (38,325 m3/yr).

Wastewater would exit the macrophyte system at 0.24 to 6.0 mg/L TP, considerably higher than

the range tested in this study; however, within expectations of 3 mg/L values tested by Greer and

Ziebell (1972) at aquarium scale using Corbicula. Under this maximum phosphorus loading rate









expected in the macrophyte filter outflow, a clam raceway system having up to 38,325 m2 of

treatment area would be needed to remove the remaining phosphorus in the system.

Losses of phosphorus around 1166 mg P/m2/yr would be expected from the 38,325 m2 of

clam raceway system area and would require additional treatment prior to discharge into surface

waters since effluent TP concentration would be around 0.05 to 1.24 mg/L at a volume of 105

m3/day in a properly managed clam system under these circumstances. This value may be

improved by increasing the retention time or deceasing water usage. Application of clams or

macrophytes at such a scale is preposterous considering the limited size scale demonstrated by

Sooknah and Wilkie (2004) and the lack of large-scale success of Corbicula-based raceways at

the DRU that prevents an accurate system sizing from being made.

Other large-scale harvested aquatic plant systems such the Algal Turf Scrubber (ATS)

raceways described by Craggs et al. (1996) may also need to be employed prior to introduction

to clam systems, in addition to macrophyte-based raceways to lower TP to an optimum level. TP

in clam raceway influent should be between the 0.04 and 3 mg/L range shown by Greer and

Ziebell (1972) to have some phosphorus removal potential as determined in aquarium-based

studies. Small-scale algal turf scrubber systems have been used by Pizzaro et al. (2002), Wilkie

and Mulbry (2002) and Kebede-Westhead et al. (2003) on diluted dairy wastewater at the DRU

to remove an estimated 29,200 to 120,000 mg P/m2/yr. Assuming the lower phosphorus removal

value of 29,200 mg P/m2/yr, the system would need to have at least 7,876 m2 of treatment area,

much larger than the largest ATS wastewater treatment system tested by Craggs et al. (1996) at

1,018 m2, but not impossible for large scale implementation at the DRU. Using ATS systems to

treat dairy wastewater at a quartenary or lower level, as suggested by these calculations, is









certainly more feasible than clam raceways requiring a treatment area 1,774 times larger than the

area afforded by the 3-raceway system tested in this study.

Concerns over salinity and DO levels in dairy wastewater remediation (Sooknah and

Wilkie 2004) may also impact clam raceways. Corbicula can tolerate at least some salinity in

the natural environment (Deaton 1980, McCorkle and Dietz 1980), however low tolerance for

dissolved oxygen levels in the 3-5 mg/L range reported by Belanger (1985), suggests that oxygen

may be the more critical factor. Dissolved oxygen levels of less than 3 mg/L and ammonia

levels from 0.29 to 3.53 mg/L TAN in various macrophyte system effluents reported by Sooknah

and Wilkie (2004) will need to be addressed prior to application of clam systems since they are

not within the tolerable ranges for Corbicula.

Sustainability

Application of clam raceway systems as a dairy wastewater treatment mechanism will need

further investigation, primarily since raceway systems were not able to operate under dairy

effluent addition at 5 or 10 % by volume due to high ammonia concerns. Keeping both ammonia

and temperature within a tolerable level will be critical to implementing Corbicula as a

wastewater treatment mechanism. High levels of undesirable materials such as ammonia in

source water can be addressed by dilution of effluents; however, increased constraints on

freshwater usage may limit the use of clam raceways alone to solve the phosphorus management

of any operating dairy, especially in Florida. Successful implementation of clam raceways or

other high water demand/low aereal phosphorus sequestration potential for the treatment of dairy

wastewater on a large scale may be limited unless coupled with other technologies such as

harvested plant systems as seen in the mariculture industry by Shpigel et al. (1993), Shpigel et al.

(1996) and Neori et al. (2000).









Corbicula-based raceway systems may be more applicable to aquaculture waste scenerios

since ammonia, dissolved oxygen and temperature are typically monitored and managed in

commercial aquaculture systems. While salt tolerance in Corbicula affords the animal some

expansion into mariculture effluent phosphorus treatment, species selection in these systems will

probably be limited to saltwater species traditionally cultured in larger scale as a food crop, such

as oysters (Jones et al. 2001). Bivalve-based systems may be able to perform for longer

durations between harvest efforts by utilizing other species due to the shorter life spans found in

Corbicula (normally about 2 years, from McMahon and Bogan 2001).

The lack of marketability for Corbicula cultured in wastewater conditions may also be an

obstacle to implementation since phosphorus treatment effectiveness is not only gauged by

phosphorus removal and sequestration potential, but also cost effectiveness, energy costs and

water consumption. Corbicula has been proposed as a desirable protein source by Iritani et al.

(1979) and has been recommended as a product to offset operational costs in clam-based systems

by (Mattice 1977), Greer and Zeibell (1972) Stanley (1974), and Haines (1977). Even though a

market for live Corbicula or Corbicula-based products may exist in some places (Chen 1976,

Phelps 1994), clams cultured on dairy or municipal wastewater may not be acceptable in any

market due to human health concerns. Biomass produced on wastewater can sequester toxins

such as metals (Marcussen 2007) and pesticides (Barber 2006), although worries over choliform

bacteria may be unfounded (Islam 2004). Stanley (1974) cautions about potential problems with

poisoning from cyanobacteria toxins in wastewater aquaculture systems as well. The lack of

market for Corbicula as a food resource domestically is due in part to the filter-feeding ability of

freshwater bivalves that captures various harmful organisms easily transferred to humans such as

Cryptosporidium (Graczyk et al. 2003 and Izumi et al. 2004), Giardia (Graczyk et al. 2003) and









Cyclospora (Graczyk e al 1998). Other species such as pearl oysters can provide profit

opportunities other than food or aggregate in mariculture waste treatment systems as suggested

by Gifford et al. (2004), and the organisms could possibly be adapted to raceway culture

conditions.

In an attempt to use raceway biomass in a positive way, dead clam shells removed from the

raceways in this study were spread out as an aggregate over areas surrounding the raceways that

were covered with crushed lime rock to prevent weed growth. Clam shells are typically removed

from aggregate sources excavated from river bottoms in the Applachicola River in Florida and

may not be as desirable an aggregate choice compared to gravel. Problems with decaying soft

tissue may prevent use of clam aggregates without pre-treatment to reduce odors.

Substrate removed from the raceways is rich in clam biosolids and could potentially be

composted for use as a soil amendment (Greer and Ziebell 1977). This compost may be more

valuable than the clams themselves, therefore, composting both clams and substrates

simultaneously may be the best alternative for phosphorus sequestered by Corbicula raceway

systems. Products from vegetative-based phosphorus treatment systems are often equally

unwanted and ultimately the most valuable as a soil amendment by composting the material to

offset operational costs in plant and animal-based treatment systems as suggested by Greer and

Ziebell (1974).

Even if an economically feasible use could be found for Corbicula-based treatment

raceway by-products, and adequate phosphorus removal on a large scale could be established in

a wastewater treatment stream, implementation of a system that utilizes such an invasive animal

can be met with negative responses. Introduction of Corbicula has been implicated in different

large-scale habitat modifications in natural water bodies that may not be viewed as acceptable in









all surface water systems (Ingram 1959). Other studies indicated concerns over native mussel

displacement due to the high colonization success and reproductive capacity of Corbicula

(Kraemer 1979, Darrington 2002, Cooper et al. 2005).

Inevitably, no matter what size and component design of a biologically-based system, land

and water demands, or its usefulness and desirability, environmental conditions must allow

clams to survive, grow and reproduce to sequester phosphorus effectively into shell and meat

tissue at a large scale. Potential problems with high temperatures, ammonia, food availability,

seasonality, phytoplankton production, amphipod infestations, system performance evaluation,

biofouling by vegetation, available land area and water usage need to be solved before a reliable

large-scale Corbicula-based raceway treatment system can be implemented into any wastewater

stream in Florida.









CHAPTER 5
SUMMARY

The three primary goals of this dissertation research project were; 1) To design, construct

and implement an experimental raceway system for removal of particulate phosphorus from

wastewater streams on a size scale that represents real-life applications 2) To test suitability of

the freshwater clam Corbicula as the primary active agent in phosphorus removal within a

raceway environment and 3) To determine if P-removal capabilities of the system are adequate to

deal with phosphorus loads anticipated for dairy wastewater streams.

Raceway Function and Attributes

The raceway-based systems constructed in this study were chosen and developed as a low-

cost, portable and easier-to-assemble alternative to other raceways constructed from concrete,

fiberglass or plastic. The advantage of the modular raceway design was the culture tanks are

easily scalable by length, width, depth and quantity of units to meet surface area needs of the

application, desired hydrologic regimes and experimental design criteria. The design also

maximizes physical accessibility to the bottom area for stocking, sampling, harvest and

maintenance. The raceway systems constructed at both Blountstown and Hague, FL sites

operated without failure or leaking over the study period and provided an ideal platform for the

water treatment experiments incorporated in the current study.

Adaptability of Clams to Raceway Conditions

The tag and recapture study was the primary measure of growth potential of Corbicula.

Individual tagged clam shell growth rates based on length ranged from around 0.001 mm/day on

an annual basis in the low nutrient addition treatment group to up to 0.118 mm/day in the high

nutrient addition treatment group in spring. Temporal patterns in tagged clam growth rates

showed seasonality possibly due in part to changing water temperature, level of nutrient









addition, source water phosphorus levels and possibly phytoplankton availability. Tagged clams

grew in all nutrient addition treatment groups without a consistent correlation to chlorophyll a

concentration, suggesting that either clams utilized non-chlorophyll a containing food resources,

such as bacteria and suspended detritus, or phytoplankton biomass in the source ponds was

sufficient to sustain growth.

Despite the growth observed in the surviving tagged clams, the overall clam population did

not adapt well to raceway conditions over extended periods, with over 90 % mortality during a

one-year period. Timing of mortality indicated that high summer temperatures in the raceways

may have been the major factor responsible for the severe losses. Water temperatures in the

range of 30 C and above have been implicated as a limiting factor in the success of Corbicula in

other applied studies and in natural populations (Greer and Ziebell 1972, Mattice 1977, Haines

1979, Buttner and Heidinger 1980, Buttner 1986). Major population declines took place when

water temperatures reached this level in all systems, regardless of the level of nitrogen and

phosphorus addition and chlorophyll a in the source water.

Other environmental factors present in the raceways, including increased ammonia levels

encountered during periodic execution of 6-hour recirculation trials, and potential stress from

infestation by amphipods (Hyalella azteca), may have also contributed to mortality and affected

phosphorus removal and sequestration potential as well. Interactions observed between

Corbicula and Hyalella in this study are the first to recognize amphipods as having a potential

predatory or parasitic role in clam population dynamics; however, the intricacies of this

interaction are not yet understood.

P-removal Capacity

Clam raceway systems were able to demonstrate phosphorus removal and sequestration

potential as evidenced by significant shell growth in tagged clams. Individual clams selected for









analysis were between 14.9 and 30.4 mm shell length (0.773 6.041 g total clam DW). Clams

contained an average of 0.299 mg P/g DW (SE = 0.005, range 0.141 to 0.606, n = 228), derived

from shell and meat values combined, which equates to 0.143 to 1.411 mg P/individual. The

concentration of phosphorus (mg P/g DW) in the shell and meat tissues did not change with shell

size, location collected or exposure time in the raceways. Meat tissue had a much higher average

concentration of phosphorus (7.657 mg P/ g meat DW, SE = 0.103) than the shell (0.053 mg P/g

shell DW, SE = 0.001). As a result, the majority of the total clam phosphorus was sequestered in

meat as compared to shell biomass, even though shell biomass comprised the majority

(approximately 82%) of the total clam DW biomass.

Based on the tagged clam study, phosphorus sequestration potential was estimated to range

from 0.803 to 2.884 mg P/individual/yr for adult clams. Corbicula biomass phosphorus

sequestration potential estimated in this study was similar to values given by Fuji (1979) for

biomass phosphorus sequestration in natural populations. Ideally, raceway clam populations

would be able to sustain estimated clam biomass phosphorus sequestration rates along with

successful reproduction and recruitment as seen in the Fuji (1979) model. An similar raceway

system as tested in this study under ideal conditions would be stocked with 648 mg P/m2, and

would be expected to sequester 28,000 mg P/yr (1296 mg P/m2/yr) as clam biomass. In a clam

culture scenario, this live biomass could be harvested along with most of the 708 mg P/m2/yr

expected to be produced by normal mortality and excretion retained in the raceway sediment.

Ideally, a Corbicula-based raceway would be designed and operated to maximize

sequestration of particulate clam waste phosphorus and other solids settling along with growth of

clam biomass. Phosphorus removal potential may be higher in an engineered system equipped to

deal with environmental variation and capable of further enhancing growth rate and treatment









area through sustainable harvest. By retaining and harvesting settled particulates that would

otherwise be lost in natural systems and maximizing growth and reproduction rates, an annual

phosphorus removal rate of at least 4100 mg P/m2/yr is expected for the Corbicula-based

raceway system as tested in this study. This phosphorus removal and sequestration potential is

higher than potentials calculated for other freshwater bivalve and finfish populations.

Future Applications

Limitation of reproductive success by Corbicula held in captivity in this and other studies

(McMahon and Bogan 2001) suggests that implementation of Corbicula systems at large scale

may not be attractive if juveniles need to be restocked periodically from natural populations.

Perhaps partially open systems such as power plant discharge canals that are open to a natural

water body part of the year would allow for inflow of juveniles to repopulate raceways after

harvest or mortality events may be needed to maintain adequate stocks. Other solutions for

sustaining clam stocks in raceway-based systems such as genetic selection for traits that support

growth and reproduction outside of the normal tolerances has been proposed for Corbicula by

Sickel (1986) however this would take an exceptionally long time to develop, if at all, given

homogeneity and genotypic plasticity of Corbicula.

Clam raceway, phosphorus treatment potential of anaerobically digested dairy effluent

could not be assessed directly, due to toxicity concerns over total ammonia nitrogen levels

present at > 2 mg/L NH3-N present in source ponds at 5 % and 10 % (volume of source pond

volume) wastewater addition levels tested. Low tolerance to ammonia toxicity in Corbicula

(Cherry et al. 2001) and to ammonia in combination with high temperatures have been

implicated in limiting the success of clam-based treatment systems utilizing municipal sewage

treatment plant effluentsby Haines (1979). Ammonia toxicity is of great concern in all

aquaculture systems (Harris et al. 1998), especially recirculating ones where technologies









utilizing treatment by volatilization, sequestration or conversion have been developed to

encourage higher density cultures of fish and other aquatic organisms (Timmons et al. 2002).

Lack of ammonia management technologies tested at large scale for application in clam raceway

systems will be an important goal in future development of freshwater animal aquaculture

applications for dairy waste treatment. Bivalve-based treatment of dairy-derived wastewater

phosphorus would require implementation and integration of additional treatment technologies in

order to reduce high levels of nitrogenous wastes common in dairy operations.

Application of vegetative systems such as wetlands, managed aquatic plant systems and

biofilm systems may be able to provide needed treatment of ammonia prior to clam raceway

addition in order to increase feasibility. Coupling of vegetative and clam-based systems in this

way would lower water demand while decreasing ammonia and utilizing less treatment surface

area than required by clam systems alone. Using vegetative-type filters along with bivalves has

also been demonstrated in marine mariculture wastewater treatment systems to remove dissolved

phosphorus from the outflow of bivalve-based systems (Jones et al. 2001, Borges et al. 2005) and

has been suggested for Corbicula culture in agriculture wastewater by Stanley (1974).

Consideration should be given to source pond geology, depth, sediment permeability and

use of clay or plastic liners in future systems to limit or promote possible exchange of pond

water nutrients and heat with the surrounding environment to maintain tolerable environmental

conditions. Seasonality of system function is a central issue not only from the standpoint of

summer high temperatures, but from growth and survival, issues under low winter temperatures.

Despite the lack of Corbicula success on a large scale, the raceway-based recirculation

system design demonstrated in this study provided a dependable, easy to construct and reusable

platform for testing aquaculture potential of organisms in wastewater treatment conditions at









large scale. The design should be used as a standard system to assess other organisms besides

Corbicula for comparative purposes since systems can be easily constructed to accommodate a

variety of operating parameters. The raceway system designs employed here are versatile enough

to be applied to other organisms such as bivalves, fish, algae and high plants; targeted for large-

scale water treatment/biofiltration studies in both fresh and saltwater conditions and a variety of

locations, as well as effluent sources.

From a comparative standpoint, it is important to make the observation that the phosphorus

removal capacities of many animal-based systems, which typically range from 17 to 5840

mg/m2/yr, are low compared to the best-case estimates for algae and other plant based systems,

which range from 320 to 365,000 mg/m2/yr. However, this comparison is somewhat misleading,

since the function and structure of the two systems are different in several important aspects.

Plant/algae systems provide a mechanism for removal of particulate and dissolved nutrient forms

as opposed to animal-based systems that focus on particulates. Also, plant/algae systems are

energetically dependent on sunlight, while animal systems can be independent of such a direct

requirement. In addition, the end products of plant/algae and animal systems are fundamentally

different and subject to separate end-use issues and opportunities. Ultimately, many treatment

needs may be best addressed by integration of animal and plant/algae systems, thereby allowing

for optimal processing of soluble and particulate wastes and production of a wide range of

valuable goods and services, such as food, feed, biodiesel, building materials and chemicals.









LIST OF REFERENCES


Adey, W., C. Luckett and K. Jensen. 1993. Phosphorus removal from natural waters using
controlled algal production. Restoration Ecology March 1993:29-38.

American Public Health Association (APHA). 1998. Standard methods for the examination of
water and wastewater, 20th edition. Section 4500-P B5 & E.

Anthony, J.L., D.H. Kesler and W.L. Downing. 2001. Length-specific growth rates in
freshwater mussels (Bivalvia: Unionidae): extreme longevity or generalized growth
cessation?. Freshwater Biology 46:1349-1359.

Azim M.E., M.A. Wahab and A.A. van Dam. 2001. Optimization of fertilization rate for
maximizing periphyton production on artificial substrates and the implications for
periphyton-based aquaculture. Aquaculture Research 32:749-760.

Azim, M.E., A. Milstein, M.A. Wahab, M.C.J. Verdegam. 2003. Periphyton-water quality
relationships in fertilized fishponds with artificial substrates. Aquaculture 228:169-187.

Barber, L.B., S.H. Keefe and R.C. Antweiler. 2006. Accumulation of contaminants in fish from
wastewater treatment wetlands. Environmental Science and Technology 40:603-611.

Beaver, J.R., T.L. Crisman and R.J. Brock. 1991. Grazing effects of an exotic bivalve
(Corbicula fluminea) on hypereutrophic lake water. Lake and Reservior Management
7(1):45-51.

Belanger, S.E.. 1991. The effect of dissolved oxygen, sediment and sewage treatment plant
discharges upon growth, survival and density of Asiatic clams. Hydrobiologia 218:113-
126.

Bitterman, A.M., R.D. Hunter and R.C. Haas. 1994. Allometry of shell growth of caged and
uncaged zebra mussels (Dreissenapolymorpha) in Lake St. Clair. American
Malacological Bulletin 11 (1):41-49.

Blalock, H.N. and J.J. Herod. 1999. A comparative study of stream habitat and substrate utilized
by Corbiculafluminea in the New River, Florida. Florida Scientist 62(2):145-151.

Brousseau, D.J.. 1979. Analysis of growth rate in Mya arenaria using the Von Bertalanffy
equation. Marine Biology 51(3):221-227.

Brock, R.J.. 2000. Assessment of aquatic food web alterations in the presence of the exotic
clam, Corbiculafluminea and the cichlid, Oreochromis aureus. PhD Dissertation.
University of Florida, Gainesville, FL. 217 pp.

Bunting, S.W.. 2007. Confronting the realities of wastewater aquaculture in peri-urban Kolkata
with bioeconomic modeling. Water Research 41:499-505.









Busch, R.L.. 1974. Asiatic clams Corbicula manilensis (Phillippi) as biological filters in channel
catfish, Icaluruspunctatus (Rafinesque) cultures. MS Thesis. Auburn University. 84 pp.

Buttner, J.K.. 1986. Biology of Corbicula in catfish rearing ponds. Proceedings of the Second
International Corbicula Symposium. The American Malacological Bulletin. Special
Edition No. 2:211-218.

Buttner, J.K. and R.C. Heidinger. 1980. Seasonal variations in growth of the Asiatic clam,
Corbiculafluminea (bivalvia: corbiculidae) in southern Illinois fish pond. The Nautilus
94(1):8-10.

Cataldo, D.H., D.Boltovskoy, J. Stripekis and M. Pose. 2001. Condition index and growth rates
of field caged Corbiculafluminea (Bivalvia) as biomarkers of pollution gradients in the
Parana River delta (Argentina). Aquatic Ecosystems Health and Management 4:187-201.

Cavallo, D., A. Pusceddu and A. Giangrande. 2007. Particulate organic matter uptake rates of
two benthic filter-feeders (Sabella spallanzanii and Branchiomma luctuosum) candidates
for the clarification of aquaculture wastewaters. Marine Pollution Bulletin 54(5):622-625.

Chen, T.P.. 1976. Culture of the freshwater clam Corbiculafluminea. IN: Aquaculture
Practices in Taiwan. pp. 106-111.

Copar, S.G. and N.J. Yess. 1996. Survey of elements in clams and oysters. Food Additives and
Contaminants 13(4):553-560.

Covich, A.P. and J.H. Thorpe. 2001. Introduction to the subphyllum Crustacea. IN: Ecology and
Classification of North American Freshwater Invertebrates, J.H. Thorp and A.P. Covich,
eds., Academic Press, San Diego, CA. pp. 777-809.

Cox, G.W.. 1996. Laboratory manual of general ecology, 7th edition. Wm.C. Brown Publishers,
Bubuque, IA. 278 pp.

Craggs, R.J., W.H. Adey, B.K. Jessup and W.J. Oswald. 1996. A controlled stream mesocosm
for tertiary treatment of sewage. Ecological Engineering 6:149-169.

Dame, R.F.. 1996. Ecology of marine bivalves an ecosystem approach. CRC Marine Science
Series. CRC Press, Boca Raton, FL. 254 pp.

Dao, T.H., A. Lugo-Ospina and J.B. Reeves III. 2006. Wastewater chemistry and fractionation
of bioactive phosphorus in dairy manure. Communications in Soil Science and Plant
Analysis 37:907-924.

DeBusk, T.A., K.A. Grace and F.E. Dierberg. 2004. An investigation of the limits of phosphorus
removal in wetlands: a mesocosm study of a shallow periphyton-dominated treatment
system. Ecological Engineering 23:1-14.









Debrosse, G.A. and S.K. Allen, Jr.. 1994. The suitability of land based evaluations of
Crassostrea gigas as an indicator of performance in the field. Abstracts, National
Shellfisheries Association, 1994 Annual Meeting, Charleston SC. p. 277.

Dempster, P., D.J. Baird and M.C.M. Beveridge. 1995. Can fish survive by filter-feeding on
microparticles? Energy balance in tilapia graxing on algal suspensions. Journal ofFish
Biology 47:7-17.

Drapcho, C.M. and D.E. Brune. 2000. The partitioned aquaculture system: impact of design and
environmental parameters on algal productivity and photosynthetic production.
Aquacultural Engineering 21:151-168.

Edwards, D.D. and R.V. Dimock, Jr.. 1988. A comparison of the population dynamics of
Unionicolaformosa from anodontine bivalves in North Carolina farm pond. Journal of
Elisha Mitchell Science Society 104:70-78.

Enes, P. and M.T. Borges. 2003. Evaluation of microalgae and industrial cheese whey as diets
for Tapes decussates (L.) seed: effects on water quality, growth, survival, condition and
filtration rate. Aquaculture Research 34:299-309.

Eng, L.L. 1976. A note on the occurrence of a symbiotic oligochaete, Chaetogaster limnaei in
the mantle cavity of the Asiatic clam, Corbicula manilensis. The Veliger 19(2):208.

Epifanio, E.E. and R. Sma. 1975. Toxicity of ammonia, nitrite ion, nitrate ion and
orthophosphate to Mercenaria mercenaria and Crassostrea virginica. Marine Biology
33:231-236.

Fisher, G.R., R.E. Kuhn and R.V. Dimock, Jr.. 2000. The symbiotic watermite Unionicola
formosa (Acari-Unionicolidae) ingests mucus and tissue of its molluscan host. Journal of
Parasitology 86:1254-1258.

Foe, C. and A. Knight. 1985. The effect of phytoplankton and suspended sediment on the
growth of Corbiculafluminea (Bivalvia). Hydrobiologia 127:105-115.

Fuji, A.. 1979. Phosphorus budget in natural population of Corbiculajaponica prime in
poikilohaline lagoon, Zyusan-ko. Bulletin of the Faculty of Fisheries, Hokkaido
University 30(1):34-49.

Gainey, L.F.. 1978. The response of the Corbiculidae (Mollusca: Bivalvia) to osmotic stress: the
organismal response. Physiological Zoology 51(1):68-78.

Gardner, J.A., Jr., W.R. Woodall, Jr and A.A. Staats, Jr.. 1976. The invasion of the Asiatic clam
(Corbicula manilnsis Philippi) in the Altamaha River, Georgia. Nautilus 90:117-125.

Ghaly, A.E., M. Kamal and N.S. Mahmoud. 2005. Phytoremidiation of aquaculture wastewater
for water recycling and production of fish feed. Environment International 31:1-13.









Gifford, S., R.H. Dunstan and W.O'Conner. 2004. Pearl aquaculture- profitable environmental
remediation?. The Science of the Total Environment 319:27-37.

Greer, D.E. and C.D. Ziebell. 1972. Biological removal of phosphates from water. Journal
Water Pollution Control Federation 44(12):2342-2348.

Habel, M.L.. 1970. Oxygen consumption, temperature tolerance, and filtration rate of the
introduced Asiatic clam Corbicula manilensis from the Tennessee River. MS Thesis.
Auburn University, USA. 91 pp.

Haines, C.M.C., M. Edmunds and A.R. Pewsey. 1994. The pea crab, Pinnotherespisum
(Linnaeus, 1767), and its association with the common mussel, Mytilus edulis (Linnaeus,
1758), in the Solent (UK). Journal of.hell//pli Research 13(1):5-10.

Haines, K.C.. 1977. The use of Corbicula as a clarifying agent in experimental tertiary sewage
treatment process on St. Croix, U.S. Virgin Islands. Proceedings, First International
Corbicula Symposium. Texas Christian University, USA. pp. 165-175.

Harris Jr., B., D. Morse and H.H. Head. 1990. Phosphorus nutrition and excretion by dairy
animals. University of Maryland, USA website: http. www.inform.umd.edu. 11 pp.

Hickman, R.W.. 1992. Mussel cultivation, Chapter 13. IN: The mussel Mytilus-ecology,
physiology, genetics and culture. Gosling, E., ed. Elsevier, New York, NY. 598 pp.

Huchette, S.M.H., C.S. Koh and R.W. Day. 2003. Growth of juvenile blacklip abalone (Haliois
rubra) in aquaculture tanks: effects of density and ammonia. Aquaculture 219:457-470.

Ingram, W.M.. 1959. Asiatic clams as potential pests in California water supplies. Journal of the
American Water Association 51:363-370.

Islam, M.S., M.S. Kabir and S.I. Khan. 2004. Wastewater-grown duckweed may be safely used
as fish feed. Canadian Journal of Microbiology 50:51-56.

Jakobworks, G.S., G.D. Pruder and J. Wang. 1993. Growth trial with the American Oyster
Crassostrea virginica using shrimp pond water as feed. Journal of the World
Aquaculture Society 24(3): 344-351.

Jara-Jara, R., A.J. Pazos and M. Abad. 1997. Growth of clam seed (Ruditapes decussates) reared
in wastewater effluent from a fish farm in Galicia (NW Spain). Aquaculture 158:247-262.

Jara-Jara, R., M. Abad and A.J. Pazos. 2000. Growth and reproductive patterns in Venerupis
pullastra seed reared in wastewater effluent from a fish farm in Galicia (N.W. Spain).
Journal of .\hel//i h Research 19(2):949-956.

Johnson, A.F., D.M. Vietor and M. Rouquette, Jr.. 2004. Fate of phosphorus in dairy wastewater
and poultry litter applied on grassland. Journal of Environmental Quality 33:735-739.









Jones, T. and G.K. Iwama. 1991. Polyculture of the Pacific oyster, Crassostrea gigas
(Thunberg), with Chinook salmon, Oncorhynchus tshawytscha. Aquaculture 92:313-322.

Jones, A.B. and N.P. Preston. 1999. Sydney rock oyster, Saccostrea commercials (Iredale and
Roughley), filtration of shrimp farm effluent: the effects on water quality. Aquaculture
Research 30(1):51-57.

Jones, A.B., W.C. Dennison and N.P. Preston 2001. Integrated treatment of shrimp effluent by
sedimentation, oyster filtration and macroalgal absorption: A laboratory scale study.
Aquaculture 193:155-178.

Joy, J.E. and L.E. McCoy. 1975. Comparisons of shell dimensions and viscera mass weights in
Corbicula manilensis (Philippi, 1844). Nautilus 89:51-54.

Kebede-Westhead, E., C. Pizarro, W.W. Mulbry and A.C. Wilkie. 2003. Production and nutrient
removal by periphyton grown under different loading rates of anaerobically digested
flushed dairy manure. Journal of Phycology 39:1275-1282.

Kinne, P.N., T.M. Samocha and E.R. Jones. 2001. Characterization of intensive shrimp pond
effluent and preliminary studies on biofiltration. North American Journal ofAquaculture
63:25-33.

Knight, R.L., V.M.E. Payne Jr. and R.E. Borer. 2000. Constructed wetlands for livestock
wastewater management. Ecological Engineering 15(1-2):41-55.

Kraemer, L.R.. 1979. Corbicula (Bivalvia: Sphaeriacea) vs indigenous mussels (Bivalvia:
Unionacea) in U.S. rivers: a hard case for interspecific competition. American Zoologist
19:1085-1096.

Kraemer, L.R., C. Swanson, M. Galloway and R. Kraemer. 1986. Biological basis of behavior in
Corbiculafluminea, II. Functional morphology of reproduction and development and
review of evidence for self-fertilization. Proceedings of the Second International
Corbicula Symposium. The American Malacological Bulletin. Special Edition No. 2; 193-
201.

Krom, M.D., S. Ellner, J. van Rijn and A. Neori. 1995. Nitrogen and phosphorus cycling and
transformations in a prototype 'non-polluting' integrated mariculture system, Eilat, Israel.
Marine Ecology Progress Series 18:25-36.

Kurth, J., C. Loftin and J. Zydlewski. 2007. PIT tags increase effectiveness of freshwater mussel
recaptures. Journal of the North American Benthological Society 26(2):253-260.

Lansing, S.L. and J.F. Martin. 2006. Use of an ecological treatment system (ETS) for removal of
nutrients from dairy wastewater. Ecological Engineering 28:235-245.

Lauritsen, D.D.. 1985. Filter-feeding, food utilization and nutrient remineralization by
Corbiculafluminea (Bivalvia) and its contribution to nutrient cycling in a North Carolina
river. PhD Dissertation. North Carolina State University, Raleigh, NC. 130 pp.









Lauritsen, D.D., S. Mozley. 1989. Nutrient excretion by the Asiatic clam Corbiculafluminea.
The North American Benthological Society 8:134-139.

Layzer, J.B. and J.R. Heinricher. 2004. Coded wire tag retention in ebonyshell mussels
Fusconaia ebena. North American Journal of Fisheries Management 24:228-230.

Lefebvre, S., L. Barille and M. Clere. 2000. Pacific oyster (Crassostrea gigas) feeding
responses to a fish-farm effluent. Aquaculture 187:185-198.

Lemarie, D.P., D.R. Smith and R.F. Villella. 1995. Evaluation of tag types and adhesives for
marking freshwater mussels. The Conservation and Management of Freshwater Mussels
II. Initiatives for the Future, October 16-18, St. Louis Missouri. National Biological
Service: 14pp.

Lin, J., M. Sparis and R.W. Hagood. 2001. Growing giant clam (Tridacna derasa) in
aquaculture effluent. Aquarium Sciences and Conservation 3:225-230.

Lincoln, E.P., J.W. Crawford and A.C. Wilkie. 1993. Spirulina in animal agriculture. Bulletin de
l'Institut oceanographique, Monaco 12:109-115.

Lincoln, E.P., A.C. Wilkie and B.T. French. 1996. Cyanobacterial process for renovating dairy
wastewater. Biomass andBioenergy 10(2):63-68.

Lorio, W.J. and S. Malone. 1995. Biology and culture of the northern quahog clam Mercenaria
mercenaria. USDA Southern Regional Aquaculture Center. Publication No. 433. 4 pp.

MacMillan, R.J., R.J. Cawthorn and S.K. Whyte. 1994. Design and maintenance of a closed
artificial seawater system for long-term holding of bivalve shellfish. Aquacultural
Engineering 13:241-250.

Marcussen, H., P.E. Holm and L.T. Ha. 2007. Food safety aspects of toxic element
accumulation in fish from wastewater-fed ponds in Hanoi, Vietnam. Tropical Medicine
and International Health 12(2):34-39.

Mattice, J.S.. 1977. Interactions of Corbicula sp. with power plants. Proceedings, First
International Corbicula Symposium. Texas Christian University. pp. 120-138.

Mattice, J.S., and L.L. Wright. 1986. Aspects of growth of Corbiculafluminea. Proceedings of
the Second International Corbicula Symposium. The American Malacological Bulletin.
Special Edition No. 2: 167-178.

Mazzola, A. and G. Sara. 2001. The effect of fish farming organic waste on food availability for
bivalve mollusks (Gaeta Gulf, Central Tyrrhenian, MED): stable carbon isotopic analysis.
Aquaculture 192:361-379.

McCorkel, S. and T.H. Dietz. 1980. Sodium transport in the fresthwater Asiatic clam Corbicula
fluminea. Biological Bulletin 159:325-336.









McMahon, R.F. 1983. Ecology of an invasion pest bivalve, Corbicula. IN: The Mollusca, vol 6,
W.D. Russell-Hunter, ed.. Academic Press, San Diego, CA. pp. 505-561.

McMahon, R.F. 2002. Evolutionary and physiological adaptations of aquatic invasive animals: r
selection versus resistance. Canadian Journal of Fisheries and Aquatic Science 59:1235-
1244.

McMahon, R.F. and C.J. Williams. 1986. A reassessment of growth rate, life span, life cycles
and population dynamics in a natural population and field caged individuals of
Corbicula fluminea (Muller) (Bivalvia: Corbiculacea). Proceedings of the Second
International Corbicula Symposium. The American Malacological Bulletin. Special
Edition No. 2:156-166.

McMahon, R.F., Bogan, A.E.. 2001. Mollusca: Bivalvia. IN: Ecology and Classification of
North American Freshwater Invertebrates, J.H. Thorp and A.P. Covich, eds. Academic
Press, San Diego, CA. pp. 331-429.

Mercado-Silva, N.. 2005. Condition index of the eastern oyster, Crassostrea virginica (Gmelin,
1791) in Sapelo Island, Georgia- effects of site, position on bed and pea crab parasitism.
Journal of\he//ll/i h Research (24)1:121-126.

Miller, R.C. and F.A. McClure. 1931. The fresh-water clam industry of the Pearl River.
Lingham Science Journal 10(2 & 3):307-322.

Mulbry, W.W. and A.C.Wilkie. 2001. Growth of benthic freshwater algae on dairy manures.
Journal of Applied Phycology 13:301-306.

Neori, A., M. Shpigel and D. Ben-Ezra. 2000. A sustainable integrated system for culture of
fish, seaweed and abalone. Aquaculture 186: 279-291.

Nizzoli, D., M. Bartoli and P. Viaroli. 2006. Nitrogen and phosphorus budgets during a farming
cycle of the manila clam Ruditapes philippinarum: an in situ experiment. Aquaculture
261:98-108.

National Resource Conservation Service. (NRCS). 1999. National Engineering Handbook Part
651. Agricultural Waste Management Field Handbook. US Department of Agriculture,
Natural Resources Conservation Service, Washington, DC.

National Resource Conservation Service. (NRCS). 2006. National Engineering Handbook Part
651. Agricultural Waste Management Field Handbook. US Department of Agriculture,
Natural Resources Conservation Service, Washington, DC.

Olszewski, M.. 1977. The potential use of power plant reject heat in commercial aquaculture.
ORN/TM-5663. Engineering Technology Division, Oak Ridge National Laboratory.

Patrick, R., J. Cairns, Jr. and A. Scheier. 1968. The relative sensitivity of diatoms, snails and
fish to twenty common constituents of industrial wastes. Progressive Fish Culture
30:137-140.









Phelps, H.L. 1994. Potential for Corbicula in aquaculture. Journal of.hell//ph Research
13(1):319.

Phlips, E.J., S. Badylak and T. Grosskopf. 2002. Factors affecting the abundance of
phytoplankton in a restricted subtropical lagoon, the Indian River Lagoon, Florida, USA.
Estuarine, Coastal and Shelf Science 55 (3):385-402.

Pizarro, C., E. Kebede-Westhead and W. Mulbry. 2002. Nitrogen and phosphorus removal rates
using small algal turfs grown with dairy manure. Journal of AppliedPhycology 14:469-
473.

Prein, M.. 2002. Integration of aquaculture into crop-animal systems in Asia. Agricultural
Systems 71:127-146.

Reddy, K.R. and W.H. Smith. 1987. Aquatic plants for water treatment and resource recovery.
Magnolia Publishing Inc. 1032 pp.

Rodgers Jr, J.H., D.S. Cherry, K.L. Dickson and J. Cairns. 1977. Invasion, population dynamics
and elemental accumulation of Corbiculafluminea in the New river at Glen Lyn,
Virginia. Proceedings, First International Corbicula Symposium. Texas Christian
University. pp. 99-110.

Sartory, D.P. and J.U. Grobbelaar. 1984. Extraction of chlorophyll a from freshwater
phytoplankton for spectrophotometric analysis. Hydrobiologia 144(3): 177-187.

Schmidlin, S. and B. Baur. 2007. Distribution and substrate preference of the invasive clam
Corbiculafluminea in the river Rhine in the region of Basel (Switzerland, Germany,
France). Aquatic Sciences 69:153-161.

Scinto, L.J. and K.R. Reddy. 2003. Biotic and abiotic uptake of phosphorus by periphyton in a
subtropical freshwater wetland. Aquatic Botany 77:203-222.

Sharpley, A.N., S.C. Chapra and J.T. Wedepohl. 1994. Managing agricultural phosphorus for
protection of surface waters: issues and options. Journal of Environmental Quality
23:437-451.

Shpigel, M. and R. Fridman. 1990. Propagation of the Manila clam (Tapes semidecussatus) in
the effluent of fish aquaculture ponds in Eilat, Israel. Aquaculture 90:113-122.

Shpigel, M. and R.A. Blaylock. 1991. The Pacific oyster, Crassostrea gigas, as a biological
filter for a marine fish aquaculture pond. Aquaculture 92:187-197.

Shpigel, M., A. Neori, D.M. Popper and H. Gordin. 1993. A proposed model for
"environmentally clean" land-based culture of fish, bivalves and seaweeds. Aquaculture
117:115-128.









Shpigel, M. and A. Neori. 1996. The integrated culture of seaweed, abalone, fish and claims in
modular intensive land-based systems: I. Proportions of size and projected revenues.
Aquaculture Engineering 15(5):3 13-326.

Shpigel, M., A. Gasith and E. Kimmel. 1997. A biomechanical filter for treating fish-pond
effluents. Aquaculture 152:103-117.

Sickel, J.B.. 1986. Corbicula population mortalities: factors influencing population control.
Proceedings of the Second International Corbicula Symposium. The American
Malacological Bulletin. Special Edition No. 2:89-94.

Silverman, H., E.C. Achberger and J.W. Lynn. 1985. Filtration and utilization of laboratory-
cultured bacteria by Dreissenapolymorpha, Corbiculafluminea and Carunculina
texasensis. Biological Bulletin 189:308-319.

Sindilariu, P.D.. 2007. Reduction in effluent nutrient loads from flow-through facilities for trout
production: a review. Aquaculture Research 38:1005-1036.

Sokal, R.R. and F.J. Rohlf. 1995. Biometry, the principles and practice of statistics in biological
research, 3rd Edition. W.H. Freeman and Company, New York, NY. 887 pp.

Sooknah, R.D. and A.C. Wilkie. 2004. Nutrient removal by floating aquatic macrophytes
cultured in anaerobically digested flushed dairy manure wastewater. Ecological
Engineering. 22:27-42.

Soto, D. and G. Mena. 1999. Filter feeding by the freshwater mussel, Diplodon chilensis, as a
biocontrol of salmon farming eutrophication. Aquaculture 171: 65-81.

Stanley, R.A.. 1974. Methods of biological nutrients from livestock waste: a literature review
and systems analysis. Environmental Biology Branch in cooperation with Agricultural
Resource Development Branch, Tennessee Valley Authority. Pub. Y-80. 45 pp.

Stuart, K.R., A.G. Eversole and D.E. Brune. 2001. Filtration of green algae and cyanobacteria
by freshwater mussels in the partitioned aquaculture system. Journal of the World
Aquaculture Society 23 (1): 105-111.

Swingle, H.S.. 1966. Fish kills caused by phytoplankton blooms and their prevention. FAO
World Symposium on Warm Water Fish Culture. Rome, Italy. 11 pp.

Timmons, M.B., J.M. Ebeling and F.W. Wheaton. 2002. Recirculating aquaculture systems, 2nd
Edition. USDA Northern Regional Aquaculture Cener, Pub. No. 01-002. 769 pp.

Toll, R.B., R.S. Prezant and H.B. Rollins. 2003. A novel method for locating tagged infaunal
bivalves: submersible pulse technology metal detectors. Journal of .he//Jfi/h/ Research
22(2):501-503.

Van Rijn, J.. 1996. The potential for integrating biological treatment systems in recirculating
fish culture- a review. Aquaculture 139:181-201.









Vantaram, A.. 2004. Honeycomb fiber-reinforced polymer sandwich components for
development of aquaculture raceway systems. MS thesis. West Virginia University. 107
pp.

Villadolid, D.V., and F.G. Del Rosario. 1930. Some studies on the biology of tulla (Corbicula
manillensis Philippi) a common food clam of Lagunade Bay and its tributaries. Phillipine
Agriculturalist 19:355-382.

Wargo, R.N. and S.E. Ford. 1993. The effect of shell infestation by Polydora sp. and infection
by Haplosporidium nelsoni (MSX) on the tissue condition of oysters, Crassostrea
virginica. Estuaries 16(2):229-234.

Wetzel, R.G.. 2001. Limnology, lake and river ecosystems, 3rd Edition. Academic Press, San
Diego. 1006 pp.

Wilkie, A.C. and W.W. Mulbry. 2002. Recovery of dairy manure nutrients by benthic
freshwater algae. Bioresource Technology 84:81-91.

Wilkie, A.C. 2003. Anaerobic digestion of flushed dairy manure. Proceedings- Anaerobic
Digester Technology Applications in Animal Agriculture- A National Summit. Water
Environment Federation, Alexandria, VA. pp. 350-354.

Wilkie, A.C., H.F. Castro and K.R. Cubinski. 2004. Fixed-film anaerobic digestion of flushed
dairy manure after primary treatment: wastewater production and characterization.
Biosystems Engineering 89(4):457-471.

Williams, C.J. and R.F. McMahon. 1986. Power station entrainment of Corbiculafluminea in
relation to population dynamics, reproductive cycle and biotic and abiotic variables.
Proceedings of the Second International Corbicula Symposium. The American
Malacological Bulletin. Special Edition No. 2:99-111.

Wood, J., G. Fernandez and A. Baker. 2007. Efficiency of reed beds in treating dairy
wastewater. Biosystems Engineering 98:455-469.

Zhou, Y., H. Yang and T. Zhang. 2006. Density-dependent effects on seston dynamics and rates
of filtration and biodeposition of the suspension-cultured scallop Chlamysfarreri in a
eutrophic by (northern China): An experimental study in semi-in situ flow-through
systems. Journal of Marine Systems 59:143-158.









BIOGRAPHICAL SKETCH

Lance W Riley was born at the US Navy Submarine base in Guam, Mariannas Islands in

1974 to Capt. (USN) Roy "Luke" Riley and his lovely wife, Linda Cline Riley. The family

eventually moved to Gold Hill, North Carolina, and Lance graduated from Mount Pleasant High

School, Mount Pleasant, NC in 1992. Lance received his Bachelor of Science degree in

Environmental Biology at the University of North Carolina at Charlotte in 1998. He then earned

his Master of Science degree in Environmental Engineering Sciences with a Graduate Certificate

in Wetlands at the University of Florida in 2000. Lance is currently employed at the University

of Florida Fisheries and Aquatic Sciences Department where he performs analytical and field

research using laboratory experiments and in-situ monitoring of waterways throughout Florida.





PAGE 1

1 FRESHWATER CLAMS AS A TREATMENT MECHANISM FOR PHOSPHORUS IN AGRICULTURAL WASTEWATER By LANCE W RILEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Lance W Riley

PAGE 3

3 To my parents, Captain Roy Luke Riley (United States Navy, Retired) and Linda C. Riley, Thank you for all of your love, support and confidence, I couldnt have done this without you both, I love you

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank m y committee chair, Dr. Edward Phlips, the co-chair, Dr. Ann Wilkie, and the rest of my committee (Dr. Tom Crisman, Dr. Roger Nordstedt, Dr. Shirley Baker and Dr. Patrick Baker) for their help and guidance. Specia l thanks goes out to Ivan Mish, Jon Mish, the students and administration at Alee Academy (Umatilla, FL), all of the other volunteers for helping with clam stocking, raceway monitoring and amphipod interaction investigation and Bill Lindberg. Special thanks also to Dr. Phil Barkley, Dr. Kelly Foote and all of the staff at the UF Student Health Center and Shands Neurosurge ry Department, look at the bionhick man go! Most of all, I want to give a ve ry special thank you to my parent s, all of my family and friends that made this possible. Funding for this re search was provided by the following entities: United States Department of Agriculture (USDA-CSREES Special Research GrantFreshwater clams as tertiary tr eatment for agriculture wastewater. E. Phlips, S. Baker, P. Lazur. 2001-2004. $80,000) United States Department of Agriculture (USDA-CSREES Special Research GrantIntegrating clams into a dairy wastewater treatment train. E. Phlips and P. Baker. 20032004. $79,824) University of Florida Department of Fisher ies and Aquatic Sciences (Project Facility Construction Grant. R. Riley 2002. $30,000)

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4 LIST OF TABLES ...........................................................................................................................8 LIST OF FIGURES .......................................................................................................................10 ABSTRACT 12 CHAP TER 1 INTRODUCTION .................................................................................................................. 14 2 RACEWAY-BASED RECIRCULATING WA STEWATER TREATMENT SYSTEM DESIGN AND CONSTRUCTION ........................................................................................ 20 Introduction .................................................................................................................. ...........20 Raceway Design .....................................................................................................................25 Blountstown Facility .......................................................................................................... .....29 Hague Facility .........................................................................................................................32 System Design Summary ........................................................................................................35 3 ADAPTABILITY OF Corbicula TO TREATMENT RACEWAYS ..................................... 39 Introduction .................................................................................................................. ...........39 Methods ..................................................................................................................................41 Raceway-based Treatment System .................................................................................. 42 Source ponds ............................................................................................................42 Raceways .................................................................................................................. 44 Water analysis ..........................................................................................................45 Clam Population Dynamics In The Raceway Environment ............................................ 46 Stocking clam raceways ........................................................................................... 47 Clam raceway population sampling ......................................................................... 49 Tagged clams ............................................................................................................51 Clam survival ........................................................................................................... 53 Biomass changes ...................................................................................................... 53 Reproduction and recruitment ..................................................................................57 Health ....................................................................................................................... 57 Results .....................................................................................................................................60 Raceway System Environmental Parameters .................................................................. 60 Clam Population Dynamics In Treatment Raceways ...................................................... 66 Survival .................................................................................................................... 66 Growth ......................................................................................................................68 Reproduction and recruitment ..................................................................................77 Health ....................................................................................................................... 78 Amphipod infestation ............................................................................................... 79

PAGE 6

6 Discussion .................................................................................................................... ...........81 Ammonia Concerns .........................................................................................................82 Temperature ................................................................................................................... ..84 Food Availability .............................................................................................................86 Dissolved Oxygen ........................................................................................................... 88 Multiple Stressors ............................................................................................................ 88 Parasites and Predation ....................................................................................................89 Reproductive Success ...................................................................................................... 93 Clam Stock Assessment Issues ........................................................................................ 94 General Conclusion .........................................................................................................96 4 PHOSPHORUS REMOVAL AND SEQUEST RATION IN CLAM RACEWAYS ............. 98 Introduction .................................................................................................................. ...........98 Methods ................................................................................................................................102 Raceway-based Treatment System ................................................................................ 102 Source water ponds ................................................................................................103 Raceways ................................................................................................................ 105 Water quality monitoring ....................................................................................... 105 Raceway clam populations ..................................................................................... 107 P Removal From Source Water By Clam Raceways .................................................... 108 Raceway through-flow trials ..................................................................................108 Raceway water recirculation ..................................................................................112 Sequestration of phosphorus by cl am s in treatment raceways ...............................114 Results ...................................................................................................................................116 Raceway Environmental Conditions ............................................................................. 116 Raceway clam populations ..................................................................................... 118 Phosphorus Uptake In Clam Raceways ......................................................................... 118 Raceway through-flow input/outpu t measurements ............................................... 118 Raceway recirculation measurements .................................................................... 120 Sequestration Of Phosphorus By Clam s In Treatment Raceways .................................124 Phosphorus allocation in clam biomass ................................................................. 125 Treatment raceway clam population phosphorus ................................................... 127 Discussion .................................................................................................................... .........128 Distribution of Phosphorus Taken Up By Clams .......................................................... 128 Estimates of Phosphorus Uptake Rates ......................................................................... 131 Comparison Of Phosphorus Removal By Cla m Raceways And Other Systems .......... 133 Problems With Measuring Shor t-term Phosphorus Uptake .......................................... 136 Dairy Application Demands And Issues ....................................................................... 139 Sustainability ................................................................................................................ .143 5 SUMMARY ....................................................................................................................... ...147 Raceway Function and Attributes ......................................................................................... 147 Adaptability of Clams to Raceway Conditions .................................................................... 147 P-removal Capacity ..............................................................................................................148 Future Applications ..............................................................................................................150

PAGE 7

7 LIST OF REFERENCES .............................................................................................................153 BIOGRAPHICAL SKETCH .......................................................................................................163

PAGE 8

8 LIST OF TABLES Table page 2-1 Raceway dimensions and capacities as tes ted, availabl e capacities adjusted for standpipe presence .............................................................................................................27 3-1 Source pond and raceway numerical designa tions for the treatm ent systems at the Hague site...........................................................................................................................42 3-2 Raceway (RW) stocking and population sa m pling schedule for the low, medium and high nutrient addition treatment systems ........................................................................... 46 3-3 Number of clams stocked in each raceway estim ated using the volumetric method ......... 66 3-4 Number of live clams found alive at each sam pling interval estimated using the spatial technique .................................................................................................................67 3-5 Actual number of live clams stocked in each raceway and at the end of the study ........... 68 3-6 Mean shell lengths measured from clams in each raceway (RW) both at stocking and at each sampling in terval ................................................................................................... 70 3-7 Shell growth rates for tagged clams in each nutrient addition treatm ent for each seasonal time interval .........................................................................................................72 3-8 Shell size information on clams sampled for tissue biomass analysis from each nutrient addition treatm ent raceway system ...................................................................... 72 3-9 Mean and range of ash content values fo r m eat, shell and total clam tissues pooled for all clams sampled ......................................................................................................... 73 3-10 Results of the meat, shell and total cl am tissue dry weight (DW) to shell length correlation analysis ............................................................................................................73 3-11 Dry weight (DW) biomass vs length regression relations hips, significance differences and variability for whole clam shell and meat tissues from each nutrient addition treatment ............................................................................................................ ..74 3-12 Mean, standard error (SE) and range of condition indices values (CI(WT) and CI(VOL)) calculated for the medium nut rient addition treatment at Interval 1 compared to values calculated at all other treatment/interval combination ............................................ 78 3-13 Individual shell length a nd biom ass dry weight (DW) growth rates reported for Corbicula and other bivalves occupying differe nt fresh and saline environments ............ 81 4-1 Source pond and raceway numerical designa tions for the treatm ent systems at the Hague site.........................................................................................................................103

PAGE 9

9 4-2 Raceway source water input flow rates for the p eriod of July 1 to August 24, 2002 ...... 109 4-4 Monthly mean input total phosphorus (TP) concentrations in raceways and standard error (SE), at time 0 in the recirculation trials for each nutrient addition treatment pond groups ......................................................................................................................120 4-5 Raceway system monthly mean input to tal dissolved phosphorus (TDP) at tim e zero in the recirculation trials for each nutrient addition treatment ......................................... 121 4-6 Total dissolved phosphorus (TDP) removal rates calculated from TDP slopes in the recirculation trials for each nutrient add ition treatment system during April and May 2003..................................................................................................................................122 4-7 Raceway total dissolved phosphorus (TDP) va lues at tim e 0 for covered raceways in the low and high nutrient addition treatments .................................................................. 122 4-8 Raceway chlorophyll a (chl a) values at tim e 0 for raceways in the low and high nutrient addition treatments during April and May 2003 ................................................ 123 4-9 Raceway chlorophyll a (chl a) values at tim e 0 for covered raceways in the low and high nutrient addition treatments ..................................................................................... 124 4-10 Mean shell length, clam wet weight (WW) m eat and shell tiss ue dry weights (DW) and condition index (CI) values for the sample population of clams used to determine clam biomass phosphorus content ................................................................................... 125 4-11 Mean and range of ash content values fo r m eat, shell and total clam tissues pooled for all clams sampled ....................................................................................................... 125 4-12 Summary statistics for phosphorus concentr ations [P] found in m eat, shell and clam tissue types pooled for all clams sampled ........................................................................ 126 4-13 Amounts of phosphorus (P) contained in m eat, shell and clam tissues along with percentages of total clam phosphorus allocated to meat and shell tissues for individual clams ...............................................................................................................126 4-14 Comparison of meat and shell phosphorus concentrations [P] in dry weight (DW ) biomass of Corbicula versus other fresh and saltwater clams ......................................... 130 4-15 Estimated annual phosphorus removal in various biological treatm ent systems applied to different effluent types usi ng systems of varying design and scale ................ 134

PAGE 10

10 LIST OF FIGURES Figure page 2-2 Blountstown system plumbing diagram .............................................................................31 2-3 Hague system plumbing diagram ....................................................................................... 34 3-1 Linear regression relationshi p of glass sphere volum e to glass sphere weight used to estimate clam shell cavity volume for the vo lume-based condition index calculation ...... 58 3-2 Air temperature readings at the Dair y Research U nit in Hague, FL over the study period .................................................................................................................................61 3-3 Input water temperatures in the low, me dium and high nutrient addition treatments ....... 61 3-4 Raceway dissolved oxygen (DO) readings in the low, m edium and high nutrient addition treatments ........................................................................................................... ..62 3-5 Raceway pH in the low, medium and high nutrient addition treatm ent systems ............... 62 3-6 Total phosphorus (TP) in the low, medium and high nutrient addition treatments. .......... 64 3-7 Total dissolved phosphorus (TDP) in the low, medium and high nutrient addition treatm ents .................................................................................................................... .......64 3-8 Total nitrogen (TN) in the low, medium and high nnutrient addition treatm ent source water ......................................................................................................................... ..........65 3-9 Chlorophyll a (chl a) in the low, m edium and high nutrient addition treatment source ponds ......................................................................................................................... .........65 3-10 Number of live clams in each nutrient addition treatment ................................................. 69 3-11 Cumulative number of dead found on the substrate surface in the low, medium and high nutrient addition treatments ....................................................................................... 69 3-12 Changes in shell lengths of tagged clam s captured alive in each nutrient addition treatm ent ..................................................................................................................... .......71 3-12 Regression relationships for shell length vs actual and predicted (Table 3-11) whole clam dry weight (DW) values for each nutrient addition treatment .................................. 75 3-13 Regression relationships for shell length vs actual and predicted (Table 3-11) shell dry weight (DW ) values for each nutrient addition treatment ........................................... 75 3-14 Regression relationships for shell length vs actual and predicted (Table 3-11) m eat dry weight (DW) values for clams ..................................................................................... 76

PAGE 11

11 3-15 Estimated clam dry weight (DW) biomass over time in the low medium and high nutrient addition treatm ents ...............................................................................................77 4-1 Raceway input total phosphorus (TP) in th e low nutrient addition treatment during the through-flow trials for July and August of 2002 ........................................................ 119 4-2 Frequency distribution of changes in total phosphorus (TP) from the input to the output in the low nutrient addition treatm ent raceways from July through August 2002..................................................................................................................................120 4-3 Amount of phosphorus (P) sequestered in clam biomass for the low, medium and high nutrient addition treatm ents over the study period .................................................. 127

PAGE 12

12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FRESHWATER CLAMS AS A TREATMENT MECHANISM FOR PHOSPHORUS IN AGRICULTURAL WASTEWATER By Lance W Riley August 2008 Chair: Edward J. Phlips Co-chair: Ann Wilkie Major: Fisheries a nd Aquatic Sciences The objective of this study was to determine th e potential of using a recirculating raceway system to remove phosphorus-containing material from agricultural wastewater streams. The focus of the research was on the biol ogical and physical characteristics of Corbicula populations and monitoring of various water quality parameters within the system, with special emphasis on phosphorus dynamics. A prototype raceway syst em was designed and constructed at the University of Florida Dairy Research Unit at Hague, Florida to test the adaptability and phosphorus removal capacity of the cl ams in wastewater treatment. The ability of freshwater clams to captu re, sequester and retain phosphorus-containing material from varying amounts of fertilizer additions was demonstrated in this study. Clam biomass contained an average phosphorus concen tration of 0.299 mg P/g of whole clam DW (SE = 0.005), similar to other bivalves. Tagged clam s recaptured alive over the course of the study showed growth rates of up to 0.117 mm/day in shell length (0.0024 g clam DW/day), yielding phosphorus removal rates up to 0.0079 mg P/individua l/day. Overall, race way clam populations were subject to high mortality and were unable to demonstrate significantly long-term removal of total phosphorus, dissolved phosphorus or chlorophyll a from overlying source water. High

PAGE 13

13 temperatures and possible impacts from amphi pod infestations may have affected clam populations. Even though some clams in this st udy did survive and grow, use of Corbicula culture for phosphorus treatment in Florida agriculture operations may requi re creative solutions to temperature and parasite problems. Despite thes e issues, the raceway-based recirculation system design demonstrated in this study provided a dependable, easy to construct and reusable platform for testing aquaculture potential of a variety of organisms in wa stewater treatment conditions at large scale. The ultimate goal of this study was to provide an effective biological remediation mechanism for removal of phosphorus from dairy waste streams; however, toxicity of dairy effluent, even at high dilutions, may prohibit ap plication of clam-based aquaculture systems without additional treatment mechanisms.

PAGE 14

14 CHAPTER 1 INTRODUCTION Phosphorus (P) produced in waste m anagement of agricultural opera tions can result in degradation of water quality in surface and groun dwater flows. An addition of nutrients to aquatic systems by human activit ies has resulted in the growth of nuisance macrophytes, alga bacteria and periphyton in natural and manmade systems (Sharpley 1994, Johnson et al. 2004, Dao et al. 2006, Wetzel 2001, Phlips et al. 2002, Timmons et al. 2002). This unwanted growth can also cause contamination of drinking water supplies, degradation of aquatic habitat for desirable species, fouling of engineered wate r systems, limitation of navigation and other negative impacts to commercial and recreational ac tivities (Sharpley 1994, Johnson et al.. 2004, Dao et al. 2006). In warmer climates such as Florida, aquatic plant growth is magnified due to the nearly year-round growing season (Scinto and Reddy 2003, DeBusk et al. 2004). The focus of the current study was applica tion of a clam-based biofiltrat ion approach to remove of particulate phosphorus from agri culture wastewater streams, with special emphasis on dairy systems. Excess nutrients can enter surface waters from both point and non-point sources associated with concentrated farming activ ities (Sharpley et al. 1994, Kni ght et al. 2000, Johnson et al. 2004, Dao et al. 2006). Manure and wastes produced in dairy operations are much different than other land-based agriculture operations since mu ch of the waste products generated have high water content due to milk parlor operations and bedding waste handling (NRCS 1999, Wilkie 2003). High liquid fractions in da iry wastes are encouraged by the removal of solids through settling and through microbial decomposition by anaerobic digestion (Wilkie 2003). Management practices that targ et nutrients such as phosphorus from animal manure include oxidation ponds, facultative la goons and storage ponds in conj unction with land application,

PAGE 15

15 constructed wetlands and composting (NRCS 20 06). In order to manage Florida dairy wastewaters, liquids are stored in short-term retention ponds prior to land application that supplies water and fertilizer to forage crops (Wilkie 2003). There are severe limitations to the technologies currently available to reduce, effectively and economically, nutrient levels in agricultural wastewater str eams. Some nutrients are bound in the terrestrial environment, but the remainde r end up in surface water bodies (Sharpley et al. 1994) and can also enter groundwater (Johnson et al 2004). Phosphorus is of special concern because it is the primary limiting factor for growth of algae and other plants in most freshwater systems (Wetzel 2001), and it cannot be remove d by volatilization, as in the case of other important nutrients, such as nitrogen and carbon (Wetzel 2001, Timmons 2002). Dairy wastes subjected to anaerobic diges tion can pose some special probl ems for aquatic environments because of their high liquid content (Wilkie et al. 2004) and high dissolved phosphorus content, making them readily available for uptake by aquatic macrophytes and algae (Sooknah and Wilkie 2004). In response to the special demands of nutrien t removal from wastewater streams, recent research has focused on integration of aquaculture systems. Use of aquaculture for wastewater treatment is designed to convert phosphorus into a solid form that can be harvested as a potentially useful commodity. In order to be su ccessful, these culture systems must be capable of promoting growth and reproducti on of the organisms that are th e end product of the process. However, organisms capable of desirable wate r treatment functions, su ch as filtration of particulate phosphorus-containing matter, in the natural envi ronment may not conform to aquaculture conditions. Theref ore, the experimental system design should provide as many

PAGE 16

16 natural habitat conditions for the target or ganisms as possible, including hydrology, food resources, substrate and temperature. Unlike production aquaculture, organisms in wast ewater treatment systems are selected for their ability to manipulate water quality parameters within source water and not necessarily for their specific value as a consumer commodity. Development of experimental and eventually commercial wastewater treatment systems is therefore driven more by public opinion on environmental issues through government legi slation than by consumer product demand and profit normally associated with traditional aq uaculture species. However, like traditional aquaculture systems, wastewater designs must also be cost eff ective, not too land intensive and have low water demands to be economically and environmentally feasible. Various aquatic organisms have been evaluated for potential applications in the treatment of agricultural wastewater, including dairy. Plant-based systems have received the most attention; however, some animal based system s have also been proposed. Aquatic macrophyte systems remove phosphorus from dairy effluents in small and large-scale systems (Reddy and Smith 1987, Sooknah and Wilkie 2004, Lansing and Martin 2006, Wood et al. 2007). Periphyton-based systems similar to the Alga l Turf Scrubber (ATS) technology developed by Adey and Hachney (1989) have been demonstrated to remove phosphorus from dairy wastewaters at experimental s cales (Pizarro et al. 2002) (M ulbry and Wilkie 2001). Algae suspended in large outdoor tanks containing dairy wastewater have also demonstrated phosphorus treatment potential (Sooknah and Wilkie 2004), as have phytoplankton in laboratory flasks (Lincoln et al. 1993, Lincoln et al. 1996). Freshwater, pond-scale studies of aquaculture-based wastewat er treatment have focused mostly on fish (Greer and Ziebell 1974, Demp ster et al. 1995, Van Rijn 1996, Drapcho and

PAGE 17

17 Brune 2000, Prein 2002, Azim et al. 2003, Ghaly 2005, Sindilariu 2007) and to a lesser extent on filter-feeding bivalves (Busch 1974, Buttner and Heidinger 1980, Buttner 1986). Use of bivalves in polyculture with other organisms has also be en demonstrated as a treatment mechanism in freshwater aquaculture operations in fish polyculture ponds (Buttner and Heidinger 1980, Buttner 1986, Soto and Mena 1999). Use of bi valves for wastewater treatment is more pronounced in the mariculture industry, where mo re elaborate systems have been used in conjunction with phytoplankton and/or seaweed to remove nutrient s generated from finfish and shrimp culture (Shpigel 1993, Shpigel and Neor i 1996, Lefebvre et al. 2000, Jones et al. 2001, Mazzola and Sara 2001). Freshwater clams were proposed and for use nutrient reduction by Gree r and Ziebell (1972) and Stanley (1974) who made calculations based on available literature; however, no large-scale systems have been tested. Litt le is known about phosphorus seque stration by these organisms or their potential for large-scale culture using dair y wastewater. Freshwater clams may provide an ideal P-removal vector for dairy wastewater beca use of their ability to remove and sequester phosphorus from the overlying water column (Fu ji 1979). In a typical clam-based system, dissolved phosphorus from wastewater is converted to a particulate form through a phytoplankton intermediary and is coupled with ot her wastewater particulates to feed clam populations (Greer and Ziebell 1972 Stanley 1974). Filter feeding allows the clam to pump overlying water through its siphon and into the mantle cavity where particulates are then removed by the gills and converted to biom ass (McMahon and Bogan 2001). Sequestered phosphorus in clam biomass and sediment depositio ns can be periodically removed at harvest intervals to remove phosphorus perman ently from the treatment system.

PAGE 18

18 One of the clam species that has been the focu s of past efforts in treatment systems is the well known invasive, Corbicula This organism is a recent invader to North America as described in McMahon and Bogan (2001), and is a significant cont ributer to fouling in power plants and industrial raw water systems Williams and McMahon (1986). The success of this organism as a biofouling agent is due to its high reproductive fecundity stemming from selffertilization no complex life cycle needing wate r-born gametes and intermediate hosts, unlike freshwater mussels that posse ss all of these traits (McMah on and Bogan 2001, McMahon 2002). A statewide distribution of Corbicula has been noted in most Florida waterways (Blalock and Herod 1999), so the clam can be considered as a naturalized species in stead of a potential invasive species. Its high reproductive poten tial makes the clam an ideal candidate for propagation under aquaculture cond itions, since it should repopulate rapidly following harvest of only a few individuals. Corbicula can be found throughout most of North Am erican are all expected to be very similar due to reproductive characteristics such as hermaphrodism and self-fertilization that result in the production of exact copies of the parent lineage as examined in McMahon and Bogan (2001). Exact taxonomic determ ination of clams in the genus Corbicula is subject to intense debate and ontogenetic variation can lead to improper usage of different species designations such as fluminea and japonica that are commonly used in the literature to describe most Corbicula clams found in freshwater environments. For this reason, clams in this study are only referred to as Corbicula. The goal of this study was to analyze the phos phorus removal potentia l of an engineered raceway system containing populations of the freshwater clam Corbicula The potential for using freshwater clams as a mechanism for phosphorus removal was based on its ability to

PAGE 19

19 remove and sequester phosphorus through active biofiltration. This study had the following objectives: Design, construct and operate a large-scale r aceway-based treatment system for examining the performance of freshwater clams as a P removal mechanism Determine the adaptability of the freshwater clam, Corbicula, to wastewater treatment conditions Determine the ability of clam raceways to remove and sequester phosphorus containing material from agricultural waste streams In this study, a raceway-based recirculati ng system was developed in order to study phosphorus removal rates using clam populations under simulated wastewater conditions in raceway systems. Design and construction of sy stems that can evaluate the performance of organisms in wastewater aquaculture at a commerc ial level are critical for developing nutrient management strategies for future applications. Systems must account for problems not only with the aquaculture practices, but with the conditions unique to dairy wa stewater effluent that may be remedied by dilution (Sooknah and Wilkie 2004) The following questions were posed for investigation in this study: What are the growth rates of clams and surviv al and recruitment rates of clam populations exposed to different concentrations of nutrients? Does the physiological conditi on of raceway clam vary over time and does it correlate with nutrient addition, environmental pa rameters or mortality events? Are clam raceway systems able to capture, sequ ester and retain P-containing material from varying concentrations of dairy wastewater effluent? How do seasonality, clam population dynamics temperature, algal density and P availability affect the rem oval of P-containing material by clam raceway systems? How is P sequestered and allocated by clam s into soft tissue and shell biomass? Is this technology suitable for use as a m echanism for P-removal by agricultural operations in Florida?

PAGE 20

20 CHAPTER 2 RACEWAY-BASED RECIRCULATING WASTEW ATER TREATMENT SYSTEM DESIGN AND CONS TRUCTION Introduction Most biologically-based wast ewater treatm ent systems involve use of ponds, tanks or raceways as steps in the removal of solids, nut rients and contaminants (Buttner 1986, Shpigel 1993, MacMillan et al. 1994, Shpige l et al. 1997, Jara-Jara et al. 1997, Jones and Preston 1999, Jones et al. 2002, Sooknah and Wilkie 2004). The focus of this design effort was removal of nutrients using filter-feeding bivalves as the activ e agent in the final stage of a process beginning with conversion of soluble nutri ents into particulate forms vi a production of plankton. The design had to meet several key criteria in term s of both experimental and operational demands. From an experimental standpoint, the system had to incorporate the ability to deal with multiple treatment groups in a replicated manner. Operationa lly, the system had to be of sufficient size to provide a reasonable measure of potential success in real life applications. Many elements of traditional aquaculture system designs were incorporated into the design process of the raceway-based treatment system s used in this study. Aquaculture system hydrology flows either flow-through or recirculating water flow regimes, depending upon the extent of water reuse and residence time (Van Rijn 1996). Recirculating systems offer the distinct advantages of lower wa ter consumption and confinement of wastes, reducing potential harm to the natural aquatic environment (Tim mons et al. 2002). These attributes make recirculting systems ideal for study of wastewat er treatment mechanisms because they do not involve discharge into the environment. Raceway-based aquaculture systems have been used to cultivate a variety of aquatic organisms including fish, bivalves algae and plants (Shpigel and Neori 1996, Adey et al. 1993). The most common large-scale raceways are usually associated with the production of finfish,

PAGE 21

21 such as salmonids (Timmons et al. 2002). Thes e structures typically range in size from 3-5.5 m in width, 24-46 m in length and 0.8-1.1 m deep (Ti mmons et al. 2002). This larger scale limits construction materials most often to concrete, plastic, or earthen structures with plastic liners (Sindilariu 2007, Van Rijn 1996). Reinforced fiberg lass panels have also been used in finfish culture to construct race ways using a modular design as an alternative to concrete (Vantaram 2004). Raceways used in the commerc ial rearing of Quahog clams ( Merceneria merceneria ) employ long, sand-bottomed, flow-through plastic troughs to raise juve niles prior to placement in estuarine farm sites for grow out (Lorio a nd Malone 1995). In addition to widespread commercial applications, raceways have also been used in a variety of experimental aquaculture systems targeting organisms associated with biof iltration or bioaccumulation, such as bivalves, benthic microalgae and macrophytes (S hpigel 1993, Craggs et al. 1996). The majority of experimental raceway designs using bivalves are involved in wastewater remediation. In these systems, commercially valuable bivalve species are commonly cultured as a secondary commodity on the effluent of primary culture organisms such as finfish and shellfish (Buttner 1986, Shpigel 1993, MacMillan et al. 1994, Shpigel et al. 1997, Jara-Jara et al. 1997, Jones and Preston 1999, Jones et al. 2002, Zhou et al. 2006). Examples of raceways for the culture of bivalve species incl ude: 14.4 L fiberglass tanks (Shpi gel et al. 1997), 34 L plastic tanks (Jones and Preston 1999), 340 L plastic tank s (Huchette et al. 2003), 1500L concrete tanks (Jones et al. 2002), 1500 L fiberglass tanks (Mac Millan et al. 1994), 2240 L fiberglass tanks (Jara-Jara et al. 1997), 2 m3 V-bottom fiberglass ta nks (Shpigel 1993), and 15,000 m3 concrete tanks (Zhou et al. 2006). A variety of raceway designs have also used algae or higher plants as the active treatment agent. Possibly the most notable vegetative raceway system is the Algal Turf Scrubber, which

PAGE 22

22 consists of an artificial stream used to culture periphytic algae. Periphyton is grown on plastic mesh screens placed in shallow rectangular flum es. Source water is supplied using a pulse-flow regime down the length of the raceway (Adey et al. 1993). This design has been adapted for use at various size scales from small-scale laborat ory systems (Mulbry and Wilkie 2001, Pizzarro et al. 2002, Wilkie and Mulbry 2002, Kebebe-Westhead 2003), to 1021 m2 raceways consisting of landfill liners between concrete sidewalls used fo r tertiary treatment of municipal wastewater (Craggs et al. 1996). Drawing on many elements of the aforementi oned aquaculture technologies, a two stage treatment system was designed for this study. The first stage involved growth of phytoplankton in ponds supplemented with either nutrients fr om a dairy wastewater stream or inorganic fertilizer. The ponds served as the source of wate r for a series of recirculating raceways. The freshwater clam was used as the primary agen t for nutrient removal from the source water, through filtration of plankton and conve rsion into harvestable biomass. The raceway design was chosen for this app lication since it mimics the small stream environment widely occupied by Corbicula in North Florida (Blalock and Herod 1999). Like a stream system, water is constantly supplied to th e raceways, and the channel-like shape (width to length ratio= 1 : 7.5 in this study) induces a plug-flow hydrology that ha s little back mixing. This hydrology is maintained by recirculating wate r between the source ponds and raceways in this system, while replenishing food partic les and dissolved oxygen and removing waste products. A coarse sand substrate was chosen for this application since it is an intermediate aggregate size preferred by Corbicula in small stream environments (Blalock and Herod 1999, Schmidlin and Baur 2007).

PAGE 23

23 The raceway design can be constructed of a vari ety of materials and is scalable, adaptable to a variety of hydrologic regimes and can be easily manipulated for sampling, cleaning and maintenance. Raceways are also versatile, in that the structure can be used to culture a variety of aquatic organisms, including other bivalves, fish, algae and plants using a variety of substrates. The design used in this study is easy to disassemble and transport for reuse at different effluent source locations, can be assembled in remote locati ons and can be used in short-term evaluations of source waters without leaving a significant footprint. The m odular components used in this design can be prefabricated and assembled quickly on site without intensive labor requirements needed to construct other large-scale systems. Integrating Corbicula into wastewater streams vi a phytoplankton production has been suggested by Stanley (1974); however no studies have been performed to evaluate the viability of using large-scale engineered systems that emul ate features that may be applicable in a fullsized wastewater treatmen t system. The study of Corbicula biofiltration potential has been limited to smalland medium-scale applications such as laboratory-bas ed bench scale flow chambers (Lauritsen 1985), aerated 37 L aquari a with sand substrate (B eaver et al. 1991, Brock 2000), 150 L aquaria (Greer and Zi ebell 1972) and 515 L rectangular fiberglass tanks containing mesh trays (Haines 1977). Cultivation in larger systems has focused on shallow ponds used to produce monoculture Corbicula as a food crop in Taiwan (Phelps 1994). Use of Corbicula in a pond-based polyculture scenario has been stud ied in catfish rearing ponds using benthic sediments and suspended cages as substrates (B uttner and Heidinger 1980, Buttner 1986). The clams have also been cultured in cages suspe nded within power plant discharge canals (Mattice 1977). The aquaculture system described in this study is meant to be an intermediate size scale

PAGE 24

24 between the smaller experimental systems and th e much larger pond and canal systems used in previous Corbicula studies. Investigations using experimental large-scale engineered systems are an important step in developing commercial size treatment systems, es pecially when targeting an organism like Corbicula that has not been traditionally cultured at such a scale. Predicting the adaptability of these organisms to large-scale culture scenarios cannot be accomplished sufficiently from smallscale experiments since the behavior of the orga nism in these systems may not coincide with observations in larger systems. The systems designed for this study provide an opportunity for replication of treatment groups with out sacrificing the structural elements of real-world treatment systems. These systems were designed using the following basic considerations necessary for the implementation of any experimental, biological-based water treatment system: Must be capable of maintain ing environmental conditions ne cessary to promote survival, growth and reproduction of the target or ganism such as hydrology, substrate, food resources, waste removal and aeration System design intended for scientific manipulat ion must be able to conform to desired experimental treatments and replic ation for statistical analysis System must be scalable to provide an ade quate surface area for the desired outcome in a real-world treatment application System design and operation must be applicab le to various land topographies and source water body layouts found at different site locations Design must maximize energy efficiency by using gravity flow to reduce pumping requirements To test the efficacy of the basic pond-race way design, two recirculating systems were designed and constructed at different locations in northern Florida. The first system was completed in October 2002 at the Sam Mitc hell Aquaculture Demonstration Facility in Blountstown, Florida and consiste d of nine raceways supplied by two source ponds. This system was operated from November 2002 until January 2003 when the entire facility in Blountstown

PAGE 25

25 was closed permanently due to university budget cuts. Parts of this system were excavated, dismantled and transported to the Dairy Research Unit in Hague, FL where a second system was constructed and operated from June 2003 to Oct ober 2004. The Hague facility consisted of 3 separate systems, each with 2 ponds and 3 raceways. Multiple ponds were used in both locations to provide an alternative to sustain phytoplan kton populations in case of an unfavorable pond condition, while multiple raceways were used fo r statistical rigors. Differences in the topography and source pond design for each location re quired the use of different water delivery system configurations; however, the individual raceway design remained the same for both locations. Multiple raceways were assembled at the Blountstown and Hague facilities for this study using different source pond layout s and water delivery configur ations. The water delivery configurations used for these two systems were : 1) Source water was gravity-fed through the raceways and pumped back to the source pond and 2) Source water was pumped to the raceways and gravity-fed back to the source pond. The gravity-fed source water op tion was applied to the Blountstown system, while the Hague system incorporated the pump-fed source water configuration due to the elevations of the availa ble areas for raceway construction in relation to the source ponds. Raceway Design The central com ponents of the raceway syst em were closed-ended rectangular tanks. Raceways were formed from a 1.0 m wide by 7.4 m long channel, assembled from 10 individual panels constructed with pressure-treated wood framing and plywood backing. These frame sections were joined together using galvanized lag screws and a framing board (3.8 cm (1.5) thick x 8.9 cm (3.5) wide) was placed across the width of the raceway at the bottom of each

PAGE 26

26 framing section joint to help maintain the rectangular shape. The ba sic layout and components of the raceways used in this study are illustrated in Figure 2-1. Figure 2-1. Components and de sign of individual raceways. A custom made, black 20-mil thick ABS-PVC lin er was then placed inside the assembled raceway and attached using trea ted wood furring strips fastened to the outside of the framing. The substrate was added before the furring strips were attached to minimize stretching as a result of sand settling. Coarse grade SiO2 filtration sand (0.6-1.0 mm par ticle size), available from Feldspar, Incorporated in Edgar, Florida, was used as a substrate. Some slack was left in the liners around the raceway sides in order to allow for expansion and contraction with changes in temperature. A 0.31 m thick layer of fill dirt was placed around each raceway frame before the substrate was added to help support the raceway frames. Raceways were filled with sand to a

PAGE 27

27 depth of 0.20 m so the substrate surface would be level with the raceway drain valves located on each output standpipe elbow (Figure 2-1). Fill dirt placed around each of the raceways was used in conjunction with the interior sand substrate in order to stabili ze the sides framing and prevent buckling. The finished raceway dimensions and capacities are listed below in Table 2-1. The supply plumbing to each raceway consisted of a 10.2 cm (4) inside diameter (ID) poly-vinyl-chloride (PVC) line stemming from th e 15.2 cm (6) ID main supply line from each source pond. Supply lines had a brass gate valve that allowed selection of the desired source pond. After the valve, supply lines are reduced to 5.1 cm (2) ID, a nd the two were joined into a single raceway input line equipped with a brass gate valve to regulate input flow and a Pitot-tube type flow meter to help balance input water fl ow between the raceways in the system. Source water was fed into the raceway through a slotted water distribution bar that dissipated the energy of the falling water over the width of the raceway, initiating a laminar-type plug-flow hydrological pattern. Threaded caps were used at the ends of th e distribution bars to allow easy cleaning of slotted portions to prevent blockage from biofouling. Water exited the raceways via a standpipe system that acted as a type of weir structure to govern water column height and channel outpu t water to the appropr iate source pond. The Table 2-1. Raceway dimensions and capacities as tested, ava ilable capacities adjusted for standpipe presence. Specification Parameter Value Inner Dimensions (empty) Width 1.0 m Length 7.4 m Height 0.6 m Available Volume 4200 L (4.2 m3) Substrate Capacity Depth 0.2 m Available Substrate Surface Area 7.2 m2 Volume 1432 L (1.4 m3) Water Capacity Depth 0.2 m Available Volume 1432 L (1.4 m3)

PAGE 28

28 standpipes were made from 15.2 cm (6) ID PVC 90o elbows with threaded male adapters on the output ends. The threaded ends passed through holes cut in the back plywood panels of the raceways and the liners. The threaded female adap ters on the outside of th e panel were tightened to hold the standpipe system in place, and a bul khead fitting was formed over the liner where output lines passed through by using an aluminum washer on the inside of the raceway to prevent leakage around the pipes. For this study, raceway water depth was main tained at 0.20 m; however, depth could be adjusted by extending the standpipe height using additional segments of pipe. Raceway output water was routed to the desired source pond by placing a 20 cm (8) long removable diverter pipe with a rubber no-hub conn ector over one of the output standpipes to divert water to the appropriate source pond return plumbing. Small, 1.9 cm (3/4) ID, drain valves were added to each of the 90o standpipe elbows level with the substr ate surface inside the raceway to allow removal of overlying water for periodic substr ate sampling or observation, and they could be used to drain the raceway for batch-fed experiments. The raceways had a maximum input flow rate of 303 Liters per minute (LPM) (80 gallons per minute (GPM)) due to the gravity flow cap acity of the 15.2 cm (6) diameter output standpipe. This flow rate lim ited theoretical retention time to no less than 6.3 minutes at the 0.20 m raceway water depth used in this study. Race way flow rate was maintained at 227 LPM (60 GPM) in this study yielding a retention time of 9.5 minutes. The one-meter width was chosen since it is approximately the limiting distance for accessing the entire bottom area by hand from the sides of the raceways. Raceway length was estimated from target stocking amounts of 7,000 to 10,000 adult clams pe r raceway at population densities similar to the high densities ( > 1,000 clams/m2) sometimes found in naturally

PAGE 29

29 occurring Corbicula populations (McMahon and Bogan 2001). Ultimately, the exact length and height dimensions of the raceways were determin ed according to the dimensions of the standard size for the plywood used in the r aceway side framing, to expedite construction. In this case, the 7.4 m length is a result of using 3 lengths of a standard sheet of plywood, while the 0.6 m raceway side height equals half of the width of a standard sheet of plywood. The low width to length ratio of the raceway design makes the bottom area more physically accessible than circular tanks of the same volume or surface area. When widths of 1m or less are used, bottom area can be manipulated easily by hand provided both sides of the raceway are accessible. By scaling width and length, the am ount of culture area can be expanded without losing the plug-flow hydrology. Multiple raceways can also be employed to increase the scale of the culture area as well as to conform to the statistical demands of experimental research. Blountstown Facility The raceway system constructed at the freshw ater fish aquaculture farm in Blountstown, FL site (30o35.5 North, 85o02.6 West) consisted of 2 source ponds supplying a group of 9 raceways arranged side-by-side. Raceways were positioned at the same elevation as the source pond bottoms with ponds located to the north an d west. Source ponds had an approximate area of 0.20 ha (0.5 acre), and water de pth was maintained at 1.5 m, yi elding an estimated volume of 3084 m3 (108,900 ft3). A 15.2 cm (6) ID PVC supply line was installe d through each pond berm using a concrete anti-seep collar. Sour ce water from the ponds was gravity fed to the raceways, and the flow rate was regulated using a brass gate valve installed on each supply line. A gravity fed inflow was chosen to minimize ope rational and equipment costs associated with pumping water both to and from the raceways. Flow rate of the incoming water was determined using a 15.2 cm (6) ID turbine-type, in-lin e flow meter for each raceway supply line.

PAGE 30

30 Maximum system flow rate was limited to 950 LPM at 1.5 m source pond depth by the gravityfed design. The design of the plumbing for the entire raceway set is illustrated in Figure 2-2. Supply lines from the ponds were connected to a 15.2 cm (6) ID PVC manifold for each. Each manifold was plumbed with a single 10.2 cm (4) ID PVC feed line for ea ch raceway with a 10.2 cm (4) ID brass gate valve in stalled on each raceway feed line to select for the desired source pond inflow. Raceway feed lines from the manifo lds were then reduced to 5.1 cm (2) ID PVC lines and joined together to form the raceway input plumbing. The output plumbing from each raceway consis ted of two separate 15.2 cm (6) ID PVC output lines that joined corresponding manifolds made of 20 cm (8 ) ID PVC. Each manifold emptied into a separate 4,542 L (1,200 gallon) concrete sump tank buried underground. Vertical vent pipes were installed on the ends of the manifolds to prevent a suction effect in the standpipes for the raceway furthest from the sump tanks. Each sump tank was equipped with a 120-volt, 5-horsepower centrifugal pump rated at 787 LPM (208 GPM) that was cycled using a float switch. The pump suction lines drew water from the botto m of the sumps using a 10.2 cm (4) ID PVC line equipped with a PVC foot valv e to prevent the need fo r priming of the pumps at the startup of each run cycle. A 10.2 cm (4 ) ID PVC outflow line leading to each pond was plumbed from each pump and an array of 10.2 cm (4) ID brass gate valves was used to divert water to the desired pond. This plumbing system was designed so that both source ponds could be used simultaneously by supplying several raceways without any mixing of the two water bodies.

PAGE 31

31 Figure 2-2. Blountstown sy stem plumbing diagram.

PAGE 32

32 Hague Facility The raceway system constructed in Hague, FL (29o48.0 North, 82o25.1 West) was made of three independent test systems, each cons isting of two source ponds and three raceways. Source ponds had an approximate area of 0.05 hect are, and depths were maintained at 1.9 m, yielding an estimated volume of 970 m3. Each pond set was fitted with aeration supplied by a 1.5-horsepower continuous-duty centripetal blow er, and a 5.08 cm (2) ID PVC main line reduced to a polyethylene line exte nding to the center of the ponds. A 1.9 cm () ID brass ball valve was installed at each pond to balance the airflow to weighted 15.2 cm (6) long, stone diffusers on the pond bottoms. In each triplicate test system, raceways were positioned near the banks of the source ponds at the north end of each pond set. Source wate r was supplied by continuously pumping from the south end of the ponds to the raceways where it ex ited through a standpipe and was returned to the north end of the pond by gravity-feed. A diagram of the plumbing used for each raceway system at the Hague facility is shown in Fi gure 2-3. A single 120-volt, 5 HP centrifugal pump rated at 787 LPM (208 GPM) was located on a conc rete pad between the ponds on the south end of each treatment pond pair. The suction side of each pump was plumbed using 7.6 cm (3) ID PVC pipe with a PVC foot valve at the pond end to prevent loss of prime. A strainer made from plastic 64 mm () mesh screen was installed ov er the intake to prevent the passing of large particles that might have been ha rmful to the pump. A pair of br ass 7.6 cm (3) ID gate valves was used to isolate the desired supply pond. Th e pump intakes were suspended from steelframed piers to one meter above the bottom of the pond bottom. A union was also added to each suction line at the pier to enab le removal of the submersed por tion for regular cleaning of the strainer screen.

PAGE 33

33 Water output from each pump was delivered to the raceways using a 10.2 cm (4) ID PVC line and regulated through a set of overpressure relief valves locat ed near the raceways that sent the overpressure water back to the source pond through a 10.2 cm (4) ID PVC return line. The relief system was necessary because the continu ous-duty pump configurati on introduces water at a constant flow. Therefore, in order to reduce the amount of flow to the raceways without cycling the pump, some water vol ume must be relieved from th e pump output line. Each pump was also fitted with a 75-pounds per inch2-rated pressure relief valve as an emergency feature in the event of a line blockage. After the overpressure water was relieved, the source water passed through a 10.2 cm (4) ID turbine flow meter be fore entering a 15.2 cm (6) ID PVC raceway supply manifold similar to the one used in the Blountstown system. The manifold is reduced to a 10.2 cm (4) ID PVC fitting at each raceway and further reduced to a 5.08 cm (2) ID PVC raceway feed line that empties into the distribution bar. Each raceway feed line flow was measured using a P itot-tube flow meter located between the valve and the spreader bar. These flow meters, along w ith the valves at each raceway, were used to regulate the flow balance between individual raceways, while the flow meter and relief valves before the manifold regulated available water input to the raceway set. As in the Blountstown system, the standpipe plumbing from each raceway consisted of 2 separate 15.2 cm (6) ID PVC out put lines that joined to manifo lds made of 20 cm (8) ID PVC corresponding to the source pond recei ving the outflow. Vertical vent pipes were installed on the ends of the manifolds to prevent a suction effect in the standpipes for the raceway fartherest from the outflow, as used in the Blountstown system.

PAGE 34

34 Figure 2-3. Hague system plumbing diagram.

PAGE 35

35 System Design Summary The Blountstown facility source ponds were three tim es larger than the Hague ponds; however, only two ponds were available as opposed to the six at the Hague facility. The Hague facility allowed for three diffe rent pond treatments to be performed simultaneously since the raceways were distributed into three separate pond/raceway systems each consisting of three raceways and two source ponds. This layout allowe d for an alternate source pond in the event of a catastrophic event such as phytoplankton population crash or water quality issues. The Blountstown system could only support two simu ltaneous pond treatments and only one if an alternate pond was to be incorporated. Using nine raceways per pond treatment at the Blountstown system increased th e ability to test different raceway conditions with more statistical rigor than the three raceways per pond treatment. Cons equently, the Hague system had a 1 : 224 raceway set volume to pond volume ratio, similar to the 1 : 236 ratio found in each Blountstown pond/raceway set when all nine raceways were bei ng fed from a single source pond. The gravity inflow/pump outfl ow configuration constructe d at Blountstown may be the more energy efficient choice; however, in the event of pump failure, the source pond will continue to flow, flooding the ra ceway site and evacuating the s ource pond. In order to prevent this catastrophic failure, the only options are large self-actuating va lves for the main input lines or an emergency power generator, both of wh ich are expensive propositions. A pneumatic or mechanical self-actuating valve triggered by a ba ttery-back up float switch circuit from the sump tanks is the ideal mechanism since pump failure may occur without power outage. Another drawback of the gravity-fed raceway infl ow is that flow rate is dependent on water depth in the source ponds. In the Blountstown syst em, the input flow rate to the raceway varied with changes in pond depth. As a result of the ma in input flow variation, the individual raceway

PAGE 36

36 flow rates would become unequal, requiring consta nt adjustment of the individual raceway input valves and pond refill rate to maintain consistent flow regime. This constant monitoring was not needed at the Hague facility since the pump input delivered a constant, inflow volume balanced among the raceway set. Consequently, the maximum input flow rate availa ble to the raceway set was limited by the source pond depth and the input flow line diameter in the gravity-fed design, whereas a larger capacity pump can be installe d to increase flow in the pump-fed design to achieve the individual racew ay maximum of 227 LPM. The raceway-based system design used in this study provided important elements of flexibility, which allows for a broad range of app lications. Integration of raceways into various source and receiving water system configurations can be accomplished by modifying the water delivery component to adapt to the layout of the water bodies, area topography and raceway hydrological demands. The most en ergy efficient design is the gr avity inflow/gravity outflow configuration that eliminates costs and energy consumption associated with multiple pumps; however, this can only be applied in a flow-throu gh system with the proper raceway elevations. In most circumstances, at least one pump will be required to feed and/or evacuate the raceway systems. Other components, such as settling pon ds, could also be integrated into the source water stream to remove excess particulates (Krom et al. 1995, Shpi gel and Fridman 1990 Van Rijn 1996). Changing input flow rate or adjusting the wa ter depth by modifying standpipe height or substrate volume allows for mani pulation of raceway hydrology. Ch anges in these settings could be used to adjust water retention time or linear velocity in constantly flowing systems. Water retention time is often shortened in large commercial raceway systems either to maintain high dissolved oxygen concentrations or to avoid ammonia build-up (Tim mons et al. 2002).

PAGE 37

37 Retention time can be markedly increased to increase exposure of the water for treatment purposes by using a batch-fed regime where race way inflow is stopped and raceway water is circulated or agitated to mix and re-aerate it. In this study, retention time was increased to six hours using a batch-fed configurat ion that incorporated a 1/3 hor sepower submersible pump to circulate and aerate raceway water. Another con cern with long retention tim es is heat exchange, since temperature increases in shallow raceways are due in part to radiant heat exposure from the sun as well as evaporative heating/cooling. The raceway structures in this study incorporated a modular design allowing for transport and reuse at multiple sites. Portability makes th e raceways ideal for appli cation in the scientific arena since: 1) experiments are of ten short term, 2) systems must be cleared from the site at the end of the study and 3) sites may be located in remote areas. Another modular design has been applied to reinforced fiberglass panels by Vant aram (2004) as a lightweight and less permanent alternative to concrete. The Vantaram (2004) system requires no bracing around the sides of the raceway to maintain the structural integrity such as in metal, fiberglass or wood-framed raceways that need buttresses or dirt to maintain the desi red shape. The raceway design used in this study required soil backfill for structural support. Th is made each raceway fully accessible from all sides, allowing for easy and comfortable phys ical manipulation, cleaning and sampling, an advantage in the experimental systems. The a dvantage of this design over metal, fiberglass and the Vantaram-type panels is that the materials can be readily purchased locally and assembled without molding structures, curing time, special t ools and health concerns over solvents and dust, all of which impact budgetary demands. The raceway-based systems constructed in this study were chosen and developed as a lowcost, less permanent and easy to assemble alte rnative to other raceways constructed from

PAGE 38

38 concrete, fiberglass or plastic. The raceway system designs employed here are versatile enough to be applied to other organism s targeted for large-scale water treatment/biofiltration studies in both fresh and saltwater condi tions, a variety of locations and effluent sources.

PAGE 39

39 CHAPTER 3 ADAPTABILITY OF Corbicula TO TREATMENT RACEWAYS Introduction The potential for using the freshwater clam Corbicula in dairy wastewater treatment ultimately depends on the ability of the organism to adapt successfully to that environment. Clam-based treatment raceways are essentiall y production aquaculture systems designed to function using phytoplankton biomass grown on wastewater as the main food source. Growth and harvest of clam tissue biomass represent the accumulation and removal of wastewater derived nutrients. As in traditional aquaculture systems, growth, recruitment and health of the clams in the treatment raceway population are importa nt elements in assessment of the potential success of large-scale systems. Studies based on the production of filter-feeding organisms as a mechanism for wastewater treatment has been applied mainly in the mari culture industry, where commercially desirable species are cultured using finfish and shellfish farm effluents in both natural and engineered systems (Shpigel and Blaylock 1991, Jakob et al. 1993, Shpigel et al. 1997, Lin et al. 2001). No large-scale commercial markets fo r freshwater filter feeders curren tly exist. The freshwater clam Corbicula may be an ideal candidate for such a wastewat er treatment system in freshwater due to its high filtration and growth rates under eutroph ic conditions (Greer and Ziebell 1972, Mattice 1977, Buttner 1986, Beaver et al. 1991, Brock 2000). Corbicula is also known for its ability to form high-density populations (Gardner 1976) an d maintain high rates of reproduction (Rodgers et al. 1977, McMahon and Williams 1986, McMahon and Bogan 2001). In freshwater aquaculture applications, Corbicula utilizes uneaten feed and f eces, in addition to phytoplankton, produced from the polyculture of other organisms (Buttner 1986).

PAGE 40

40 Sustained aquaculture of freshwater Corbicula for human consumption has been reported in open pond-based polyculture systems (Villadolid and Del Rosario 1930, Miller and McClure 1931, Ingram 1965, Chen 1976). These open systems utilize constant exchange of water from natural bodies that can contain juveniles and are therefore not dependent upon reproduction from individuals within the system. Little is known about large-scale Corbicula aquaculture practices using engineered wastewater treatment systems. Other freshwater species of bivalves have al so been considered for wastewater treatment including Lampsilus clairbornensis (Swingle 1966), Diplodon chilensis (Soto and Mena 1999) and Eliptio complanata (Stuart et al. 2001); however, Corbicula may be better suited for aquaculture since it does not requ ire an intermediate fish hos t for reproduction and can selffertilize (Kraemer et al. 1986, McMahon and B ogan 2001). This reproductive advantage over freshwater mussels indicates that Corbicula populations should in theo ry be self-sustaining and should not require re-stocking from external sources or polyculture with proper fish intermediates to balance recruitment with biomass removed by mortality and harvest. An application of Corbicula aquaculture to large-scale trea tment of agriculture effluents using phytoplankton as an intermediary has been proposed by Greer and Zeibell (1972), Stanley (1974) and Olszewski et al. (1977). Greer and Ziebell (1972) used Corbicula in short-term aquarium-based filtration experiments (16 days) to evaluate the clams treatment potential for waters enriched with inorganic nitrogen and phosphorus. However, the short duration and small scale of this study may not accurately reflect th e organisms ability to provide treatment over longer time periods and in a larger system needed for real-life applicatio ns. Long-term studies utilizing Corbicula for wastewater aquaculture have target ed freshwater finfish effluent (Habel 1970, Busch 1974, Buttner and Heidinger 1980, Buttn er 1986) and municipal wastewater

PAGE 41

41 (Haines 1977), but not agricu ltural wastewaters. In a ddition to nutrient reduction, Corbicula based systems decreases turbidity (Habel 1970, Busch 1974, Haines 1977, Buttner 1986) and increase dissolved oxygen at dawn (Buttner 1986). The extent that thes e treatment actions can be performed is primarily a function of the amount of biomass present and is strongly influenced by population dynamics. In this study, a large-scale raceway-based treatment system was used to examine the adaptability of Corbicula to the raceway conditions that may be encountered in northern Florida agriculture operations. It was hypothesized that Corbicula populations introdu ced to treatment raceways would survive, grow and reproduce under wastewater conditions. Changes in clam population were compared to raceway water qua lity, nutrient concentrations and phytoplankton biomass parameters to assess the impact of e nvironmental conditions an d food availability on population dynamics. Assessment of populati on parameters along with monitoring of environmental conditions was used to determine the potential for the succ essful application of Corbicula in the large-scale treatm ent of agricultural efflue nts in northern Florida. Methods Growth of clam s in the treatment raceways was assessed as a function of changes in shell size, tissue biomass and health, while survival an d recruitment of juvenile clams was used to assess changes in the numbers of individuals in the raceways over time. Tagging studies are more precise, but they introduce additional hand ling stress resulting from tag application. Population sampling limits handling stress but is imprecise and requires an independent confirmation of growth from tagging studies. The combination of these two methods was intended to provide a measure of quality assurance/quality control.

PAGE 42

42 Raceway-based Treatment System The previously described (Chapter 2) racewa y-based recirculating sy stem constructed at the University of Florida Dairy Research Unit in Hague, Florida was used to test the adaptability of Corbicula populations to large scale culture u nder exposure to simulated agricultural wastewater conditions. Three independent pond/raceway system s were constructed in June 2002 to compare source waters with low, medium and high levels of nutrients. Each pond/raceway system consisted of two earthen source water ponds and three wood-framed, PVC-lined raceways. Table 3-1 shows the numerical desi gnations for the ponds and raceways in each nutrient addition treatment group. Table 3-1. Source pond and raceway numerical designations for the treatment systems at the Hague site. Nutrient addition treatment Source pond Raceways Low 1, 2 1 3 Medium 3, 4 4 6 High 5, 6 7 9 Source ponds Source ponds had an approxim ate area of 0.05 hect ares (ha), with depths of approximately 2 m and volumes approximately 1000 m3. Ponds were enriched with a blend of nitrogen and phosphorus fertilizer or anaerobica lly digested dairy farm wastewat er, to simulate possible water conditions associated with tertiary wastewater treatment. The low nutrient addition treatment received no external nutrient addi tion. A 5 % and 10 % addition of anaerobically digested dairy farm effluent was added to Pond 3 from the medium nutrient group and Pond 5 from the high nutrient addition treatment group, respectively. For effluent physical and chemical characteristics see Wilkie et al. (2004). Efflue nt was pumped from the digester to the source ponds and metered through a 2.54 cm (1) turbine-type flowmeter.

PAGE 43

43 Pond 4 in the medium nutrient addition treat ment was dosed with 1.1 kg of triple super phosphate (9Ca(H2PO4)2 N-P-K = 0-45-0) and 6.8 kg ammonium nitrate (NH4NO3, N-P-K = 15-0-0) resulting in a total a ddition of 0.23 kg of phosphorus a nd 1.02 kg of nitrogen per pond in October 2002, one month prior to clam stocking. Pond 6 in the high nutri ent addition treatment was dosed with 2.2 kg of triple super phosphate (0.46 kg P) a nd 13.6 kg of ammonium nitrate (2.04 kg N) in January 2002, also one month before introduction of clam populations. Nutrient additions were de signed to enhance phytoplankt on biomass, which was the putative source of particulate nutrition for the clams. Fertilizer loading levels were targeted at increasing phosphorus and nitrogen levels by 0.2 3 mg/L TP and 1020 g/L TN in the medium nutrient treatment source pond and 0.46 mg/L TP and 2040 g/L TN in the high nutrient treatment source pond. Fertlizer was introduced to ponds by placing it into a burlap bag suspended in the water column by a 0.5 m x 0.5 m floating frame constructed from 5.08 cm (2) ID PVC pipe. Neither of the ponds supplied with dairy effluent were exposed to the clam raceways due to excessively high ammonia levels (2.0 mg/L or greater as NH3-N), which represented a direct threat to the health of the clams. NH3-N levels in the effluent ponds and input and output water from each operating raceway were monitore d monthly using Aquacheck brand Ammonia Nitrogen (NH3-N) Test Strips (Hach Incorporate d, Colorado, USA) commonly used for aquaculture and aquarium a pplications. Levels of NH3-N in the raceways supplied by ponds 4 and 6 never reached the 0.25 mg/L (as NH3-N) minimum value of the test kit. Source ponds were circulated through the raceway s for 10 to 20 days prior to clam addition. Supply ponds were aerated at night throughout the study to help with mixing, maintenance of nighttime dissolved oxygen and ev aporative cooling. Floating macrophytes were

PAGE 44

44 cleared by hand from each source pond several times at the start of the experiment. In addition, six juvenile triploid grass carp ranging between 15 and 20 cm in length were stocked in each pond in June 2002 to help reduce vegetation in po nds. No fish mortality was evident in ponds not exposed to dairy wastewater effluent. In contrast, ponds treated with wastewater all experienced 100 % fish mortality in less than one week after effluent addition. Fish mortality events corresponded with high ammonia levels ( > 0.25 mg/L NH3-N). Raceways Raceways were 1 m in width by 7.4 m in length, w ith an available substr ate surface area of 7.2 m2, after subtracting standpipe area. Water de pths were maintained at 0.2 m, yielding a raceway water capacity of 1.4 m3 each. Source water inflow to the raceways was maintained at 227 liters per minute (LPM) (60 GPM) during normal operating conditions. The flow rates yielded retention times of approximately 9.5 minutes, with a linear veloc ity of 1.17 m/min. Flow rates were calculated assuming near laminar flow through the raceway structure. Raceways were filled to 0.2 m depth with a coarse grade SiO2 filtration sand (0.6-1.0 mm pa rticle size), that was purchased from Feldspar Incor porated in Edgar, Florida. Aquatic plants such as Chara sp and filamentous algae (mainly Spirogyra sp .) growing on the raceway substrate and liners were removed by hand from each raceway at least weekly to reduce fouling from increased water retention time. The co ntrol of plants, especially Chara, on the substrate made accurate estimation of clam biodeposit sedimentation impossible. This was due to the constant resuspension of sediment deposits associated with disturbance from removal of the vegetation that was often rooted below th e sediment surface. Filamentous algae tended to utilize the sides of th e raceways where PVC liners were submersed.

PAGE 45

45 Water analysis Source water ponds and raceways w ere monito red once per week at 6:00 and 18:00 for temperature, dissolved oxygen, pH, chlorophyll a total nitrogen and total phosphorus. Measurements in the ponds were taken at the end of sampling piers near the intake pipe leading to the raceways. Monitoring intervals were ch anged to monthly in June 2003, after 75 % or more of the clams were assumed dead from nu mbers of cumulative dead clams found on the sediment surface. Temperature and dissolved oxygen was measured using a YSI model DO550, and pH was measured using a Fisher model AP63 meter. Water temperature, dissolved oxygen and pH readings were compared using a paired t-test (Microsoft Excel) at the raceway input and output, as well as between nutrient addition treatments. Water samples were collected from the sampling piers in each pond for nutrients, using a pole sampler designed especially for this ex periment. The sampler used a plunger-type mechanism to collect a 1 L water sample from in front of the intake pipe. When the unit was lowered to the desired depth, the plunger was ac tuated by the operator via a spring-loaded handle at the opposite end of the pole. After the sample was collected, the plunger was released, sealing a 1 L plastic (Nalgene Incorporat ed, USA) bottle and raised for retrieval. The sample bottle was unscrewed from the sampler and cappe d for transport to the laboratory. Water samples collected from the source ponds were analyzed at the laboratory for phosphorus, nitrogen and phytoplankton biomass in terms of chlorophyll a Total phosphorus (TP) and total dissolved phosphor us (TDP) were determined using the potassium persulfate digestion method (APHA 1998) with a Hitachi spectrophotometer. TDP determination involved pre-filtering through a 0.7 m glass fiber filter. Total nitrog en was determined using potassium persulfate digestion method (APHA 1998) with colorimeteric analysis performed using a BranLuebbe auto analyzer. Phytoplankton biomass was estimated using chlorophyll a (chl a),

PAGE 46

46 measured by filtering 250 mL of water onto a 0.7 m glass fiber filter, followed by an ethanol extraction (Sartory and Grobbelaar 1984) and spectrophotometr ic determination (APHA 1998) using a Hitachi spectrophotometer. Micros cope observations of phytoplankton species composition were obtained periodically to desc ribe dominant organisms with help from Mary Cichra at the University of Florida Fisheries and Aquatic Sciences Depa rtment. Data obtained from these phytoplankton species observat ions were not assessed quantitatively. Clam Population Dynamics In The Racew ay Environment The adaptability of clams to the raceway environment was determined by assessing survival, recruitment, growth and health. Clam s were sampled at the time of stocking and at specified intervals over a period of 440 days to evaluate population density, shell size and tissue biomass. A tagging study was employed to vali date survival and growth determined from monitoring of raceway populations. A three-m onth interval was chosen for sampling time duration der to reduce handling stre ss, while retaining the ability to assess changes in clams on a seasonal basis. The raceway stocking and sampling time schedule is shown in Table 3-2. Table 3-2. Raceway (RW) stocking and population sampling schedule for the low, medium and high nutrient addition treatment systems. Time interval 0 corresponds to the time of stocking, and treatments were not stocked during the same season due to the time required to obtain the large numbers of clams needed. Year Season Time interval Low nutrient RW 1-3 Med. nutrient RW 4-6 High nutrient RW 7-9 2002 Summer 0 2002 Fall 1 0 2003 Winter 2 1 0 2003 Spring 3 2 1 2003 Summer 4 3 2 Shell length was used as the primary indicato r of clam size, sinc e it can be measured quickly, is not subject to the variability exhibite d by soft tissue, and is stable over time. Shell

PAGE 47

47 length was defined as the greatest distance anterior to posterior measured perpendicular to the hinge line using a caliper meas ured to the nearest 0.01 mm. Shell length may be the best measurable size variable compared to height and width in Corbicula because it is the largest size variable, making it less sensitive to measurement e rror. Shell length also encompasses areas of the shell that are less susceptible to erosi on. Other studies of Corbicula have used shell length as a descriptor of size, including Mattice and Wright (1986) and McMahon and Williams (1986). In order to confirm that lengt h provides the most dependable measurement of clam size, an allometric analysis was performed on a sample of 500 clams obtained from the Santa Fe River (29o51.1 North, 82 o37.9 West) in March 2002. Each clam was measured for shell length, width and height. A regre ssion analysis using SAS (PROC REG) (SAS Institute Cary, NC) yielded shell length as the measuremen t with the highest R-square value (r2 = 0.96) compared to the width and height (r2 > 0.94), thus making this size variable the most consistent over the size range used in this study (shell lengths 9.4 mm to 28.4 mm). Length may al so be the preferred variable because shell erosion wa s apparent in the umbo region in larger clams, thereby affecting height and width measurement values. Measur ement error was determined by repeating the length measurements three times on twenty ra ndomly selected clams from the allometric analysis. The maximum variance for the mean shell length was + 0.1 mm. Stocking clam raceways Clam s for stocking the raceways were obtained from populations in three different natural water bodies under permit number FNC-04-022 is sued by the Florida Fish and Wildlife Commission. Clams for the low nutrient group (Raceways 1-3) were collected in June 2002 from a 0.5 km stretch of the Santa Fe River near the State Road 49-bri dge in Gilchrist County, Florida (29o54.2 North, 82 o52.0 West). By November 2002, the rising water level of the river

PAGE 48

48 made further clam excavation impossible; theref ore, animals for Raceways 4-9 were collected from lakes located in Lake County, Florida that had accessible populations of clams. Clams for the medium nutrient group (Raceways 4-6) were collected from the southwest shore of Lake George (29o12.2 North, 81o35.7 West) in November 2002. Clams for the high nutrient group (Raceways 7-9) were obtained from the west shore of Lake Dalhousie (28o54.0 North, 81o36.8 West) in February 2003. Possibl e adaptability issues with ontogenetic differences in the populations obtained from di fferent locations were not addressed due to the difficulties inherent in locating and obtaining such high numbers of clams from systems in a timely manner. All three collection sites had coarse sand sedi ments similar to the substrate used in the raceways. Clams were excavated by shoveling botto m material into weighted baskets made from plastic mesh with 0.635 cm2 ( in2) perforations. Clams were also excavated by hand using trowels or a commercial clam rake m odified for the small size of the Corbicula by affixing similar plastic mesh on the inside of the co llection basket. Periodi c excavation of bottom sediment using a 0.25 m2 PVC sampling quadrat was used to determine population densities for the clams in their natura l habitat. Densities ra nged from 48 and 864 clams/m2 with a mean of 272 clams/m2 (standard error (SE) = 23, n = 47 obser vations) for all locations combined. After excavation, clams were enumerated and divided into mesh bags. The clams were then placed into coolers packed with wet newspa pers and kept out of di rect sunlight to help minimize heat stress and dessication. They were tr ansported directly to the aquaculture facility in Hague, FL and scattered evenly throughout each raceway. Stocking of each raceway took up to 15 days involving 2 to 6 people working per day. Raceways 1 were stocked from June 17 to 28, 2002, Raceways 4 from November 4 to 14, 2002, and Raceways 7-9 from February 11 to 26, 2003.

PAGE 49

49 Stocking densities were estimated using a vol umetric method of enumeration. The method employed a 0.5 L plastic container that was used to transfer the clam s collected in the field to the mesh shipping bags. Ten mesh bags of clams were added to each racew ay. Each mesh bag contained approximately 1,000 clams. This method was chosen as opposed to counting each individual or bulk weighing in order to minimize handling stress, time and equipment needed to enumerate the large numbers of clams needed for this study. Shell length measurements were taken on 270 cl ams per raceway, selected at random from mesh basg just before stocking. A sample of clams was obtained from each mesh bag by scooping a sample from the middle of the bag usi ng the 0.5 L plastic container mentioned above. A total of 27 clams were selected for the bioma ss analysis per raceway by keeping the tenth clam out of every 270 clams sampled for shell length determination. Clam raceway population sampling Sam ple sites within the raceways were defined using a submersible quadrat grid and sampling sleeve. The quadrat grid consisted of an aluminum frame divided inside into 10 cm x 10 cm squares using 2 mm thick nylon line. In order to define 100 cm2 surface areas more accurately for excavation, a 10 cm x 10 cm ID square metal tube, 20 cm in length was used as a sampling sleeve. The sleeve was inserted into the raceway sediment, and clams were extracted to a depth of approximately 10 cm. The entire grid measured 1.2 m x 0.9 m. Gr ids were deployed at the input, middle and output of each raceway to cover the length of each raceway. The 10 cm wide area was not sampled to reduce the risk of puncturing the r aceway liner. Areas around the raceway edges, under the input spreader bars, ar ound the outflow standpipes were not sampled. Placement of grid in all three sections allowed for 85 % covera ge of the raceway bottom, providing a total of

PAGE 50

50 567 possible quadrat locations per ra ceway. The grid rested on stainless steel pegs buried in the raceway sediment to ensure repeatability in placement. The number of sampling quadrats needed pe r event was determined in December 2001 using a power analysis with equations from Sokal and Rolhf (1995) and performed using Microsoft Excel on a population of 500 clams intr oduced to the Blountstown raceway system. Clams were distributed over a 1 m2 area of active raceway substrate and left for 5 days. A 100 cm2 PVC square quadrat was used to divide sa mple areas into 10 cm x 10 cm increments. Quadrats were excavated along a 1m transect in the middle of the raceway, and the number of live clams recorded. The power analysis on the cl am density data yielded a sample size of at least 19 quadrats to achieve a 95 % confidence interval for clam density. A sample number of 27 quadrats per raceway were chosen since the sample grid consisted of 9 sample coordinates over the width of the raceway and 3 grid placements per raceway. Clam sampling followed a stratified random de sign without repetition. One length division was chosen at random for each width division within the grid to stratify the sample areas over the width of the raceway. Stratificat ion of raceway sampling allowed for spatial analysis of data over both the length and width of the raceways. Quadrat positions were repeated within each grid placement at the input, middle a nd output regions of the raceways. Live clams in each quadrat were counted, measured for shell size and returned to the same location in the raceway. Clams us ed in the biomass analysis were selected from clams included in the density analysis by retaining the third cl am excavated from each sample quadrat. In the event that there were fewer than 3 clams, the first clam excavated was kept for biomass analysis. This procedure yielded a maximu m possible sample size of 27 clams per raceway at each time interval. This sample size was chosen to mi nimize the impact on raceway populations due to the

PAGE 51

51 destructive nature of biomass de termination. Individuals for biom ass determination were placed in numbered plastic bags and frozen until analysis. In order to estimate raceway population densi ties using the spatial technique, the clam densities obtained during each sampling event we re used to calculate the average number of clams per 100 cm2 quadrat sampled (n = 27 quadrat samples per raceway). Average quadrat densities were then converted to the number of clams per m2 and multiplied by 7.2 m2 of available substrate area per raceway to obtain the estimated number of clams per raceway. Tagged clams Clam s used in the tagging study were obtaine d from a random sampli ng of individuals at the time of stocking. The clams were marked with EZ-Code brand wire markers (Thomas & Betts Incorporated, USA) which are self-adhesive numerical tags applied to the shells that minimized the handling stress to the animals. Th is type of tag was chosen because it caused less damage and reduced the risk of injury when compared to engraving (Mattice and Wright 1986, McMahon and Williams 1986, Lemarie et al. 1995) or insertion of passive integrated transponder tags into the shell cavity (Kur th et al. 2007). The pre-applie d adhesive expedited the tagging process by reducing drying time and increasing tag r eadability of the liquid adhesive traditionally used for affixing numerical tags (Lemarie et al. 1995), brass washers (Toll et al. 2003), coded wires (Layzer and Heinricher 2004 ) or monofilament tethers that anchor the animals to the substrate (Foe and Knight 1986). This type of tag was chosen because of their small size, readability, adhesive strength and low cost compared to traditional numerical bee tags used in bivalve resear ch. These vinyl decals are generally used in electrical applications and have 4 mm-tall black numbers with a white background. Tags were trimmed to 5 mm x 5 mm squares before application to towel dried shells. Tagged clam shell lengths were meas ured as described earli er and placed into the

PAGE 52

52 raceways using the sampling grid described abov e. Using the randomly generated coordinates that were sampled over the study, one tagged clam was placed in each location using at least 33 locations at the input, middle and output sections of each raceway. In order to test the durability of the tags, a sample of 20 live clams was obtained from the Santa Fe River (29o51.1 North, 82 o37.9 West) in December 2001. Tags were affixed and clams placed in a 1 L plastic (Nalgene Incorporated, USA) bottle along with 200 mL of coarse sand and 200 mL of water. The mixture wa s capped and shaken vigorously by hand for 15 minutes, after which the clams were removed and rinsed with water for inspection. Only one clam lost its tag and was remove d while the remaining numbered clams were placed back in the bottle. Then the clams were again shaken for fifteen minutes and removed for inspection. Tags on six of the clams had come off during this tr eatment, and the shells of all of the clams exhibited chipping around the margins. All tags remained legible afte r both treatments. The conditions that these clams were exposed to are certainly harsher than the raceway environment because of the lack of significant water moveme nt to produce the same tumbling effect but neglects dissolution and bacterial decay that may also account for loss of adhesive strength with longer exposure times in aquatic systems. A total of 108 clams were randomly selected fr om each raceway at the time of stocking for the application of tags. Specimens were tagged and distributed in a stratified random fashion by using the sampling grid to place 36 clams in the input, middle and output sections of each raceway. Shell length, clam number and initial quadrat coordinates we re recorded for each tagged clam at stocking and at each 3-month samp ling interval when found alive. Tagged clams that were later found dead were not used even though growth may have been evident by comparing measurements of final shell size to that at stocking.

PAGE 53

53 Clam survival Clam survival was assessed by counting live i ndividuals in the raceways at stocking and at designated sampling intervals. Total densities were determin ed at the end of the study by counting the live and dead clams remaining in the raceway, along with cumulative counts of dead clams removed from each raceway over th e course of the study. Raceway population densities at stocking were estimated using the volumetric technique and at each sampling interval using the spatial technique, as described above. Clam stocking estimates were later verified using information on cumulative counts of dead clams removed over the course of the study, along with counts of dead and live clams at the end of the study. In the case of visible clam mortality events, dead clam shells were removed and counted. Counts of dead clams removed from the substrate surface were used as an overall indication of population mortality. No attempts were made to evaluate the dead shells buried in the substrate after the events because the removal of dead clam shells caused rususpension of sediment deposits. Biomass changes Changes in clam biomass were assessed usi ng tissue weight and length data taken from clams obtained at stocking and at each designated sampling interv al. Tissue weight to shell length relationships were developed from a subs et of clams collected at stocking and sampling intervals. Sample clams were frozen prior to dry weight (DW) analysis (Copar and Yess 1996). Freezing provided an alternative to live shucking, since gaping of frozen clams occurs naturally when the clam is removed from the freezer and pl aced at room temperature for about 15 minutes. Shucking can also chip shell va lve edges resulting in measurement error in shell-dependent variables.

PAGE 54

54 Clam dry weights were determined for meat and shell of freshly thawed clams. Soft tissues and shell material were separated and placed in to individual dried and pre-weighed aluminum drying dishes. Whole clam, soft tissue and shel l wet-weights were reco rded and placed in a drying oven at 80 oC for a period of 24 hours. This period of time was used due to the results of preliminary tests to establish a drying time needed to attain constant weight in clams sampled from the Santa Fe River (29o51.1 North, 82 o37.9 West) in June 2002. After 24 hours, the dried tissue samples were removed from the oven and allowed to cool to room temperature in a dessicator. Meat and shell ti ssue dry weights were recorded to the nearest 0.001 g and tissues stored in a dessicator before ashing. Dried shell and meat tissues of selected indi viduals were ashed to obtain ash weight, from which ash free dry weight (AFDW) was calculate d. Dried meat tissue was collected from the drying dishes and placed into ce ramic crucibles. Shell was prep ared for ashing using a stainless steel grinding device powered by a rotary hammer. The device consisted of a 7.62 cm (3) tall x 2.54 cm (1) ID chamber made by welding a segment of pipe onto a 7.62 cm2 (3)2 x 0.635 cm () thick plate. A plunger made from a stainless steel billet was machined on one end to accept a standard 3/8 square drive adapter attached to a Hilti T52 (Hilti Incorpor ated, Germany) rotary hammer. Each specimen was placed in the cham ber and the piston lowered down on top of the shell. The rotary hammer was engaged for 10 sec onds or less, long enough to pulverize the shell. The resulting powder was collected in pre-weighed ceramic cruc ibles, dried for 24 hours at 80 oC and weighed to the nearest 0.001 g. Shell biomass recovered from the grinding device averaged 94.9 % (SE = 0.3, n = 238) of the original shell DW. Tissue samples were then placed in a muffle furnace at 550 oC for 6 hours and cooled to room te mperature in a dessicator prior to

PAGE 55

55 weighing. The ash values were used to cal culate percentage ash composition and AFDW (Wetzel 2001). Dry weight values were primarily used to assess clam biomass since there was less variability and a larger sample size (n = 453) than for AFDW (n = 238). However, AFDW/DW relationships did provide insight into organic content and biom ass allocation. The coefficient of variation was very similar for DW (0.54) and wet weight (0.52) for the animals in this study (n = 453). Wet weight measurements taken from whole fr ozen clams at the time of analysis were also used to help describe biomass allocation. Shell, meat and whole clam DW measuremen ts were performed on the low nutrient group at stocking, interval 1 and interv al 2, the medium nutrient raceways at stocking and at interval 1, and on the high nutrient raceways at stocking. The AFDW of shell, meat and whole clams were taken from the low nutrient raceways at stocking and at interval 1 and the medium nutrient raceways at stocking only. The DW and AFDW analysis were discontinued after these time periods due to the establishment of strong linear regressi on relationships (r2 > 0.90) between biomass and shell. Tissue biomass allocation was used to evaluate biomass distribution in shell and meat, as well as to examine biomass. Biomass allocation was also used to examine water and ash content of shell and meat tissue, for co mparison with other clam studies. Clam wet and dry weights of shell and meat were used to calculate percent shell tissue and percent water content for the whole clam. In order to understand variability in clam tissue biomass allo cation, an ANCOVA (SAS PROC MIXED procedure, SAS Institute Cary, NC) was performe d using both the percent shell tissue and the percent water co ntent as the response variables a nd nutrient level as the factor with covariates shell length, time interval, nutr ient level and season. The least squared means

PAGE 56

56 (LSMEANS) procedure was applied to percent shell tissue and the percent water content (Microsoft Excel) for pair-wise comparisons. Pair-wise comparisons of the means were performed using Tukeys method to control the expe riment-wise error rate. Data from individual raceways within each nutrient ad dition treatment were pooled in this analysis since no blocking effect was found in either analysis. The mean percentage of shell tissue and wate r content was determined for each nutrient addition treatment. Mean ash content was determ ined for clam, shell and meat tissues for each nutrient addition treatment. The coefficient of variation was calculated for the DW, and AFDW values determined for whole clam biomass to in dicate the least variable biomass parameter. Coefficient of variation was cal culated by dividing the standard deviation by the mean for each variable in the sample population. A relationship between tissue DW and shell length was used to provide a means of estimating biomass using measurements of sh ell size. Tissue DW and shell length were transformed using natural logarithm (ln) to best fit the polynomial re gression calculated by the SAS PROC REG procedure (SAS Institute Cary, NC). An ANCOVA (SAS PROC MIXED procedure, SAS Institute Cary, NC) was performed on the DW and shell length data using the natural logarithm (ln) of the va lues for the clam, shell and meat tissue DW. Tissue DW values were used as the response variables versus she ll length, while nutrient level was the factor with covariates time and season. Non-significant e ffects were removed from the ANCOVA model. The tissue DW and shell length values were also analyzed using a correlation procedure (SAS PROC CORR, SAS Institute Cary, NC). Clam biomass was estimated using shell lengt h to DW relationships developed in the regression analyses. Raceway clam biomass wa s calculated using the regression equations to

PAGE 57

57 convert shell measurements to tissue biomass. Changes in clam biomass over time were plotted for each nutrient addition treatment at each sa mpling interval to assess biomass production and clam growth. Reproduction and recruitment Clam populations were assessed for reproductio n and recruitment by determining whether juveniles were present in the system at any time Adult clams were used to stock the raceways, and therefore, any clams found with a shell leng th less than the smallest clam measured at stocking should indicate successful reproduction and recruitment. In this study, juvenile clams were defined as having a shell le ngth less than 5 mm. Source ponds were also drained at the end of the study to check for the presen ce of clams that may have been released as juveniles from the raceway populations but did not successfully recr uit to the raceway. Water samples from the source ponds used for phytoplankton analysis were also inspected for juveniles suspended in the water column. Health Changes in clam health were tested to eval uate the physiological condition of clams in the raceway environment over time. The goal was to relate changes in condition to changes in raceway environmental parameters. Other studie s have used percent meat content to assess population health (Haines 1977). A similar approach to the condition indices was used in this study, which also relies on the amount of meat and shell tissues pres ent in the clams. Condition index (CI) was estimated using gravimetric and volum etric indices, based on dry meat : dry shell weight and dry meat : shell cavity volume, respectively. Both indices are defined by the ratio of a sensitive numeratortissue dry weight, to relatively insensitive denominators-shell weight and sh ell cavity volume. The resulting values were compared for

PAGE 58

58 each index and used to describe changes in ra ceway clam populations. CI is a measure of the nutritive status of the cl am (Rainer and Mann 1992). Clams were collected at stoc king and during the seasonal time intervals described in the biomass sampling section. Meat DW and shell DW, along with shell cavity volume values obtained from biomass sampling, were us ed to calculate the gravimetric (CI(WT)) and volumetric (CI(VOL)) indices The following relationships, af ter Rainer and Mann (1992), were used to estimate CI values: CI(WT) = (dry meat weight (g) x 100 / dry shell weight (g)) CI(VOL) = (dry meat weight (g) x 100 / shell cavity volume (mL)) Figure 3-1. Linear regression relatio nship of glass sphere volume to glass sphere weight used to estimate clam shell cavity volume for the vo lume-based condition index calculation. Shell cavity volume was calculated by f illing the empty shell valves with 100 m diameter glass beads and weighing the beads. This method was selected since the small size of Corbicula makes it very difficult to assess volume accurately using the water displacement method (Rainer and Mann 1992). In order to obtain a weight to volume conversion factor, a volumetric analysis y = 0.1404x R2 = 0.9887 0 0.5 1 1.5 2 2.5 3 024681012141618 Bead Volume (ml)Bead Weight (g)

PAGE 59

59 was performed by weighing glass beads. Three samples of beads were weighed at each 1 mL increment up to 18 mL (n = 54), and a linear regression relationship was developed between bead volume and weight (Figure 3-1) usi ng the PROC REG procedure (SAS Institute Cary, NC). The regression equation from this anal ysis yielded the followi ng conversion formula: Clam Shell Valve Cavity Volume (mL) = Bead Weight (g) / 0.1414 A dry weight and shell volume analysis wa s then performed on a sample of 20 clams obtained from the Santa Fe River (29o51.1 North, 82 o37.9 West) to investigate possible differences in volume between the two valves of each clam. Three bead weight samples were taken from the two shell valves of each clam, and a paired, two-sample t-test for means was performed on the measurements using Micros oft Excel. The twosided p-value was not significant (p = 1), therefore, no difference in shell cavity volu me was found between the 2 shell valves in each animal. Therefore, by modifying the above conversion factor, the following equation was used to calculate the shell cavity volume for each clam: Clam Shell Cavity Volume (mL) = 2 x (Bead Weight (g) / 0.1414) Two different comparative an alyses were performed on CI(WT) vs CI(VOL) values to determine the best index for assessing clam health. A linear regression analysis was performed on each index versus length and CI(WT) vs CI(VOL) using the SAS PROC REG procedure, (SAS Institute, Cary, NC), and the coefficient of variation was calculated for both CI(WT) and CI(VOL). Variation was similar for the two indices; although, CI(WT) had the lower coefficient value. An ANOVA (SAS PROC MIXED procedure, SAS Institute Cary, NC) was then performed using the condition values for both indices as the response variable and shell length as the factor to determine the e ffect of clam size on condition for each index. Shell length was subsequently removed from the model, and an ANCOVA was performed with the CI(WT) and

PAGE 60

60 CI(VOL) indices as response variables, nutrient addition treatment as the factor with covariates time interval, raceways within each nutrient level and season. The least squared means (LSMEANS) procedure was applied to CI(WT) and CI(VOL) indices (Microsoft Excel). Pair-wise comparisons of the means were performed using Tukeys method to control the experiment-wise error rate. Results Raceway System Environmental Parameters Air tem peratures at the racewat y site at the Dairy Research Unit in Hague, FL ranged from 2 to 41 oC (Figure 3-2). Recorded air temperat ures varied diurnally as much as 15 oC. Raceway water temperatures ranged from 10.1 to 32.6 oC (Figure 3-3) and displayed a similar seasonal pattern as air temperature. However, water temperatures only differed diurnally by a maximum of 2 oC. Water temperatures reached 30 oC or greater just after the beginning of the experiment in July 2002 through October 2002 and from May through the end of th e study in August 2003. Raceway dissolved oxygen (DO) averaged 8.79 mg/L (SE = 0.07) and ranged from 6.10 to 12.03 mg/L (Figure 3-4). DO was higher in the afternoons by up to 4.26 mg/L with greater diurnal differences in the warmer months. Higher DO values were observed during the November 2002 to April 2002 period corresponding to lower air and water temperatures. Raceway pH averaged 7.76 (SE = 0.02) and ranged from 6.87 to 8.81. Diurnal fluctuations in pH ranged from .76 to 0.89. Raceway pH was significantly higher (p > 0.05) in the low and high nutrient addition treatments (Figure 3-5); however, water temperature and DO values did not differ significantly between the low, medium and high nutrient addition treatments. No significant differences (p > 0.05) were detected at the input and output of each raceway or between raceways in each nutrient addition treatment for the water temperature, dissolved oxygen and pH parameters.

PAGE 61

61 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateTemperature ( oC) Figure 3-2. Air temperature readings at the Dairy Research Unit in Hague, FL over the study period. Air temperatures varied both s easonally and diurnally with afternoon values exceeding 40 oC in the summer of 2002. 5.0 10.0 15.0 20.0 25.0 30.0 35.0 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateTemperature ( oC) Low Med High Figure 3-3. Input water temperat ures in the low, medium and hi gh nutrient addition treatments. Water temperatures fluctuated both di urnally and seasonally, often exceeding 30 oC in the afternoons and did not differ between treatments.

PAGE 62

62 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateDO (mg/L ) Low Med High Figure 3-4. Raceway dissolved oxygen (DO) re adings in the low, medium and high nutrient addition treatments. Values did not differ between treatments. 6.00 6.50 7.00 7.50 8.00 8.50 9.00 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DatepH Low Med High Figure 3-5. Raceway pH in the low, medium and high nutrient addition treatment systems. Raceway pH values fluctuated both sesa sonally and diurnally and tended to be elevated in the low and high nutrient addition treatments.

PAGE 63

63 Increases in phosphorus and ni trogen levels in the source ponds did not correspond to fertilizer addition or clam mortality events. Source pond total phosphorus (TP) ranged between 0.061 and 0.211 mg/L in the low nutrient additi on treatment, 0.047 and 0.471 mg/L in the medium nutrient addition treatment and 0.043 a nd 0.386 mg/L in the high nutrient addition treatment (Figure 3-6). Sour ce pond total dissolved phosphorus (TDP) ranged from 0.002 to 0.091 mg/L in the low nutrient addition treatment 0.007 to 0.223 mg/L in the medium nutrient addition treatment and 0.028 to 0.300 mg/L in the high nutrient addition treatment (Figure 3-7). A major portion of the source pond TP was made up of the dissolved form as TDP in all of the nutrient addition treatments as TDP followed a similar pattern as TP with sharp increases in the spring 2003. Source pond total nitrogen (TN) ranged from 0.177 to 9.034 g/L in the low nutrient addition treatment, 0.328 to 12.069 g/L in the medium nutrient addition treatment and 1.066 to 2.786 g/L in the high nutrient addition treatment (Fi gure 3-8). Total N values fluctuated in the low and medium nutrient addition treatments and only slightly increased in the high nutrient addition treatment over the experimental period. Chlorophyll a (chl a) ranged from 3.218 to 27.511 mg/m3 in the low nutrient addition treatment, 4.505 to 26.397 mg/m3 in the medium and 19.789 to 147.299 mg/m3 in the high (Figure 3-9). Chlorophyll a in all treatments displayed an increase after February of 2003 with peaks from April to August of 2003. Phyt oplankton communities in the source ponds were dominated by diatoms in the low nutrient additi on treatment, diatoms and cyanophytes in the medium and chlorophytes in the high. Diatom s were present in all ponds throughout the study period.

PAGE 64

64 Figure 3-6. Total phosphorus (TP) in the low, medium and high nutrient addition treatments. TP increased during the summer of 2003 in the medium and high treatments while values fluctuated with no apparent seasonality in the low treatment. Figure 3-7. Total dissolved phosphorus (TDP) in the low, medium and high nutrient addition treatments. Total phosphorus was comprise d largely of TDP ther efore patterns were similar to trends in TP. 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateTP (mg/L) Low Med High 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateTDP (mg/L) Low Med High

PAGE 65

65 0.00 2.00 4.00 6.00 8.00 10.00 12.00 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateTN (g/L) Low Med High Figure 3-8. Total nitrogen (TN) in the low, medium and high nnutrient addition treatment source water. TN fluctuated in the low and medi um treatments and steadily increased in the high nutrient addition treatment. Figure 3-9. Chlorophyll a (chl a) in the low, medium and high nutrient addition treatment source ponds. All treatments show an increase in chl a beginning in Febuary 2003. 0 20 40 60 80 100 120 140 160 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateChl a (mg/m3 ) Low Med High

PAGE 66

66 Clam Population Dynamics In Treatment Raceways Survival Clam stocking densities were estimated at between 6,336 to 10,686 individuals per raceway using the volumetric method (Table 3-3), which was near the target stocking density of 10,000 clams per raceway. The number of live clam s in each raceway declined at each sampling interval for all raceways in each of the nutrient addition treatment systems (Table 3-4). All nine raceways had decreasing densities and low number s of surviving live clams at the end of the study period (Table 3-5). The actual number of clams stocked was reduced by 99 % by the end of the study in all raceways; however, a sma ll number of clams survived (Table 3-5). The actual number of clams stocked in each raceway was overestimated using the volumetric method of estimation at the time of stoc king. Therefore, the va lues given in Table 33 were not used in any further analysis. The fi nal live clam densities for each raceway estimated by the spatial method (Table 3-4) were more simi lar to the actual number of live clams found in each raceway at the end of the study (Table 3-5). Table 3-3. Number of clams stocked in each raceway estimated using the volumetric method. The estimated number of clams per raceway (RW) and standard errors (SE) were calculated from the mean number of clam s measured per transfer cup and the number of full cups added to each raceway. The number of clams sacrificed for biomass analysis was subtracted to yield the estimated clams/raceway. RW Cups/RW Mean clams/cup SE n Estimated clams/RW SE 1 52 206 4 10 10,686 208 2 51 207 6 10 10,547 306 3 51 190 4 10 9,670 204 4 50 192 3 10 9,575 150 5 50 196 7 10 9,790 350 6 50 196 4 10 9,775 200 7 99 64 6 10 6,336 594 8 69 102 12 10 7,004 828 9 109 78 13 10 8,448 1417

PAGE 67

67 Table 3-4. Number of live clams found alive at each sampling interval estimated using the spatial technique. Clams per raceway (RW) and standard errors (SE) were calculated using the average number of clams found per 100 cm2 quadrat sampled, (n = 27 samples per raceway) converted to the number of clams per m2,which was multiplied by the 7.2 m2 of available substrate area per raceway. Summer 2002 Fall 2002 Winter 2002 Spring 2003 RW Clams/ RW SE Clams/ RW SE Clams/ RW SE Clams/ RW SE 1 4000 306 560 150 427 110 320 117 2 4400 340 907 179 1013 259 880 208 3 4267 295 827 166 613 251 80 44 4 n/a n/a 4282 387 1973 292 400 129 5 n/a n/a 4128 286 2747 356 640 155 6 n/a n/a 4377 563 2827 354 400 111 7 n/a n/a n/a n/a 2476 294 27 27 8 n/a n/a n/a n/a 3225 369 133 67 9 n/a n/a n/a n/a 4665 424 107 74 There were two indications that the spatial technique may have also overestimated the actual number of clams per raceway at each time interval : 1) raceway densities estimated at the end of the study were much higher than the actu al number of live clams found in each raceway at the end of the study and 2) the increase in clam density in raceway 9 from stocking (Table 3-5) from the first sampling interval (Table 3-4) c ould not be substantiated since 123 dead clams had been removed from the raceway during that time period and no indications of reproduction were observed. The number of live clams in each nu trient addition treatment (Figure 3-10) was calculated using the actual number of clams stocke d (Table 3-5), the actual number of live clams found at the end of the study (Table 3-5) and the number of clams at each sampling interval in between (Table 3-4). No attempt at biomass harvest was made due to the high losses of clams over the study period. Clams were only removed from the raceways as needed for periodic mortality counts and at seasonal sampling intervals for biomass determin ation. There were no i ndications of predation by wildlife as evidenced by the lack of obvious disturbance to the raceway substrate and lack of

PAGE 68

68 tracks that would have been left by mammalian pr edators around the raceway site. Large birds such as cattle egrets often congregated on the si des of the raceways; however, predation of clams in the raceway by avian predators was not directly observed. Table 3-5. Actual number of live clams stocked in each raceway and at the end of the study. Actual number of clams stocked was dete rmined from cumulative counts of live and dead clams over time as well as live a nd dead clams present at the end of the study. The number of clams removed re presents losses due to mortality and destructive sampling but not predation. Raceway Actual clams stocked Clams removed Live clams present at finish 1 7083 7047 65 2 9251 9210 47 3 6794 6771 23 4 6356 6301 55 5 4663 4601 62 6 4478 4434 38 7 4880 4871 9 8 5340 5323 17 9 6338 6319 19 Clam survival could not be accurately veri fied using periodic counts of dead shells removed from the substrate surface of each racew ay because dead clams were also found buried in the substrate, making them inaccessible for enumeration until the entire raceway could be excavated at the end of the experiment. The cu mulative number of dead clams collected on the substrate surface after large mortality events (F igure 3-11) reflects trends in reductions of raceway live clams (Figure 3-10). Dead clams removed over the course of the study and at the end of the study accounted for more than 90 % of the total number of clams stocked. Growth There were no clear patterns in sh ell size distributions at stocking or at each sampling interval; therefore, a cohort-based analysis of changes in shell size wi thin the raceway population could not be performed. Instead, the average shell le ngth of the individuals found over time in each raceway is given (Table 3-6). No consistent long-term trends in shell size data were observed

PAGE 69

69 over time in raceways 1-6. All raceways showed some increase in mean shell length at the first sampling interval, and clams in raceways 7-9 show ed a continued increase in mean shell length at Interval 2; however, any trends present must be viewed with caution due to the decreasing sample size and high mortality rates. Figure 3-10. Number of live clams in each nutri ent addition treatment. No increases in live clam density were found in any of the nutrient addition treatments. Figure 3-11. Cumulative number of dead found on the substrate surface in the low, medium and high nutrient addition treatments. Mortality was greater during the warmer periods in all three nutrient addition treatments.

PAGE 70

70 Table 3-6. Mean shell lengths measured from clams in each raceway (RW) both at stocking and at each sampling interval. Mean shell length displayed a constant increase in the high nutrient addition tr eatment raceways; however, decreases in sample sizes (n) over time in each raceway prevented positive identification of any discernable trends in shell size distributions and mean shell length over time. RW Shell Length Time Interval Stocking Interval 1 Interval 2 Interval 3 Interval 4 1 Mean 18.6 20.3 20.1 19.5 21.2 SE 0.2 0.2 0.5 0.6 0.5 Range 10.9-30.6 16.6-30.0 13.2-24.1 14.4-23.3 18.7-23.3 n 270 150 21 16 12 2 Mean 18.0 20.4 15.9 16.4 20.1 SE 0.2 0.2 0.7 0.5 0.4 Range 8.0-27.1 8.9-28.7 10.2-24.6 11.0-23.6 11-24.1 n 270 165 34 38 35 3 Mean 18.0 20.8 17.5 17.4 19.6 SE 0.2 0.2 0.8 0.4 0.3 Range 8.5-28.5 13.2-24.1 8.8-24.4 15.0-22.4 19.1-20.0 n 270 160 31 23 3 4 Mean 19.2 20.0 22.5 22.6 SE 0.2 0.2 0.2 0.8 Range 9.21-28.9 10.6-29.0 16.3-29.2 13.9-25.6 n 270 200 74 15 5 Mean 19.5 19.8 22.6 20.0 SE 0.2 0.2 0.3 1.0 Range 12.1-27.4 12.8-33.5 10.0-28.0 9.7-24.8 n 270 206 102 24 6 Mean 19.1 18.9 22.0 20.3 SE 0.2 0.2 0.3 1.2 Range 11.1-26.5 11.7-31.3 8.4-30.0 10.2-25.1 n 270 203 105 15 7 Mean Length 22.2 25.8 31.4 SE 0.5 0.6 0 Range 10.2-34.1 16.4-34.3 31.4 n 270 130 1 8 Mean 24.9 25.0 29.2 SE 0.4 0.4 1.5 Range 10.3-38.8 17.3-33.3 25.0-33.0 n 270 168 5 9 Mean 22.7 26.5 27.0 SE 0.4 0.4 1.1 Range 10.3-34.0 5.8-41.3 23.7-28.7 n 270 222 4

PAGE 71

71 Of the 972 total tagged clams introduced to the raceways, only 25 were recaptured alive over the course of the study. All of the recaptured live tagged clams showed increased shell length (Figure 3-12). The largest increases in shell length were observed in clams recovered from the high nutrient group, and clams in the me dium nutrient addition treatment appeared to have a larger increase in shell growth during warmer months (Table 3-7). The range in shell length for the individuals sampled for biomass an alysis was 10.2 to 33.7 mm (Table 3-8). The high nutrient addition treatment contained indivi duals with the largest shell size (31-33.7 mm) compared to the low and medium nutrient addition treatment. Figure 3-12. Changes in shell lengths of tagge d clams captured alive in each nutrient addition treatment. Growth was observed in all liv e tagged clams recove red with the largest increases found in the high and medium nutrient groups. 10.0 15.0 20.0 25.0 30.0 35.0 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateShelll Length (mm) Low Med High

PAGE 72

72 Table 3-7. Shell growth rates for tagged cl ams in each nutrient addi tion treatment for each seasonal time interval. Growth rates were calculated from initial and recovered shell lengths of live, tagged clams. Shell growth rates were similar for live, tagged clams found at any time in the low nutrient addition treatment, while growth rates during the spring were much higher than the winter for the medium nutrient addition treatment. Nutrient treatment Season Mean growth rate (mm/day) SE (mm/day) Growth rate range (mm/day) Initial shell size range (mm) n Low All 0.0119 0.0036 0.0009-0.0363 13.3-27.7 10 Med Winter 2002 0.0031 0.0008 0-0.0056 17.5-20.5 7 Med Spring 2003 0.0556 0.0158 0.0258-0.0856 17.023.9 4 High Spring 2003 0.0830 0.0168 0.0427-0.1173 12.2-28.6 4 Table 3-8. Shell size information on clams sampled for tissue biomass analysis from each nutrient addition treatment raceway system The largest individuals were found in the high nutrient addition treatment. Nutrient treatment Average length (mm) SE (mm) Shell size range (mm) n Low 19.9 0.2 10.2 30.4 210 Med 20.3 0.2 13 28.4 162 High 24.8 0.7 12.6 33.7 81 The mean water content of whole clams was 31. 5 % (SE = 0.4) with a range of 25.0 45.1 %, n = 453 observations. Water cont ent did not vary significantly (p > 0.05) with shell length, between raceways, over time or between nut rient addition treatments based on the ANOVA analysis. The wide range of water content values observed pr evented the use of tissue wet weights to assess biomass even though clam wet we ight values were highly correlated with the whole clam DW values (corr coeff = 0.90, n = 453 observations). Ash content of the meat tissue was much lower than the shell tissue, and shell ash made up a greater portion of the total clam ash than meat tissue (Table 3-9). As h content did not vary significantly (p > 0.05) with shell length, be tween raceways, over time or between nutrient addition treatments based on the ANOVA analysis.

PAGE 73

73 Table 3-9. Mean and range of ash content valu es for meat, shell and total clam tissues pooled for all clams sampled. Shell tissue had higher ash content an d made up a greater portion of the total clam as h than meat tissue. Tissue Mean ash (%) SE (%) Ash content range (%) n Clam 94.71 0.06 90.45 97.39 238 Shell 97.48 0.02 96.06 98.23 238 Meat 13.74 0.55 2.50 17.00 238 Overall, clams sampled in this study allocate d a mean of 95.9 % (SE = 0.1) of total clam DW biomass as shell tissue with a range of 88.4 99.7 %, n = 453. Meat biomass accounted for the other 4.1 % of the dry weight on average (SE = 0.1) with a ra nge of 0.3 to 11.6 %, n = 453. The amount of DW biomass allocated to shell tissue did not va ry significantly (p > 0.05) with shell length, between individual raceways within each nutrient a ddition treatment, over time in each nutrient level or between nutri ent addition treatments. Dry wei ght values for the meat, shell and whole clam varied significantly (p < 0.05, n = 453) with shell length. Correlation analysis showed a strong relationship be tween shell and whole clam ti ssue biomass and shell length (Table 3-10). Table 3-10. Results of the meat, shell and tota l clam tissue dry weight (DW) to shell length correlation analysis. Shell and whol e clam tissue biomass had a stronger correlation to length than to meat tissue Length Meat DW Shell DW Clam DW Length 1 x x x Meat DW 0.626 1 x x Shell DW 0.966 0.599 1 x Clam DW 0.967 0.626 0.999 1 No statistically significant differences (p > 0.05) in the lengt h-weight regression relationships were found between raceways, over time or seasonality in each nutrient addition treatment using the ANOVA. Howe ver, significant differences (p > 0.05) were found in each nutrient level for all meat, shell and whole clam tissue. Tissue DW and shell length values were

PAGE 74

74 then pooled for the clams in each nutrient addi tion treatment, and length was removed in the ANOVA. Regression relationships between shell length and DW are presented in Table 3-11 for each tissue type and each nutrient addition treatmen t. Intercepts and slopes of the regression lines for ln (clam DW) and ln (shell DW) for th e high nutrient group were significantly different (p < 0.05) than the low and medi um nutrient levels. This relati onship did not differ significantly (p > 0.05) between the low and medium nutri ent addition treatments; however, high nutrient addition treatment did differ significantly (p = 0.03) from the low and medium levels. Table 3-11. Dry weight (DW) biomass vs length regression rela tionships, significance differences and variability for whole clam, shell and meat tissues from each nutrient addition treatment. This relationship did not diffe r significantly between the low and medium nutrient addition treatments The relationship fo r the high nutrient addition treatment was significantly diffe rent from both the low and medium nutrient addition treatments. Tissue Nut. level Regression Equations denotes significant difference (p < 0.05) r2 n Clam Low ln (clam DW) = -3.6884 + 0.2893 (length) -0.0038 (length2) 0.9394 243 Med ln (clam DW) = -3.6723 + 0.2873 (length) -0.0038 (length2) 0.9614 161 High ln (clam DW) = -4.9450* + 0.3949*(length) -0.0059*(length2) 0.9738 81 Shell Low ln (shell DW) = -3.8920 + 0.3042 (length) -0.0041 (length2) 0.9421 243 Med ln (shell DW) = -3.9270 + 0.3056 (length) -0.0042 (length2) 0.9642 161 High ln (shell DW) = -5.0245* + 0.3990*(length) -0.0060*(length2) 0.9718 81 Meat Low ln (meat DW) = -4.0291 + 0.0190 (length) + 0.0021 (length2) 0.2903 243 Med ln (meat DW) = -2.8659 0.0716 (length) + 0.0042 (length2) 0.2447 161 High ln (meat DW) = -6.7690* 0.2476 (length) -0.0029 (length2) 0.8895 81 Shell length values were plotted against the actual and predicted DW values derived from the regression equations (Table 3-11). These re lationships are plotted for the whole clam, shell and meat tissues in Figure 3-12, Figure 3-13 and Figure 3-14, respectively. The mean percentage shell tissue calculated using th e regression equations for the predicted whole clam biomass values was similar to the tissue DW and wet weight (WW) values.

PAGE 75

75 Figure 3-12. Regression relations hips for shell length vs actual and predicted (Table 3-11) whole clam dry weight (DW) values for each nut rient addition treatment. This relationship for the high treatment was significantly diffe rent than both low and medium nutrient addition treatments. Figure 3-13. Regression relations hips for shell length vs actual and predicted (Table 3-11) shell dry weight (DW) values for each nutrient addition treatment. The relationship was significantly different in the high nutrient addition trea tment than both the low and medium nutrient addition treatments. 0 1 2 3 4 5 6 7 81318232833 Shell length (mm)Shell DW (g) Actual Low Predicted Med Predicted High Predicted 0 1 2 3 4 5 6 7 81 31 82 32 83 3 Shell length (mm)Whole clam DW (g ) Actual Low Predicted Med Predicted High Predicted

PAGE 76

76 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 81 31 82 32 83 3 Shell length (mm)Meat DW (g) Actual Low Predicted Med Predicted High Predicted Figure 3-14. Regression relations hips for shell length vs actual and predicted (Table 3-11) meat dry weight (DW) values for clams. Re lationships for the high nutrient addition treatment were significantly different than the low and medium nutrient addition treatments. These regression relationships can only be applied to clams in the 10.2 mm to 33.7 mm shell length range for the high nutrient addition treatment and the 10.2 30.4 mm range for the medium and low treatments that were used to ca lculate them (Table 3-8). Interpolation of the equations (Table 3-11) for clams beyond 34 mm upper limit results in reduced or logarithmically increased tissue dry weight estimates due to the shape of the extended regression line. Biomass at stocking and at proceeding time inte rvals was estimated using length vs weight relationships, similar to Joy and McCoy (1975). Calculations of biomass were made for each nutrient addition treatment (Figure 3-15) using average shell length (Table 3-6). Shell length to dry weight regression equations (Table 3-11) and live clam densiti es (Figure 3-10) were similar to those found by Cataldo et al. ( 2001) who were not able to dis cern population cohorts. It was

PAGE 77

77 acceptable to apply the regression equations to ca lculate dry weight biomass from the average shell lengths at each interval since shell lengt h values fell within the ranges required by each nutrient addition treatment level (Table 3-8). Estimated populati on biomass decreased markedly from stocking to the end of the study for each nutrient addition treatment. Figure 3-15. Estimated clam dry weight (DW) biomass over time in the low medium and high nutrient addition treatments. Increases in estimated DW biomass may not be indicative of tissue production due to the variability in the clam density measurements. Reproduction and recruitment No evidence of successful reproduction and recruitment was found in any of the nutrient addition treatments, as indicated by declines in clam populations (Figure 3-10), the lack of individuals less than 7 mm in sh ell length (Table 3-6), and no evidence of juveniles in any of the water samples used in the phytoplankton speciation survey.

PAGE 78

78 Health Linear regression analysis yi elded a correlation coefficient of 0.85 between the condition indices based on CI(WT) and CI(VOL), indicating that they were very similar. The CI(WT) index had a slightly lower coefficient of variation (0.53) than CI(VOL) (0.56) (n = 452). Shell length was not a significant factor (p > 0.05) in the values of either index, so it was removed from the analysis. Results for the ANOVA for both indices were similar, showing significantly higher condition values for both indices (p = 0.01) in the medium nutrient leve l, but only at sampling Interval 1. CI(WT) and CI(VOL) ranged from 0.302 to13.210 and 1.435 to 15.831 per individual, respectively (Table 3-12). C ondition indices calculated for cl ams in the medium nutrient addition treatment at interval 1 tended to be higher than all other nutrient addition treatment/interval combinations. This trend was probably due to slightly higher meat dry weight values in the medium nutrient addition treatment at interval 1. Consequently, meat DW values were not significantly different at interval 1 compared to the ot her intervals within the medium nutrient addition treatment; therefore, the CI is ap parently highly susceptible to differences in meat tissue quantity that may not be a pparent using the regression analysis. Table 3-12. Mean, standard error (SE) and range of condition indices values (CI(WT) and CI(VOL)) calculated for the medium nutrient addition treatment at Interval 1 compared to values calculated at all other treatment/interval combination. Values for both indices tended to be lo wer in all other treatment/interval combination. Condition index values Population Index Mean SE Range n Medium nutrient CI ( WT ) 7.758 0.238 2.798-13.1280 81 Interval 1 CI ( VOL ) 8.931 0.262 3.406-14.9269 81 All other CI ( WT ) 3.601 0.080 0.302-10.033 372 combinations CI(VOL) 4.581 0.131 1.435-15.831 372

PAGE 79

79 Amphipod infestation Amphipods (Hyallela azteca) began appearing in dead clam soft tissue and empty shell valves removed from the raceways on July 26, 2002 (38 days after introduction of clams). A sample of five live clams on the substrate surface was randomly removed for inspection on August 9, 2002 and yielded one clam with an am phipod. The clams were immediately shucked to reveal six additional live amphipods enclosed in the shell cavity of one clam. A notched clam specimen was removed from the sediment surface. The clam was shucked, treated with a 90 % ethanol solution and examin ed under a dissecting scope (10x 30x) fitted with a digital camera to reveal nine amphipods. Amphipods were identified and photographed by Gary Warren at the Florida Fish and Wildlif e Conservation Commission in Gainesville, Florida. Amphipods were ruled out as a direct ca use of the chipping due to their lack of hard mouthparts capable of damaging clam she ll material (Covich and Thorpe 2001). In an attempt to quantify the extent of th e amphipod infestation in the raceway clam populations, collections from the following three categories of clams were made: fresh dead clams, live clams on top of the sediment and live clams buried in the sediment. Fresh dead clams are defined as having gaped valves with soft tissue intact. Six live clams were removed from the sediment surface on August 9, 2002, and a sample of 10 live clams was excavated from the substrate on August 18, 2002. All clams were co llected at random from the same raceway. Amphipods contained in fresh dead clams were counted immediately after removal from the raceway, and any shell notching was noted. Live clams were placed in Ziploc bags, spun tight, sealed and wrapped with a rubbe r band to prevent clams from opening during transport to the laboratory for shucking and amphipod count. All clams sampled had a shell length greater than 20 mm.

PAGE 80

80 The observations showed that 100 % of the fresh dead clams sampled contained live amphipods. Amphipod abundances in these clams av eraged 12 individuals per clam and ranged from 2 to 23 individuals per clam, and 36 % of th e dead shell valves sampled had notches. All live clams taken from the substrate surface were notched, and only 50 % of these individuals contained amphipods. Amphipod abundances within the shell cavities in these live, notched clams ranged from 4 to 9 individuals per clam None of the live clams found beneath the raceway substrate surface had notches, and only on e of these clams contained amphipods within the shell cavity. This live clam, however, had the highest nu mber of amphipods recorded per live clam sampled, nineteen. After the initial observations on amphipod i nvasion and shell chipping in raceway clams were made, a small scale experimental system was constructed in an indoor laboratory at Alee Academy in Eustis, FL to test the repeatability of this phenome non in smaller-scale aquaculture environments. A total of 9 27-L glass aquaria were stocked with between 10, 20 or 30 clams (200 to 600 clams/m2) and 50, 75 or 100 amphipods (1075 to 2150 amphipods/m2) collected by hand from clam raceways and transported in larg e aerated coolers to the laboratory in Eustis, Florida. Lights were maintained on a 12 hou r-on/12 hour-off cycle and water temperature was regulated at 23-25 oC, using submersible heaters. Sand for the aquari um substrate was obtained from the same retailer as the Hague, FL facility. Aquaria were fed with 1 gram per week per aquarium of dried, powdered Spirogyra sp. algae obtained from the Hague raceways. No indications of shell chipping, pene tration of live clam shell cavities or high clam mortality were recorded. The lack of clam/am phipod interaction in these small-scale systems indicates that the Hague observations could be an isol ated incident or that the negative interactions are more likely to occur at larger scale or higher air temperatures.

PAGE 81

81 Discussion Use of the freshwater clam, Corbicula, as the primary active agent in wastewater treatment raceway systems yielded mixed results. Si gnificant growth and phosphorus sequestration was observed in tagged clams, which survived thr ough the study period. Growth in tagged clams appeared to be substantial during warmer pe riods in the medium and high nutrient addition treatment systems, with growth rates ranging fr om 0.043 to 0.117 mm/day (Table 3-7). These rates are similar to those reported for ot her freshwater bivalves (Table 3-13). Table 3-13. Individual shell length and biomass dry weight (DW) growth rates reported for Corbicula and other bivalves occupyi ng different fresh and saline environments. Some values were not reported by the refere nce author and are noted as (n/a). Organism Shell Length Growth (mm/day) Biomass Growth (g DW/day) Environment Type Reference Corbicula 0.117 0.0024 Agriculture nutrient treatment raceways This Study Corbicula 0.058 0.0023 Laboratory at optimum temp Foe & Knight (1986) Corbicula 0.180 n/a Natural river McMahon & Williams (1986) Corbicula 0.085 n/a Power plant effluent canals Mattice & Wright (1986) Corbicula japonica n/a 0.0077 Natural estuary Fuji (1979) Dreissena polymorpha 0.080 n/a Natural lakes and rivers McMahon & Bogan (2001) Dreissena polymorpha 0.095 0.0150 Great Lakes Bitterman et al. (1994) Elliptio complanata 0.001 n/a Natural lakes Anthony et al. (2001) Venerupis pullastra 0.189 0.0189 Mariculture effluent treatment raceways Jara-Jara et al. (2000) Tapes semidecussatus n/a 0.0422 Mariculture effluent treatment ponds Shpigel & Fridman (1990) Ruditapes decussatus 0.333 0.0083 Mariculture effluent treatment raceways Jara-Jara et al. (1997) Ruditapes philippinarum 0.377 0.0730 Natural estuary mariculture Nizzoli et al. (2006)

PAGE 82

82 Despite the positive results with tagged clams, the long-term performance of the broader clam population in the raceways was poor, and dairy wastewater effluent could not tested due to the high ammonia levels present in the effluent addition ponds throughout the study. The latter problem highlights an important issue in the us e of an animal-based system for wastewater treatment: ammonia toxicity. From a broader perspective, the high rates of mortality of clams in the raceway systems focus attention on a range of environmental and design issues that must be dealt with in future research efforts, such as temperature, food availability, dissolved oxygen, the effect of multiple environmental stressors, parasitism and predation, depressed reproduction in captivity and issues with enumeration of raceway clam biomass. Ammonia Concerns Use of Corbicula or other freshwater bi valves in large-scale raceway-based systems as a phosphorus treatment mechanism for dairy wastewater may be limited foremost by ammonia toxicity concerns that could sk ew normal water usage and land requirements. Ammonia in the toxic, unionized form ammonium (NH4 +), is a serious concern in the aquatic environment. Levels of ammonium of 0.2 mM as ammonium-nitrogen (NH4 +-N) negatively affect other bivalves and finfish (Patrick et al. 1968, Epifanio and Srna 1975). Amounts of the toxic form are directly proportional to pH and temperat ure and increases in ammonia-nitrogen (NH3-N) are generally used to gauge ammonia toxicity. Ammonia levels harmful to Corbicula and other freshwater musse ls have been suggested by Cherry et al. (2005) to be 0.11 to 0.8 mg/L total amm onia nitrogen (NH3-N). Levels of 0.041 to 0.158 mg/L NH3-N are harmful to bivalves in marine aquaculture systems and may even be harmful at concentrations as low as 0.006 mg/L NH3-N (Harris et al. 1998). The presence of ammonia concentrations in excess of 0.25 mg/L NH3-N in the source ponds receiving

PAGE 83

83 anaerobically digested dairy effluent in this study indicate that ev en effluent addition rates as low as 5 % to 10 % of source pond volum e are not acceptable for production of Corbicula. High ammonia levels in commercial aquacultur e systems decimate stocks if not managed correctly. This is especially true in recircul ating systems where the lack of water exchange necessitates removal of ammonia to achieve sa fe levels (Timmons et al. 2002). Keeping ammonium within a tolerable ra nge will be critical to using Corbicula as a wastewater treatment mechanism (Haines 1977). Additional treatments would need to be added to the dairy wastewater stream to manage am monium. Coupling vegetation with Corbicula into polyculture systems has also been proposed for nutrient uptake in agricultural waste streams (Stanley 1970, Greer and Ziebell 1974, Mattice 1977). This type of system may provide an additional ammonia management strategy in future applications; however, no large-scale experiments have been developed to demonstrate the appl icability of such technologies. Anaerobically digested dairy effluents may also present additional ammonia concerns compared to non-digested effluents, since dige ster systems convert nitrogen bound in solids to ammonia by microbial degradation, which adds to the ammonia al ready present in the barn and milking operation wastewaters (Wil kie et al. 2004). Potential pr oblems with introduction of harmful levels of ammonia were circumvented in this study by simulating algal concentration under dairy wastewater conditions using nitrogen and phosphorous fe rtilizer applied at different loading rates. Among the raceways supplied by pond water of di fferent nutrient concentrations, all were subject to high rates of clam mort ality over time, particularly in summer. It is possible to form several hypotheses about the environmental factors responsibl e for these losses other than ammonia, including high temperature, limite d food availability, low dissolved oxygen and

PAGE 84

84 parasite interactions. It is usef ul to explore each of these cons iderations in an effort to find possible solutions and future directi ons for research and development. Temperature Temperature probably had the most impact on growth of Corbicula populations in the clam raceways. Tagged clams recaptured alive in this study showed growth in water temperatures between 10 and 30 oC with higher growth rates occu rring at temperatures above 18 oC. Growth rate in Corbicula increases with temperature as Matti ce (1977) observed a maximal growth rate of clams in power plant canals at 24 oC. Other studies using lab experiments by Mattice (1977) and Foe and Knight (1986) suggest that temperatures above 25 oC cause a decrease in growth. Buttner and Heidinger (1980) observed the highest growth rates in the temperature range of 11.2 to 24.7 oC and minimal growth at 4.2 oC. Conversely, exposure to temperatures exceeding 30 oC coincided with increased populat ion mortality in this study. McMahon and Williams (1986) suggest an upper lethal limit of 36 oC for naturally occurring populations of Corbicula. Habel (1970) reports 98 % mortality when temperature exceeds 35 oC, and Busch (1974) observed no growth a nd high mortality when temperatures exceed 32 oC in polyculture ponds shared with ca tfish. Buttner (1986) suggests that temperatures above 30 oC be avoided when culturing Corbicula, which is supported by the results of this study since water temperatures greater than 30 oC were observed through much of the study (Figure 3-2), and losses of biomass (Figure 3-11) were consistent with exposure to water temperature above 30 oC. Site selection based on temperature may be the key to successful application of Corbicula to wastewater treatment. High temperatures have been implicated in limiting success of other Corbicula-based experimental wastewat er treatment systems in arid and tropical areas prone to high temperatures such as Ar izona (Greer and Ziebell 1972) and the US Virgin Islands

PAGE 85

85 (Haines1977). Pond-based experiments examining use of Corbicula for aquaculture effluent treatment in temperate Illinois have also repor ted problems with high te mperatures (Habel 1970, Busch 1974, Buttner 1986). Implementation of Corbicula-based treatment systems may be restricted to locations where wa ter temperatures do not exceed 30 oC as suggested by Greer and Ziebell (1972) and may require geographical loca tions even farther north or in traditionally colder climates than Florida or even Illinois. Corbicula-based systems may be better suited to a pplication in colder areas because there are no real effective techniques fo r cooling water in systems of th is scale or larger. Source ponds at the Hague site were aerated at night to induce evaporative cooling, but water temperatures still reached 30 oC and above in the afternoon throughout th e spring, summer and fall. Industrialsized water chillers are not a practical soluti on due to equipment and energy costs that would exceed the value of the treatment potential. Sh ading of open water areas may help reduce heat buildup, but it also lowers light availability for phytoplankton productivity. Shading of small raceways with plywood coverings has been used by Haines (1977) to prevent nuisance vegetation growth in Corbicula systems and greenhouse netting is also commonly used in commercial aquaculture and agriculture operations but the effects that covering of raceways may have on clam populations through changes in temperature, oxygen exchange and disruption of possible circadian rhythms are not known. Heat exchange with cold water drawn deep from aquifers or surface water reservoirs may provide solutions, however, the energy required for pumping and lack of specific locations with available cold water sources may prohibit implementation. Application of Corbicula to systems in colder climates may result in limited biomass production at temperatures below 4.2 oC (Buttner and Heidinger 1980) but cold-related

PAGE 86

86 mortalities as well (McMahon and Williams 19 86). Ice cover can depress dissolved oxygen below tolerable levels in north ern areas (McMahon and Bogan 2001) Temperatures in this study did not fall into the low temperature ranges (< 5 oC) identified by McMahon and Williams (1986) as lethal; however lower clam growth rates were evident during the winter months. Waste heat generated from power plants has been proposed by Mattice (1977) for encouraging biomass production in Corbicula aquaculture in colder climates, however, this resource may not be applicable to dairy operations, since locations of power plants would determine available sites for clam raceways. Land use requirements for coupling power plants and dairy farms may be the ultimate barrier to waste heat utilization of this type due to the potential contamination by agriculture practices in the vicinity of surface water reservoirs needed for power plant cooling towers. The need for tran sfer of wastewater ove r long distances in the volumes produced by typical dairy farms (avera ge milking herd of 359 cows = wastewater production of 502 m3/day at Dairy Research Unit in Hague, FL (Wilkie et al. 2004), decreases the attractiveness of this type of waste heat us age at any appreciable scale in dairy wastewater treatment. Food Availability Food limitation may also have been a problem in the survival of clams in this study. Phytoplankton availability has been implicated as a limiting factor for Corbicula biomass production in natural systems (Foe and Knight 1986, Mattice and Wright 1986), and low particle concentrations have been li nked to poor performance in Corbicula-based wastewater treatment systems (Haines 1977). Chlorophyll a levels remained below the 100 to 300 g/L range reported by Cohen et al. (1986) to sust ain growth and survival in natural systems. Source pond phytoplankton production was probably limited to factors other than phosphorus availability.

PAGE 87

87 As expected, increased in source pond wa ter phosphorus in the high nutrient addition treatment system corresponde d with an increase in phyt oplankton biomass; however, phytoplankton biomass remained low in the medium nutrient addition treatment system (Figure 4-3). Utilization of phosphorus by phytoplankton is the first stage in this clam raceway system; however, performance of this phase was poor since there was a large portion of dissolved phosphorus in all of the system s indicating that phytoplankton production was probably not limited by phosphorus. In order to identify possibl e limiting factors, study of nutrient dynamics beyond nitrogen and phosphorus may be needed in large-scale systems when utilizing phytoplankton incubation ponds for clam raceway source water. Phosphorus and nitrogen levels in source pond water after the one-month incubation period following fertilizer additions were similar to pre-addition levels in both fertilizer addition treatments. Source pond total phos phorus concentrations of 0.23 and 0.46 mg P/L were expected in the medium and high nutrient addition treatments repectively, if the fertlilizer P would have been retained in the water column. These total phosphorus levels were not achieved until February 2003 after the clam raceways had been operational for at least one month. Initial investigation of phytoplankton prod uctivity with wastewater additi on, as demonstrated in clambased treatment system scaling by Borges et al. (2005), should be empl oyed to enhance largescale clam raceway food availability prior to clam exposure to identify and solve potential problems. Although Corbicula can utilize a wide vari ety of particles as a f ood (Silverman et al. 1995, McMahon and Bogan 2001), the potential for its use in phosphorus treatment depends on the extent that phosphorus is transformed into a pa rticulate form for clam uptake. Phytoplankton species management could potentially impact clam food resources since food value, and algal

PAGE 88

88 phosphorus content can vary with speciation and environmental conditions (Wetzel 2001). This may be ideal in theory, but it may not be practical at a large scale due to the susceptibility of the outdoor systems to influx of local genetic material. Maintaini ng desired phytoplankton composition in large-scale wastewater treatment through inoculation may not be effective in governing algal speciation in outdoor, larger-scal e systems as suggested by Greer and Ziebell (1972) and Haines (1977). Management of phytopl ankton nutrients in larg e-scale clam raceway systems may also require additional engineering solutions such as pond liners to minimize losses of nutrients due to th e interaction with sedi ments and groundwater. Dissolved Oxygen Low dissolved oxygen levels associated with receiving waters downstream from municipal wastewater treatment plant discharges have also been associated with high mortality and low growth rates in Corbicula (Belanger 1991). In this study, di ssolved oxygen never fell below 6.0 mg/L, which is well over the 3.0 mg/L minimum suggested by Buttner (1986) for Corbicula under aquaculture wastewater conditions. Main tenance of adequate dissolved oxygen may be attributed to the design of the raceway input spre ader bar that was insta lled to help volatilize ammonia and aerate incoming raceway water. Multiple Stressors Temperature alone may not have been respons ible for the poor surv ival of clams in raceways since other chemical and physical st ressors in the environment can decrease the threshold temperature for survival. Fo r example, Buttner (1986) attributed Corbicula biomass losses to temperatures up to 33 oC, along with depressed oxygen levels (< 40 % saturation) at temperatures above 25 oC. Haines (1977) attributes lack of growth, high mortalities and temperature stress, in combination with th e possible presence of ammonium, at 30-32 oC in municipal wastewater conditions. Other studies have attributed no growth along with substantial

PAGE 89

89 stocked biomass loses (Habel 1970, Haines 1977) to a combination of high temperatures and low dissolved oxygen (Buttner and Heidinger 1980, Buttner 1986). Interactions of these concerns, along with possible amphi pod interactions observed in th is study, may also inhibit implementation of Corbicula-based systems in dairy wastewater treatment. Parasites and Predation Potential problems with amphipods that infe sted clam raceways at the Hague, Florida facility were unforeseen and may have contributed to the limited adaptability of freshwater clams to the large-scale raceway aquaculture system s. The previously undescribed phenomenon of Corbicula infestation by the amphipod Hyalella azteca may have contributed to population declines, decreased filtration and lack of growth. This phenomenon may have been the result of the substrate choice and other environmental pa rameters unique to the raceway systems that could be encountered in ot her engineered systems. Negative impacts on Corbicula may be due to direct parasitic interactions or tissue damage from movement and occupation within the shell cav ity; or possibly its incurrence of stress from constant valve closure stimulated by amphi pod movements over the raceway substrate. Repetitive valve closure as a result of the clams defensive response to outside stimuli, such as the constant activity of an amphipod infestation, may deplete energy reserves and limit its ability to respire, expel waste products and obtain food (Gainey 1978). Shell chipping appeared to be related directly to the ability of the amphipod to invade clam shell cavities. Th ere was, however, evidence that amphipod invasion of sh ell cavities can also occur in the absence of a chipped shell area; therefore, amphipod invasion needed to be verified by shucking li ve, freshly collected organisms. The chipping of shell material was probabl y related to the amphipod infestation even though the breech in shell tissue may not have been the only entrance mode. Chipping was

PAGE 90

90 probably caused as a result of repetitive valve closure and the inability of the clams to expel sand grains as the valves were closing. Shell marg in chipping was not evident in the systems exposed to fertilizer addition, which may have caused a difference in substrate characteristics with the accrual of soft sediments as a result of settleable material such as phytopl ankton and clam wastes over the sand substrate. This change in sedi ment structure may have helped to deter shell chipping and the phenomenon of inter-cavity invasion by amphipods since live tissue samples from these raceways did not contain amphipods. Clams undergoing physiological stress, such as the mortality events in this study indicated, may be more susceptible to shell cavity invasion by amphipods, even without the occurr ence of shell chipping. This idea is supported by the low number of live clams without evidence of she ll chipping that were found containing amphipods. Engineered systems are still in their experimental infancy an d may be subject to unforeseen parasitism and predation by organisms that are not normally pests in the natural environment as systems achieve larger and larger scale. Ne gative interactions be tween invertebrates and bivalves are more common in the mariculture industry, where organisms are maintained in high densities using either closed engineered (Dame 1996, Lin et al. 2001) or open-water natural systems (Hickman 1992, Haines et al. 1994, Dame 1996). Application of bi valve-based biofilter technology in the freshwater envi ronment may be susceptible to similar problems that are much less understood at this time due to the lack of large-scale culture expe rience beyond the marine environment. Parasitic and/or amensal interactions have b een described between na turally occurring and cultured open-water marine bivalves by inverteb rates such as pea crabs and other crustaceans (Hickman 1992, Haines et al. 1994, Dame 1996, Mercado-Silva 2005). In estuarine populations of Corbicula, attachment of barnacles has led to declin es in clam health and increased mortality

PAGE 91

91 (Foe and Knight 1986). Lin et al (2001) reported mortality due to parasitic intera ctions between Pyramidellid snails and the giant clam Tridacna derasa in an engineered mariculture effluent treatment system. Oysters (Crassostrea spp.) in both natural and culture systems are susceptible to negative effects of shell-boring polycheates (Wargo and Ford 1993, Debrosse and Allen Jr 1994). Other microscopic parasites have also ha ve a detrimental effect on oysters, including the endoparasite known as MSX (Wargo and Ford 1993), however, no effects from organisms of this nature have been documented in Corbicula. Observations and negative effects of inter-she ll cavity parasites in naturally occurring freshwater clams and mussels have been reported in a variety of systems. Parasitism of the freshwater clam Pisidium amnicum in Finish lakes by tremetodes (Bunodera luciopercae, Palaeorchis crassus and Phyllodistomum elongatum) has been reported by Holopainen et al. (1997). Parasitism of freshwater unionid mussels by unionicolid mites in the United States has been reported by Edwards and Dimock, Jr. (1988) and Fisher et al. (2000) Freshwater mussels inhabiting Chilean salmon farms can be preyed upon by freshwater crabs and shrimp when the valves of the mussels are open (Soto and Mena 1999), but these predators did not inhabit the inside of the shell valves as did the amphipods in this study. Corbicula populations inhabiting natural systems in the United States may be impacted by parasitism of soft tissue by the oligochaete Chaetogaster limnaei, as described by Eng (1976). Corbicula stocked in fish polyculture ponds have been subj ect to predation by terrestrial mammals and fish (Chen 1976, Buttner and Heidinger 1980), but no repor ts of amphipods negatively affecting Corbicula populations have been made. This is likely the first observed report of the effects of Hyalella azteca presence as a possibly negative aspect in Corbicula populations as well as infiltra tion of the mantle cavity area

PAGE 92

92 by amphipods. Understanding the Corbicula and Hyallela interaction observed here will require further study in order to accurately describe understand and document the extent of the interaction and the effect that it may have on the animals health and survival. The extent that amphipods may affect freshwater clam and musse l populations under culture conditions can be magnified by substrate choice as well as other en vironmental stressors such as temperature that can pre-stress populations prior to amphipod infe station. Potential problems with amphipods identified here would need to be solved either by polyc ulture with other organisms as suggested by Soto and Mena (1999) and Lin et al. (2001) or pesticides tolera ble to clam populations. Even with fish polyculture, invertebrate pests may s till be problematic in freshwater bivalve-based treatment systems as indicated by Soto and Mena (1999). Incidental pa rasites and other pests would need to be managed in any future freshwater bivalve-based wastewater treatment system, which is difficult since organisms such as Hyalella azteca that occur with the clam in the natural environment may have an undescribed impact on the target organisms, especially under largescale monoculture conditions. Fish/clam polyculture may also help eliminat e potential problems w ith amphipods in clam raceways through predation, provided that the fish do not impact clam populations as well. Other problems from nuisance organisms encountered in the raceways, such as biofouling from plant growth, may also be solved through polycu lture of clams with ot her organisms such as herbivorous snails that have been used to c ontrol biofouling by macroalgae in clam-based mariculture systems (Lin et al. 2001). In natural and engineered systems, the presen ce of larger organisms may affect survival and growth of Corbicula; for example, predation of Corbicula in culture ponds by muskrats has been reported by Buttner and Heidinger (1980), but no indications of such disturbances were

PAGE 93

93 apparent at the Hague site. Ch en (1976) reported predation of Corbicula by fish in large-scale polyculture systems. In estuarine populations of Corbicula, attachment of barnacles has led to declines in clam health and increased mortalit y (Foe and Knight 1986). Parasitism of freshwater bivalves mussels by unionicolid mites in natura l water bodies has been reported by Edwards and Dimock, Jr. (1988) and Fisher et al. (2000), but no negative rela tionships between freshwater bivalves and amphipods has been reported in the literature. Reproductive Success An important part of Corbicula biomass production is the clams ability to reproduce and recruit new individuals to the raceways. The lack of juveniles (defined as having a shell length less than 7 mm) observed in the raceways and water samples in this study indicates that reproduction and subsequent recruitment did not take place. The 5 mm shell length threshold was chosen, as opposed to the 10 mm size used by Buttner (1986), since th e presence of clams 7 to 10 mm were recorded during stocking and significant growth in the raceways was questionable. Clams were expected to repopulate th e raceways readily as a result of Corbicula-specific attributes such as high fecundity in monoecious individuals, brooding of la rvae with no need for intermediate hosts, rapid growth (B yrne et al. 2000) and their ability to recruit new individuals in a wide variety of habitats (Sickel 1986). The lack of adapted predators has also been attributed to recolonization success in natu rally occurring populations of Corbicula (Sickel 1986). This is advantageous for raceway-based monocultures in closed systems that can be engineered to keep larger predators out, such as fish, raccoons and birds that may prey on smaller individuals with softer shells (Buttner 1986). Smaller omnivorous predators such as flatworms or amphipods are more difficult to control with mechanical me thods and may prevent clam recruitment in raceways even if successful repr oduction and larvae expulsion into the water column takes place.

PAGE 94

94 The lack of reproduction or recruitment in clam raceways is odd since clams readily repopulate even obscure engineered systems such as power plant reactor plumbing (Hakenkamp and Palmer 1999). These systems are subject to influx of larvae from natural population reproduction where clam raceways are closed to su ch influx of genetic material. Clams have also been reported to reproduce successfully at p ond-scale in polyculture w ith fish for treatment of aquaculture wastewater (Buttner 1986). Even with a lack of in flux of natural genetic material, high temperatures and the presen ce of potential clam predators ( catfish), culture ponds exhibited clam reproduction and recr uitment (Buttner 1986) As McMahon and Bogan (2001) pointed out, successful Corbicula reproduction may not be achievable in engineered culture or experimental systems especially at smaller scales. The lack of evidence supporting the idea that successful in cidents of clam reproduction took place in the raceways supports this theory; however, the clam raceway systems in this study were presumed to be of large enough scale so re production is not inhibited. E nvironmental stressors such as high temperatures, ammonia levels, amphipod infest ation and food limitation that may have been responsible for growth limitation may have al so reduced the clams reproductive capacity. Clam Stock Assessment Issues Another problem with comparison and implementation of Corbicula raceways at large scale may be that population sampling can prevent an accurate assessment of phosphorus sequestration via raceway clam biomass and settle d waste products. In this study, live clam density values estimated from seasonal spatia l sampling differed greatly from clam density estimations using actual live counts at the last sampling event of the stud y period. Clam biomass losses due to mortality were not verified in this study since dead clams were not measured. Measurement error in the raceway population spa tial sampling technique probably resulted in an over estimation of live clam density even t hough 3.8 % of each raceway was sampled during

PAGE 95

95 each event. In the raceway design used in this study, there is an open, uniform area that maximizes conditions for growth with far less spatial variability in biomass accrual and hydrologic conditions than expected in natural systems. Any atte mpt at large-scale aquaculture of Corbicula should pay particular attention to clam population monitoring and disturbance of culture substrate due to the variability encountered with the methodology used in this study. Methods for assessing clam populations at stocki ng also need to be a ddressed in order to estimate population growth and bioma ss phosphorus sequestration accurately. Clam size can have a stro ng negative relationship on Corbicula growth rate (Buttner and Hiedinger 1980, Foe and Knight 1986, McMahon and Williams 1986). This relationship is similar to the von Bertalanfy growth model used to describe differences in growth rate with age in clam of the saltwater clam Mya arenia (Brousseau 1979). However, age is difficult to estimate in Corbicula due to the absence of growth rings related to annual shell deposition as found in saltwater clams, such as Mya arenia (Brousseau 1979). No conclusions could be made on size/growth relationships since no discernable trends were found in population shell length distributions, and low recovery of live tagged individuals did not al low for statistical analysis of growth rate trends. Counts of dead clams on the substrate surface ma y be the best indication of mortality and actual clam biomass than the volumetric and spatial estimations used in this study. The cumulative number of dead clams collected ove r time and excavated at the end of the study accounted for up to 90 % of the clams stocked in th e raceways. Stocking densities were grossly over-estimated by the volumetric method and othe r techniques should be considered for future clam stocking estimation. Raceway clam densities estimated by the spatial technique may have also overestimated the actual live clam bioma ss compared to the cumulative number of dead

PAGE 96

96 clams found at each sampling time. Since there was no indication of reproduction by clams, it is assumed that the number of dead clams is the best reflection of actual live density remaining in the raceway. General Conclusion Overall clam population bioma ss phosphorus sequestration di d not occur due to high mortality, even though significant growth rates were observed for tagged clams. Timing of noticeable mortality events indicated that hi gh seasonal temperature was the major factor limiting the ability of clam r aceway populations to adapt to treatment raceways. Water temperatures in the range of 28 to 30 oC and above have been implicated as the limiting factor to success of Corbicula applications in most other wast e treatment studies and in natural populations (Greer and Ziebell 1972, Haines 1977, Mattice 1977, Buttner and Heidinger 1980, Buttner 1986). Major population declines took place when water temperatures reached this level, regardless of the le vel of nitrogen and phosphorus addition and chlorophyll a in source pond water. Bivalve-based treatment of dairy-derived wast ewater phosphorus would require, at the very least; scaling, study and implemen tation of additional treatment t echnologies in order to reduce high levels of nitrogenous wastes common in da iry operations. Other than obvious issues with implementing mechanisms unt ested at large scale, Corbicula-based treatment of dairy or any other agriculture-based wastewater stream will n eed to manage predation and parasitism from unforeseen organisms, as well as environmenta l parameters, before applications at any appreciable scale could possibly take place. Other environmental factors present in the ra ceways, including stress from the potential problems due to infestation by amphipods (Hyalella azteca), may have affected phosphorus removal and sequestration potential as well. Consequently, interactions reported between

PAGE 97

97 Corbicula and Hyalella in this study are the first to recognize amphipods as having a potential predatory or parasitic role in clam population dyn amics. The negative interactions observed in this study between amphipods and clams are typical of aquaculture and mariculture systems and will need to be managed in future systems. Possible management strategies borrowed from mariculture systems include hand removal, which is not applicable to the small size and numbers of amphipods, top-down predation by snails or other invertebrates, polyculture with higher order predators such as fish, or development of pesticides for this application.

PAGE 98

98 CHAPTER 4 PHOSPHORUS REMOVAL AND SEQUE STR ATION IN CLAM RACEWAYS Introduction For large-scale, clam-based raceway systems to be successful in removing and retaining phosphorus from wastewater, they mu st operate within the physiologi cal tolerances of the animal to ensure survival, growth and reproduction. In addition to concerns ove r adaptability of clams to a particular raceway environment, the syst em must demonstrate an ability to remove phosphorus at rates compatible w ith the needs of wastewater tr eatment systems. Naturally occurring populations of filter-feeding bivalves, such as Corbicula, can remove phosphorus from the water column and sequester it into shell and meat biomass (F uji 1979). Wastewater nutrient management in agriculture, aquaculture, muni cipal and surface water sources by freshwater bivalves, utilizing a phytoplankt on intermediary, has been pro posed by Stanley (1974), Haines (1977), Mattice (1977) and Greer and Ziebell (1979). In addition to phosphorus removal (Soto and Me na 2004), freshwater bivalves have been used to lower turbidity (Habel 1970, Busch 1974, Haines 1977, Buttner 1986), nitrogenous waste levels (Buttner 1986), particulate protein co ncentrations (Haines 1977) and seston biomass (Mattice 1977). Actual investigations into wa stewater treatment poten tial of bivalve-based systems has been largely limited to the mariculture industry, where species that are commercially desirable as a food commodity have been ta rgeted for profitable production (Shpigel and Fridman 1990, Jones and Iwana 1991, Shpigel and Blaylock 1991, Jakobworks et al. 1993, Shpigel et al. 1993, Shpigel and Neori 1996, Shpigel et al. 1997, Jones and Preston 1999, Neori et al. 2000, Jones et al. 2001). These studi es focused on management of water clarity parameters, nitrogenous compounds and removal of wastewater-generated suspended solids including phytoplankton intermedia ries, as opposed to phosphorus removal. In mariculture and

PAGE 99

99 aquaculture waste systems, undigested feed and f eces from a primary species, such as fish or shrimp, has also been used as food for bivalves (Jakobworks et al. 1993, Jones et al. 2001). Some bivalve-based treatment systems in the mariculture industry utilize multiple stages of biofiltration including filter-f eeding organisms and harveste d macrophytes (seaweed) (Shpigel and Neori 1996, Neori et al. 2000, Jone s et al. 2001, Kinne et al. 2001). Adaptation of mariculture tec hnologies to freshwater wast e streams from agriculture, aquaculture and municipal wastewat er treatment facilities requires use of freshwater organisms, such as Corbicula, that are far less desirable commercially than other bivalves. These organisms are potentially well-suited for wast ewater treatment applications b ecause of their high filtration, growth and reproductive abilities that can contribute to phosphorus uptake and sequestration. Other freshwater bivalves, including the mussel Diplodon chilensis (Soto and Mena 1999), to reduce phosphorus in small aquarium-type syst ems in wastewater streams generated by freshwater aquaculture; however, freshwater musse l species are typically sexually dimorphic and require a fish as an intermediate host for larval development. These traits can limit the geographical distribution and biomass production potential of mussels. Corbicula, on the other hand, can rapidly repopulate an ar ea as a result of a number of lif e history traits including high fecundity, hermaphroditic sexuali ty, self-fertilization and marsupial incubation of larvae (McMahon 1983, McMahon and Bogan 2001). Corbicula are also able to reproduce over longer time spans due to their bivoltine (occurs twi ce annually) reproductive effort periodicity, as opposed to most mussel species, which tend to be univoltine (occurs only once each year) (McMahon 1983, McMahon and Bogan 2001). High population densities achievable by Corbicula, exceeding 1000 individuals/m2 (McMahon 1983), can minimize the space needed for treatment, compared to freshwater mussels.

PAGE 100

100 Utilizing Corbicula for treatment of phosphorus from agri cultural wastewater effluents has been proposed by Greer and Ziebell (1972), Stan ley (1974) and Mattice (1977). Similar to bivalve-based treatment systems used in the ma riculture industry (Shp igel 1993, Shpigel and Neori 1996, Lefebvre et al. 2000 and Mazzola and Sara 2001), this approach uses a phytoplankton intermediary to convert dissolved phosphorus into particulat e matter that can be used for food. Wastewater effluent is used to generate phytoplankton biomass, which in turn provides clams with an unlimited food source. Spar city of food resources has been indicated as a limitation to Corbicula growth in natural systems (Foe and Knight 1985), however, wastewater effluent addition decreases the likelihood that phosphorus or nitrogen become s a limiting factor in phytoplankton production. It is necessary to recognize that source water supply redundancy must be designed into any engineered syst em to handle catastrophic events such as phytoplankton population crashes or harmful al gae blooms, which may occur under normal operation. Even though dairy e ffluent addition can support alga l growth (Wilkie and Mulbry 2002), phytoplankton production can be limited by a range of factors commonly found in aquatic systems such as micronutrient limitation, season ality of light availability and temperature. Corbicula utilizes a wide variety of phosphorus-containing particulates. Greer and Ziebell (1972) observed that Corbicula is not only able to remove phos phorus added to source waters from municipal wastewater effl uent through consumption of phytopl ankton biomass, but can also remove dissolved forms of phosphorus when they are converted to colloid al hydroxyl-apatite at high pH levels. This phenomenon appears to require elevated dissolved phosphorus concentration (5 to 15 mg/L PO4 3-), along with a pH > 9, which can be achieved through addition of lime and the diurnal reduction in CO2 due to phytoplankton meta bolism (Greer and Ziebell 1972). Although sequestration of co lloidal phosphorus has been obser ved only in the laboratory,

PAGE 101

101 achieving removal of colloidal phosphorus from wastewater streams by Corbicula adds to the organisms appeal as a phosphorus removal mechanism. In the Fuji (1979) model for Corbicula in a natural population, phosphorus consumed by clams is sequestered within the system either by sedimentation of feces and pseudofeces or ingestion and incorporation into new clam tissue and shell. Filtration by Corbicula is characterized by high rates of phytoplankt on clearance, up to 500 mL/day/clam in hypereutrophic lake water (Beaver et al. 1991). Like growth rate, filtra tion rate is primarily dependent upon temperature and size (McMahon and Williams 1986), but can also be governed by particle concentrations (Hai nes 1977). Fuji (1979) suggested that, of the total amount of phosphorus consumed annually by filtration, 62 % is ingested, while the rest is rejected and deposited as pseudofeces. The particulate phosp horus ingested from the water column can be converted directly into soluble form as me tabolic waste product ex cretions by the clams (Lauritsen and Mozley 1989). The rest of th e total amount of phosphorus ingested is exported from the system through mortality (13 %), eject ion of gametes (9 %), excretion (8 %) and juveniles that were exported from the system by water currents and theref ore not recruited into the population (1 %) (Fuji 1979). Phosphorus (P) content (mg P/i ndividual) of clam biomass increases with body weight over time in size-based age groups (Fuji 1979). Phosphorus content for clams in each age group also varies with the time of the year due to ejec tion of gametes. Fuji (1979) estimate phosphorus content of clams between 0.011 g and 3.093 g total clam dry weight to be from 0.008 mg P to 1.500 mg P per individual, yieldi ng biomass concentrations in th e range of 0.506 mg P/g to 0.800 mg P/g of total clam dry weight biomass. Us ing phosphorus content of clam biomass derived from tissue analysis, Fuji (1979) estimated th at clam populations can sequester 130 mg P/m2 in

PAGE 102

102 one year as biomass growth (without accounting for mortality ) in a population originally containing 65 mg P/m2. Data from Fuji (1979 ) showed amounts of phosphorus in meat tissue (8.293 22.667 mg P/g meat dry weight) was up to 14 times higher than in shell tissue biomass. In this study, a large-scale raceway-based wast ewater treatment system was used to assess the phosphorus removal potential of Corbicula populations. The system was based on phytoplankton as an intermediary to convert dissolved nutrients into particulate form (Greer and Ziebell 1972, Stanley 1974 and Haines 1977, Ma ttice 1977). Source ponds that fed the raceways in this system acted as incubation ponds for phyt oplankton. The potential for using a clam-based raceway system was assessed by the two most important mechanisms used to determine phosphorus treatment capacity. First, direct re moval of phosphorus from the overlying water within the raceway was measured in terms of change in phosphorus concentration in the water column. Second, the capacity of the raceways to retain and sequester phosphorus as clam biomass was determined by changes in clam growth over time. Corbicula populations were exposed to high, medium and low wastewater nutri ent addition treatments to test the ability of the clam systems to remove and sequester phosphor us over time in system s of different nutrient load. Methods Raceway-based Treatment System A raceway-based recirculating water treatment system constructed at the University of Florida Dairy Research Unit in Hague, FL (See Ch apter 2) was used to test the P-removal and sequestration potential of Corbicula. Three independent pond/raceway systems were constructed in June 2002 to compare source waters with low, medium and high levels of nutrients. Each nutrient addition treatment group utilized two earthen source water ponds in conjunction with

PAGE 103

103 three wood-framed, plastic-lined raceways. Tabl e 4-1 shows the numerical designations for the ponds and raceways in each nutri ent addition treatment group. Table 4-1. Source pond and raceway numeri cal designations for th e treatment systems at the Hague site. Nutrient addition treatment Source Pond Raceways Low 1, 2 1 3 Medium 3, 4 4 6 High 5, 6 7 9 Source water ponds Source ponds had an approximate area of 0.05 h ectares (ha), with depths of approximately 2 m and volumes approximately 1000 m3. Ponds were enriched with a blend of nitrogen and phosphorus fertilizer or anaerobica lly digested dairy farm wastewater, to simulate possible water conditions associated with tertiary wastewater treatment. The low nutrient addition treatment received no external nutrient addition. A 5 % a nd 10 % addition of anaerobically digested dairy farm effluent was added to Pond 3 from the medium nutrient group and Pond 5 from the high nutrient addition treatment group, respectively. For efflue nt physical and chemical characteristics at the University of Florida Dairy Research Unit, see Wilkie et al. (2004). Effluent was pumped from the digester to the source ponds and metered through a 2.54 cm (1) turbine-type flowmeter. Pond 4 in the medium nutrient addition treatme nt was dosed with 1.1 kg of triple super phosphate (9Ca(H2PO4)2, N-P-K = 0-45-0) and 6.8 kg ammonium nitrate (NH4NO3, N-P-K = 150-0) resulting in a to tal addition of 0.23 kg of phosphorus and 1.02 kg of nitrogen per pond in October 2002, one month prior to clam stocking. Pond 6 in the high nut rient addition treatment was dosed with 2.2 kg of triple super phosphate (0.46 kg P) and 13.6 kg of ammonium nitrate (2.04 kg N) in January 2002, also one month before introduction of clam populations.

PAGE 104

104 Nutrient additions were de signed to enhance phytoplankt on biomass, which was the putative source of particulate nutrition for the clams. Fertilizer loading levels were targeted at increasing phosphorus and nitrogen levels by 0.23 mg/L TP and 1.02 mg/L TN in the medium nutrient treatment source pond and 0.46 mg/L TP and 2.00 mg/L TN in the high nutrient treatment source pond. Fertlizer was introduced to ponds by placing it into a burlap bag that was suspended in the water column from a floating frame (0.5 m x 0.5 m) constructed from 5.08 cm (2) ID PVC pipe. Neither of the ponds supplied with dairy effluent were exposed to the clam raceways due to excessively high ammonia levels (2.0 mg/L or greater, as NH3-N), which represented a direct threat to the health of the clams. Ammonia (NH3-N) levels in the effl uent ponds and input and output water from each operating raceway were monitored monthly using Aquacheck brand Ammonia Nitrogen (NH3-N) Test Strips (Hach Incorporated, Colorado, USA) advertised for use in aquaculture and aquarium applications. Levels of NH3-N in the operating raceways never reached the 0.25 mg/L (as NH3-N) minimum value of the test kit. Source ponds were circulated through the raceway s for 10 to 20 days prior to clam addition. Supply ponds were aer ated at night throughout the study to help with mixing, maintenance of nighttime dissolved oxygen and eva porative cooling. Floating macrophytes were cleared by hand from each source pond several times at the start of the experiment. In addition, six juvenile triploid grass carp ranging between 15 and 20 cm in length were stocked in each pond in June 2002 to help reduce vegetation in po nds. No fish mortality was evident in ponds not exposed to dairy wastewater effluent. In contrast, ponds treated with wastewater all experienced 100 % fish mortality in less than one week after effluent addition. Fish mortality events corresponded with high amm onia levels (> 0.25 mg/L as NH3-N).

PAGE 105

105 Raceways Raceways were 1 m in width by 7.4 m in length with an available s ubstrate surface area of 7.2 m2, after subtracting standpipe area. Water de pths were maintained at 0.2 m, yielding a raceway water capacity of 1.4 m3 each. Source water inflow to the raceways was maintained at 227 liters per minute (LPM) (60 GPM) during normal operating conditions. The flow rates yielded retention times of approximately 9.5 minutes, with a linear velo city of 1.17 m/min. Turnover time for each supply water pond was a pproximately 24 hours. Flow rates were calculated assuming near laminar flow through the raceway structure. Ra ceways were filled to 0.2 m depth with a coarse grade SiO2 filtration sand (0.6-1.0 mm particle size), that was purchased from Feldspar Incorporated in Edgar, Florida. Nuisance aquatic plants such as Chara sp. and filamentous algae (mainly Spirogyra sp.) growing on the raceway substrate and liners were removed by hand from each raceway at least weekly to reduce fouling. The cont rol of nuisance plants, especially Chara, on the substrate made accurate estimation of clam biodeposit sediment ation impossible. This is due the constant resuspension of sediment deposits associated wi th disturbance from removal of the vegetation that was often rooted below the sediment surface. Filamentous algae tended to utilize the sides of the raceways where PVC liners were submersed. Water quality monitoring Source water ponds and raceways were monitore d weekly in the morning and evening for temperature, dissolved oxygen, pH, chlorophyll a, total nitrogen and total phosphorus. Measurements in the ponds were taken at the end of sampling piers near the intake pipe leading to the raceways. Monitoring was carried out mo nthly after 75 % or more of the clams were assumed dead. Temperature and dissolved oxygen was measured using a YSI model DO550, and pH was measured using a Fisher model AP63 me ter by submerging the probe tips to mid-water

PAGE 106

106 column depth at the raceway input and output. Water temperature, dissolved oxygen and pH readings were compared using a paired t-test (Microsoft Excel) at the raceway input and output, as well as between source ponds in the nutrient addition treatments. Water samples were collected from the samp ling piers in each pond for nutrients using a pole sampler designed especially for this ex periment. The sampler used a plunger-type mechanism to collect a 1 L water sample from in front of the intake pipe. When the unit was lowered to the desired depth, the plunger was ac tuated by the operator via a spring-loaded handle at the opposite end of the pole. After the sample was collected, the plunger was released, sealing a 1 L plastic bottle (Nalgene Incorporated, USA) bottle and raised for retrieval. The sample bottle was unscrewed from the sampler and capped for transport to the laboratory. Water samples collected from the source ponds were analyzed at the laboratory for phosphorus, nitrogen and phytoplankton biomass in terms of chlorophyll a. Total phosphorus (TP) and total dissolved phosphor us (TDP) were determined using the potassium persulfate digestion method (APHA 1998) with a Hitachi (J apan) spectrophotometer. TDP determination involved pre-filtering through a 0.7 m glass fiber filter. Total nitrogen was determined using potassium persulfate digestion method (APHA 1 998) with colorimeteric analysis performed using a Bran-Luebbe (Germany) auto analyzer. Phytoplankton biomass was estimated using chlorophyll a (chl a). Samples were measured by filtering 250 mL of water onto a 0.7 m glass fiber filter, followed by an ethanol extr action (Sartory and Grobbelaar 1984) and spectrophotometric determination (APHA 1998) us ing a Hitachi (Japan) spectrophotometer. Microscope observations of phytoplankton specie s composition were obtained periodically to describe dominant organisms with help from Mary Cichra at the Universi ty of Florida Fisheries

PAGE 107

107 and Aquatic Sciences Department. Data obtaine d from these phytoplankton species observations were not assessed quantitatively. Raceway clam populations Clams for stocking the raceways were obtained from populations in three different natural water bodies under permit number FNC-04-022 is sued by the Florida Fish and Wildlife Commission. Clams for the low nutrient group (Raceways 1-3) were collected in June 2002 from a 0.5 km stretch of the Santa Fe River near the State Road 49-bridge in Gilchrist County, Florida (29o54.2 North, 82 o52.0 West). By November 2002, the rising water level of the river made further clam excavation impossible. Ther efore, animals for Racew ays 4-9 were collected from lakes located in Lake County, Florida that had accessible populations of clams. Clams for the medium nutrient group (Racew ays 4-6) were collected from the southwest shore of Lake George (29o12.2 North, 81o35.7 West) in November 2002. Clams for the high nutrient group (Raceways 7-9) were obtained from the west shore of Lake Dalhousie (28o54.0 North, 81o36.8 West) in February 2003. All three collection sites had coarse sand sedi ments similar to the substrate used in the raceways. Clams were excavated by shoveling botto m material into weight ed baskets made from 0.635 cm () plastic mesh. Clams were also excavated by hand using tr owels or a commercial clam rake modified for the small size of the Corbicula by affixing plastic mesh on the inside of the collection basket. Periodic exca vation of bottom sediment using a 0.25 m2 PVC sampling quadrat was used to determine population densitie s for the clams in their natural habitat. Densities ranged from 48 and 864 clams/m2 with a mean of 272 clams/m2 (standared error (SE) = 23, n = 47 observations) for all locations combined. After excavation, clams were enumerated and divided into mesh bags. The clams were then placed into coolers packed with wet newspa pers and kept out of di rect sunlight to help

PAGE 108

108 minimize heat stress and dessication. They were tr ansported directly to the aquaculture facility in Hague, FL and scattered evenly throughout each raceway. Stocking of each raceway took up to 15 days involving 2 to 6 people working per day. Raceways 1 were stocked from June 17 to 28, 2002, Raceways 4 from November 4 to 14, 2002, and Raceways 7-9 from February 11 to 26, 2003. Ten mesh bags of clams were adde d to each raceway. Each mesh bag contained approximately 1,000 clams. This method was chos en as opposed to counting each individual or bulk weighing in order to minimize handling stress, time and equipment needed to enumerate the large numbers of clams needed for this study. Clam populations in the raceways were monitored for density, size and biomass at approximately 3-month intervals following stocki ng, as described in Chapter 3. A tag and recapture study was also employed to verify clam growth (Chapter 3). Live clam density, shell size and growth data from clam population surv eys and tag and recapture data were used to assess clam biomass, phosphorus removal a nd changes in clam phosphorus sequestration. P Removal From Source Water By Clam Raceways The ability of the clam raceway system to remove phosphorus and P-containing material from overlying source water was estimated from three components: phosphorus removal from water within a single raceway pass (through-flow): phosphorus removal from water recirculated within the raceways and phosphorus removal from water in the source ponds through sequestration of phosphorus in clam biomass over time. Raceway through-flow trials P removal from water using raceway through-fl ow measurements was only assessed in the low nutrient raceways beginning afte r stocking in July 2002. Raceway flows were maintained at 227 LPM (60 GPM) from the time of stocking un til August 2002, when input flow rates were changed to increase raceway water retention time (Table 4-2). The low flow condition of 151

PAGE 109

109 LPM (40 GPM) and a higher flow condition of 227 LPM yielded linear water velocities of 0.78 m/min and 1.17 m/min and water retention times in the raceway of 9.5 minutes and 6.3 minutes, respectively. Table 4-2. Raceway source water input flow rates for the period of July 1 to August 24, 2002. Flow rates were lowered in August to increase water retention time for raceway water phosphorus removal trials using through-flow conditions Raceway Raceway Inflow Rate (LPM) # 7/1-7/31 8/5-8/8 8/12-8/15 8/19-8/22 1 227 227 227 151 2 227 227 189 151 3 227 227 151 151 Samples were taken from the input and output of each raceway to estimate the changes in TP concentrations resulting fr om each pass of water over the cl am population in the raceways. The values were used to estimate instantaneous phosphorus removal rates from each raceway. Environmental parameters including water temper ature, DO and pH were also sampled at the raceway input and output at 6:00, 12:00 and 18:00 hours. The initial experiment was designed to provide baseline data on amount and variability of phosphorus removal from raceways on a short-term basis. The sample technique involved taking a water sample from the raceway input and output at each sample time of day. In this phase, samples were captured from falling water under th e input water distributio n bar and from within the output standpipe of each raceway. One liter of sample water was collected at each input and output using plastic bottles (Nal gene Incorporated, USA). Sa mples were taken from each raceway daily at 6:00, 12:00 and 18:00 for five 3-day periods from July 1-31, 2002 (high flow) and three, 3-day periods from August 5-24, 2002 (variable flow). The differences between TP and TDP concentr ations at the input and output end of the raceways were converted to area-specific removal or addition values by multiplying by 200 L/m2. Removal rates were then calculated using th e retention time estimated for each flow rate.

PAGE 110

110 Negative values indicate an uptake of P, while positive values indicate a phosphorus addition from the raceways to the water column. Rates larger than 120 mg P /m2/hr or smaller than mg P/m2/hr were considered outliers and removed fr om the analysis to normalize the data. The July and August through-flow TP removal data were analyzed using an ANOVA (SAS PROC MIXED procedure, SAS Institute, Cary, NC). The model was a randomized complete block design (Cox 1996) with raceway as the bloc k, where time of day and sampling period (the 5 different 3-day trials) were the factors with covariates inflow DO a nd water temperature. Another ANOVA (SAS PROC MIXED procedure, SAS Institute, Cary, NC) was performed on the August data in order to determine the significan ce of flow rate. This analysis consisted of a randomized complete block design with three f actors, flow rate, time of day and raceway number, and no covariates. Observations from Raceway 2 from August 12-15 were excluded from the analysis due to a low sample size, since it was the only instance wh ere input flow rate wa s not 151 or 227 LPM. Another ANOVA (PROC MIXED, SAS Institute, Cary, NC) was performed using a randomized complete block design with month as the block to determine if the P-removal values were significantly different for each month. A pair-wise comparison of the means was performed using Tukeys method to control the experiment-wise error rate in all ANOVAs. Uneaqual variances for each time period were as sumed in the ANOVA due to the significantly low variability (x2, p = 0.033) in the 6:00 samples for the month of July. Sample variability was assumed equal over both flow conditions for the flow rate comparison analysis (x2, p = 0.27); however, unequal variance was assu med in ANOVA models due to significant differences in variability between each month (x2, p < 0.001) and sampling technique (x2, p = 0.002).

PAGE 111

111 In another set of phosphorus removal tests, auto-samplers were employed to increase the sampling frequency over each 24-hour test peri od. This methodology allowed for a larger sample size spread over the entire day and night A pair of Teledyne-Is co, Inc. auto-samplers was used to collect water samples from the i nput and output of raceway s 1 through 3, in August 2002. Samplers were programmed to capture water at 2-hour time intervals over a 24-hour period. Samples were collected at two-hour time inte rvals beginning at 6:00 and ending at 6:00 the following day, over two, four-day time periods (August 5-8 and August 21-24). Each of the three raceways was only sampled for one 24-hour interval during each 4-day time period. During the first sampling interval raceway flows remained at 227 LPM; flows were changed to 151 LPM for the second interval. DO, water temp, air temp and pH were sampled at 6:00, 12:00 and 18:00. An ANOVA (PROC MIXED, SAS Institute, Cary, NC) was performed on the August auto-sampler data using a randomized complete bl ock design with two factors, flow rate and time of day, and raceway number as the block. Pair wise comparisons of means were performed using Tukeys method to control the experiment-wise error rate. Phosphorus removal rates obtained from August samples and auto-sampler techniques were compared to determine any differences in the two sampling met hods. An ANOVA (PROC MIXED, SAS Institute, Cary, NC) was performed on the 6:00, 12:00 and 18:00 data using a randomized complete block design with two factor s, flow rate and time of day and sampling method as the block. A pair-wise comparison of the means was performed using Tukeys method to control the experiment-wise error rate. Only data collected at 6:00, 12:00 and 18:00 sampling

PAGE 112

112 periods were used in the analysis, since there were far more 24-hour values. Samples taken at 189 LPM flow rate were also withheld from the comparisons. Raceway water recirculation To provide additional information on phosphorus uptake rates, a second method was applied to extend the probability of exposure to th e clams. This involved recirculation of water within the raceway to extend the total exposure period. Phosphorus removal was measured in terms of total phosphorus (TP), total di ssolved phosphorus (TDP) and chlorophyll a (chl a). A submersible pump placed within a plastic co llar was used to simulate standpipe water flow characteristics. This unit was placed just be fore the output standpipes, and water was piped to the input water distribution bar at about 189 LPM within the normal operation range of 151 to 227 LPM. Raceway recirculati on was started 10 minutes prior to water sampling. Periphyton and dead clams were removed prior to recirculation. Water samples for nutrient analysis were taken hourly at the raceway input using an ISCO auto sampler. The auto sampler intake line was submerged approximately 3 cm from the wate r surface. Water temperature, DO and pH measurements were taken at the raceway input at the beginning and end of each recirculation time period. Raceways 1-3 and 7-9 were utilized in this ex periment. Raceways 4-6 were excluded from the study due to high clam mortality. Clam popula tions were stable in raceways 1-3 and 7-9 although 7-9 had higher densities of clams. R aceway recirculation test s were conducted during two separate experimental periods, April 11, 2003 to June 10, 2003 and June 20, 2003 to June 27, 2003. Tests were carried out for 6-hour in tervals in the morning from 6:00 to 12:00 hours and again in the evening from 18:00 to 24:00 hours. To investigate the impact that particle deposition may have on phosphorus removal, a control group was set up by covering the bottom of the raceways with weighted plastic covers to

PAGE 113

113 eliminate interaction between the clams and the overlying water column. The control tests were only run for 3 hours so clams would not be impacted. P removal potential was estimated from regr ession relationships using TP, TDP and chl a values in normally operating and covered raceway s. Slope values formed by regression lines were used to calculate removal rate s according to the following formulae: TP and TDP: Regression slope (mg/L/hr) 200 (L/m2) = removal rate (mg/m2/hr) Chl a: Regression slope (mg/m3/hr) 0.20 (m3/m2) = removal rate (mg/m2/hr) Six-hour recirculati on TP, TDP and chl a removal values from the uncovered raceways were used to perform an ANOVA (PROC MIXED, SAS Institute, Cary, NC) using a randomized complete block design, with slope of the line formed by TP, TDP and chl a values over each 6-hour time period as the block, and month (April/May) and nutrient enrichment (high/low) as the factors to determine if remova l values were greater than zero. Slopes from each am/pm time period were pooled since raceways within each treatment were not factored into the ANOVA, due to the lack of significant differences (p > 0.05) found between am and pm trials or raceways within each treatment. A pa ir-wise comparison of the means was performed using Tukeys method to contro l experiment-wise error rate. TP, TDP and chl a values from the three-hour recirc ulation experiments in the covered raceways were used to perform a sim ilar ANOVA (PROC MIXED, SAS Institute, Cary, NC) using a randomized complete block design with slope of the line formed by TP measurements over each 3-hour time period as the block and nutrien t enrichment (high/low) as the factors. Slope values from all raceways in each triplicat e for each sample interval were pooled for each

PAGE 114

114 nutrient addition treatment as in the uncovered raceways. A pair-wise comparison of means was also performed on the covered data using Tukey s method to control expe riment-wise error rate. An ANCOVA for repeated measures (PROC REPEATED, SAS Institute, Cary, NC) with a covariance structure AR-1 model was then ut ilized to determine any covariance between the TP, TDP and chl a slopes for the low and high nutrient leve ls. This procedure was chosen due to differences in variance between the nutrient addition treatments and lack of a significant difference between the am and pm time periods in each nutrient additi on treatment from the uncovered analysis found in the ANOVA; therefore, ti me can be treated as a continuous variable. Sequestration of phosphorus by clams in treatment raceways Another method used to estimate phosphorus removal from the treatment water was determination of increases in clam biomass over time, in combination with analysis of the phosphorus content of clam tissue. Clams were randomly selected from raceways (as described in Chapter 3) for the dry weight (DW) and ashfree dry weight (AFDW) analysis. Tests were performed using the low and medium nutrien t addition treatment raceways in June and November 2002. A total of 26 clams were sampled per raceway, yielding a sample size of 79 clams per nutrient add ition treatment group. Shell length, wet weight, meat and shell DW and AFDW values were measured for each clam using methods described in Chapter 3. Phos phorus content of the meat and shell tissues in the selected clams was determined using an ignition and hydrochloric digestion method (Andersen 1976). In th is method, tissue material was drie d, finely ground and ignited at 550 oC until a constant weight is achieved. The ash wa s then dissolved by adding 1N HCl to achieve a pH < 2, heated and diluted with DI water to neutralize the solution prior to phosphorus determination using inductively-coupled plasma spectrometer (Copar and Yess 1996).

PAGE 115

115 Meat tissue samples were prepared by dissolvi ng the ashed material in 1 mL of 1N HCl, heating and diluting to 50 mL. For shell material up to 1 g DW of shell ash material was taken from each clam, dissolved in 3 mL of 1N HCl, heated and diluted to 50 mL. Phosphorus analysis was carried out on the sample solu tions using an inductively coupled plasma spectrometer and was performed by the Analytical Research Laboratory of the Soil Science Department at the University of Florida. Phosphor us concentrations in the sample solutions were then converted to mg P per g of DW for the meat shell and total clam tissues using the ash DW and tissue DW values obtained prior to digestion with the following formulae: Meat = [P] in sample solution x _50 mL dilution volume_ x sampled meat ash DW 1000 sampled meat ash DW Shell = [P] in sample solution x _50 mL dilution volume_ x crushed shell ash DW 1000 sampled shell ash DW crushed shell DW Clam = (meat mg P/g meat DW x meat DW) + (shell mg P/g shell DW x shell DW) (total meat DW + total shell DW) Meat, shell and clam phosphorus co ncentrations were then comp ared to shell length values using two different regression analyses (SAS PROC MIXED, SAS PROC GLM), where meat, shell and clam phosphorus concentrations were the response variables wi th length and length2 as factors to obtain the best-fit regression line. A polynomial regression was chosen since both length and length2 factors were significant (p < 0.05). Bo th analyses yielded no significant (p > 0.05) relationships between any of the tissue ph osphorus concentrations and shell length. An ANOVA (SAS PROC MIXED pr ocedure, SAS Institute, Cary, NC) was then performed using the meat, shell and clam tissue phosphorus concentrati ons as response variables and nutrient addition treatment as the factor with covariates, 3-month time interval, individual raceways within each nutrient level and season in each analysis. The least squared means (LSMEANS) procedure was applied to meat shell and clam tissue phosphoru s concentrations (Microsoft

PAGE 116

116 Excel) for pair-wise comparisons. Pair wise comparisons of the means were performed using Tukeys method to control the experiment-wise error rate. Data from individual raceways within each nut rient addition treatmen t were pooled in the analysis since no blocking effect was found. No relationship between meat, shell and clam phosphorus concentrations (mg P per g dry weight ) and shell length could be established using the regression analysis; therefore, shell length as a factor was also removed from the ANOVA. A group of 12 outliers, defined as phosphorus con centration values higher than the range of 3 standard deviations away from the m ean, were removed from the analysis. Sequestration of phosphorus by raceway clam populations was estimated using population DW biomass estimates for each nutrient additi on treatment regime both at stocking and each sampling interval. Estimated clam biomass at each time interval was multiplied by the average phosphorus concentration of the clams, yielding the amount of phosphorus a llocated in the total raceway clam biomass. Clam biomass phosphorus values were applied to raceway population and tagged clam results to assess removal of phosphorus by clams. Results Raceway Environmental Conditions Air temperatures at the racewat y site at the Dairy Research Unit in Hague, FL ranged from 2 to 41 oC (Figure 3-2). Recorded air temperatures varied diurnally as much as 15 oC. Raceway water temperatures ranged from 10.1 to 32.6 oC (Figure 3-3) and displayed a similar seasonal pattern as air temperature. However, water temperatures only differed diurnally by a maximum of 2 oC. Water temperatures reached 30 oC or greater just after the beginning of the experiment in July 2002 through October 2002 and from May through the end of th e study in August 2003. Raceway dissolved oxygen (DO) averaged 8.79 mg/L (SE = 0.07) and ranged from 6.10 to 12.03 mg/L (Figure 3-4). DO was higher in the afternoons by up to 4.26 mg/L with greater

PAGE 117

117 diurnal differences in the warmer months. Higher DO values were observed during the November 2002 to April 2002 period corresponding to lower air and water temperatures. Raceway pH averaged 7.76 (SE = 0.02) and ranged from 6.87 to 8.81. Diurnal fluctuations in pH ranged from .76 to 0.89. Raceway pH was significantly higher (p > 0.05) in the low and high nutrient addition treatments (Figure 3-5); however, water temperature and DO values did not differ significantly between the low, medium and high nutrient addition treatments. No significant differences (p > 0.05) were detected at the input and output of each raceway or between raceways in each nutrient addition trea tment for the water temperature, dissolved oxygen and pH parameters. Increases in phosphorus and ni trogen levels in the source ponds did not correspond to fertilizer addition or clam mortality events. Source pond total phosphorus (TP) ranged between 0.061 and 0.211 mg/L in the low nutrient additi on treatment, 0.047 and 0.471 mg/L in the medium nutrient addition treatment and 0.043 a nd 0.386 mg/L in the high nutrient addition treatment (Figure 3-6). Sour ce pond total dissolved phosphorus (TDP) ranged from 0.002 to 0.091 mg/L in the low nutrient addition treatment 0.007 to 0.223 mg/L in the medium nutrient addition treatment and 0.028 to 0.300 mg/L in the high nutrient addition treatment (Figure 3-7). A major portion of the source pond TP was made up of the dissolved form as TDP in all of the nutrient addition treatments as TDP followed a similar pattern as TP with sharp increases in the spring 2003. Source pond total nitrogen (TN) ranged from 0.177 to 9.034 g/L in the low nutrient addition treatment, 0.328 to 12.069 g/L in the medium nutrient addition treatment and 1.066 to 2.786 g/L in the high nutrient addition treatment (Fi gure 3-8). TN values fluctuated in the low

PAGE 118

118 and medium nutrient addition treatments and only slightly increased in the high nutrient addition treatment over the experimental period. Chlorophyll a (chl a) ranged from 3.218 to 27.511 mg/m3 in the low nutrient addition treatment, 4.505 to 26.397 mg/m3 in the medium and 19.789 to 147.299 mg/m3 in the high (Figure 3-9). Chl a in all treatments displayed an incr ease after February of 2003 with peaks from April to August of 2003. Phytoplankton co mmunities in the source ponds were dominated by diatoms in the low nutrient addition treatmen t, diatoms and cyanophytes in the medium and chlorophytes in the high. Diatoms were pr esent in all ponds throughout the study period. Raceway clam populations Raceway live clam biomass, based on populat ion shell length and survivorship data decreased over time in all the nutrient addition treatment raceways (Figure 3-13). Populations were subject to large mortality events during expe rimental periods (Figure 3-10), as described in Chapter 3. Phosphorus Uptake In Clam Raceways Raceway through-flow input/output measurements Temporal patterns in low nutrient addition raceway TP concentrations during the throughflow trials from July 1-August 21, 2002 (Fi gure 4-1), were similar to patterns of TP concentrations in the source ponds (Figure 3-6). Raceway input TP values varied up to 0.07 mg/L per day with no apparent diurnal trends. Input TP was slightly higher in August 2002 than in July 2002, but did not differ significantly (p < 0.05) between the three triplicate raceways. The results of the ANOVA showed that neith er phosphorus removal (negative input/output delta TP) nor addition (positive input/output delta) were significantly different than zero (p > 0.05) for the different times of day, flow rates, months or sampling techniques. By contrast, time of day did have a significant effect (p = 0.02) on the magnitude of the rate of change in TP

PAGE 119

119 concentrations in July trials. The ANOVA showed that the 6:00 sampling time had an estimated phosphorus removal significantly (p = 0.001) hi gher than the 12:00 and 18:00 hours of day; however, this was probably due to the lowe r variability for the 6:00 sampling time. Figure 4-1. Raceway input total phosphorus (TP) in the low nutrient addition treatment during the through-flow trials for July and August of 2002. Time of day did not significantly affect chan ges in TP between the input to the output regardless of flow rate, month, input TP or sa mpling technique. Input temperature, DO, and sampling date also had no effect (p > 0.05) on the change in TP in any of the trials. Change in TP observations did not significantly differ (p < 0.05) between raceways or between grab sampling and ISCO auto-sampler techniques. Du e to the lack of significant differences in experimental variables, changes in TP ( TP) were pooled for all through-flow trials. Raceway TP values ranged between .585 and 0.119 mg/L, yielding a removal rate range of -116.206 mg P/m2/hr to an addition rate of 99.035 mg P/m2/hr (Figure 4-2). The 25th percentile for removal had TP values ranging between 116.206 and -11.427 mg P/m2/hr (n = 255). TP values appeared to be normally distributed in all trials with a slight kurtosis; however, the TP frequency data is not kurtotic enough (Figure 42) to be considered

PAGE 120

120 unreasonable for assuming normality for the ANOVA an alyses. The results of these trials determined that a raceway water retention time longer than 10 minutes is needed to evaluate significant removal of TP by the clam raceways. 0 10 20 30 40 50 60 -120-100-80-60-40-20020406080100120 TP (mg P/m2/hr)Frequency Figure 4-2. Frequency distribution of changes in total phos phorus (TP) from the input to the output in the low nutrient addition treatm ent raceways from July through August 2002. Raceway recirculation measurements Raceway input TP values at time 0 of each r ecirculation trials decreased significantly (p < 0.001) from April to May 2003 in the both high and low nutrient addition treatments (Table 4-4). Input TP was significantly higher (p < 0.001) in the high nutrient addition treatment compared to the low (Table 4-4). Initial TP values measured at time 0 of the recirculation trials (Table 4-4) were higher than source pond TP values (Figure 4-1), in both the low and high nutrient addition systems. Table 4-4. Monthly mean input total phosphorus (TP) concentrations in raceways and standard error (SE), at time 0 in the recirculation trials for each nutrient addition treatment pond groups. Month Nutrient treatment Time 0 mean TP (mg/L) SE (mg/L) n April High 1.229 0.024 6 April Low 0.853 0.044 6 May High 0.245 0.013 6 May Low 0.067 0.030 6

PAGE 121

121 Regardless of TP level at time zero, slopes of the regression lines calculated for TP values calculated for the April-May trials did not i ndicate any significant de crease or addition of phosphorus by the raceways. The ANOVA analysis showed that none of the mean TP slopes calculated for the six-hour recirc ulation period were significantly different from zero in any of the treatments during the two-month period. TP slopes calculated for the covered raceways did not differ significantly (p > 0.05) from zero as well. Raceway TDP values at time 0 of each recirculat ion trials decreased significantly (p < 0.01) from April to May 2003 in both the high and low nutrient addition treatment pond groups (Table 4-5). Input TDP was significantly higher (p < 0.01) in the high nutrien t addition treatment pond group compared to the low (Table 4-5). Init ial TDP values measured at time 0 of the recirculation trials (Table 4-5) were higher than source pond TDP va lues (Figure 4-2), in both the low and high nutrient addition systems. Table 4-5. Raceway system monthly mean input total dissolved phosphorus (TDP) at time zero in the recirculation trials for each nutrient addition treatment. TDP at time zero decreased in low and high nutrient addition treatments from April to May 2003. Month Nutrient treatment TDP mean (mg/L) SE (mg/L) n April High 1.119 0.024 6 April Low 0.631 0.014 6 May High 0.044 0.003 6 May Low 0.045 0.003 6 The analysis showed a significant difference (p = 0.05) in TDP over time for recirculation trials. Significant TDP decrease was obser ved in both the high and low nutrient addition treatment systems during April 2003, while TDP incr eased during the 6-hour recirculation trials conducted in May 2003 (Table 4-6).

PAGE 122

122 Table 4-6. Total dissolved phosphorus (TDP ) removal rate s calculated from TDP slopes in the recirculation trials for each nut rient addition treatment system during April and May 2003. TDP removal wa s observed during April but not in May for both treatments. Month Nutrient treatment TDP intercept (mg/L) SE (mg/L) TDP slope (mg/L/hr) TDP removal (mg/m2/hr) n April High 1.227 0.146 -0.172 -22.566 6 April Low 0.648 0.066 -0.001 -0.131 6 May High 0.033 0.009 0.002 0.262 6 May Low 0.044 0.009 0.0001 0.123 6 Slopes of the regression lines formed by TDP va lues were not significantly different from zero (p > 0.05) in any of the covered raceway trials in the low and high nutrient addition treatments. TDP values at time 0 of the covere d raceway trials were si gnificantly (p < 0.01) higher in the high nutrient system compared to th e low (Table 4-7); however, TDP values at time zero in the covered raceways (Table 4-7) were much highe r than source pond TDP values (Figure 4-3). Table 4-7. Raceway total dissolved phosphor us (TDP) values at time 0 for covered raceways in the low and high nutrient addition treatments. Raceway TDP at time 0 was much higher than source pond TDP values in Figure 4-3. Month Nutrient treatment Mean TDP at Time 0 (mg/L) SE (mg/L) n April High 0.615 0.008 3 May Low 0.554 0.008 3 Raceway chl a values at time 0 of each recirculati on trials decreased significantly (p < 0.01) from April to May of 2003 in the low nutrient addition treatm ent pond group but not the high (Table 4-8). In April, chl a in the low nutrient addition treatment pond group was significantly (p < 0.05) higher than the high nutrient addition treatme nt group (Table 4-8). Input chl a was significantly higher (p < 0.01) in the hi gh nutrient addition treatment compared to the

PAGE 123

123 low in May. Chl a values in the low nutrient addition treatment pond raceways at time zero during the April 2003 recirculati on trials (Table 4-8) were higher than the source pond chl a values during the same month (Figure 4-8). Chl a values in the high nut rient addition treatment raceways at time 0 (Table 4-8) were lower th an source pond values in both April and May 2003 (Figure 4-3). Table 4-8. Raceway chlorophyll a (chl a) values at time 0 for raceways in the low and high nutrient addition treatments during April and May 2003. Chl a at time 0 was much higher in the low nutri ent raceways during April. Month Nutrient treatment Chl a mean (mg/m3) SE (mg/m3) n April High 76.682 2.440 6 April Low 112.452 2.785 6 May High 72.050 5.075 6 May Low 19.598 1.095 6 The ANOVA performed on the 6-hour recircula tion trial data showed a significant (p = 0.02) reduction in chl a over time, and there were no signifi cant interactions of time with any other variabe. This means that the chl a slopes were the same in each experimental treatment groups even though chl a values at time zero were significantly (p < 0.01) different between nutrient addition treatments duri ng each month. A mean chl a removal rate of .190 mg/m2/hr was estimated from the mean slope (-1.450 mg/m3/hr) formed for chl a values over the 6-hour recirculation time for nutrient additio n treatments and months by the ANOVA. Chl a values in the covered raceway trials involving the low nutrient group were much higher at time zero (Table 4-9) than in the source ponds (Figure 4-3) Conversely, source pond chl a values (Figure 4-3) we re higher than values at time zero (Table 4-9) in the high nutrient addition treatment. No chl a slopes were significantly different from zero (p < 0.05) in any of

PAGE 124

124 the covered raceway treatmen ts; therefore, it is assu med that removal of chl a in the raceways was a result of exposure to clam populations. Table 4-9. Raceway chlorophyll a (chl a) values at time 0 for covered raceways in the low and high nutrient addition treatments. Chl a at time 0 was much higher in the low nutrient raceways. Month Nutrient treatment Chl a mean (mg/m3) SE (mg/m3) n April High 54.291 2.831 3 May Low 108.280 2.831 3 Significant reductions in raceway water chl a were detected after 6 hours of water recirculation; however, these removal values are may not represent removal under normal operation. TDP removal was onl y significant in the high treatme nt over the April 2003 period and was estimated at nearly 22 mg P/m2/hr. Otherwise, TDP removal rates were no different than 0, indicating that this high mean removal rate based on 3 trials may have been coincidental. A chlorophyll a removal rate of 0.190 mg chl a /m2/hr was estimated from recirculation trials in both low and high nutrient addition tr eatments since no difference in chl a removal rates could be found between them. Some difference in removal rates was expected since both live clam biomass and initial chl a were considerably higher in th e high nutrient addition treatment compared to the low during the recirculation trials. Sequestration Of Phosphorus By Clams In Treatment Raceways Shell length and DW biomass characteristics of the clam population sampled for this analysis are shown in Table 4-10. Ash cont ent and AFDW values for the sampled clam population are given in Table 4-11. Shell contained the majority of the total clam ash indicating a larger inorganic content as opposed to meat tissue.

PAGE 125

125 Table 4-10. Mean shell length, clam wet wei ght (WW), meat and shell tissue dry weights (DW) and condition index (CI) values for the sample population of clams used to determine clam biomass phosphorus conten t. Much of the clam DW biomass was contained in shell tissue. Parameter Shell length (mm) Clam WW (g) Meat DW (g) Shell DW (g) Clam DW (g) CI(wt) CI(vol) Mean 20.5 3.056 0.064 1.887 1.951 3.42 4.64 SE 0.2 0.079 0.002 0.047 0.050 0.07 0.16 Range 14.930.4 1.2009.280 0.0100.220 0.7315.900 0.7736.041 1.000.22 1.1415.83 n 228 228 228 228 228 228 228 Table 4-11. Mean and range of ash content va lues for meat, shell and total clam tissues pooled for all clams sampled. Shell tissue had higher ash content and made up a greater portion of the total clam ash than meat tissue. Tissue Mean ash (%) SE (%) Ash content range (%) n Clam 94.71 0.06 90.45 97.39 228 Shell 97.48 0.02 96.06 98.23 228 Meat 13.74 0.55 2.50 17.00 228 Phosphorus allocation in clam biomass No significant differences in phosphorus concentr ations for the meat, shell and clam tissues could be found between the 2 nutrient addition trea tments at stocking or between time intervals for the low treatment using the ANOVA. No discernable relationships in phosphorus concentrations for each tissue type (meat, shell and whole clam) were found between the tested variables (shell length, meat DW, shell DW, clam DW, condition indices, meat AFDW, shell AFDW, clam AFDW) in the ANOVA correlation. Values for tissue phosphorus concentrations were then pooled for all clams used in the experiment to obtain the meat, shell and clam tissue phosphorus concentration data (Table 4-12). The values in Table 4-13 indicate that cl ams of shell lengths between 13.9 and 30.4 mm contain from 0.143 to 1.411 mg P per individual. They also indicate that th e majority of the total

PAGE 126

126 phosphorus is allocated into meat tissue even though shell material constitutes the majority of the total clam DW (Table 4-10). Table 4-12. Summary statistics for phosphorus concentratio ns [P] found in meat, shell and clam tissue types pooled for all clams sampled. Phosphorus was found in much higher concentrations in the meat tissue than shell. Tissue type Mean tissue [P] (mg P/ g DW) SE tissue [P] (mg P/ g DW) Range tissue [P] (mg P/ g DW) n Meat 7.657 0.103 2.650 12.853 228 Shell 0.053 0.001 0.022 0.136 228 Clam 0.299 0.005 0.141 0.606 228 Table 4-13. Amounts of phosphorus (P) contained in meat, sh ell and clam tissues along with percentages of total clam phosphorus allocated to meat and shell tissues for individual clams. The majority of the total clam phosphorus is a result of meat phosphorus and not shell phosphorus. Parameter Amount of P contained in meat (mg) Amount of P contained in shell (mg) Amount of P contained in clam (mg) % of Total clam P allocated to meat % of Total clam P allocated to shell Mean 0.480 0.095 0.575 82.2 17.8 SE 0.014 0.003 0.016 0.4 0.4 Range 0.087 1.137 0.042 0.347 0.143 1.411 50.7 93.1 6.9 49.3 n 228 228 228 228 228 No statistically significan t conclusions related to s easonality and clam phosphorus concentration could be assessed due to the drastically decreasing sample sizes at each sampling interval after stocking as a result of the high ove rall population mortality th at took place in all of the raceways. Shell phosphorus values may have been slightly higher than expected since internal cavity area was not cleaned of re sidual pallail fluid and meat from shucking. Phosphorus detected from the inner clam shell was not accounted for in the analysis and is considered minimal, therefore no further investigation is needed

PAGE 127

127 Treatment raceway clam population phosphorus No significant reductions in net phosphorus we re documented in the raceways, probably due to high overall clam mortality. The amount of phosphorus contained in the total raceway live clam populations was calculated using the mean clam tissue phosphorus concentration values in Table 4-12 and the clam population DW biomass es timation in Figure 3-13, as shown in Figure 4-3. High clam population mortal ity prevented accurate assessm ent of raceway clam population phosphorus on a temporal scale. Figure 4-3. Amount of phosphorus (P) sequestered in clam biomass for the low, medium and high nutrient addition treatments over the study period. Raceway clam phosphorus decreased with clam mortality and biom ass loss, and no significant accumulation of phosphorus occurred over any of seasonal time periods. In the high nutrient addition treatment pond raceways from February to May 2003 (Table 3-7), clams measuring 12.2 to 28.6 mm in shell length grew from 0.042 to 0.118 mm/individual/day (Figure 3-11). This equate s to a phosphorus sequestration rate of 0.0022 to 0.0079 mg P/individual/day calcula ted from the clam DW to sh ell length relationship for the high nutrient addition treatment pond group (Table 3-11) and the clam biomass phosphorus concentration of 0.299 mg P/g DW (SE = 0.005) (T able 4-12). Individual tagged clams in the 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 05/14/0208/17/0211/20/0202/23/0305/29/0309/01/03 DateClam P (mg ) Low Med High

PAGE 128

128 medium nutrient addition trea tment were able to sequester from 0.0015 to 0.0071 mg P/individual/day during February to May 2003 time period. Due to low individual clam growth rates (0.026 to 0.086 mm/day) during winter (November 2002 to February 2003) (Table 3-7), phosphorus sequestration potenti al was limited to less than 0.0004 mg P/individual/day. Discussion The goal of this phase of the clam treatme nt raceway study was to evaluate phosphorus removal potential and limitations. The three majo r components of the evaluation incuded, 1) distribution of phosphorus take n up by clams, 2) rate of phosphrous removal by the clam raceways and 3) sustainability of phosphorus rem oval. The results of the study revealed some promising best-case scenarios for the system in terms of phosphorus removal, but also a number of major challenges in terms of sustainability. Distribution of Phosphoru s Taken Up By Clams Clams allocate phosphorus in both meat and shell biomass. Mean clam biomass phosphorus concentration determined from this study was 0.299 mg P/g of clam DW (SE = 0.005, range 0.141 to 0.606, n = 228). Phosphorus c oncentration did not differ with clam size or between populations of stocked clams from the Sa nta Fe River and Lake George. Clam biomass phosphorus was allocated in greater amounts to clam meat tissue, as opposed to shell, as suggested by Fuji (1979). Meat tissue had much higher average concen trations of phosphorus (7.657 mg P/ g meat DW, SE = 0.103) than shell (0.053 mg P/g shell DW, SE = 0.001), and as a result, more clam phosphorus was seques tered in meat than shell biomass. On an individual clam basis, clams containe d 0.143 1.411 mg P/individual based on clam phosphorus concentration and dry wei ght values. This is slightly higher than values reported by Fuji (1979) for Corbicula in an estuarine lagoon (0.0081.500 mg P/individual). Clam meat phosphorus ranged from 0.087 to 1.137 mg P/individua l in this study, similar to the range of

PAGE 129

129 0.007 to 1.433 mg P/individual reported by Fuji (1979). Shell phosphorus ranged from 0.042 to 0.347 mg P/individual in this st udy, slightly higher compared to values ranging from 0.001 to 0.067 mg P/individual shell found by Fuji (1979). The relationship between shell size and individual clam phosphorus was suggested to be exponential by Fuji (197 9). In this study, individual clam phosphorus to clam size followed a curvilinear regression relationship developed in Chapter 3 for clam biomass to clam size. Decreasing slope with increasing size in clams larger than 28 mm in shell length was most likely due to shell erosion evident in the larger clam specimens from Lake Dalhousie, FL as explained in Chapter 3. Other studies of biomass phosphorus in Corbicula (Fuji 1979) and saltwater Manilla clams, Ruditapes (Nizzoli et al. 2006) have suggested no effect of clam si ze or sample location on biomass phosphorus concentration; however, these studies did find that biomass phosphorus concentration differs seasonally. It was suggested that increases in clam phosphorus concentrations during warmer months are related to increases in phytoplankton availability a nd reproductive activity, which did not occur in the clam raceways. Oysters contain around 1.067 mg P/g DW (New ell 2004), which is higher than observed for Corbicula in this study (Table 4-14). Saltwater Manilla clams cultured in an open estuary (Nizzoli et al. 2006) had simila r meat phosphorus concentration as clams in this study; however, phosphorus concentrations in clam shell material in this study (0.053 mg P/g shell DW) were on the low end of the range observed for Manilla clams. Phosphorus con centrations reported by Yamamuro et al. (2000) for the sh ells of other saltwater clams (Musculista, Ruditapes and Anadara), are also higher (0.101 to 0.880 mg P/g shell DW) than phosphorus concentrations reported here. The large difference between the values reported by Yamamuro et al. (2000) and

PAGE 130

130 other studies may be due to chemical absorption from the overlying water column, since Yamamuros values were based on analysis of dead shell material. Table 4-14. Comparison of meat and shell phos phorus concentrations [P] in dry weight (DW) biomass of Corbicula versus other fresh and saltwater clams. Organism Meat [P] (mg P/g of meat DW) Shell [P] (mg P/g of shell DW) Total [P] (mg P/g of total DW) Total P per individual (mg) Reference Corbicula (freshwater) 2.65012.853 0.0220.136 0.1410.6060.1431.411This study Corbicula (estuarine) 8.29322.667 0.0240.056 0.5060.8000.0081.500Fuji (1979) Manilla clams 3.000 > 0.150 0.373 n/a Nizzoli et al. (2006) Hardshell clams 1.100 + 0.250 (SE) n/a n/a n/a Capar &Yess (1996) Softshell clams 1.400 + 0.500 (SE) n/a n/a n/a Capar & Yess (1996) Eastern oysters 1.100 + 0.300 (SE) n/a n/a n/a Capar & Yess (1996) Pacific oysters 2.000 + 0.550 (SE) n/a n/a n/a Capar & Yess (1996) Saltwater clams n/a 0.101 0.880 n/a n/a Yamamuro et al. (2000) Even though meat contains higher amounts of P, it is more mobile af ter death than shell material and can more easily return to the overlying water column or substrate (Fuji 1979, Nizzoli et al. 2006). Shell phosphorus, therefor e, may be a better long-term sequestration product if phosphorus is in fact in corporated into the organic porti ons of the shell material and not just loosely bound to shell surfaces. Elemen ts bound in clam shells may require additional steps to degrade since the mate rial is so tightly bound. The higher levels of phosphorus sequestered into shell biomass in this study may, in part, have been due to meat biomass and fluid retained by the shell during processing (McMahon and Bogan 2001).

PAGE 131

131 Estimates of Phosphorus Uptake Rates Estimated clam biomass phosphorus sequestration rates from the tagged clam study ranged from 0.0022 to 0.0079 mg P/individual/day Estimated annual phosphorus sequestration potential of 0.803 to 2.884 mg P /individual/yr calculated from da ily rates may be overestimated due to lower winter growth of clams. Therefor e, an individual clam se questration rate between 0.201 to 0.721 mg P/individual/yr may be a mo re accurate estimation of clam phosphorus sequestration, since it is based on longer-t erm growth estimates. Using the phosphorus sequestration rate for clam population biomass based on tagged clam growth, a theoretical clam population of 16,560 individuals per 21.6 m2 of raceway area in northern Florida should be able to sequester on average 354 mg P/m2/yr, with a range of 154 to 553 mg P/m2/yr, assuming a 3month cessation in growth during winter. Clam raceway phosphorus sequestration would be expected to increase beyond rates estimated for tagged clam growth if successful reproduction and recruitment occur and clams are not limited by environmental conditions in th e raceway. High reproductive capacity of the animals in large natural systems suggests, that un der ideal conditions, raceways should be able to reach population densities of over 2000 clams/m2 as reported for natural systems by Gardner et al. (1976), Eng (1979), Sickel ( 1986) and McMahon and Bogan (2001) This is roughly 3 times the stocked density of 766 clams/m2. Theoretically, the ideal raceway population would be expected to follow similar phosphorus dynamics as presented in the Fuji (19 79) model, who suggested that an estuarine Corbicula population of 65 mg P/m2 can sequester clam biomass phos phorus at an annual rate of about 130 mg P/m2/yr as growth and recruitment when not limited by food availability. The stocked clam densities in this study represent a total phosphorus level of 14,000 mg P (648 mg P/m2), translating into 28,000 mg P (1296 mg P/m2) of harvested phosphorus annually using the

PAGE 132

132 Fuji (1979) model. This value is much highe r than the tagged clam phosphorus removal rates given in this study, since it does not account fo r mortality, which would ideally be minimized by strategic harvest of clam biomass that would also maintain younger clams capable of higher growth rates. The Fuji (1979) model lends itself well to application in an ideal clam raceway calculation since food resources did not appear to be limited and phosphorus sequestration was influenced by seasonality as it is in the natural environment. Adding to the phosphorus removal potential, phosphorus accumulated in raceways from clam waste would be periodically removed with raceway sediment, thereby minimizing losses to the surrounding environment. The stocked clam popul ation densities used in this study would be expected to produce nearly 4000 mg P/m2/yr in biosolids as feces and pseudofeces according to the Fuji (1979) model. Ideally, a Corbicula-based raceway could be designed and operated to maximize removal of clam wastes and other solid s settling on the substrat e along with growth of clam biomass. A similar approach in bivalv e-based mariculture waste treatment systems has been employed by Shpigel et al. (1993), Shpigel et al. (1996) and Neori et al. (2000) in which settling ponds are used to increase treatment poten tial. Settling of organic constituents may lead to other problems such as dissolved oxygen demand and ammonia production due to decay (Dame 1996). Dissolved phosphorus loss from accu mulated sediments has also been observed in freshwater bivalve-based treatment systems (Soto and Mena 1999). Harvest and substrate removal/replacement times would need to be optimized to minimize these stressors. An annual phosphorus uptake rate of at at least 5300 mg P/m2/yr is expected for the Corbicula based raceway system in this study unde r ideal conditions and solids management. The theoretical clam-based raceway system would also be expected to lose phosphorus at a rate of 1166 mg P/m2/yr as excreted wastes based on the Fu ji (1979) model, that may be more

PAGE 133

133 difficult to capture and are often lost in natural systems with outflowing water. Based on these theoretical values, clam raceway P removal potential would be e xpected to be approximately 4100 mg P/m2/yr if solids capture simply by manageme nt of settling through optimizing raceway hydrology or the addition of a settling stage following clam raceways. Comparison Of Phosphorus Removal By Clam Raceways And Other Systems Estimates for phosphorus sequestration by clams at the stocking rates used in this study are 154 to 553 mg P/m2/yr based on tagged clam growth rate s. The clam phosphorus sequestration values in this study were lower than potential values of a theoretical cultured population of Corbicula, estimated by Mattice (1977) to be between 1200 and 1400 mg P/m2/yr (Table 4-15). The latter range is based on annual clam wet weight biomass growth rates of naturally occurring populations and potential maricu lture density assumptions by Mattice (1977), converted to phosphorus removal capacity using average % wate r and phosphorus content values. However, the biomass phosphorus accumulation range estima ted by Mattice (1977) may be unrealistically high due to population biomass density assumptions of over 10,000 clams/m2, which may not be possible even under ideal culture conditions. In another study, estimated clam biomass phosphorus sequestration for Corbicula exposed to municipal wastewater by Gr eer and Ziebell (1974) was calcu lated using biomass phosphorus content values, yielding 877 mg P/m2/yr (Table 4-15). However, the small scale of the experimental system used by Greer and Ziebell (1974), short time duration of the study and lack of data given on the actual su rface area in the experimental culture, casts some doubt on the value of these estimates for large scal e systems. Estuarine populations of Corbicula have been estimated to sequester phosphorus at 134 mg P/m2/yr (Fuji 1979), wh ile other phosphorus sequestration values calculated by Fuji (1979) for Corbicula populations from other studies

PAGE 134

134 range from 13 to 250 mg P/m2/yr compared to some other freshwater bivalves that ranged from 37 to 77 mg P/m2/yr. Rates reported by Nizzoli et al. (2006) indi cate that phosphorus removed from an estuary through harvest of Manilla clams can range from 1300 to 2600 mg P/m2/yr using 3-month growth intervals between harvests (Table 4-15). Phosphorus sequestration may be possible using raceway-based Manilla clam culture in saltwater systems and could be increased significantly by capture of waste materials lost to tidal outflow and sediment deposition in natural systems. Phosphorus sequestration rates for bivalve cultu re systems are much less than the 98,000 mg P/m2/yr rate given by Dame (1996) for an intert idal oyster reef comm unity containing many different organisms; therefore, this value is not indicative of bivalve phosphorus removal potential. It is difficult to co mpare phosphorus sequestration in Corbicula to mariculture effluent phosphorus treatment systems using other bivalves since treatment potential in these systems is usually assessed on the basis of ni trogen and phytoplankton removal. Table 4-15. Estimated annual phosphorus removal in various biological treatment systems applied to different effluent types using systems of varying design and scale. Estimated phosphorus removal by the Corbicula-based system in this study is similar to other harvested systems, while harvested aquatic plantbased systems are capable of much hi gher removal rates than animal-based systems.* denotes phosphorus removal rate based on periodic harvest of system Effluent type Target organism P removal (mg/m2/yr) System type and scale Reference author(s) Agriculture N and P fertilizer Corbicula 154 to 553 Raceway tagged clam growth Measurements from this study Agriculture N and P fertilizer Corbicula 4100 *Large-scale raceways, theoretical a This study, ideal conditions Nutrient enriched water Corbicula 877 Theoretical aquaculture Greer and Ziebell (1974) Nutrient enriched water Corbicula 1200-1400 Theoretical aquaculture Mattice (1977)

PAGE 135

135 Table 4-15 Continued. Estuary, natural Brackish water clams, Corbicula and Musculista 17-60 Natural tidal estuary Fuji (1979), Yamamuro et al. (2000) Estuary mariculture Manilla clam, Ruditapes 1300-2600 as biomass *Natural tidal estuary Nizzoli et al. (2006) Estuary, natural Oyster reef ecosystem 98,000 Natural tidal estuary Dame et al. (1989) in Dame (1996) Tertiary-treated municipal wastewater Freshwater finfish, Tilapia and catfish 194 to 5840 biomass & wastes *Large-scale pond culture b Hallock and Ziebell (1970), b Greer and Ziebell (1974), b Bunting (2007) Agriculture runoff-enriched surface water Periphyton 36,550 to 47,450 *Large-scale algal turf scrubber a Adey (1993) Tertiary-treated municipal wastewater Periphyton 91,250 to 266,450 *Large-scale algal turf scrubber a Craggs et al. (1996) Diluted dairy wastewater Periphyton 29,200 to 120,000 *Small-scale algal turf scrubber c W ilki e an d Mulbry (2002), Pizzaro et al. (2002) Diluted dairy wastewater Assorted aquatic plants and algae 58,000 *Med.-scale circular tanks a Soo k na h an d W ilki e ( 2004 ) Aquaculture wastewater Lettuce grown hydroponically 365,000 *Med.-scale conveyor belt a A dl er et a l ( 2003 ) Agriculture runoff-enriched surface water Periphyton dominated, mixed vegetation 320 Large-scale wetland-type raceways a DeBus k et a l ( 2004 ) Dairy effluent Corn silage forage crops 6900 *Land application d W ilki e an d Mulbry (2002) Dairy effluent Grassland forage crops 18,800 to 59,000 *Land application Jo h nson et a l ( 2004 ) Surface water systems Natural wetland vegetation 146 to 803,000 General Wetlands Do dd s ( 2003 ) a P removal based on annual P removal values for target organisms from seasonal-based experiments at an appreciable scale b P removal estimated using fish P content fr om Greer and Ziebell (1974) and annual fish production rates from Hallock and Ziebel l (1970) and Greer and Ziebell (1974) c Lower P removal value from Wilkie and Mulbry (2002), higher value from Pizzaro et al. (2002), includes Kebede-Westhead et al. (2003) P removal estimate under the same conditions d P removal estimated by reference author using annual values from other studies Even under ideal culture conditions, estimated removal of phosphorus by Corbicula and other bivalves pales compared to those de termined for aquatic periphyton and macrophyte-

PAGE 136

136 dominated systems (Table 4-15). Harvested syst ems appear to be capable of higher phosphorus removal rates, especially in plant-based treatmen t systems. Natural wetlands show the largest phosphorus removal values; however, they also have the largest range, with values as low as 146 mg P/m2/yr being reported by Dodds (2003) (Table 4-15). Wetlands are subject to capacity limitations after several years whereas harvested systems are theoretically more sustainable. Periodic harvesting increases phosphorus rem oval potential of the or ganisms by minimizing losses due to mortaility, maintaining high grow th rates and increasing the longevity of the system. Vegetative systems may be be tter suited for management of phosphorus in a farm-scale system due to their ability to withstand a wider variety of environmental stressors and decreased water use compared to clam raceways as seen in this study. Vegetative systems have been added to bivalve-based systems in the mariculture industry to provide treatment following exposure to bivalve populations (Shpigel et al. 1993, Shpigel et al. 1996, Neori et al. 2000). Periphyton may also be used to remove phosphorus following exposure to Corbicula-based systems since harvested types of these systems have been suc cessful in rapidly sequestering phosphorus from agricultural applications especi ally at low water phosphorus c oncentrations (Scinto and Reddy 2003). Raceway systems used in this study may be applicable in vegetative-based biofilters as well. Problems With Measuring Short-term Phosphorus Uptake There were no indications that the raceway systems were ab le to provide a noticeable reduction in water phosphorus or chlorophyll a during the normal through-flow operation at a range of 151 to 227 L/minute. Reduction of phos phorus by the clam raceways was not observed even after being converted to a temporary recirc ulation mode to increas e retention time to 6 hours from 9.5 minutes and 6.3 minutes. High temperature, low dissolved oxygen and ammonia

PAGE 137

137 toxicity concerns prevented longer retention times by recircula tion in an attempt to show phosphorus reduction especially during daylight hours. There wa s no significant reduction of total phosphorus levels in th e covered raceways, indicating that raceway water phosphorus reduction may not be affected by clams or settling in the raceway systems. The most plausible explanation for the lack of large-scale phosphorus uptake may be the obvious environmental and physiological stress th at clams were undergoing over the entire study period including during the thr ough-flow and recirculation trials. High clam mortality was evident throughout the study, especially in wa rmer months when water phosphorus removal trials were taking place even though condition indicies did not indicate any decreases in clam health over time. Haines (1977) attributes similar mortality and d ecreased water treatment capacity in Corbicula cultured on municipal wastewater to high temperature and possibly ammonia. Also, potential stress from amphipod infestation may have severely affected the raceway clam populations ability to filter feed and, therefore, remove phosphorus-containing material as described in Chapter 3. Similarly, Kinne et al. (2001) was unable to show that TP wa s lowered significantly in a medium-scale raceway system using oysters to tr eat shrimp farm effluent, possibly because of soluble phosphorus excretion meeting phosphorus uptake or ammonia concer ns. Conversion of particulate phosphorus to dissolved phosphorus by excretion of soluble phosphorus by Corbicula waste products as described by Lauritsen and Mozley (1989) may not explain the lack of phosphorus removal in this system. The extent of phosphorus conversion within the raceways is unknown since significant TDP addition was never de tected during any recirculation trial, and TP removal was no different in raceways contai ning substantial amounts of clams compared to those with depleted stocks. Clam density is expected to limit treatm ent capacity for raceway

PAGE 138

138 clam populations. However, raceway systems in the high nutrient additi on treatment contained approximately 460 clams/m2 (10,000 clams per raceway) more th an the low nutrient system, but still did not provide a notic eable phosphorus removal. Chlorophyll a uptake values in the recirculati on trials may have been induced by disturbance of substrate materials related to switching raceway hydrology from through-flow to recirculation for experimental purposes. Dist urbance of raceway subs trate did occur in the recirculation trials as indicated by the higher TP, TDP and chl a values as opposed to source pond levels at the same time. Haines (1977) attributed increased treatment potential of Corbicula-based systems to increased particle concen tration. Therefore, any disturbance and increase in particle availability may result in removal by clam populations as well as settling. The apparent significant chlorophyll a removal seen in both high and low nutrient addition treatments may have been due to such a disturba nce. However, removal rates were probably not due to settling, since covered raceways failed to provide any definitive indication of chlorophyll a uptake. Given the large differences in cl am population density between the high and low nutrient addition treatments during the recirculation trials, actual removal of chlorophyll a due to clam filtration is also doubtful. Potential pr oblems encountered with the raceway recirculation modification suggests that phosphorus uptake was probably not achievable under normal through-flow operating conditions by the clam raceways, as expected. Problems such as high temperature, low dissolved oxygen and dangerous levels of ammonia may result with the sixhour retention times, as seen in this study. Raceway system hydrology can also be modified to accommodate full or partial raceway water recirculation as demonstrated by the additional pump placement in this study to extend retention time. Water retention time would need to be managed clos ely with real-time mon itoring of temperature,

PAGE 139

139 ammonia and dissolved oxygen to avoid levels th at may negatively affect clam populations. System water retention time greater than 10 minutes and less than 6 hours is recommended to improve treatment potential, while minimizing envir onmental stress. More conclusive testing of TP, TDP and chl a uptake by clam populations at large scale is needed before re liable short-term removal rate estimates and system op erating tolerances can be assessed. Dairy Application Demands And Issues Issues with water consumption may limit widesp read use of this technology since digested dairy wastewater effluent additi on less than 5 % (by volume) was needed in clam raceway source water ponds to overcome ammonia concerns. So oknah and Wilkie (2004) have demonstrated that a 1 : 1 dilution of anaerobi cally digested wastewater enhances biological uptake of phosphorus and nitrogen in aquati c plant-based systems, substa ntially decreasing amounts of ammonia in effluent water. Application of th ese kinds of vegetative systems may provide pretreatment of phosphorus and harmful nitrogenous compounds prior to clam raceway addition to increase feasibility at a large sc ale. Coupling of vegetative and clam-based systems in this way would lower water demand, while decreasing a mmonia and utilizing less treatment surface area than required by clam systems alone. Application of clam raceway t echnology in a real-world scenario is perhaps best analyzed on an individual farm basis. The University of Florida Dairy Research Unit (DRU) in Hague, Florida is described by Wilkie et al. (2004) as having an average milking herd of 359 cows, a wastewater production of 502 m3/day and a daily freshwater water usage of 52.25 m3/day. A large scale, fixed film anaerobic di gester is in place at the DRU and is not expected to change the amounts of P-loading to wastewater handling systems that are restri cted primarily to a secondary settling lagoon until land application. Normal digester outflow would be similar to values

PAGE 140

140 reported by Sooknah and Wilkie (2004) for water qua lity in a system accepting digester effluent from the DRU, including 24 mg/L TP a nd 136 mg/L total ammonia nitrogen (TAN). Phosphorus and ammonia levels in the macrophyte system output are expected to be 0.24 to 6.0 mg/L TP with a significan t portion of phosphorus as dissolved phosphorus and 0.29 to 3.53 mg/L TAN (Sooknah and Wilkie 2004). Insertion of air-stripping or biofilm filtration systems similar to ones used in the a quaculture industry (Timmons et al 2001) or bacteria-based systems under assessment for dairy wastewater, as indicated by Sooknah and Wilkie (2004), could possibly be inserted followi ng macrophyte treatment to lower TAN levels. Another possible solution to the ammonia problem may be anothe r vegetative treatment phase. However, these biological-type systems may not be as dependable as mechanical ones used in commercial aquaculture due to seasonality and temperature dependence of biological systems. Application of aquaculture technology should help to decrease TAN and increas e DO levels with little or no excess water demand except for increased evapora tion with increased treat ment surface area. The continuous growth of nuisance plants evident during raceway operation may indicate the need for a vegetative system to be incor porated into the clam system design to get a noticeable decrease in raceway water phosphorus. Fuji (1979) suggested that clam deposition of organic materials provides a food source for plant growth in detrita l systems; therefore, harvested plant growth in polyculture with Corbicula may provide a way to increase phosphorus sequestration potential of the raceway system. Us e of vegetative-type filters along with bivalves has been demonstrated in marine mariculture wa stewater treatment systems to remove dissolved phosphorus excreted from bivalves (Jones et al. 2001, Borges et al. 2005) and has been suggested for Corbicula culture in agriculture wastewater (Sta nley 1974); but no app lications of this technology coupling have been demonstrated usi ng dairy effluent. Nuisance plant growth did

PAGE 141

141 not appear to be inhibited by the high water temperatures encountered in this study, unlike clam populations, adding to the appeal of plant-based systems. Using the 1:1 dilution and phosphorus rem oval potentials suggested by Sooknah and Wilkie (2004) for an aquatic macrophyte-based sy stem at the DRU, I would expect freshwater usage to increase from 502 m3/day, to 554 m3/day just for the macrophyte system. This would increase wastewater volume to 105 m3/day and require an active system capable of sequestering 655 mg TAN/m2/day (239,075 mg TAN/m2/yr). This value is based on TAN removal of 99.6 % from diluted wastewater containing 68 mg/L TA N and over a 31-day period using a 0.5 m x 0.36 m raceway at a water depth of 0.3 m (Sooknah and Wilkie 2004). The resulting macrophyte system sized for TAN removal at the DRU would need 68 m2 of treatment area per day of retention. With the suggested hydraulic reten tion time of 31 days to remove incoming TAN, raceway treatment area would need to be at least 2116 m2 to meet the TAN treatment needs for Corbicula systems using anaerobically dige sted wastewater at the DRU. Removal of ammonia would be the primary goa l of vegetative treatment prior to clam raceway introduction; however, macrophyte system s have been suggested for treatment of phosphorus as well (Sooknah and Wilkie 2004). The macrophyte system phosphorus removal potential would be 58,000 mg P/m2/yr removing 96.5 % of the incoming phosphorus under ideal conditions as indicated by Sooknah and Wilkie (2 004). Assuming that ammonia levels could be maintained without removal of phosphorus or further dilution, phosphorus loading from the macrophyte system is expected to be up to 230 kg P/yr for the 105 m3/day (38,325 m3/yr). Wastewater would exit the macrophyte system at 0. 24 to 6.0 mg/L TP, considerably higher than the range tested in this study; how ever, within expectati ons of 3 mg/L values tested by Greer and Ziebell (1972) at aquarium scale using Corbicula. Under this maximum phosphorus loading rate

PAGE 142

142 expected in the macrophyte filter outflow, a clam raceway system having up to 38,325 m2 of treatment area would be need ed to remove the remaini ng phosphorus in the system. Losses of phosphorus around 1166 mg P/m2/yr would be expected from the 38,325 m2 of clam raceway system area and would require addi tional treatment prior to discharge into surface waters since effluent TP concentration woul d be around 0.05 to 1.24 mg/L at a volume of 105 m3/day in a properly managed clam system under these circumstances. This value may be improved by increasing the retention time or deceas ing water usage. Appl ication of clams or macrophytes at such a scale is preposterous co nsidering the limited size scale demonstrated by Sooknah and Wilkie (2004) and the lack of large-scale success of Corbicula-based raceways at the DRU that prevents an accurate system sizing from being made. Other large-scale harvested aquatic plant sy stems such the Algal Turf Scrubber (ATS) raceways described by Craggs et al. (1996) may al so need to be employed prior to introduction to clam systems, in addition to macrophyte-based r aceways to lower TP to an optimum level. TP in clam raceway influent should be between the 0.04 and 3 mg/L range shown by Greer and Ziebell (1972) to have some phosphorus removal potential as determined in aquarium-based studies. Small-scale algal turf scrubber systems have been used by Pizzaro et al. (2002), Wilkie and Mulbry (2002) and Kebede-Westhead et al. (2003) on diluted dairy wa stewater at the DRU to remove an estimated 29,200 to 120,000 mg P/m2/yr. Assuming the lower phosphorus removal value of 29,200 mg P/m2/yr, the system would need to have at least 7,876 m2 of treatment area, much larger than the largest AT S wastewater treatment system te sted by Craggs et al. (1996) at 1,018 m2, but not impossible for large scale implemen tation at the DRU. Using ATS systems to treat dairy wastewater at a qua rtenary or lower level, as s uggested by these calculations, is

PAGE 143

143 certainly more feasible than clam raceways requi ring a treatment area 1,774 times larger than the area afforded by the 3-raceway system tested in this study. Concerns over salinity and DO levels in dairy wastewater remediation (Sooknah and Wilkie 2004) may also impact clam raceways. Corbicula can tolerate at least some salinity in the natural environment (Deaton 1980, McCorkle and Dietz 1980), however low tolerance for dissolved oxygen levels in the 3-5 mg/L range reported by Belanger (1985), suggests that oxygen may be the more critical factor. Dissolved oxygen levels of less than 3 mg/L and ammonia levels from 0.29 to 3.53 mg/L TAN in various macrophyte system effluents reported by Sooknah and Wilkie (2004) will need to be addressed prior to application of clam systems since they are not within the tolerable ranges for Corbicula. Sustainability Application of clam raceway sy stems as a dairy wastewater treatment mechanism will need further investigation, primarily since raceway systems were not able to operate under dairy effluent addition at 5 or 10 % by volume due to high ammonia concerns. Keeping both ammonia and temperature within a tolerable level will be critical to implementing Corbicula as a wastewater treatment mechanism. High levels of undesirable materials such as ammonia in source water can be addressed by dilution of effluents; however, in creased constraints on freshwater usage may limit the use of clam raceways alone to solve the phosphorus management of any operating dairy, especially in Florida. Successful impl ementation of clam raceways or other high water demand/low aereal phosphorus sequestration potential for the treatment of dairy wastewater on a large scale may be limited unles s coupled with other technologies such as harvested plant systems as seen in the mariculture industry by Shpigel et al. (1993), Shpigel et al. (1996) and Neori et al. (2000).

PAGE 144

144 Corbicula-based raceway systems may be more app licable to aquaculture waste scenerios since ammonia, dissolved oxygen and temperatur e are typically monitored and managed in commercial aquaculture systems. While salt tolerance in Corbicula affords the animal some expansion into mariculture effluent phosphorus tr eatment, species selection in these systems will probably be limited to saltwater sp ecies traditionally cultured in larger scale as a food crop, such as oysters (Jones et al. 2001). Bivalve-base d systems may be able to perform for longer durations between harvest efforts by utilizing other species due to the sh orter life spans found in Corbicula (normally about 2 years, from McMahon and Bogan 2001). The lack of marketability for Corbicula cultured in wastewater conditions may also be an obstacle to implementation since phosphorus tr eatment effectiveness is not only gauged by phosphorus removal and sequestration potential, but also cost e ffectiveness, energy costs and water consumption. Corbicula has been proposed as a desirabl e protein source by Iritani et al. (1979) and has been recommended as a product to o ffset operational costs in clam-based systems by (Mattice 1977), Greer and Zeib ell (1972) Stanley (1974), and Haines (1977). Even though a market for live Corbicula or Corbicula-based products may exist in some places (Chen 1976, Phelps 1994), clams cultured on dairy or municipal wastewater may not be acceptable in any market due to human health concerns. Biomass produced on wastewater can sequester toxins such as metals (Marcussen 2007) and pesticid es (Barber 2006), although worries over choliform bacteria may be unfounded (Islam 2004). Stanley (1974) cautions about potential problems with poisoning from cyanobacteria toxins in wastewater aquaculture systems as well. The lack of market for Corbicula as a food resource domestically is due in part to the filterfeeding ability of freshwater bivalves that capture s various harmful organisms easily transferred to humans such as Cryptosporidium (Graczyk et al. 2003 and Izumi et al. 2004), Giardia (Graczyk et al. 2003) and

PAGE 145

145 Cyclospora (Graczyk e al 1998). Other species such as pearl oysters can provide profit opportunities other than food or a ggregate in mariculture waste tr eatment systems as suggested by Gifford et al. (2004), and the organisms c ould possibly be adapted to raceway culture conditions. In an attempt to use raceway biomass in a posi tive way, dead clam shells removed from the raceways in this study were spread out as an ag gregate over areas surrounding the raceways that were covered with crushed lime rock to prevent w eed growth. Clam shells are typically removed from aggregate sources excavated from river bo ttoms in the Applachicola River in Florida and may not be as desirable an aggregate choice compared to gravel. Problems with decaying soft tissue may prevent use of clam aggregates without pre-treatment to reduce odors. Substrate removed from the racew ays is rich in clam bioso lids and could potentially be composted for use as a soil amendment (Greer a nd Ziebell 1977). This compost may be more valuable than the clams themselves, therefore, composting both clams and substrates simultaneously may be the best altern ative for phosphorus sequestered by Corbicula raceway systems. Products from vegetative-based phos phorus treatment systems are often equally unwanted and ultimately the most valuable as a soil amendment by composting the material to offset operational costs in plant and animal-bas ed treatment systems as suggested by Greer and Ziebell (1974). Even if an economically feasible use could be found for Corbicula-based treatment raceway by-products, and adequate phosphorus rem oval on a large scale could be established in a wastewater treatment stream, implementation of a system that utilizes such an invasive animal can be met with negative responses. Introduction of Corbicula has been implicated in different large-scale habitat modifications in natural water bodies that may not be viewed as acceptable in

PAGE 146

146 all surface water systems (Ingram 1959). Other studies indicated concer ns over native mussel displacement due to the high colonizati on success and reproductive capacity of Corbicula (Kraemer 1979, Darrington 2002, Cooper et al. 2005). Inevitably, no matter what size and component design of a biologica lly-based system, land and water demands, or its usefulness and desi rability, environmental conditions must allow clams to survive, grow and reproduce to sequest er phosphorus effectively into shell and meat tissue at a large scale. Potent ial problems with high temperatur es, ammonia, food availability, seasonality, phytoplankton producti on, amphipod infestations, system performance evaluation, biofouling by vegetation, available land area and water usage need to be solved before a reliable large-scale Corbicula-based raceway treatment system can be implemented into any wastewater stream in Florida.

PAGE 147

147 CHAPTER 5 SUMMARY The three prim ary goals of this dissertation re search project were; 1) To design, construct and implement an experimental raceway system for removal of particulate phosphorus from wastewater streams on a size scale that represents r eal-life applications 2) To test suitability of the freshwater clam Corbicula as the primary active agent in phosphorus removal within a raceway environment and 3) To determine if P-re moval capabilities of the sy stem are adequate to deal with phosphorus loads anticipa ted for dairy wastewater streams. Raceway Function and Attributes The raceway-based systems constructed in this study were chosen and developed as a lowcost, portable and easier-to-assemble alternative to other raceways constructed from concrete, fiberglass or plastic. The a dvantage of the modular raceway design was the culture tanks are easily scalable by length, width, depth and quantity of units to meet surface area needs of the application, desired hydrologic regimes and e xperimental design criter ia. The design also maximizes physical accessibility to the botto m area for stocking, sampling, harvest and maintenance. The raceway systems constructed at both Blountstown and Hague, FL sites operated without failure or leaki ng over the study period and provi ded an ideal platform for the water treatment experiments incorp orated in the current study. Adaptability of Clams to Raceway Conditions The tag and recapture study was the prim ary measure of growth potential of Corbicula. Individual tagged clam shell growth rates based on length ranged from around 0.001 mm/day on an annual basis in the low nutrient addition tr eatment group to up to 0.118 mm/day in the high nutrient addition treatment group in spring. Tem poral patterns in tagged clam growth rates showed seasonality prossibly due in part to changing water te mperature, level of nutrient

PAGE 148

148 addition, source water phosphorus levels and possibly phytoplankton availability. Tagged clams grew in all nutrient addition treatment groups without a consistent co rrelation to chlorophyll a concentration, suggesting that e ither clams utilized non-chlorophyll a containing food resources, such as bacteria and suspended detritus, or phytoplankton biomass in the source ponds was sufficient to sustain growth. Despite the growth observed in the surviving tagged clams, the overall clam population did not adapt well to raceway conditions over extend ed periods, with over 90 % mortality during a one-year period. Timing of mortality indicated that high summer temperatures in the raceways may have been the major factor responsible for the severe losses. Water temperatures in the range of 30 oC and above have been implicated as a limiting factor in the success of Corbicula in other applied studies and in na tural populations (Greer and Ziebell 1972, Mattice 1977, Haines 1979, Buttner and Heidinger 1980, Buttner 1986). Major population declines took place when water temperatures reached this level in all systems, regardless of the level of nitrogen and phosphorus addition and chlorophyll a in the source water. Other environmental factors present in the ra ceways, including incr eased ammonia levels encountered during periodic execution of 6-hour r ecirculation trials, and potential stress from infestation by amphipods (Hyalella azteca), may have also contribute d to mortality and affected phosphorus removal and sequestration potential as well. Interactions observed between Corbicula and Hyalella in this study are the first to recognize amphipods as having a potential predatory or parasitic role in clam populati on dynamics; however, the intricacies of this interaction are not yet understood. P-removal Capacity Clam raceway systems were able to dem onstrate phosphorus removal and sequestration potential as evidenced by significant shell growth in tagged clams. Individual clams selected for

PAGE 149

149 analysis were between 14.9 and 30.4 mm shell length (0.773 6.041 g total clam DW). Clams contained an average of 0.299 mg P/g DW (SE = 0.005, range 0.141 to 0.606, n = 228), derived from shell and meat values combined, which e quates to 0.143 to 1.411 mg P/individual. The concentration of phosphorus (mg P/g DW) in the shel l and meat tissues did not change with shell size, location collected or exposur e time in the raceways. Meat tissue had a much higher average concentration of phosphorus (7.657 mg P/ g meat DW, SE = 0.103) than the shell (0.053 mg P/g shell DW, SE = 0.001). As a result, the majority of the total clam phosphorus was sequestered in meat as compared to shell biomass, even though shell biomass comprised the majority (approximately 82%) of the total clam DW biomass. Based on the tagged clam study, phosphorus sequestration potential was estimated to range from 0.803 to 2.884 mg P/indivi dual/yr for adult clams. Corbicula biomass phosphorus sequestration potential estimated in this study was similar to values given by Fuji (1979) for biomass phosphorus sequestration in natural populations. Ideally, raceway clam populations would be able to sustain estimated clam biomass phosphorus sequestration rates along with successful reproduction and recruitment as seen in the Fuji (1979) model. An similar raceway system as tested in this study under ideal conditions would be stocked with 648 mg P/m2, and would be expected to sequester 28,000 mg P/yr (1296 mg P/m2/yr) as clam biomass. In a clam culture scenario, this live biomass could be harvested along with most of the 708 mg P/m2/yr expected to be produced by normal mortality and excretion retained in the raceway sediment. Ideally, a Corbicula-based raceway would be designed and operated to maximize sequestration of particulate clam waste phosphorus and other solids settling along with growth of clam biomass. Phosphorus removal potential may be higher in an engineered system equipped to deal with environmental variati on and capable of further enhancing growth rate and treatment

PAGE 150

150 area through sustainable harvest. By retaining and harvesting se ttled particulates that would otherwise be lost in natural systems and maxi mizing growth and reproduction rates, an annual phosphorus removal rate of at least 4100 mg P/m2/yr is expected for the Corbiculabased raceway system as tested in this study. This phosphorus removal and seque stration potential is higher than potentials calculated for other freshwater bivalve a nd finfish populations. Future Applications Limitation of reproductive success by Corbicula held in captivity in this and other studies (McMahon and Bogan 2001) suggests that implementation of Corbicula systems at large scale may not be attractive if juveniles need to be restocked periodically from natural populations. Perhaps partially open systems such as power plan t discharge canals that are open to a natural water body part of the year would allow for inflow of juveniles to repopulate raceways after harvest or mortality events may be needed to maintain adequate stocks. Other solutions for sustaining clam stocks in raceway-based systems su ch as genetic selection for traits that support growth and reproduction outside of the normal tolerances has been proposed for Corbicula by Sickel (1986) however this would take an exceptionally long time to develop, if at all, given homogeneity and genotypic plasticity of Corbicula. Clam raceway, phosphorus treatment potential of anaerobically digested dairy effluent could not be assessed directly, due to toxicity con cerns over total ammonia nitrogen levels present at > 2 mg/L NH3-N present in source ponds at 5 % and 10 % (volume of source pond volume) wastewater addition levels tested. Low tolerance to ammonia toxicity in Corbicula (Cherry et al. 2001) and to ammonia in combin ation with high temperatures have been implicated in limiting the success of clam-based treatment systems utilizing municipal sewage treatment plant effluentsby Haines (1979). A mmonia toxicity is of great concern in all aquaculture systems (Harris et al. 1998), espe cially recirculating ones where technologies

PAGE 151

151 utilizing treatment by volatiliza tion, sequestration or convers ion have been developed to encourage higher density cultures of fish and other aquatic organisms (Timmons et al. 2002). Lack of ammonia management technologies tested at large scale for application in clam raceway systems will be an important goal in future development of freshwater animal aquaculture applications for dairy waste treatment. Bivalv e-based treatment of dairy-derived wastewater phosphorus would require implementation and integr ation of additional treat ment technologies in order to reduce high levels of nitroge nous wastes common in dairy operations. Application of vegetative systems such as wetlands, managed aquatic plant systems and biofilm systems may be able to provide needed treatment of ammonia prior to clam raceway addition in order to increase feasibility. Coupling of vegetative and clam-based systems in this way would lower water demand while decreasing ammonia and utilizing less treatment surface area than required by clam systems alone. Using vegetative-type filters along with bivalves has also been demonstrated in marine mariculture wastewater treatment systems to remove dissolved phosphorus from the outflow of bivalve-based systems (Jones et al. 2001, Borges et al. 2005) and has been suggested for Corbicula culture in agriculture wa stewater by Stanley (1974). Consideration should be given to source pond geology, depth, sediment permeability and use of clay or plastic liners in future system s to limit or promote possible exchange of pond water nutrients and heat with the surrounding envi ronment to maintain tolerable environmental conditions. Seasonality of system function is a central issue not only from the standpoint of summer high temperatures, but from growth and su rvival, issues under low winter temperatures. Despite the lack of Corbicula success on a large scale, th e raceway-based recirculation system design demonstrated in this study provided a dependable, easy to construct and reusable platform for testing aquaculture potential of organisms in wastewater treatment conditions at

PAGE 152

152 large scale. The design should be used as a st andard system to assess other organisms besides Corbicula for comparative purposes since systems can be easily constructed to accommodate a variety of operating parameters. The raceway sy stem designs employed here are versatile enough to be applied to other organisms such as bivalves, fish, algae a nd high plants; targeted for largescale water treatment/biofiltration studies in both fresh and saltwater condi tions and a variety of locations, as well as effluent sources. From a comparative standpoint, it is important to make the observation that the phosphorus removal capacities of many animal-based syst ems, which typically range from 17 to 5840 mg/m2/yr, are low compared to the best-case estima tes for algae and other plant based systems, which range from 320 to 365,000 mg/m2/yr. However, this comparison is somewhat misleading, since the function and structure of the two system s are different in several important aspects. Plant/algae systems provide a mechanism for remova l of particulate and dissolved nutrient forms as opposed to animal-based systems that focus on particulates. Also, plant/algae systems are energetically dependent on sunlight, while animal systems can be independent of such a direct requirement. In addition, the end products of plant/algae and animal systems are fundamentally different and subject to separa te end-use issues and opportuniti es. Ultimately, many treatment needs may be best addressed by integration of animal and plant/algae systems, thereby allowing for optimal processing of soluble and particul ate wastes and production of a wide range of valuable goods and services, such as food, feed, biodiesel, bu ilding materials and chemicals.

PAGE 153

153 LIST OF REFERENCES Adey, W ., C. Luckett and K. Jensen. 1993. Phosphorus removal from natural waters using controlled algal production. Restoration Ecology March 1993:29-38. American Public Health Association (APHA). 1998. Standard methods for the examination of water and wastewater, 20th edition. Section 4500-P B5 & E. Anthony, J.L., D.H. Kesler and W.L. Downi ng. 2001. Length-specific growth rates in freshwater mussels (Bivalvia: Unionidae): extreme longevity or generalized growth cessation?. Freshwater Biology 46:1349-1359. Azim M.E., M.A. Wahab and A.A. van Dam. 2001. Optimization of fertilization rate for maximizing periphyton production on artificial substrates and the implications for periphyton-based aquaculture. Aquaculture Research 32:749-760. Azim, M.E., A. Milstein, M.A. Wahab, M.C. J. Verdegam. 2003. Periphyton-water quality relationships in fertilized fis hponds with artificial substrates. Aquaculture 228:169-187. Barber, L.B., S.H. Keefe and R.C. Antweiler. 20 06. Accumulation of contaminants in fish from wastewater treatment wetlands. Environmental Science and Technology 40:603-611. Beaver, J.R., T.L. Crisman and R.J. Brock. 1991. Grazing effects of an exotic bivalve (Corbicula fluminea) on hypereutrophic lake water. Lake and Reservior Management 7(1):45-51. Belanger, S.E.. 1991. The effect of dissolved oxygen, sediment and sewage treatment plant discharges upon growth, survival and density of Asiatic clams. Hydrobiologia 218:113126. Bitterman, A.M., R.D. Hunter a nd R.C. Haas. 1994. Allometry of shell growth of caged and uncaged zebra mussels (Dreissena polymorpha) in Lake St. Clair. American Malacological Bulletin 11(1):41-49. Blalock, H.N. and J.J. Herod. 1999. A compara tive study of stream habita t and substrate utilized by Corbicula fluminea in the New River, Florida. Florida Scientist 62(2):145-151. Brousseau, D.J.. 1979. Analysis of growth rate in Mya arenaria using the Von Bertalanffy equation. Marine Biology 51(3):221-227. Brock, R.J.. 2000. Assessment of aquatic food web alterations in the presence of the exotic clam, Corbicula fluminea and the cichlid, Oreochromis aureus. PhD Dissertation. University of Florida, Gainesville, FL. 217 pp. Bunting, S.W.. 2007. Confronting the realities of wastewater aquaculture in peri-urban Kolkata with bioeconomic modeling. Water Research 41:499-505.

PAGE 154

154 Busch, R.L.. 1974. Asiatic clams Corbicula manilensis (Phillippi) as biological filters in channel catfish, Icalurus punctatus (Rafinesque) cultures. MS Th esis. Auburn University. 84 pp. Buttner, J.K.. 1986. Biology of Corbicula in catfish rearing ponds. Proceedings of the Second International Corbicula Symposium. The American Malacological Bulletin. Special Edition No. 2:211-218. Buttner, J.K. and R.C. Heidinger. 1980. Seasona l variations in growth of the Asiatic clam, Corbicula fluminea (bivalvia: corbicu lidae) in southern Illinois fish pond. The Nautilus 94(1):8-10. Cataldo, D.H., D.Boltovskoy, J. Stripekis and M. Pose. 2001. Condition index and growth rates of field caged Corbicula fluminea (Bivalvia) as biomarkers of pollution gradients in the Parana River delta (Argentina). Aquatic Ecosystems Health and Management 4:187-201. Cavallo, D., A. Pusceddu and A. Giangrande. 200 7. Particulate organic matter uptake rates of two benthic filter-feeders (Sabella spallanzanii and Branchiomma luctuosum) candidates for the clarification of aquaculture wastewaters. Marine Pollution Bulletin 54(5):622-625. Chen, T.P.. 1976. Culture of the freshwater clam Corbicula fluminea. IN: Aquaculture Practices in Taiwan. pp. 106-111. Copar, S.G. and N.J. Yess. 1996. Survey of elements in clams and oysters. Food Additives and Contaminants 13(4):553-560. Covich, A.P. and J.H. Thorpe. 2001. Introductio n to the subphyllum Crustacea. IN: Ecology and Classification of North American Freshwater Invertebrates, J.H. Thorp and A.P. Covich, eds., Academic Press, San Diego, CA. pp. 777-809. Cox, G.W.. 1996. Laboratory manual of general ecology, 7th edition. Wm.C. Brown Publishers, Bubuque, IA. 278 pp. Craggs, R.J., W.H. Adey, B.K. Jessup and W.J. Oswald. 1996. A controlled stream mesocosm for tertiary treatment of sewage. Ecological Engineering 6:149-169. Dame, R.F.. 1996. Ecology of marine bivalves an ecosystem approach. CRC Marine Science Series. CRC Press, Boca Raton, FL. 254 pp. Dao, T.H., A. Lugo-Ospina and J.B. Reeves II I. 2006. Wastewater ch emistry and fractionation of bioactive phosphorus in dairy manure. Communications in Soil Science and Plant Analysis 37:907-924. DeBusk, T.A., K.A. Grace and F.E. Dierberg. 2004. An investigation of the limits of phosphorus removal in wetlands: a mesocosm study of a shallow periphyton-dominated treatment system. Ecological Engineering 23:1-14.

PAGE 155

155 Debrosse, G.A. and S.K. Allen, Jr.. 1994. The suitability of land based evaluations of Crassostrea gigas as an indicator of performance in the field. Abstracts, National Shellfisheries Association, 1994 Annual Meeting, Charleston SC. p. 277. Dempster, P., D.J. Baird and M.C.M. Beveridge 1995. Can fish survive by filter-feeding on microparticles? Energy balance in tilapia graxing on algal suspensions. Journal of Fish Biology 47:7-17. Drapcho, C.M. and D.E. Brune. 2000. The partit ioned aquaculture system: impact of design and environmental parameters on algal produc tivity and photosynth etic production. Aquacultural Engineering 21:151-168. Edwards, D.D. and R.V. Dimock, Jr.. 1988. A comparison of the population dynamics of Unionicola formosa from anodontine bivalves in North Carolina farm pond. Journal of Elisha Mitchell Science Society 104:70-78. Enes, P. and M.T. Borges. 2003. Evaluation of microalgae and industrial cheese whey as diets for Tapes decussates (L.) seed: effects on water qualit y, growth, survival, condition and filtration rate. Aquaculture Research 34:299-309. Eng, L.L. 1976. A note on the occurr ence of a symbiotic oligochaete, Chaetogaster limnaei in the mantle cavity of the Asiatic clam, Corbicula manilensis. The Veliger 19(2):208. Epifanio, E.E. and R. Srna. 1975. Toxicity of ammonia, nitrite ion, nitrate ion and orthophosphate to Mercenaria mercenaria and Crassostrea virginica. Marine Biology 33:231-236. Fisher, G.R., R.E. Kuhn and R.V. Dimock, Jr.. 2000. The symbiotic watermite Unionicola formosa (Acari-Unionicolidae) ingests mucus and tissue of its molluscan host. Journal of Parasitology 86:1254-1258. Foe, C. and A. Knight. 1985. The effect of phytoplankton and suspended sediment on the growth of Corbicula fluminea (Bivalvia). Hydrobiologia 127:105-115. Fuji, A.. 1979. Phosphorus budget in natural population of Corbicula japonica prime in poikilohaline lagoon, Zyusan-ko. Bulletin of the Faculty of Fisheries, Hokkaido University 30(1):34-49. Gainey, L.F.. 1978. The response of the Corbiculid ae (Mollusca: Bivalvia) to osmotic stress: the organismal response. Physiological Zoology 51(1):68-78. Gardner, J.A., Jr., W.R. Woodall, Jr and A.A. Staat s, Jr.. 1976. The invasi on of the Asiatic clam (Corbicula manilnsis Philippi) in the Altamaha River, Georgia. Nautilus 90:117-125. Ghaly, A.E., M. Kamal and N.S. Mahmoud. 2005 Phytoremidiation of aquaculture wastewater for water recycling and production of fish feed. Environment International 31:1-13.

PAGE 156

156 Gifford, S., R.H. Dunstan and W.OConner. 2004. Pearl aquaculturepr ofitable environmental remediation?. The Science of the Total Environment 319:27-37. Greer, D.E. and C.D. Ziebell. 1972. Biological removal of phosphates from water. Journal Water Pollution Control Federation 44(12):2342-2348. Habel, M.L.. 1970. Oxygen consumption, temperat ure tolerance, and f iltration rate of the introduced Asiatic clam Corbicula manilensis from the Tennessee River. MS Thesis. Auburn University, USA. 91 pp. Haines, C.M.C., M. Edmunds and A.R. Pewsey. 1994. The pea crab, Pinnotheres pisum (Linnaeus, 1767), and its associ ation with the common mussel, Mytilus edulis (Linnaeus, 1758), in the Solent (UK). Journal of Shellfish Research 13(1):5-10. Haines, K.C.. 1977. The use of Corbicula as a clarifying agent in experimental tertiary sewage treatment process on St. Croix, U.S. Virgin Islands. Proceedings, First International Corbicula Symposium. Texas Christian University, USA. pp. 165-175. Harris Jr., B., D. Morse and H.H. Head. 1990. Phosphorus nutrition and excretion by dairy animals. University of Maryland, USA website: http./www.inform.umd.edu. 11 pp. Hickman, R.W.. 1992. Mussel cultiv ation, Chapter 13. IN: The mussel Mytilus-ecology, physiology, genetics and culture. Gosling, E., ed. Elsevier, New York, NY. 598 pp. Huchette, S.M.H., C.S. Koh and R.W. Day. 2003. Growth of juvenile blacklip abalone (Haliois rubra) in aquaculture tanks: effect s of density and ammonia. Aquaculture 219:457-470. Ingram, W.M.. 1959. Asiatic clams as poten tial pests in Californ ia water supplies. Journal of the American Water Association 51:363-370. Islam, M.S., M.S. Kabir and S.I. Khan. 2004. Wastewater-grown duckweed may be safely used as fish feed. Canadian Journal of Microbiology 50:51-56. Jakobworks, G.S., G.D. Pruder and J. Wang. 19 93. Growth trial with the American Oyster Crassostrea virginica using shrimp pond water as feed. Journal of the World Aquaculture Society 24(3): 344-351. Jara-Jara, R., A.J. Pazos and M. Abad. 1997. Growth of clam seed (Ruditapes decussates) reared in wastewater effluent from a fi sh farm in Galicia (NW Spain). Aquaculture 158:247-262. Jara-Jara, R., M. Abad and A.J. Pazos. 2000. Growth and reproductive patterns in Venerupis pullastra seed reared in wastewater effluent from a fish farm in Galicia (N.W. Spain). Journal of Shellfish Research 19(2):949-956. Johnson, A.F., D.M. Vietor and M. Rouquette, Jr.. 2004. Fate of phosphorus in dairy wastewater and poultry litter applied on grassland. Journal of Environmental Quality 33:735-739.

PAGE 157

157 Jones, T. and G.K. Iwama. 1991. Polyculture of the Pacific oyster, Crassostrea gigas (Thunberg), with Chinook salmon, Oncorhynchus tshawytscha. Aquaculture 92:313-322. Jones, A.B. and N.P. Preston. 1999. Sydney rock oyster, Saccostrea commercialis (Iredale and Roughley), filtration of shrimp farm effluent: the effects on water quality. Aquaculture Research 30(1):51-57. Jones, A.B., W.C. Dennison and N.P. Preston 2001. Integrated treatment of shrimp effluent by sedimentation, oyster filtration and macroa lgal absorption: A laboratory scale study. Aquaculture 193:155-178. Joy, J.E. and L.E. McCoy. 1975. Comparisons of shell dimensions and viscera mass weights in Corbicula manilensis (Philippi, 1844). Nautilus 89:51-54. Kebede-Westhead, E., C. Pizarro, W.W. Mulbry and A.C. Wilkie. 2003. Production and nutrient removal by periphyton grown under different lo ading rates of anaerobically digested flushed dairy manure. Journal of Phycology 39:1275-1282. Kinne, P.N., T.M. Samocha and E.R. Jones. 2001. Characterization of intensive shrimp pond effluent and preliminary studies on biofiltration. North American Journal of Aquaculture 63:25-33. Knight, R.L., V.M.E. Payne Jr. and R.E. Bore r. 2000. Constructed wetlands for livestock wastewater management. Ecological Engineering 15(1-2):41-55. Kraemer, L.R.. 1979. Corbicula (Bivalvia: Sphaeriacea) vs indigenous mussels (Bivalvia: Unionacea) in U.S. rivers: a hard case for interspecific competition. American Zoologist 19:1085-1096. Kraemer, L.R., C. Swanson, M. Galloway and R. Kraemer. 1986. Biological basis of behavior in Corbicula fluminea, II. Functional morphology of re production and development and review of evidence for self-fertilization. Proceedings of the Second International Corbicula Symposium. The American Malacological Bulletin. Special Edition No. 2;193201. Krom, M.D., S. Ellner, J. van Rijn and A. Neori. 1995. Nitrogen and phosphorus cycling and transformations in a prototype non-polluting in tegrated mariculture system, Eilat, Israel. Marine Ecology Progress Series 18:25-36. Kurth, J., C. Loftin and J. Zydlewski. 2007. PI T tags increase effectiven ess of freshwater mussel recaptures. Journal of the North Amer ican Benthological Society 26(2):253-260. Lansing, S.L. and J.F. Martin. 2006. Use of an ecological treatment syst em (ETS) for removal of nutrients from dairy wastewater. Ecological Engineering 28:235-245. Lauritsen, D.D.. 1985. Filter-feeding, food utilization and nutrient remineralization by Corbicula fluminea (Bivalvia) and its contribution to nutrient cycling in a North Carolina river. PhD Dissertation. North Carolina State University, Raleigh, NC. 130 pp.

PAGE 158

158 Lauritsen, D.D., S. Mozley. 1989. Nu trient excretion by the Asiatic clam Corbicula fluminea. The North American Benthological Society 8:134-139. Layzer, J.B. and J.R. Heinricher. 2004. C oded wire tag retention in ebonyshell mussels Fusconaia ebena. North American Journal of Fisheries Management 24:228-230. Lefebvre, S., L. Barille and M. Clere. 2000. Pacific oyster (Crassostrea gigas) feeding responses to a fish-farm effluent. Aquaculture 187:185-198. Lemarie, D.P., D.R. Smith and R.F. Villella. 1995. Evaluation of tag types and adhesives for marking freshwater mussels. The Conservation and Management of Freshwater Mussels II: Initiatives for the Future, October 16-18, St. Louis Missouri. National Biological Service: 14pp. Lin, J., M. Sparis and R.W. Ha good. 2001. Growing giant clam (Tridacna derasa) in aquaculture effluent. Aquarium Sciences and Conservation 3:225-230. Lincoln, E.P., J.W. Crawford and A.C. Wilkie. 1993. Spirulina in animal agriculture. Bulletin de lInstitut oceanographique, Monaco 12:109-115. Lincoln, E.P., A.C. Wilkie and B.T. French. 1996. Cyanobacterial process for renovating dairy wastewater. Biomass and Bioenergy 10(2):63-68. Lorio, W.J. and S. Malone. 1995. Biology and culture of the northern quahog clam Mercenaria mercenaria. USDA Southern Regional Aquaculture Center. Publication No. 433. 4 pp. MacMillan, R.J., R.J. Cawthorn and S.K. Whyt e. 1994. Design and maintenance of a closed artificial seawater syst em for long-term holding of bivalve shellfish. Aquacultural Engineering 13:241-250. Marcussen, H., P.E. Holm and L.T. Ha. 2007. Food safety aspects of toxic element accumulation in fish from wastewater-fed ponds in Hanoi, Vietnam. Tropical Medicine and International Health 12(2):34-39. Mattice, J.S.. 1977. Interactions of Corbicula sp. with power plants. Proceedings, First International Corbicula Symposium. Texas Christian University. pp. 120-138. Mattice, J.S., and L.L. Wright. 1986. Aspects of growth of Corbicula fluminea. Proceedings of the Second International Corbicula Symposium. The American Malacological Bulletin. Special Edition No. 2: 167-178. Mazzola, A. and G. Sara. 2001. The effect of fish farming organic waste on food availability for bivalve mollusks (Gaeta Gulf, Central Tyrrhen ian, MED): stable carbon isotopic analysis. Aquaculture 192:361-379. McCorkel, S. and T.H. Dietz. 1980. Sodium transport in the fresthwater Asiatic clam Corbicula fluminea. Biological Bulletin 159:325-336.

PAGE 159

159 McMahon, R.F. 1983. Ecology of an invasion pest bivalve, Corbicula. IN: The Mollusca, vol 6, W.D. Russell-Hunter, ed.. Academic Press, San Diego, CA. pp. 505-561. McMahon, R.F. 2002. Evolutionary and physiologica l adaptations of aquatic invasive animals: r selection versus resistance. Canadian Journal of Fisheries and Aquatic Science 59:12351244. McMahon, R.F. and C.J. Williams. 1986. A reassessm ent of growth rate, life span, life cycles and population dynamics in a natural populat ion and field cage d individuals of Corbicula fluminea (Muller) (Bivalvia: Corbiculacea). Proceedings of the Second International Corbicula Symposium. The American Malacological Bulletin. Special Edition No. 2:156-166. McMahon, R.F., Bogan, A.E.. 2001. Mollusca: Bivalvia. IN: Ecology an d Classification of North American Freshwater Invertebrates, J. H. Thorp and A.P. Covich, eds. Academic Press, San Diego, CA. pp. 331-429. Mercado-Silva, N.. 2005. Condition index of the eastern oyster, Crassostrea virginica (Gmelin, 1791) in Sapelo Island, Georgiaeffects of site, position on bed and pea crab parasitism. Journal of Shellfish Research (24)1:121-126. Miller, R.C. and F.A. McClure. 1931. The fr esh-water clam industry of the Pearl River. Lingham Science Journal 10(2 & 3):307-322. Mulbry, W.W. and A.C.Wilkie. 2001. Growth of benthic freshwater algae on dairy manures. Journal of Applied Phycology 13:301-306. Neori, A., M. Shpigel and D. Be n-Ezra. 2000. A sustainable inte gerated system for culture of fish, seaweed and abalone. Aquaculture 186: 279-291. Nizzoli, D., M. Bartoli and P. Viaroli. 2006. Nitrogen and phosphorus budgets during a farming cycle of the manila clam Ruditapes philippinarum: an in situ experiment. Aquaculture 261:98-108. National Resource Conservation Service. (NRCS). 1999. National Engineering Handbook Part 651. Agricultural Waste Management Field Handbook. US Department of Agriculture, Natural Resources Conservation Service, Washington, DC. National Resource Conservation Service. (NRCS). 2006. National Engineering Handbook Part 651. Agricultural Waste Management Field Handbook. US Department of Agriculture, Natural Resources Conservation Service, Washington, DC. Olszewski, M.. 1977. The potential use of power plant reject heat in comm ercial aquaculture. ORN/TM-5663. Engineering Technology Division, Oak Ridge Na tional Laboratory. Patrick, R., J. Cairns, Jr. and A. Scheier. 1968. The relative sensitivity of diatoms, snails and fish to twenty common constituents of industrial wastes. Progressive Fish Culture 30:137-140.

PAGE 160

160 Phelps, H.L. 1994. Potential for Corbicula in aquaculture. Journal of Shellfish Research 13(1):319. Phlips, E.J., S. Badylak and T. Grosskopf 2002. Factors affecting the abundance of phytoplankton in a restricted su btropical lagoon, the Indian River Lagoon, Florida, USA. Estuarine, Coastal and Shelf Science 55(3):385-402. Pizarro, C., E. Kebede-Westhead and W. Mulbry. 2002. Nitrogen and phosphorus removal rates using small algal turfs grown with dairy manure. Journal of Applied Phycology 14:469473. Prein, M.. 2002. Integration of aquacu lture into crop-animal systems in Asia. Agricultural Systems 71:127-146. Reddy, K.R. and W.H. Smith. 1987. Aquatic plants for water treatment and resource recovery. Magnolia Publishing Inc. 1032 pp. Rodgers Jr, J.H., D.S. Cherry, K.L. Dickson an d J. Cairns. 1977. Invasion, population dynamics and elemental accumulation of Corbicula fluminea in the New river at Glen Lyn, Virginia. Proceedings, First International Corbicula Symposium. Texas Christian University. pp. 99-110. Sartory, D.P. and J.U. Grobbelaar. 1984. Extraction of chlorophyll a from freshwater phytoplankton for spectrophotometric analysis. Hydrobiologia 144(3):177-187. Schmidlin, S. and B. Baur. 2007. Distribution an d substrate preference of the invasive clam Corbicula fluminea in the river Rhine in the region of Basel (Switzerland, Germany, France). Aquatic Sciences 69:153-161. Scinto, L.J. and K.R. Reddy. 2003. Biotic a nd abiotic uptake of phosphorus by periphyton in a subtropical freshwater wetland. Aquatic Botany 77:203-222. Sharpley, A.N., S.C. Chapra and J.T. Wede pohl. 1994. Managing agricultural phosphorus for protection of surface waters: issues and options. Journal of Environmental Quality 23:437-451. Shpigel, M. and R. Fridman. 1990. Propagation of the Manila clam (Tapes semidecussatus) in the effluent of fish aquacu lture ponds in Eilat, Israel. Aquaculture 90:113-122. Shpigel, M. and R.A. Blaylock. 1991. The Pacific oyster, Crassostrea gigas, as a biological filter for a marine fish aquaculture pond. Aquaculture 92:187-197. Shpigel, M., A. Neori, D.M. Popper and H. Gordin. 1993. A proposed model for environmentally clean land-based cultu re of fish, bivalves and seaweeds. Aquaculture 117:115-128.

PAGE 161

161 Shpigel, M. and A. Neori. 1996. The integrated culture of seaweed, abalone, fish and claims in modular intensive land-based systems: I. Proportions of size and projected revenues. Aquaculture Engineering 15(5):313-326. Shpigel, M., A. Gasith and E. Kimmel. 1997. A biomechanical filter for treating fish-pond effluents. Aquaculture 152:103-117. Sickel, J.B.. 1986. Corbicula population mortalities: factors in fluencing population control. Proceedings of the Second International Corbicula Symposium. The American Malacological Bulletin. Speci al Edition No. 2:89-94. Silverman, H., E.C. Achberger and J.W. Lynn. 1985. Filtration and utili zation of laboratorycultured bacteria by Dreissena polymorpha, Corbicula fluminea and Carunculina texasensis. Biological Bulletin 189:308-319. Sindilariu, P.D.. 2007. Reduction in effluent nutri ent loads from flow-through facilities for trout production: a review. Aquaculture Research 38:1005-1036. Sokal, R.R. and F.J. Rohlf. 1995. Biometry, the principles and practice of statistics in biological research, 3rd Edition. W.H. Freeman and Comp any, New York, NY. 887 pp. Sooknah, R.D. and A.C. Wilkie. 2004. Nutrient removal by floating aquatic macrophytes cultured in anaerobically digested flushed dairy manure wastewater. Ecological Engineering. 22:27-42. Soto, D. and G. Mena. 1999. Filte r feeding by the freshwater mussel, Diplodon chilensis, as a biocontrol of salmon farming eutrophi cation. Aquaculture 171: 65-81. Stanley, R.A.. 1974. Methods of biological nutrien ts from livestock waste: a literature review and systems analysis. Environmental Biology Branch in cooperation with Agricultural Resource Development Branch, Tennessee Valley Authority. Pub. Y-80. 45 pp. Stuart, K.R., A.G. Eversole and D.E. Brune. 2001. Filtration of green algae and cyanobacteria by freshwater mussels in the partitioned aquaculture system. Journal of the World Aquaculture Society 23 (1):105-111. Swingle, H.S.. 1966. Fish kills caused by phyt oplankton blooms and their prevention. FAO World Symposium on Warm Water Fish Culture. Rome, Italy. 11 pp. Timmons, M.B., J.M. Ebeling and F.W. Wheaton. 2002. Recirculating aquaculture systems, 2nd Edition. USDA Northern Regional Aquaculture Cener, Pub. No. 01-002. 769 pp. Toll, R.B., R.S. Prezant and H.B. Rollins. 2003. A novel method for locating tagged infaunal bivalves: submersible pulse technology metal detectors. Journal of Shellfish Research 22(2):501-503. Van Rijn, J.. 1996. The potential for integrati ng biological treatment sy stems in recirculating fish culturea review. Aquaculture 139:181-201.

PAGE 162

162 Vantaram, A.. 2004. Honeycomb fiber-reinforced polymer sandwich components for development of aquaculture ra ceway systems. MS thesis. West Virginia University. 107 pp. Villadolid, D.V., and F.G. Del Rosario. 1930. Some studies on the biology of tulla (Corbicula manillensis Philippi) a common food clam of Lagunade Bay and its tributaries. Phillipine Agriculturalist 19:355-382. Wargo, R.N. and S.E. Ford. 1993. The effect of shell infestation by Polydora sp. and infection by Haplosporidium nelsoni (MSX) on the tissue condition of oysters, Crassostrea virginica. Estuaries 16(2):229-234. Wetzel, R.G.. 2001. Limnology, lake and river ecosystems, 3rd Edition. Academic Press, San Diego. 1006 pp. Wilkie, A.C. and W.W. Mulbry. 2002. Rec overy of dairy manure nutrients by benthic freshwater algae. Bioresource Technology 84:81-91. Wilkie, A.C. 2003. Anaerobic digestion of flushed dairy manure. ProceedingsAnaerobic Digester Technology Applic ations in Animal Agricu ltureA National Summit. Water Environment Federation, Alex andria, VA. pp. 350-354. Wilkie, A.C., H.F. Castro and K.R. Cubinski. 2004. Fixed-film anaerobic digestion of flushed dairy manure after primary treatment: wast ewater production and characterization. Biosystems Engineering 89(4):457-471. Williams, C.J. and R.F. McMahon. 1986. Power station entrainment of Corbicula fluminea in relation to population dynamics, reproductive cycle and biotic and abiotic variables. Proceedings of the Second International Corbicula Symposium. The American Malacological Bulletin. Speci al Edition No. 2:99-111. Wood, J., G. Fernandez and A. Baker. 2007. Efficiency of reed beds in treating dairy wastewater. Biosystems Engineering 98:455-469. Zhou, Y., H. Yang and T. Zhang. 2006. Density -dependent effects on seston dynamics and rates of filtration and biodeposition of the suspension-cultured scallop Chlamys farreri in a eutrophic by (northern China): An experi mental study in semi-in situ flow-through systems. Journal of Marine Systems 59:143-158.

PAGE 163

163 BIOGRAPHICAL SKETCH Lance W Riley was born at the US Navy Submarine base in Guam, Mariannas Islands in 1974 to Capt. (USN) Roy Luke Riley and his love ly wife, Linda Cline Riley. The family eventually moved to Gold Hill, North Carolina, and Lance graduated from Mount Pleasant High School, Mount Pleasant, NC in 1992. Lance re ceived his Bachelor of Science degree in Environmental Biology at the Univer sity of North Carolina at Charlo tte in 1998. He then earned his Master of Science degree in Environmental E ngineering Sciences with a Graduate Certificate in Wetlands at the University of Florida in 2000. Lance is currently empl oyed at the University of Florida Fisheries and Aquatic Sciences Department where he performs analytical and field research using laboratory experiments and in-situ monitoring of waterw ays throughout Florida.