<%BANNER%>

Biogeochemical Transformations of Phosphorus in Wetland Soils

Permanent Link: http://ufdc.ufl.edu/UFE0019610/00001

Material Information

Title: Biogeochemical Transformations of Phosphorus in Wetland Soils
Physical Description: 1 online resource (137 p.)
Language: english
Creator: Fisher, Millard M, III
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: burial, everglades, peat, phosphorus, storage, usjrb, wetlands
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Phosphorus (P) is an essential element required by all organisms and is a principal nutrient influencing agricultural productivity worldwide. Surface runoff and subsurface leaching of P to surface waters has led to nutrient enrichment of aquatic ecosystems. Nutrient enrichment is the third largest cause of impairment of surface waters in the U.S. Excess nutrients, particularly P, have been linked to changes in flora and fauna, such as encroachment of invasive species and nuisance algal blooms. Eutrophication of historically low-nutrient lakes and wetlands alters trophic structure, favoring species that are more adapted to higher nutrient conditions. Determining pre-disturbance nutrient loading regimes is a major challenge for aquatic ecosystem researchers. Wetland soils and lake sediments that accrete materials preserve a partial record of ecosystem history that can be used to determine some pre-impact ecosystem characteristics. However, biogeochemical processes continually alter soil properties. Describing the extent of these changes was the central goal of this dissertation. Results from both P and carbon (C) characterization suggest movement of P and C into more recalcitrant organic compounds with time. I examined changes in soil phosphorus pools over a 5000-year period in two subtropical wetlands using chemical fractionation. Results indicate an increase in the relative proportion of recalcitrant P over time, with minimal change in humic- and fulvic-P fractions. Long-term wetland P accretion was estimated to be approximately 2 mg m2 yr-1, a rate much lower than has been reported for wetland soils. Two thermal methods were developed to better resolve the extent of P recalcitrance. Autoclave-extractable P declined throughout the soil profile from 40% of total P in surface soils to 10% in soils 1 m deep. Soils thermally extracted under N2 atmosphere showed both increasing extractable P with increasing thermal energy, and declining P recovery with depth for any given temperature. Both techniques suggest that susceptibility to thermal degradation may be a useful method to characterize organic P recalcitrance. Spectrophotometric investigation of organic matter extracts showed little change in average molecular weight with soil depth, though fluorescence and 13C-NMR analyses indicate increasing humic and phenolic character with increasing soil depth. The lignin content of soil fiber separated from peat increased from 35% in surface soil, to 75% at a depth of 1 m. Isotopic characterization of soil fiber seemed to indicate rapid ecological changes in the marsh that occurred at approximately 1000AD.
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 Millard M Fisher.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Reddy, Konda R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0019610:00001

Permanent Link: http://ufdc.ufl.edu/UFE0019610/00001

Material Information

Title: Biogeochemical Transformations of Phosphorus in Wetland Soils
Physical Description: 1 online resource (137 p.)
Language: english
Creator: Fisher, Millard M, III
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: burial, everglades, peat, phosphorus, storage, usjrb, wetlands
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Phosphorus (P) is an essential element required by all organisms and is a principal nutrient influencing agricultural productivity worldwide. Surface runoff and subsurface leaching of P to surface waters has led to nutrient enrichment of aquatic ecosystems. Nutrient enrichment is the third largest cause of impairment of surface waters in the U.S. Excess nutrients, particularly P, have been linked to changes in flora and fauna, such as encroachment of invasive species and nuisance algal blooms. Eutrophication of historically low-nutrient lakes and wetlands alters trophic structure, favoring species that are more adapted to higher nutrient conditions. Determining pre-disturbance nutrient loading regimes is a major challenge for aquatic ecosystem researchers. Wetland soils and lake sediments that accrete materials preserve a partial record of ecosystem history that can be used to determine some pre-impact ecosystem characteristics. However, biogeochemical processes continually alter soil properties. Describing the extent of these changes was the central goal of this dissertation. Results from both P and carbon (C) characterization suggest movement of P and C into more recalcitrant organic compounds with time. I examined changes in soil phosphorus pools over a 5000-year period in two subtropical wetlands using chemical fractionation. Results indicate an increase in the relative proportion of recalcitrant P over time, with minimal change in humic- and fulvic-P fractions. Long-term wetland P accretion was estimated to be approximately 2 mg m2 yr-1, a rate much lower than has been reported for wetland soils. Two thermal methods were developed to better resolve the extent of P recalcitrance. Autoclave-extractable P declined throughout the soil profile from 40% of total P in surface soils to 10% in soils 1 m deep. Soils thermally extracted under N2 atmosphere showed both increasing extractable P with increasing thermal energy, and declining P recovery with depth for any given temperature. Both techniques suggest that susceptibility to thermal degradation may be a useful method to characterize organic P recalcitrance. Spectrophotometric investigation of organic matter extracts showed little change in average molecular weight with soil depth, though fluorescence and 13C-NMR analyses indicate increasing humic and phenolic character with increasing soil depth. The lignin content of soil fiber separated from peat increased from 35% in surface soil, to 75% at a depth of 1 m. Isotopic characterization of soil fiber seemed to indicate rapid ecological changes in the marsh that occurred at approximately 1000AD.
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 Millard M Fisher.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Reddy, Konda R.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0019610:00001


This item has the following downloads:


Full Text





BIOGEOCHEMICAL TRANSFORMATIONS OF PHOSPHORUS INT WETLAND SOILS


By

MILLARD M. FISHER, III


















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

2007





























O 2007 Millard M. Fisher, III


























For Santa A. Fisher









ACKNOWLEDGMENTS

I am indebted to many who made this dissertation possible. Without the support and

encouragement of Dr. Lawrence Keenan, this dissertation would not have been possible. I am

very appreciative of the guidance and patience of my supervisory committee; Dr. Mark Brenner,

Dr. Bill DeBusk, Dr. Ping Hsieh, Dr. Jim Jawitz, Dr. Lawrence Keenan and Dr. Ramesh Reddy.

I owe a great debt to Mrs. Yu Wang, who tirelessly assisted me in many phases of this study. Dr.

Ben Turner was essential for assisting with both carbon and phosphorus nuclear magnetic

resonance analysis and interpretation of results. Special thanks go to Dr. Sue Newman for

arranging transportation into Water Conservation Area 2A. The research described in this

dissertation was funded by the St. Johns River Water Management District. Lastly, most of my

professional development was under the mentorship of Dr. Ramesh Reddy. I can never repay

him for the guidance, patience, and trust he has shown me during this dissertation and over the

last twenty-five years.












TABLE OF CONTENTS


page

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


LIST OF TABLES ................ ...............7............ ....


LIST OF FIGURES .............. ...............8.....


AB S TRAC T ............._. .......... ..............._ 10...


CHAPTER


1 INTRODUCTION ................. ...............12.......... ......


Background ................... ... ......... ......... .............1
Phosphorus in Natural Ecosystems ................. ...............12...
Estimating Historical Ecosystem Loading Rates .............. ...............13....
Characterizing Changes in Soil Phosphorus................ ..............1
Overview of the Problem ................. ...............18................

Hypotheses and Obj ectives ................. ...............20................
Dissertation Format .............. ...............20....


2 LONG-TERM CHANGES IN SOIL PHOSPHORUS FORMS IN TWO
SUBTROPICAL WETLANDS .............. ...............25....


Introducti on ................. ...............25.................
M ethods .............. ...............28....
Site Description .............. ...............28....
Sampling and Analysi s.................. .. .......... ...............29......
Sequential Organic Phosphorus Fractionation .............. ...............30....
Results and Discussion .............. ...............32....
Soil Properties ................ ...............32...
Soil Fractionation Results........................... ...............3

Long-Term Declines of Total Phosphorus in BCMCA............... ...............38.
Mathematical Model of Changes in Phosphorus Fractions ................. ......._.._........40
Historical Accumulation of Phosphorus ....._.._._ ........___ .....__ ...........4
Conclusions............... ..............4


3 ESTIMATING STABILITY OF ORGANIC PHOSPHORUS IN WETLAND SOILS........64


Introducti on ........._....._ ...._. __ ...............64....
M ethods .............. ...............68....
Site Description .............. ...............68....
Sampling and Analysis ........._.._ ..... ._._ ...............69....
Results and Discussion .............. .. ...............72...
Autoclave Extractable Phosphorus ................. ....__. ....._. ............7












Pyrolytic Extractable Phosphorus................ ..............7
Enzymatic Hydrolysis .............. ...............75....
Conclusions............... ..............7


4 CHARACTERIZATION OF LONG-TERM CHANGES IN ORGANIC
PHOSPHORUS IN A SUBTROPICAL WETLAND ................ ............... ......... ...84

Introducti on ................. ...............84.................
Methods .............. ... ...............87.
Organic Extract Fluorescence............... ..............8
Spectrophotometric Absorbance .................... ...............89.
13C Nuclear Magnetic Resonance (13C-NMR) .............. ...............89....
31P Nuclear Magnetic Resonance (31P-NMR) .............. ...............90....
Isotopic Analysi s of Plant Fibers ................. ...............91._.__ ....
S oil A ge ................ ...............92....... ......
Results and Discussion ..................... .. .................9
Fluorescence and Spectrophotometric Properties .............. ...............92....
13C Nuclear Magnetic Resonance (13C-NMR) .............. ...............95....
31P Nuclear Magnetic Resonance (31P-NMR) .............. ...............98....
Isotopic Analysis of Plant Fiber ................. ...............99...........
Conclusions............... ..............10


5 CONCEPTUAL MODEL OF PHOSPHORUS BURIAL IN WETLANDS,
CONCLUSIONS, AND FUTURE RESEARCH NEEDS ....___ .............. .... ........._..115


Conceptual Model of Soil Phosphorus Transformations. ....._____ .......___ ...............115
Conclusions............... ..............11
Future Research ................. ...............118......... ......


APPENDIX PHYSICAL SOIL DESCRIPTION ................ ...............123...............

LIST OF REFERENCE S ................. ...............130................


BIOGRAPHICAL SKETCH ................ ............. ............ 137...










LIST OF TABLES


Table page

2-1 Basic physico-chemical properties of surficial (0-50 cm) soils collected in July 2002
from WCA-2A and BCMCA. .............. ...............48....

2-2 Long-term depth, mass, and phosphorus accretion at two locations in Blue Cypress
Marsh Conservation Area. ............. ...............48.....

2-3 Model parameters for long-term soil accretion and phosphorus accretion at two
stations in Blue Cypress Marsh Conservation Area. ............. ...............48.....

2-4 Results of phosphorus fraction model calibration and validation at two stations in
Blue Cypress Marsh Conservation Area. ................. ....._._ ...._._ ...........4

2-5 Rate constants, equations and initial values used in StellaTM model of long-term P
declines at nutrient unimpacted station in Blue Cypress Marsh Conservation Area.........49

3-1 Basic physico-chemical properties of surficial (0-50 cm) soils collected in July 2002
from WCA-2A and BCMCA. .............. ...............78....

3-2 Regression statistics for autoclave extractable P verses depth in soil samples from
Blue Cypress Marsh Conservation Area and Water Conservation Area 2A .....................78

3-3 ANOVA General Linear Model results of pyrolysis temperature and soil depth
effects. .............. ...............78....

4-1 Results of 14C AMS dating .............. ...............106....

5-1 Components of conceptual model depicting phosphorus pools and transfers in peat
soil s. ............. ...............120....

A-1 Physical description of soil samples collected in July, 2002 from Water Conservation
Area 2A and Blue Cypress Marsh Conservation Area. ............. .....................2

A-2 Physical description of soil samples collected in March 2004 from Blue Cypress
Marsh Conservation Area. ............ .............129......










LIST OF FIGURES


Figure page

1-1 Conceptual diagram of phosphorus cycle in wetland soils (adapted from Stewart and
Tiessen 1987). .............. ...............22....

1-2 Study location in Water Conservation Area 2A, Florida, USA. .............. ...................23

1-3 Study location in Blue Cypress Marsh Conservation Area, Florida, USA. .......................24

2-1 Sequential fractionation technique used in first extraction, adapted from Ivanoff et al.
1998............... ...............50..

2-2 Soil bulk density (g cm-3) at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A..............................51

2-3 Soil loss on ignition (%) at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A..............................52

2-4 Soil total phosphorus at nutrient impacted and unimpacted stations in Blue Cypress
Marsh Conservation Area and Water Conservation Area 2A............... ...................5

2-5 Soil carbon to nitrogen ratio at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A..............................54

2-6 Soil carbon to phosphorus ratio at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A..............................55

2-7 Ratio of inorganic phosphorus to organic phosphorus at nutrient impacted and
unimpacted stations in Blue Cypress Marsh Conservation Area and Water
Conservation Area 2A ................. ...............56........... ....

2-8 Results of sequential chemical fractionation. Values are % of total P extracted. ..............57

2-9 Ratio of humic to fulvic acid in sequentially extracted soils..........._. ... ......_. ........58

2-10 Rate of increase in the % residual (unextractable) phosphorus fraction with respect to
time in Blue Cypress Marsh Conservation Area............... ...............59..

2-11 Relationship between soil depth and age in Blue Cypress Marsh Conservation Area. ....60

2-12 Relationship between soil age and total phosphorus concentration in Blue Cypress
Marsh Conservation Area .............. ...............61....

2-13 Dynamic model of storage and transfers from maj or phosphorus fractions. .................. .62

2-14 Results of model calibration and validation. ........ ......... ................ ...............63










3-1 Hot water extractable P (HEP) of Blue Cypress March Conservation Area and Water
Conservation Area 2A soils .............. ...............79....

3-2 Hot water extractable phosphorus (HEP) of Blue Cypress March Conservation Area
and Water Conservation Area 2A vs. water soluble (WSP) and microbial biomass
phosphorus (M BP). .............. ...............80....

3-3 Hot water extractable P (HEP) of two model compounds, glycerophosphate (GP) and
phytic acid (PA) .............. ...............8 1....

3-4 Percent of total phosphorus extracted from dried, ground peat using pyrolytic
extraction under nitrogen atmosphere. ....___.................. ........_.. ........ 8

3-5 Thermal extraction of organic P .............. ...............83....

4-1 The humification index (HIX) of NaOH-extracted peat samples ................. .................107

4-2 Ratio of absorbance at 465 and 665 nm (E4:E6) of NaOH peat extracts at stations Cl
and B4 in BCMCA............... ...............108.

4-3 Results of solid-state 13C-NMR spectroscopy of dried, ground peat samples collected
from station B4 on February 11, 2003 ................ ...............109........... .

4-4 Results of solution 31P-NMR spectroscopy ofNaOH-extracted samples collected
from station B4 (unimpacted) on February 11, 2003 ................. ......... ................110O

4-5 The 31P-NMR spectra of 0.5M NaOH extracts of soil samples from Blue Cypress
Marsh Conservation Area, station B4 (unimpacted) ................. .......... ................1 11

4-6 Results of lignin fractionation of plant fiber. Fiber was separated from bulk peat
samples that were collected from BCMCA stations B4 and C1 ................ ................. 112

4-7 Stable isotopic composition of plant fiber that was separated from bulk peat samples
collected from BCMCA stations B4 (unimpacted) and Cl (impacted). ................... .......113

4-8 Uncalibrated 14C age of plant macrofossils from selected depths at stations B4 and
Cl in BCM CA ................ ...............114...............

5-1 Conceptual model of dynamic exchanges among soil phosphorus pools in wetland
soil s. ............. ...............12 1....

5-2 Results of mathematical modeling of long-term declines of five phosphorus fractions
in Blue Cypress March Conservation Area. ............. ...............122....









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


BIOGEOCHEMICAL TRANSFORMATIONS OF PHOSPHORUS INT WETLAND SOILS

By

Millard M. Fisher, III

December 2007

Chair: K. Ramesh Reddy
Major: Soil and Water Science

Phosphorus (P) is an essential element required by all organisms and is a principal nutrient

influencing agricultural productivity worldwide. Surface runoff and subsurface leaching of P to

surface waters has led to nutrient enrichment of aquatic ecosystems. Nutrient enrichment is the

third largest cause of impairment of surface waters in the U.S. Excess nutrients, particularly P,

have been linked to changes in flora and fauna, such as encroachment of invasive species and

nuisance algal blooms. Eutrophication of historically low-nutrient lakes and wetlands alters

trophic structure, favoring species that are more adapted to higher nutrient conditions.

Determining pre-disturbance nutrient loading regimes is a maj or challenge for aquatic

ecosystem researchers. Wetland soils and lake sediments that accrete materials preserve a partial

record of ecosystem history that can be used to determine some pre-impact ecosystem

characteristics. However, biogeochemical processes continually alter soil properties. Describing

the extent of these changes was the central goal of this dissertation. Results from both P and

carbon (C) characterization suggest movement of P and C into more recalcitrant organic

compounds with time.

I examined changes in soil phosphorus pools over a 5000-year period in two subtropical

wetlands using chemical fractionation. Results indicate an increase in the relative proportion of









recalcitrant P over time, with minimal change in humic- and fulvic-P fractions. Long-term

wetland P accretion was estimated to be approximately 2 mg m2 -1l, a rate much lower than has

been reported for wetland soils. Two thermal methods were developed to better resolve the

extent of P recalcitrance. Autoclave-extractable P declined throughout the soil profile from 40%

of total P in surface soils to 10% in soils 1 m deep. Soils thermally extracted under N2

atmosphere showed both increasing extractable P with increasing thermal energy, and declining

P recovery with depth for any given temperature. Both techniques suggest that susceptibility to

thermal degradation may be a useful method to characterize organic P recalcitrance.

Spectrophotometric investigation of organic matter extracts showed little change in

average molecular weight with soil depth, though fluorescence and 13C-NMR analyses indicate

increasing humic and phenolic character with increasing soil depth. The lignin content of soil

fiber separated from peat increased from 35% in surface soil, to 75% at a depth of 1 m. Isotopic

characterization of soil fiber seemed to indicate rapid ecological changes in the marsh that

occurred at approximately 1000AD.









CHAPTER 1
INTTRODUCTION

Background

Phosphorus in Natural Ecosystems

Phosphorus (P) is one the principal macronutrients influencing agricultural production

worldwide. It is an essential element required by all organisms and is the 12th most abundant

element in the lithosphere. It constitutes 0. 1% of the earth' s crust (Reddy and DeLaune 2007) as

well as 0. 1% of plant and animal tissue, and tends to maintain a proportion to other

macronutrients of approximately 106: 16: 16:1, C: N: Si: P in biomass (Redfield 1963).

Phosphorus export from anthropogenic sources is the third largest cause of impairment of surface

waters in the U.S. (USEPA 1998), and in Florida approximately 35% of large lakes do not meet

their designated use due to excess nutrients (FDEP 2004). Excess nutrients, particularly P, have

been linked to changes in flora and fauna, such as encroachment of invasive species and nuisance

algal blooms. A major cause of nutrient impacts to surface waters is increased fertilizer use in the

United States (U. S.) (Sharpley 2000). Application of phosphorus fertilizers increased in the U.S.

from 2.6 million tons in 1960, to a maximum of 5.6 million tons in 1977, and has declined since

to 4.6 million tons in 2005 (Fertilizer Institute 2007). Advanced treatment of point sources of P,

as well as removal of P from many detergents has increased the relative contribution of P from

agriculture. The result is that agriculture now constitutes the leading source of P exports to

surface waters (USEPA 1996). Additionally, long-term applications of P in excess of crop and

livestock removal have led to build up in soil P storage as well as regional imbalances of P in

surface soils, particularly in developed countries (Sharpley 2000).

Lakes and wetlands often occupy the lowest position on the landscape. They are thus

closely linked with activities on adj acent uplands, and tend to accumulate nutrients and other









materials associated with surface runoff. Eutrophication of historically low-nutrient lakes and

wetlands alters trophic structure, favoring species that are better adapted to higher nutrient

conditions. In particular, phosphorus (P) has often been shown to be the principal nutrient

influencing ecosystem composition (Steward and Ornes 1975, Davis 1991, Urban et al. 1993). It

is thus critical to the preservation of wetland trophic structure to both estimate and maintain the

nutrient loading regimes that favored the establishment of the region's historic community.

Estimating Historical Ecosystem Loading Rates

The determination of pre-disturbance nutrient loading regimes is a maj or challenge for aquatic

ecosystem managers and researchers. Water and soil monitoring programs for most lakes and

wetlands in the U.S. were initiated only within the last four to five decades; perhaps none predate

large-scale agricultural and urban development. Wetland soils and lake sediments that accrete

materials preserve a partial record of ecosystem history. Proxy variables that have been used to

indirectly estimate redevelopment conditions include investigations of pollen, diatoms, algal

pigments, zooplankton exoskeletons, isotopes, and soil nutrient content. Estimates of pre-

developmental water quality in lakes have been determined through analysis of invertebrate and

diatom remains (Crisman 1978, Dixit et al. 1992, Shumate et al. 2002, Whitmore 1989). These

fossil assemblages can be well-preserved in sediment and reflect water conditions at the time

they were deposited.

Changes in sediment chemistry have also been used in lakes to infer previous water

column conditions, as well as historical nutrient loading rates (Brezonik and Engstrom 1998,

Engstrom et al. 2006, Brenner et al. 2001, Craft and Richardson 1998). Craft and Richardson

(1998) estimated an eight-fold increase in P loading to the Everglades, while Brezonik and

Engstrom (1998) estimated a four-fold increase in P loading to Lake Okeechobee, Florida based

on increases in down-core sediment total P. Brenner et al. (2001) estimated recent P









accumulation rates to be from two to seventeen times greater than accumulation rates in 1920 for

a marsh in east central Florida. Wetland nutrient assimilation models have been adapted to

calculate sustainable nutrient loading rates (Lowe and Keenan 1997). Typically, such models

rely on an empirically determined settling coefficient that accounts for loss to the sediments.

Estimation of this term is critical to nutrient loading applications, especially when used to

determine acceptable thresholds of nutrient loading to natural wetlands.

Temporal Changes in Soil Phosphorus.

One assumption of using temporal (calculated) changes in nutrient accumulation to infer

past water column conditions is that the estimated net accumulation rate for a particular time in

the past truly reflects the nutrient loading rate that prevailed at that time. However,

biogeochemical processes progressively alter both organic and inorganic P compounds, perhaps

confounding accurate estimates of past P accumulation. Transformations of soil inorganic and

organic P compounds are highly dynamic and are influenced by complex transfers among

biological and mineral pools (Stewart and Tiessen 1987). For example, once labile inorganic P is

depleted, the microbial and plant communities are capable of producing extracellular enzymes

that are responsible for hydrolysis of mono- and diester phosphorus in organic matter (Freeman

et al. 1996, Wright and Reddy 2001) thus mineralizing soil organic P. It is also suspected that

predominantly abiotic reactions alter soil organic matter (and thus organic P), resulting in

increasing humic character with time (Stevenson 1994) (Fig 1-1). Factors associated with the

formation of soils, such as climate, landscape position, time, biological processes, and parent

material all influence the rate and extent of these transformations (Jenny, 1943). It is therefore

likely that the floral and faunal precursors of the soil organic matter pool are transformed into

compounds that are progressively less similar to their parent components. In a P-limited









environment, such as many pristine aquatic ecosystems, it might also be expected that these

processes would alter materials at the expense of P, leading to a decline in soil P concentration

with greater depth, or time. It is difficult to estimate the extent of these alterations if total P is

measured.

Phosphorus fractionation schemes are capable of distinguishing between inorganic and

organic pools of P, and labile and recalcitrant P. Phosphorus fractionation has been used to

describe changes in P chemistry across horizontal surficial soil chronosequences. For example,

Walker and Syers (1976) chemically fractionated soil P across a landscape that spanned

approximately one hundred thousand years of soil pedogenesis. They demonstrated that P

fractions were not static, and that soil-forming factors continuously acted on mineral and organic

forms of P. They postulated that the distribution of soil P fractions might be a useful pedogenic

tool for dating soils. Their results were duplicated by Crews et al. (1995) for a soil

chronosequence spanning four million years in Hawaii. In both studies, apatite P in surface soil

horizons declined due to weathering, while organic P initially increased, then slowly declined in

older landscapes. Hedin et al. (2003) examined the same Hawaiian soils and determined that as

soils aged, export of soluble reactive P declined while the proportion exported as dissolved

organic P increased. This slow leakage of P was found to be sufficient to maintain older forests

in a state of constant P limitation. Thus for upland soils, soil P has been shown to undergo

continual change due to soil weathering processes. In organic wetland soils, redistribution of

both total and soil P fractions could also be expected to occur, though soil weathering processes

differ. In submerged soils, oxidative processes are much slower due to limited diffusion of

oxygen and anaerobic conditions (Reddy and Patrick 1975). Energy yield of alternate terminal

electron acceptors is low relative to oxygen, thus slowing the mineralization of organic matter









(McLatchey and Reddy 1998, DeBusk and Reddy 1998). This is the primary reason for accretion

of organic soils in peat soil wetlands. Walker and Syers (1976) compared phosphorus

fractionation results from surface soils taken across a landscape of progressively older soils to

estimate temporal changes in P pools. The same approach is useful for describing stratigraphic

changes in P fractions for progressively deeper peat samples.

Characterizing Changes in Soil Phosphorus

Characterization of the availability of soil and sediment P has been a principal goal of soil and

environmental scientists. Chemical extraction schemes are used to classify soil P into

operationally defined pools (Chang and Jackson 1957, Hedley et al. 1982, Hieljtes and Lijklema

1980, Ivanoff et al. 1998). Mild acids extract inorganic P that is associated with calcium,

magnesium, iron, and aluminum, while bases such as sodium hydroxide can be used to extract a

portion of organically-bound P. The un-extractable residue that remains after fractionation is

considered to be recalcitrant, though little is known about its chemical composition. The

resultant P fractions can be further subdivided into varying levels of labile and refractory P with

the goal being to estimate the short and long-term availability of P to biota. Organic P

fractionation derives in part from attempts to characterize soil organic matter. Late in the 18th

century, scientists extracted peat with alkaline extractants and obtained a dark fluid that formed a

precipitate upon acidification, which was termed humicc acid" (Stevenson, 1994). They observed

that more humic acid could be extracted from deeper, older peat. By 1900, it was widely

recognized that "humus" was a complex mixture of organic substances that were mostly

colloidal in nature and had weakly acidic properties. It was also recognized that soil organic

matter was comprised of over 40 compounds in the broad categories of organic acids,

hydrocarbons, fats, sterols, aldehydes, and carbohydrates. Waksman (1936) argued that the terms

humicc" and "fulvic" acids should be abandoned, because "these labels designate...certain









preparations which have been obtained by specific procedures," rather than identifying specific

compounds. Stevenson (1994) points out "reference to the acid-insoluble material as humic acid

is considerably less cumbersome than repeated reference to "alkali-soluble, acid-insoluble

fraction." Soil humus contains most, if not all, of the biochemical compounds synthesized by

living organisms and many additional compounds that are largely the product of abiotic

secondary synthesis reactions that occur in the soil environment. These compounds are thus

dissimilar to the biopolymers of microorganisms and higher plants.

Few procedures have been developed to evaluate compounds at the opposite end of the

availability continuum, i.e. P-containing compounds that have become resistant to

mineralization. Organic P in soils (especially peat) is present in chemical compounds of varying

environmental stability. Reddy et al. (1998) have shown that in Water Conservation Area 2A in

the Florida Everglades, refractory P increased from 33% of total P in surface soils, to 70% in

deeper strata. Stratigraphic trends in refractory P, similar to those observed in the northern

Everglades, were found in soils from marshes of the upper St. Johns River, Florida. Olila and

Reddy (1995) showed generally increasing residual (or recalcitrant) P content, increasing from

20% (of TP) in surface horizons to approximately 80% at a depth of 35 cm. DeBusk and Reddy

(1998) showed that the proportion of refractory carbon in soil profiles (from standing dead litter

to deeper soil strata) increases significantly with depth. In their study, lignin content increased

from approximately 12% of total dry weight for standing dead plant litter, to 50% for peat at a

depth of 30 cm. There are three principal mechanisms that could lead to enrichment in refractory

compounds in the soil profile. First, the plant community that forms the peat soil may have

changed to a community that is less enriched in refractory compounds. This could be expected if

a shrub community was succeeded by a slough community dominated by herbaceous plants.









Second, changes in external loading may have resulted in enrichment of labile compounds in

surficial peat. Lastly, over long periods of time, there may have been mobilization and loss of

labile fractions and enrichment of refractory compounds, such as lignin in deeper deposits. The

compounds associated with the more labile fractions are thus subj ect to mineralization,

mobilization, and subsequent loss from the soil profile. Pristine wetland regions that have not

been impacted by nutrient loading may therefore serve as experimental controls in that

chronological (depth) changes in P fractions would be expected to reflect primarily the effects of

organic matter decomposition.

Further insight into long-term changes in soil organic P may be gained by characterizing

both organic P and soil organic matter as a function of age or depth. Molecular level techniques

may provide insight into long-term alterations in soil Po and organic matter. Techniques such as

31P-Nuclear Magnetic Resonance (NMR) and 13C-NMR provide this information for broad

classes of both organic P and organic C compounds (Turner et al. 2003). Fluorescence and

absorbance properties of soil organic matter extracts have also been shown to be related to extent

of soil humification (Cox et al. 2000, Chen et al. 1977). Thermal techniques have been used to

characterize polyphosphates in microbial and algal biomass (Kenney et al. 2000) as well as the

extent of peat decomposition (Levesque and Dinel 1978). It is also possible that thermal stability

of organic P is related to its recalcitrance.

Overview of the Problem

Accelerated P loading to freshwater wetlands often causes changes in flora and fauna, and

these community changes are generally regarded as undesirable. All aquatic ecosystems receive

baseline P loads that vary depending on landscape position, region, and other factors.

Determination of this historical rate of P input is a maj or goal of lake and wetland scientists. One

approach to determining a sustainable P loading rate is to examine the soil or sediment record









and characterize vertical changes in soil P content, and in combination with the chrono-

stratigraphy, calculate historic P loading. A potential shortcoming of this approach is that

diagenetic changes may have altered soil P pools such that calculated P accumulation rates do

not accurately estimate the original nutrient loading rates. The main focus of this dissertation is

the characterization of those changes.

The experiments described in this dissertation were performed in two subtropical wetlands,

Blue Cypress Marsh Conservation Area (BCMCA), in east central Florida, and Water

Conservation Area 2A (WCA-2A), in the northern Florida Everglades (Fig. 1-2). Both wetlands

are underlain by peat soils, and both have regions that have been impacted by nutrient rich

agricultural discharges, yet still contain relatively pristine areas. This region has a subtropical

climate and receives approximately 1.5 m of rainfall annually. WCA-2A and other surrounding

WCAs are used for flood control, recreation, wildlife habitat, and water storage. The flora and

fauna of the Everglades are P-limited (Steward and Ornes 1975), though anthropogenic nutrient

loading from the Everglades Agricultural Area (EAA) has reduced some of this limitation,

particularly in regions adjacent to main canals (Reddy et al. 1998, DeBusk et al. 1994, Craft and

Richardson 1993). The peat soils are derived from sawgrass (Cladum jamnaicense Crantz)

(Davis 1943). They are approximately 2-m thick and the basal deposits were formed about 5,000

years before present (Gleason and Stone 1994). They thus present a vertical profile that grades

from recently deposited plant litter that can be presumed to be enriched in relatively labile

compounds, to deeper organic deposits that are recalcitrant after millennia of biogeochemical

weathering.

The BCMCA is a 1 16-km2 Shallow marsh in the headwaters of the St. Johns River (Fig. 1-

3). It is part of a joint St. Johns River Water Management District (SJRWMD) and United States










Army Corps of Engineers floodplain restoration proj ect whose goal is to provide both ecological

and flood protection benefits. BCMCA is also underlain by peat soils that vary from 1.5 m in the

eastern marsh to greater than 4 m near the center. The vegetation is a mosaic of open slough

communities (Nymphaea odorata, Eleocharis elongata, Utricularia spp.), Maidencane flats

(Panicum hemitomon), broad expanses of sawgrass, and tree islands (Acer rubrum, Taxodium

distichum, M~yrica cerifera). Agricultural discharges created a region of elevated soil P in the

northeastern side of the marsh, though those discharges ceased in August 1994.

Hypotheses and Objectives

The central hypothesis of this dissertation is that soil phosphorus undergoes

biogeochemical processing that favors transformation into stable, recalcitrant pools. For wetland

soils, the ultimate pool is hypothesized to be recalcitrant organic matter, with P as a part of the

organic matrix. The following obj ectives were developed to investigate this hypothesis:

* Characterize long-term changes in soil organic P using a fractionation scheme that was
developed for organic soils.

* Estimate the organic P is immobilization rate, and develop techniques to distinguish
extent of immobilization.

* Because organic P cycling and soil organic matter (SOM) are intimately linked,
characterize temporal changes in SOM pools.

Dissertation Format

I examined changes in phosphorus chemistry over a 5000-year soil record in two

subtropical wetland soils using sequential chemical extraction. Chapter one examines the impacts

of phosphorus on natural systems and methods used to characterize soil phosphorus. It also

describes methods of determining historical ecosystem nutrient loading records by examining the

soil stratigraphy. In chapter two, the sequential chemical fractionation of soils in cores from two

subtropical wetlands were compared using a method of organic P fractionation (Ivanoff et al.










1998). The fractionation data were evaluated in light of carbon-14 soil dates to determine rate of

P sequestration in wetland soil. A mathematical model of long-term phosphorus diagenesis is

proposed. Chapter three focuses on development of two new thermal P extraction techniques:

pyrolytic extraction of organic P and an autoclave extraction. An enzyme-based organic P

extraction was also investigated. Chapter four covers characterization of alkaline extracts of soil

organic matter (SOM), with the reasoning that insights into changes in this fraction will yield

similar insight into cycling of organic P. Results from both 13C-NMR and 31P-NMR are

described. Soil organic matter fluorescence and absorbance techniques were used to characterize

changes in SOM in two subtropical wetland soils. The carbon and nitrogen isotopic signature of

lignin separated from peat was analyzed to determine potential changes in plant communities.

Chapter five provides a synthesis of the results of preceding chapters and proposes a conceptual

model of organic P burial in wetland soils.













































Figure 1-1. Conceptual diagram of phosphorus cycle in wetland soils (adapted from Stewart and
Tiessen 1987).









80o50'O"W


80o40'O"W 80o30'O"W


80o20'O"W


26o40'O"N1 *26o40'O"N







Water
26o0'"N 26o30'O"N
Con se rvati on
Map Area
Area 1




Water
26o0'ON.Con se rvati on El 26o20'O"N
Area 2A

E5




26o10O'O"N.l I -26o10'O"N



Water
Con se rvati on
Area 3

26o0'O"N-( I 26o0'O"N








25o50'O"N-1 1 25o50'O"N




0 12.5 25 50 Km
SI I I I
80o50'O"W 80o40'O"W 80o30'O"W 80o20'O"W


Figure 1-2. Study location in Water Conservation Area 2A, Florida, USA.





















27o50'O"N-1 1 2I7o50'O"~

Map Area L













27o45'O"N-1 r 2 7o45'O"


BlueCyprss (Farm Discharge
Lake
L.





B4 C1



27o40'O"N-( 27o40'O"

Blue Cypress Marsh
Conservation Area





0 1.5 3 6 Km

I II

8(lo50'O"W 80o45'O"W 80o40'O"W




Figure 1-3. Study location in Blue Cypress Marsh Conservation Area, Florida, USA.









CHAPTER 2
LONG-TERM CHANGES IN SOIL PHOSPHORUS FORMS IN TWO SUBTROPICAL
WETLANDS

Introduction

Wetland soils and lake sediments tend to accumulate material, for two main reasons. First,

wetlands occupy a low position in the landscape, typically at the ecotone between terrestrial and

aquatic ecosystems (Mitsch and Gosselink 1993). Flow of water, suspended solids, nutrients and

other constituents is therefore generally towards them. This subsidy of water and nutrients leads

to high primary production, often in excess of 1000 g m-2 -1ai (Mitsch and Gosselink 1993).

Second, decomposition of organic matter is reduced due to slow diffusion of molecular oxygen

into water and saturated soil (Ponnamperuma 1972, Rowell 1981, Patrick and Reddy 1983,

DeBusk and Reddy 1998). The end result of high primary production and slow decomposition is

accumulation of organic matter, usually in the form of peat.

Nutrient loading in excess of that to which aquatic ecosystems have been conditioned

often results in undesirable changes in flora and fauna, such as algal blooms in lakes and

encroachment of pioneer or high nutrient adapted vegetation in wetlands. In the United States,

nutrient enrichment is one of the leading causes of impairment to wetlands (USEPA 2002).

Eutrophication of historically low-nutrient lakes and wetlands alters trophic structure, favoring

species that are more adapted to higher nutrient conditions. In particular, phosphorus (P) has

been shown to be the principal nutrient influencing wetland ecosystem composition (Steward and

Ornes 1975, Davis 1991, Urban et al. 1993). It is thus critical to the preservation of wetland

trophic structure to both estimate and maintain the nutrient loading regimes that favored the

establishment of the region' s historic community.

Unlike surface water samples, which reflect instantaneous conditions, soils and sediments

integrate conditions over long time scales (Frey 1969, Smol 1992). For example, estimates of










pre-developmental water quality in lakes have been determined through analysis of invertebrate

and diatom remains (Shumate et al. 2002, Whitmore 1989). These fossil assemblages can be well

preserved in sediment and reflect water conditions at the time they were deposited. Wetland soils

and lake sediments may also preserve a direct record of ecosystem nutrient loading that provides

information on predisturbance conditions. This record can then be used to help determine

acceptable lake or wetland nutrient loading rates (Brezonik and Engstrom 1998, Craft and

Richardson 1998, Brenner et al. 2001, Engstrom et al. 2006). This can be done by associating

dates with vertical changes in soil P content to derive P accumulation rates, then reasoning that

changes in P accumulation are mostly due to changes in ecosystem deposition, or loading rates.

The ages of specific soil horizons are typically obtained through radioisotopes such as cesium-

137, lead-210, or carbon-14. For example, Craft and Richardson (1998) estimated an eight-fold

increase in P accumulation since 1960 in nutrient impacted areas of the northern Everglades.

Brezonik and Engstrom (1998) estimated a four-fold increase in P accumulation since 1910 to

Lake Okeechobee, Florida based on sediment P content and overall sediment accumulation rate.

Brenner et al. (2001) estimated a 2.3 to 17 fold increase in P accumulation when comparing post-

1970 to pre-1920 rates for a marsh in east central Florida, and concluded that the increases were

due to increased external nutrient loading. However, dynamic biogeochemical forces act in

concert to alter the organic and inorganic materials deposited at the soil surface (Carignan and

Flett 1981, Walker and Syers 1976, Binford et al. 1983). For example, once labile inorganic P is

depleted, the microbial and plant communities are capable of producing extracellular enzymes

that can hydrolyze mono- and diester phosphorus in detrital material (Freeman et al. 1996,

Wright and Reddy 2001) thus mineralizing soil organic P. It is also suspected that predominantly

abiotic reactions alter soil organic matter (and thus organic P) resulting in increasing









humification with time (Stevenson 1994). It is therefore likely that the floral and faunal

precursors of the soil organic matter pool are transformed into compounds that are progressively

different from their precursor compounds. In a P-limited environment, such as many pristine

aquatic ecosystems, it might also be expected that these processes alter these materials at the

expense of P, causing a decline in soil P concentration with depth, or time. If inferences about

past nutrient loading rates are to be extrapolated from soil nutrient content, some knowledge of

the rate of P diagenesis is needed.

Sequential chemical fractionation techniques have been developed to distinguish between

forms of P in agricultural soils (Chang and Jackson 1957), for lake sediments (Hieltjes and

Liklema 1980) and organic wetland soils (Ivanoff et al. 1998). They are capable of distinguishing

between inorganic and organic pools of P, as well as extent of recalcitrance. Many researchers

have documented changes in soil P fractions along horizontal landscape chronosequences

(Walker and Syers 1976, Crews et al. 1995, Richter et al. 2006, Hedin et al. 2003). For example,

Walker and Syers (1976) sequentially fractionated soils along a horizontal chronosequence and

showed that long-term surface soil weathering caused a reduction in calcium-bound P and an

initial increase in organic P, followed by long-term declines.

The rate and extent of these changes, particularly with respect to soil organic P, have not

been well characterized. Deep peat soils provide an opportunity to investigate the nature of long-

term alterations in soil phosphorus fractions. These soils contain a record of changing

environmental conditions that in many cases spans thousands of years. Combining P

fractionation results with soil isotopic dating methods allows for estimation of rates of change in

maj or soil fractions, as well as for the calculation of historic soil P accretion rates. The two main









obj ectives of this study were to 1) characterize long-term changes in soil organic and inorganic

pools, and 2) determine rate of transfer into long-term recalcitrant P fractions.

Methods

Site Description

Field sampling was conducted in two subtropical wetlands, Water Conservation Area 2A

(WCA-2A) in the northern Florida Everglades and Blue Cypress Marsh Conservation Area

(BCMCA) in east central Florida, USA. This region has a subtropical climate and receives

approximately 1.5 m of rainfall annually. The WCA-2A is a 447-km2 wetland underlain by peat

soils. WCA-2A and other surrounding WCA' s are used for flood control, recreation, wildlife

habitat, and water storage. The flora and fauna of the Everglades are P-limited (Steward and

Ornes 1975), though anthropogenic nutrient loading from the Everglades Agricultural Area

(EAA) has reduced some of this limitation, particularly in regions adj acent to main canals

(Reddy et al. 1998, DeBusk et al. 1994, Craft and Richardson 1993). The peat soils are derived

from sawgrass (Cladium jamnaicense Crantz) (Davis 1943). They are approximately 2 m in

thickness and the basal deposits were formed approximately 5,000 years before present (Gleason

and Stone 1994). They thus present a vertical profile that grades from recently deposited plant

litter that can be presumed to be relatively labile, to deeper organic deposits that are recalcitrant

due to millennia of biogeochemical weathering. Two sampling locations were selected in WCA-

2A, long-term research stations El and E5 (Fig. 1-2). Station El is located in a nutrient-impacted

region, approximately 1.4-km south of South Florida Water Management District' s (SFWMD)

inflow structure S-10C. Station E5 is located approximately 10 km south of this structure, in the

unimpacted central marsh. Vegetation at station El was predominantly cattail (Typha

domingensis Pers.), and at station ES, sawgrass.









The BCMCA is a 1 16-km2 Shallow marsh in the headwaters of the St. Johns River (Fig. 1-

3). It is part of a joint St. Johns River Water Management District (SJRWMD) and United States

Army Corps of Engineers floodplain restoration proj ect whose goal is to provide both ecological

and flood protection benefits. BCMCA is also underlain by peat soils that vary from 1.5 m in the

eastern marsh to greater than 4 m near the center. The vegetation is a mosaic of open slough

communities (Nymphaea odorata, Eleocharis elongata, Utricularia spp.), Maidencane flats

(Panicum hemitomon), broad expanses of sawgrass, and tree islands (Acer rubrum, Taxodium

distichum, M~yrica cerifera). Agricultural discharges have created a region of elevated soil P in

the northeastern area of the marsh, though those discharges ceased in August 1994. One

sampling location was selected from this region of the marsh, station C1, and one station was

selected from the interior unimpacted region of the marsh, station B4.

Sampling and Analysis

Soil samples were collected in triplicate from both marshes on July 17 and 18, 2001. Soil

cores were obtained with a 2-m long x 7.3-cm diameter stainless steel soil corer. The extent to

which the soil was compacted during sampling was noted, and tended to be less than 15 cm for a

1.5 m core sample. Soil samples were extruded in the field into ZiploCTM balgs at 10 cm intervals.

The samples were immediately chilled to 4oC and were kept on ice for transport to the

laboratory. Laboratory processing was performed at the University of Florida' s Wetland

Biogeochemistry Laboratory. Wet weight was recorded for each sample and samples were

transferred to rigid polyethylene containers and thoroughly mixed by hand. A 10 20 g sub-

sample was removed for pH determination. For samples that were insufficiently moist to make a

pH determination, deionized water was added until the sample was sufficiently wet. A separate

50 100 g sub-sample was dried at 80oC to constant weight for percent moisture determination.










Sequential Organic Phosphorus Fractionation

The soil P extraction technique closely followed the technique of Ivanoff et al. (1998) (Fig.

2-1). The analysis was performed on a field-moist sample (0.5 gram dry weight equivalent), with

a soil to solution ratio of 1:50. The procedure was initiated by adding a field-moist sample to a

43 ml polyethylene centrifuge tube, then adding 2 ml of CHCl3 to the tube. The tubes were left

uncapped overnight and extracted the following day with 0.5 M NaHCO3. Twenty-five ml of

each extractant was added to the tube and it was end-to-end shaken on a reciprocating shaker for

16 hours. The samples were centrifuged at 6000 rpm, the supernatant extractant was removed

and vacuum filtered through 0.45-Clm polyethersulfone filters, and the tube was re-weighed.

Twenty-five milliliters of the next extractant was then added to the tube and the process was

repeated. This procedure was continued in such a manner that with each successive step, less P

remained in the soil sample. After the final step, only resistant P remained.

A separate, non-sequentially extracted wet sub-sample was extracted with 0.5 M NaHCO3,

but with no chloroform added to the sample. The difference between the P liberated in the

chloroformed and the P liberated in the non-chloroformed NaHCO3 extraction is believed to

represent the P content of the soil microbial biomass. The reported values were not adjusted to

account for incomplete extraction of microbial biomass with chloroform (Brookes et al. 1982). In

these soils, chloroform treatment may result in overestimation of biomass P due to extraction

from labile non-microbial organic P pools (Reddy et al. 1998).

The humic and fulvic acid content of the NaOH extract was operationally determined in

the following manner. After extracting with IM NaOH and centrifuging, 7 ml of the supernatant

fluid was withdrawn from each tube. Seven drops of concentrated H2SO4 were added to each

subsample to condense the acid-insoluble humic acid fraction. The sample was again centrifuged









at 6000 rpm for five minutes, and an aliquot of the supernatant was withdrawn. This aliquot was

analyzed for total P by adding 1 ml of 11N H2SO4 and 0.3 g K2S20s and digesting for 3 hours at

380oC. The P remaining in solution after the acidification step was considered to represent P

associated with fulvic acid. The total P content of an un-acidified NaOH sub sample was also

determined. Humic acid-bound P was determined as the difference between the TP content of the

un-acidified sample and the fulvic acid P. The total P content of the residual soil remaining after

the sequential extraction was determined by combustion of residue at 550oC, followed by

dissolution of ashes in boiling 6N HCL (Andersen 1976). All sequential extractions were

conducted at room temperature, and all extractions were completed within thirty days of sample

collection. Phosphorus was determined colorimetrically using the method of Murphy and Riley

(1962), and analysis was performed on a Technicon AutoAnalyzer (Tarrytown, NY) (U.S.

Environmental Protection Agency, 1983, Method 365.1).

Soil dating (BCMCA only) was determined through 14C analysis of plant macrofossils. Ten

centimeter sections of peat to a depth of approximately 1.5 m were carefully examined for plant

fragments such as stems, seeds, and other woody fragments. Care was taken to avoid selection of

root fragments. These plant macrofossils were sequentially extracted with 0.5M HCI (0.5 hours),

0. 1M NaOH (2 hours), and again with 0.5M HCI (0.5 hours). This removed carbonates and any

dissolved organic materials such as humic or fulvic acids that may have percolated down through

the soil profile. The samples were analyzed at University of California, Irvine Keck-CCAMS

laboratory by accelerator mass spectrometer (NEC 0.5MV 1.5 SDH-2, National Electrostatics

Corporation). Carbon dates were corrected to calendar dates using the computer program

CALIB, version 5.0.1(Stuiver and Reimer 1993).









Results and Discussion


Soil Properties

The soils of WCA-2A and BCMCA were similar with respect to most physical properties

(Table 2-1). The pH of soil from BCMCA was considerably more acidic at almost 2 units lower

than WCA-2A soils. Bulk density was similar at both locations, increasing from approximately

0.090 g (dry) cm-3 (wet) at the soil surface, to 0. 110 g (dry) cm-3 (wet) at a depth of 80 cm (Fig.

2-2). Below 100 cm, bulk density increases dramatically at the two WCA-2A stations due to

increasing sand content of the peat. Total nitrogen was slightly greater in BCMCA than WCA-

2A.

Total P at the nutrient impacted station in WCA-2A was greater than the nutrient impacted

site in BCMCA, though TP at the unimpacted site in WCA-2A was much lower than the

unimpacted site in BCMCA (Fig. 2-4). Total P was highest in surface soils (0-50 cm) of WCA-

2A station El (546 mg kg- ), and declined in the order E1>C1>B4>E5. As expected, the nutrient

impacted stations showed higher total P content in the 0 50 cm soil layer than the unimpacted

stations. However, the difference in total P between the impacted and unimpacted stations was

not nearly as dramatic at the BCMCA stations, 367 vs. 327 mg kg- This comparison is based on

a large depth interval, so any differences in P loading between the two stations would have been

obscured by inclusion of relatively deep soil in the sample. Also, total P at the unimpacted

station in BCMCA was 66% greater than the unimpacted Everglades station. This suggests that

BCMCA was not historically under the same degree of P limitation as the Everglades, and that

nutrient enrichment was more pronounced at the impacted WCA-2A station. Other studies have

shown that nutrient impacted surface soils of BCMCA appear to be higher in P than WCA-2A

(Grunwald et al. 2006), and that BCMCA is less P-limited (Prenger and Reddy 2004). This

difference may be principally due to differences in landscape position of the two wetlands than









to differences in anthropogenic nutrient loading. For example, the headwaters of the two river

systems that are connected to respective wetlands are only 50 km apart, separated by the narrow

Holopaw sand ridge. The difference is that BCMCA is located in the southernmost headwaters of

the St. Johns River, whereas WCA-2A is located at the downstream end of the Kissimmee River

- Lake Okeechobee Everglades ecosystem complex. Thus there are ample opportunities for P

to become sequestered by the lake-river-wetland complex upstream of WCA-2A, compared to

the more upstream position of BCMCA.

The carbon to nitrogen ratio (weight:weight) in BCMCA increased downcore, from 15:1 in

surface soils at both stations, to 19: 1 at a depth of 90 cm (Fig 2-5). For WCA-2A, the ratio was

relatively constant at the nutrient impacted station, 15.7:1 (fl SD). The ratio initially increased

at the unimpacted station, from 17.3:1 (f0.7 SD) at the surface, to 22.3:1 (f1.8 SD) at a depth of

50 cm. Thereafter, the ratio declined to 16.3:1 (f0.6 SD) near the base of the core. Carbon to P

ratio (C:P) increased consistently downcore at both BCMCA stations, from 600:1 to 900:1 in

surface soil, to approximately 6000:1 at a depth of 120 cm. Though the nutrient impacted

BCMCA core showed initially lower C:P, due to elevated surface soil P, C:P increased more

rapidly with increasing depth than at the interior unimpacted marsh station. One interpretation of

this is that the station that is now impacted by farm nutrient discharges may have historically

been a low nutrient region of the marsh. Both WCA-2A cores showed initially increasing C:P, to

a depth of 60 to 80 cm, then decreases to the base of the core. This decline is principally due to

increasing inorganic matter content at a depth of approximately 70 cm in WCA-2A, as reflected

in downcore increases in loss on ignition at that depth, particularly at the interior unimpacted

station (Fig. 2-3).










Organic P (TPo) constituted the maj ority of total P at both wetlands. The proportion of total

inorganic P (TPi) ranged from an average of 0.06 at the unimpacted station in BCMCA, to 0.43

at the nutrient impacted station in WCA-2A (Fig. 2-7). Surface soils at the unimpacted station in

BCMCA were enriched in TPi, which declined with depth to approximately 50-cm, with little

change thereafter. TPi:TPo declined slightly with greater depth in the soil profile at the impacted

BCMCA station. Greater TPi:TPo in WCA-2A soils reflects the more calcareous soils of WCA-

2A and the influence that calcium has on P biogeochemistry in these soils. Highest ratios were

observed in surface soils at the impacted WCA-2A station, with TPi approaching 50 percent of

TPo. The trend in TPi: TPo with respect to soil depth was different between the two WCA-2A

stations, with the ratio increasing with respect to depth at the unimpacted station and declining at

the impacted station. One possible explanation is that inflows to WCA-2A are high in both P and

calcium (Gleason 1974). However, these inflows have resulted in a north-south gradient in

surface soil P, but not in soil calcium (Reddy et al. 1998, Craft and Richardson 1993). It may

also be that interior water column concentration at the WCA-2A site is so low that adsorption of

P on precipitated calcite is inhibited.

Soil Fractionation Results

Labile phosphorus (NaHCO3 extractable Pi). Labile Pi accounted for <1 to 12% of soil P

(Fig. 2-8). Labile Pi was greatest in surface soils of BCMCA, accounting for as much as 12% of

TP at the unimpacted station, declining to approximately 4% at one meter. The reverse was true

of the impacted station in WCA-2A, with low values (3% of TP) in surface soils, increasing to

13% near the base of the peat deposit. At the impacted station in BCMCA and the unimpacted

station in WCA-2A there was no clear trend in this pool, with labile Pi accounting for <1 to 7%

of TP.









Labile organic phosphorus (NaHCO3 extractable Po). Labile Po was a minor fraction of

total P for both wetlands. Labile Po was frequently below the limit of detection, especially in

WCA-2A. This fraction was highest in surface soils for both wetlands, representing 2-3 % of TP

in BCMCA and 0-2% in WCA-2A.

Non-labile phosphorus (HCL extractable P). One molar HCI extracted inorganic P

bound to Ca and Mg compounds. This fraction is a relatively stable storage of soil P. Except for

the unimpacted BCMCA station, this fraction declined with depth. There were large differences

between BCMCA and WCA-2A. Non-labile Pi accounted for 2 50% of the TP pool in WCA-

2A, yet < 10 % in BCMCA. This reflects the influence that the calcareous geology of WCA-2A

has on P cycling, compared to the relatively softer water of BCMCA.

Microbial phosphorus. Microbial biomass P (MBP) typically declined slightly with

depth. On average, it represented 17% (17 SD) of TP. MBP in surface (0-10 cm) soil averaged

19 % (16 SD) of TP, and declined at most stations to an average of 13% (16 SD) at a depth of

approximately 1 m. Microbial biomass P increased slightly with depth at the impacted station in

WCA-2A, from 14% (10.6 SD) at the surface, to 17% (13 SD) at a depth of 0.8 m. The MBP

was significantly greater in surface (0-20 cm) soils at the two unimpacted stations (Students t; P

= 0.02).

Fulvic and humic P. Fulvic P accounted for 20% (15 SD) of TP. With the exception of the

impacted station in WCA-2A, fulvic P remained relatively constant throughout the soil profile.

At the impacted WCA-2A station fulvic P was slightly higher in surface soils (0-20 cm),

averaging 27% of TP (15 SD), declining to 15% of TP (14 SD) at a depth of 70 cm. Humic P

accounted for 15 % of TP (Al SD) and generally declined with greater depth. On average,

humic P represented from 6 % of TP (18 SD) at the impacted WCA-2A station, to 23 % of TP









(12 SD) at the unimpacted station in BCMCA. The ratio of humic to fulvic P increased from

approximately 0.5 in surface soils, to a maximum of 2.8 at the unimpacted station in BCMCA

(Fig. 2-9). The ratio was consistently greater at both unimpacted stations. The ratio began to

decline at a depth of 90 cm in BCMCA and 60 cm in WCA-2A. An increasing ratio of humic to

fulvic P through most of the soil profile probably reflects prevalence of more labile plant-derived

fulvic compounds in the upper regions of the soil profile, and increasing importance of more

recalcitrant humic compounds with time.

Humic P is determined by difference between the total P content of the NaOH extract and the

total P content of a subsample that has had humic materials precipitated from it. Hupfer et al.

(2004) has shown that for lake sediments, up to 49% of this fraction can consist of inorganic

polyphosphates. It is therefore possible that a significant proportion of this fraction is not due to

humic P, and thus humic P may be overestimated.

Residual phosphorus. Residual P, or unextractable P, showed the clearest change in the

soil profile. Residual P increased with respect to soil depth at all stations, representing an

average of 20 % of TP (15 SD) in surface soils, to 68 % of TP (18 SD) at the base of the peat

deposit. Both the proportion of P in this fraction and the rate of increase with depth were very

similar for the upper 60-cm of soil for both wetlands and for both impacted and unimpacted

stations. Below 60-cm, residual P increased more dramatically at the WCA-2A nutrient impacted

station (El) than the other stations, to a maximum of 88% of TP at a depth of 90 cm. Residual P

was the only fraction that showed consistent increases in pool size with increasing depth, or soil

age. It increased mostly linearly with respect to time (Figure 2-9). The proportion of P in this

fraction increased by 0.012% per year and represented approximately 27% of total P in surface

soils of BCMCA. This either reflects transfer from other fractions into this pool, or that this









fraction maintained a relatively constant soil concentration, while others declined. Results from

studies on the diagenesis of lake sediment provide evidence for the second mechanism. Penn et

al. (1995) fractionated sediments from Onondaga Lake, NY and found increasing refractory P

with depth. They chemically fractionated approximately 50 yr of deposition and assigned P to

two groups, a fast and a slow decay pool. At the sediment surface, there was a 1:1 ratio between

these pools, however, the proportion of P in the slow decay pool increased with depth. They

estimated decay constants of 0. 1 yr- and 4.8 yr- for the slow and fast pools, respectively, or a

50-fold decline in decay rate for recalcitrant P.

Total P declined dramatically with depth at all stations, while there was an increase in the

unextractable, or recalcitrant fraction. The microbial P pool generally declined with depth, as did

the fraction associated with humic materials. Because total P declined sharply with depth and

there were only modest changes in most other fractions, soil P pools were apparently maintain

the same downcore proportions as those in surface soils. The maintenance of relatively stable

proportions of P fractions indicates the stability of respective fractions i.e., a dynamic

equilibrium is maintained.

Increases in the residual fractions were mostly at the expense of humic and microbial P.

The principal difference in phosphorus fractions between the two wetlands was the proportion of

P in the HCl-extractable fraction. The HCl-extractable P in BCMCA was approximately 10%,

compared to 27% in WCA-2A. This is probably due to the fact that BCMCA is a slightly acidic,

soft-water wetland, underlain by sand, whereas WCA-2A is a slightly alkaline, hardwater

wetland, underlain by sand and marl, and therefore rich in calcium. That humic P declined with

increasing depth was not expected, since soil humification is thought to involve a progression of

soil organic matter to more phenolic, less bioavailable organic compounds. This suggests that









either the compounds that comprise soil humic matter are not closely associated with organic P,

or that the alkaline extraction was not completely effective at removing humic P from the

residual soil material. It is possible that accumulation of organic P in soils does not occur

through incorporation into humic- and fulvic-type compounds. Celi and Barberis (2005) describe

mechanisms that are largely abiotic and involve stabilization of organic P through sorption of

organic P compounds to iron, aluminum, and clay minerals. Though the proportion of P in the

fulvic pool did not decline with depth as did humic P, it remained as a mostly constant

percentage of total P.

Long-Term Declines of Total Phosphorus in BCMCA

The carbon-dated plant macrofossils in BCMCA allowed for the association of soil depth

and soil age (Fig. 2-11). A first-order model was fitted to the age and depth data (Equation 2-1).

Soil age (years before present)= Yo e (k*Z) (2-1)

where: Yo = soil age at time = 0, years before present,

k = first order decay constant, years- ,

Z = soil depth, cm

For the impacted and unimpacted stations, Yo equals 54 and 128 years, respectively. Rate

constants at both stations were 0.03 yr- The parameter Yo is the soil age at a depth of zero

centimeters. The older surface soil age for the nutrient impacted station is counterintuitive as

nutrient impacts tend promote rapid plant growth and therefore faster soil accretion rates. In

actuality, the nutrient impacted station likely does accrete material faster. However, the 14C

dating method is not particularly accurate for recent surficial material, as it is influenced by

atomic bomb testing, plant roots, and burrowing of invertebrates. Therefore, the surficial soil age

difference of 84 years between the two stations is probably not significant. Both stations appear

to accrete material over long time intervals at the same rate. The rate constant (0.03 yr- ) implies









that soil age increases at a rate of 3% for every unit increase depth. Equation 2-1 suggests that

soil age is exponentially proportional to soil depth, for example each unit increase of soil depth

at lower depths represents a greater change in soil date than surface horizons. This could be due

to either changing rates of accretion over time, or that subsurface horizons are both physically

compressed by the overburden, or that lower soil horizons have lost mass (and volume) over

time, or both. Since there were only modest increases in bulk density, changing accretion and/or

mass loss seem more likely (Fig. 2-2).

The relationship between soil age and depth can be used to express sample phosphorus

content as a function of age, rather than depth. A first order model can be fit to this data:

Total Pt = Pf + Po exp(~t (2-2)

where: Pc = long-term asymptotic soil P concentration, mg kg- ; Pc + Po = initial soil P

concentration at time = 0, mg kg- ; k = first order rate constant, year-

Fitting this model to the total P concentration at the two BCMCA stations yields a loss rate

constant of 0.006 yr- or a decline of 0.6% per year, and a soil minimum of 1 18 and 80 mg kg-l

for the unimpacted and impacted stations, respectively (Fig. 2-12; Table 2-4). Soils at both

WCA-2A and BCMCA asymptotically approached this value, and for both marshes, this lower

soil horizon dates to approximately 5,000 years before present (YBP). The exponential decline in

soil P could be due to a combination of factors, such as mobilization and loss of P from the soil

profile, or a gradual increase in P loading to the marsh. Given that the decline was very similar

for both impacted and unimpacted stations, the change is likely principally due to decomposition

of the original organic material, mineralization, and loss of P.

More support for this explanation can be found from lakes in catchments with very little

development. Lake Erken is a 24 km2 meSotrophic lake in east central Sweden that has









undergone very little change in the catchment for the past century (Rydin 2000). The drainage

basin consists of approximately 70% forest and 10% farmland. Similar to theWCA-2A and

BCMCA cores, Rydin (2000) found steep declines in sediment total P with respect to depth, and

attributed the difference in total P between surface sediment and deeper strata to P that had been

mobilized and lost during diagenesis. Other studies have demonstrated that this diffusive loss to

overlying water of mineralized P can be a maj or component of total ecosystem nutrient budgets

(Moore et al. 1998, Fisher et al. 2005). Carignan and Flett (1981) examined the influence of post-

depositional processes on porewater concentration gradients in lake sediments. They

homogenized lake sediments and placed them into aquaria, and within five weeks steep

concentration gradients had become established in their microcosms. They concluded that post-

depositional processes may obscure any relationship between P concentration in surficial

sediments and in the original sedimented material. Mineralization and flux of P from the soil to

water column may therefore account for much of the observed decline in P with depth in lake

sediments and wetland soils.

Mathematical Model of Changes in Phosphorus Fractions

Temporal changes in soil phosphorus fractions can be depicted as a system of storage and

transfers by coupling the data from the phosphorus sequential fractionation with soil age data.

However, whether changes occurred through advective losses from the wetland, or by transfers

into other P fractions cannot be ascertained from this research. In the model proposed here, all

changes (typically declines) in respective fractions result in direct loss from the system (Fig. 2-

13).

Each fraction was modeled as a first-order process, with concentration declines

proportional to the concentration at the beginning of the time step. The unimpacted station in

BCMCA was used to calibrate the model. Declines in soil P fractions were modeled as









P(t) = Pi exp(-k~t) + P* (2-3)

where Pi = P at t = 0 mg kg- k = first order rate constant, yrs- t = time, years, and P* = long-

term asymptotic P value. For model calibration and validation, the time verses depth model was

adjusted such that the surface (0-10 cm) sample represented time zero for both stations. This was

done by subtracting the model-determined age of the surface interval from each of the lower soil

intervals for respective sampling stations. For both BCMCA and WCA-2A, phosphorus

concentration asymptotically declined with age to a background concentration, then dropped

little thereafter. The P* model parameter prevents the P of each fraction from declining below

these minimum long-term levels. This model was fit to the P and time data for each fraction

using two objective functions: minimizing the sum of the differences between the model

prediction and the data (SSR), and by minimizing the root mean square error (RMSE). Both

obj ective functions were minimized using the Solver utility in Microsoft Excel. For most

fractions, the RMSE obj ective function resulted in the better correlation coefficient between

model and data (Table 2-4). Solving for model parameters by minimizing RMSE for humic P did

not result in convergence on a solution, therefore minimization of SSR was used to determine k

and Pi. Whether RSME or SSR was minimized made little difference in solving for k; the

average difference in k between the two methods was approximately 8%. Rate of decline in

respective fractions, as evidenced by rate constants was of the order NaHCO3 Pi > microbial

biaomass P > HCI Pi > fulvic P > residual P > NaHCO3 Po > Humic P. The apparently slow

decline of the NaHCO3 Po fraction is partially explained by the very low initial concentration of

this fraction, typically only 10 mg kg-l in surface soils. Also, model fit to this fraction was

relatively poor compared to other fractions, with the model explaining only 72% of variability in

the data at the unimpacted station. Humic P declined at a rate of 0.09% yr- and represented the









lowest loss rate. Humic P increased with increasing depth, relative to other fractions, at the

unimpacted station in BCMCA (Fig. 2-8). The lower loss rate of humic P, compared to other

fractions, led to enrichment in this fraction at this station, though this enrichment was less

apparent at the other three stations. The fit of the exponential model to the humic P data was the

least robust (r2 = 0.48) and this data is probably better described by a zero-order linear process.

Bicarbonate extractable inorganic Pi declined fastest, at a rate of 4.6% yr a rate over 50-fold

greater than humic P.

The model was validated using the nutrient impacted BCMCA station, with the Pi value of

that station substituted into the model (Fig. 2-14). Goodness of model fit to the actual data at the

impacted station was judged by comparing the RMSE of the station that the model was

calibrated to, against the RMSE of the other marsh station. This comparison is the RMSE ratio

given in Table 2-4. For example, the RMSE ratio of microbial biomass P is 2.2. This implies that

the error in prediction of long-term changes in microbial biomass P at the impacted station are

2.2 times greater than at the station at which the model was developed. This estimate of long-

term behavior is based only on the surficial (0-10 cm) concentration at the new, unmeasured

station, and the rate constant and long-term asymptotic concentration determined at the

calibration station. Model prediction was worst for humic P, with an RSME ratio of 45. The poor

prediction for this fraction is due to the combination of high initial value of humic P at the

unimpacted station and the very low rate constant. In essence, this ensures that the modeled

humic P always lags actual soil concentration, thus generating a high residual sum of squares for

the BCMCA validation station. This disparity indicates some sensitivity of the model to initial

conditions (Pi). It is also possible that, since the model was developed for a low nutrient region

of the marsh, it may not characterize P dynamics for a region that has been impacted by nutrient









runoff. Despite this potential shortcoming, the model seems to describe long-term declines in

most fractions reasonably well.

The rate constants (k) and long-term asymptotes (P ) were incorporated into a systems

dynamic model (StellaTM VeTSion 9, isee systems Inc., Lebanon NH). The model provides a

simultaneous description of decay kinetics for the seven P fractions (Fig. 5-3). The total P

mineralized is summed in a component termed "Mineralized". The loss from each pool reflects a

first order process in that the amount of P lost for each time step is proportional to the mass

contained in respective pools at the beginning of the time step. The transfer, or flow term for

each P fraction is simply:

P(ti+1)= (P(ti) P ) k (2-4)

Where P(ti+1) is the P content at the end of the time step, P(ti) is the P content at the beginning

of the time step, P* is the long-term asymptotic P content, and k is the first order rate constant.

The model allows for comparisons to be made of relative rates of change among P fractions (Fig.

5-4). The model components consist of a system of stocks and flows, and this system of

equations and initial values can be readily imported into other systems dynamics software (Table

2-5). To apply this model to a new site, the surficial soil total P concentration and the proportion

of P in each fraction would be specified as initial concentrations for respective fractions. For

subtropical peat wetlands, rate constants and long-term minimum concentrations are likely to be

similar to those found here, yet should be independently verified.

Historical Accumulation of Phosphorus

Soil age increased exponentially with depth. However, prior to 500 YBP, soil age (and

therefore accretion) increase was mostly linear (Fig. 2-1 1). If a linear model is assumed for ages

greater than 500 YBP, long-term soil accretion in BCMCA has been approximately 2 mm yr-









(Table 2-2). As noted previously, at both the Everglades and Blue Cypress Marsh (both

impacted and unimpacted stations) soil P levels declined asymptotically to approximately 100

mg kg- and declined very little thereafter. The nearly steady-state historical soil accumulation

rate and total P concentration can be combined to yield an estimate of long-term P accretion of

2.28 and 1.84 mg m-2 -1l for stations B4 (unimpacted) and Cl (impacted) in BCMCA (Table 2-

2). This value represents the mass transfer rate to very long-term P sinks. This estimate is very

low compared to previous accretion estimates for BCMCA and other wetlands. The calculation

of wetland P assimilation is highly influenced by the depth interval over which the accretion is

calculated, particularly during early diagenesis. For example, Brenner et al. (2001) estimated P

accretion in BCMCA prior to 1920 to be 20 mg m-2 day l, or 10-fold higher than this study. Craft

and Richardson (1998) estimated P accumulation rates in the upper 30-cm soil layer for a station

in WCA-2A to be 60 mg P m-2 -1l, Or 30-fold greater than these results. For this study, the

interval over which rate of accretion and P concentration were relatively constant was chosen.

This period spanned approximately 2,000 years and represented a depth interval of 80 150 cm

(Table 2-2). This long-term assimilation rate can be contrasted with estimates of overall wetland

nutrient loading rates. Richardson et al. (1997) examined inflow and outflow data from over 100

North American treatment wetlands, as well as a site in WCA-2A. Inflow loading rates ranged

from 0.01 to over 1000 g P m-2 -1l. Analysis of the inflow-outflow relationship suggested that

wetland loading rates in excess of 1 g P m-2 -1l appeared to overwhelm the capacity of the

wetland to move P into long-term sinks. For BCMCA, the historical rate of movement into long-

term pools is much slower than this and could not represent the principal mechanism for P

removal below this threshold loading rate. Richardson (1985) compared P accumulation in

freshwater wetlands and uplands and concluded that in general, wetlands have a relatively low










capacity to absorb P. He pointed out that even though peat wetlands store massive amounts of P,

they do so over very long time intervals. Penn et al. (1995) described the kinetic changes in P

fractions for the sediments of Onondaga Lake, New York. They suggested that for those organic-

rich sediments, below a depth of 30 cm, almost all of the P was highly resistant to further

decomposition due to losses of labile fractions. This fraction was constant with increasing depth

and therefore represents the long-term pool of accreting P.

Lowe and Keenan (1997) developed a conceptual model of nutrient assimilation in natural

wetlands that predicts areal extent of impact on wetlands due to anthropogenic nutrient loading.

The model can be thought of as a "thin leaky sponge", with the thickness of the sponge

proportional to short-term sinks such as plant biomass and other relatively labile pools. The

leakage rate represents movement to long-term pools, with the rate dependent on size of the

labile pool. The leakage rate is empirically determined from inflow-outflow concentration data in

treatment wetlands, though calculation using actual P burial rates such as found in this study may

provide better protection to natural wetlands.

Conclusions

The peat soils of BCMCA and WCA-2A were similar in terms of physical properties such

as texture, loss on ignition and total N and P content. The unimpacted station in BCMCA was

greater in total P than the unimpacted station in WCA-2A, perhaps reflecting differences in

landscape position, whereas the nutrient impacted station in WCA-2A was greater in surface

total P compared to the impacted station in BCMCA. Over very long time periods, soil total P in

both wetlands declined to approximately 100 mg kg l, and this represents an exponential annual

reduction for both wetlands of approximately 0.6% per year. At a depth of approximately 75 cm,

further reductions in soil P content are minimal. Soil P accretion at depths lower than this was










approximately 2 mg m-2 -1l for BCMCA soils. This is a very low accretion rate compared to

other estimates in BCMCA and other wetlands.

Soil fractionation results showed that WCA-2A soils are much higher in inorganic P,

probably due to the calcareous nature of these soils. The most dramatic change in P pools with

depth was the enrichment of the residual, recalcitrant pool. Residual P increased from 20% of

total P in surface soil, to 68% at the base of the peat deposit in both wetlands. Changes in P

associated with humic and fulvic compounds did not change in a consistent way among stations

and wetlands. In general, humic P declined with increasing soil depth, while fulvic P was mostly

constant.

My results cannot disprove conclusively the claim that chronological changes observed in

soil P profiles were caused by changing environmental conditions such as vegetation succession,

hydrology, and nutrient loading. However, two Eindings from this study suggest that observed

vertical changes in P are at least in part due to diagenesis. First, marsh stations far removed from

sources of anthropogenic nutrient show declines in soil P with depth, though it could be argued

that even locations remote from nutrient inflows are affected by increased 20th century P use.

Second, there is a constant downcore increase in residual P, as well as a widening in C to P ratios

at all stations. It seems unlikely that these soils were enriched in residual P at the time they were

accreted.

The operationally-defined extraction used in this study distinguished between Hyve organic

P pools: microbial, labile Po, fulvic, humic, and residual Po. The greatest changes in the

sequential fractionation were observed in the residual, unextractable fraction. This fraction is

believed to consist of refractory P compounds, and for peat soils, these are predominately in an

organic form. Little is known concerning the chemical composition of this fraction, though it is









assumed to be recalcitrant due to its resistance to solubilization with the chemicals used. These

five fractions constitute a categorical method of characterizing soil organic P stability. Non-

categorical (i.e., continuous) methods may be more useful at distinguishing the extent to which

organic P has been immobilized. More resolution of Po immobilization is needed to better predict

the nutrient release potential of wetland soils, when they are

* drained for agricultural or urban development;
* reflooded, as a consequence of restoration proj ects;
* constructed with the intention of nutrient removal;


In all of these cases, a common concern is how nutrient dynamics will change under the new

conditions. For organic soils, a principal determinant of this is the stability of Po.

Thermal techniques have been used to characterize the extent of decomposition of peat

and may also prove useful in characterizing the continuum of Po stability (Levesque and Dinel

1978). Also, enzymatic attack of wetland soil organic matter is the principal means by which Po

is broken down to more labile compounds. Thus, extraction of peat with phosphoesterases may

provide insights into the degradability of organic wetland soils. These techniques will be

investigated in Chapter 3.















Table 2-1. Basic physico-chemical properties of surficial (0-50 cm) soils
collected in July 2002 from WCA-2A and BCMCA.
BCMCA WCA-2A
Parameter B4 Cl El E5
TN, % 3.21 3.50 2.93 2.42
TC, % 47.4 48.3 45.1 45.7
Dry Weight, % 8.7 10.3 10.0 10.8
pH 5.70 5.88 7.59 7.30
Bulk Density, g cm-3 0.081 0.101 0.095 0.111
Loss on ignition, % 5.8 8.8 7.3 11.1
TP, mg kg-l 327 367 546 197


Table 2-2. Long-term depth, mass, and phosphorus accretion at two
locations in Blue Cypress Marsh Conservation Area.
Parameter B4 Cl Units
Depth interval 85-145 75-145 cm
Age interval 510-2900 965-3600 YBP
Accretion 0.019 0.025 cm yr-
Mean total P 118 80 mg kg-l
Mean bulk density 017.92 gc-3
P accumulation rate 2.28 1.84 mg m-2 -1l

Table 2-3. Model parameters for long-term soil accretion and
phosphorus accretion at two stations in Blue Cypress Marsh
Conservation Area.
Parameter B4 Cl Units
Soil accretion: Yo e (k*Z)
Accretion rate constant, k 0.029 0.026 yr-
Surficial age, Yo 54 128 YBP
Model fit, r2 0.96 0.79 units

Phosphorus accretion: C, + Co exp(-kt)
Ct=o 389 729 mg kg-l
Ct=, 118 80 mg kg-l
Rate constant -0.006 -0.006 yr-
Model fit, r2 0.99 0.99 units











Table 2-4. Results of phosphorus fraction model calibration and validation at two stations in
Blue Cypress Marsh Conservation Area. All values, except for RSME ratio, from
model calibration only.
Objective RSME
Fraction Function r2 Rate Constant Pi P* Ratio
yr' mg kg' mg kg
Microbial RMSE 0.9733 0.0235 128.7 10 2.2
SS 0.9729 0.0223 130.0
NaHCO3 Pi RMSE 0.9946 0.0462 67.8 3 0.8
SS 0.9940 0.0432 69.0
NaHCO3 Po RMSE 0.7155 0.0033 11.7 0 2.3
SS 0.7119 0.0039 10.8
Inorganic RMSE 0.9911 0.0219 46.6 2 3.5
SS 0.9890 0.0195 47.9
Fulvic RMSE 0.9516 0.0103 102.7 10 3.4
SS 0.9447 0.0079 106.8
Humic RMSE No Convergence
SS 0.4801 0.0009 38.6 10 45a
Residual RMSE 0.9683 0.0038 86.7 30 8.8
SS 0.9679 0.0040 85.0
Note: aRatio determined from SSR, not RSME.

Table 2-5. Rate constants, equations and initial values used in StellaTM model of long-term P
declines at nutrient unimpacted station in Blue Cypress Marsh Conservation Area.
Fraction Fraction (t) Initial Value Transfer Term
Fulvic-P Fulv(t dt)+(-LossFuly)*dt 0.226* TPinitial (Fulv-1 0)*0. 0103
Humic-P Hum(t dt)+(- LossHum)*dt 0.106* TPinitial (Hum-10)*0.0009
HCl-P Inorg(t dt)+(- LossInorg)*dt 0.087* TPinitial (Inorg-2)*0. 0219
NaHCO3-Pi LabPi(t dt)+(- LossLabPi)*dt 0.123* TPinitial (Lab Pi-3)*(0.0462)

NaHCO3-Po LabPo(t dt)+(-LossLabPo)*dt 0.008* TPinitial LabPo*0.0033
Microbial P MBP(t dt)+(- Loss MBP)*dt 0.248* TPinitial (MBP-10)*(0.023 5)
















Is7 .0.5M NaHCO3
A_ 05M NHCO316 hrs
+ CHCI,

B~Microbial P (TP A. TP B.)C Lail o,(PSP





1M HCI SoilI Extractant,
3 hrs
Extraction Time,
Non-labile Pi (S RP)
and Fraction Definition





0.5M NaOH
16 hrs
C. TP

Remove aliquot Fulvic P (TP)
Humic P (C. Fulvic P)
Acidify pH 2; centrifuge



Recalcitrant Po (TP)


Figure 2-1. Sequential fractionation technique used in first extraction, adapted from Ivanoff et al.
1998.




























BCMCA
Unimpacted


0.1 0.2


-20 -


-40 -

-60 -

-80 -

-100 -


-120 -

-140 -


0.1 0.2


-20
WCA-2A I I WCA-2A
Unimpacted Impacted
-40

-60


-80

-1 00


-1 20

-1 40
0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5

Bulk Density, g dry cm-3 Wet









Figure 2-2. Soil bulk density (g cm-3) at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A. Error bars
represent one standard deviation. Note different x-axis scale for WCA-2A cores.

















BCMCA
Unimpacted


BCMCA
Unimpacted


0-
-0


-20 -

-
-40 -

-10-


-640 -

0-





-80 -


-100 -


-120 -


-140 I ,
20 60 100 20 60 100

Loss on Ignition, %


Figure 2-3. Soil loss on ignition (%) at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A. Error bars
represent one standard deviation. Note different x-axis scale for WCA-2A cores

















01


-20


-40


-60


-80


-1 00

BCMCA I ~BCMCA
-10-Unimpacted I IImpacted

0 -140




-20


-40


-60


-80


-1 00

-10_ WCA-24 WCA-24
Unimpacted Impacted

-1 40
0 500 1000 1500 0 500 1000 1500

Total P, mg kg-l




Figure 2-4. Soil total phosphorus at nutrient impacted and unimpacted stations in Blue Cypress
Marsh Conservation Area and Water Conservation Area 2A. Error bars represent one
standard deviation.
































BCMCA
Unimpacted


-20 -


-40 -
-



-80 -


-100 -


-120 -


-140


-20 -


-40 -


-60


-80 -


-100 -


-120 -


-140 -


WCA-2A
Impacted


WCA-2A
Unimpacted


i I
10 20 30


10 20 30


Carbon : Nitrogen






Figure 2-5. Soil carbon to nitrogen ratio at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A. Error bars
represent one standard deviation.

















20-


-40 -

-60 -

-80 -

-100

-120 -


BCMCA
Uni mpacted


-140 I-


-40

-60

-80 -

-100

-120

-140


WCA-2At.
Uimpce


0 5000 10000 0 5000 10000

Carbon : Phosphorus





Figure 2-6. Soil carbon to phosphorus ratio at nutrient impacted and unimpacted stations in Blue
Cypress Marsh Conservation Area and Water Conservation Area 2A. Error bars
represent one standard deviation.












BCMCA Unimpacted


BCMCA Impacted


0


-20


-40




-80 -


-1 00


-120 I I I
WCA2A Unimpacted


WCA2A Impacted


-20


-40


-60


-80


-100


-120 1 ,
0.0 0.2 0.4 0.0 0.2 0.4

TPi :TPo




Figure 2-7. Ratio of inorganic phosphorus to organic phosphorus at nutrient impacted and
unimpacted stations in Blue Cypress Marsh Conservation Area and Water
Conservation Area 2A. Error bars represent one standard deviation.












BCMCA
Im pacted


BCMCA
Unimpacted


O Labile Pi
0 Non-labile Pi
Labile Po
0 MBP
FAP
0 HAP
0 Residual P


WCA2A
Im pacted


WCA2A
Unimpacted


80

90-

100

110

120 -


130


0 50 100 0 50 100 0


1


F


50 100 0


50 100


Figure 2-8. Results of sequential chemical fractionation. Values are % of total P extracted.





-160 1 I I
0 1 2 3 0 1 2 3

Humic: Fulvic



Figure 2-9. Ratio of humic to fulvic acid in sequentially extracted soils.


0


-20


-40


-60


-80


-100


-120


-140


BCM CA





SUnimpacted


Impacted






















Unimpacted


g*


20 -P


ResP = 29 + 0.012*t
S= 0.39
p = 0.038


10 1
0


1000


2000


3000 0


1000


2000


3000


Time, YBP


Time, YBP


Figure 2-10. Rate of increase in the % residual (unextractable) phosphorus fraction with respect
to time in Blue Cypress Marsh Conservation Area.
















6000


5000
r2 = 0.96
4000

3000~

2000 I

0_ 1000
F0
60000

5000- Age = 128*e( 03*depth) Impacted
r2 = 0.79
4000
3000

2000
1000


0 50 100 150
Depth, cm

Figure 2-11. Relationship between soil depth and age in Blue Cypress Marsh Conservation Area.
Error bars represent the analytical 14C COunting error of one standard deviation.
























































01
0 200 400 600 800 1000 1200 1400

Soil Age, YBP





Figure 2-12. Relationship between soil age and total phosphorus concentration in Blue Cypress
Marsh Conservation Area. Error bars represent one standard deviation of three
replicate measures of soil total phosphorus.




















MBP
Total P initial

Lab Pi Loss MBP


LossLabPi
Lab Po


Loss Lab Po
Inorg

Mineralized
Loss Inorg
Fuly


Loss Fuly

Hum


Loss Hum
Residual

Loss Res




Figure 2-13. Dynamic model of storage and transfers from maj or phosphorus fractions.


























*Validation
Humic Calibration
-Model









Validation
*Calibration
NaHCO o -Model


+ Validation
Calibration
Microbial -- Model









+Validation
Residual Calibration
-Model


*





Validation
*Calibration
NaHCO3 I Model








+ Validation
HCI P *Calibratior
SModel


150



15000

250



200

150




100

75


1 10 100
Time, years


1000 10000


1 10 100
Time, years


1000 10000


Figure 2-14. Results of model calibration and validation. Calibration and validation performed at
unimpacted and impacted BCMCA stations, respectively. Solid line indicates
modeled P fractions at impacted station.









CHAPTER 3
ESTIMATING STABILITY OF ORGANIC PHOSPHORUS IN WETLAND SOILS

Introduction

Fractionation of soil P into organic, inorganic, labile and recalcitrant pools is useful for

estimating effects of crop management practices in agricultural settings, and to determine soil

nutrient status in natural ecosystems. Numerous procedures have been developed to divide soil P

into respective pools. Chang and Jackson (1957) developed a sequential inorganic P

fractionation procedure that used sequential extraction with NH4C1, NaOH, and H2SO4 to

characterize P according to its association with soil minerals. Sommers et al. (1972) used a

sequential extraction to characterize organic P in lake sediments, followed by anion exchange

chromatography of the extracts. Their sequential extraction consisted of a 1N HCI extract, a cold

0.3N NaOH, and finally a 0.3 N hot (90oC) NaOH extract. The extractant solutions were eluted

through an anion exchange column to identify P fractions based on their interactions with the

column material. The second (hot) NaOH extract removed an additional 30% of the organic P

present in one of the highly organic lake sediments in their study. They believed the hot NaOH

fraction consisted of biologically resistant organic P forms. Their overall obj ective was to

characterize the type of organic P present to gain insight into the processes relating to

mineralization and immobilization. Hieltjes and Lijklema (1980) developed a similar scheme for

the fractionation of inorganic P in calcareous lake sediments. Their technique was similar to

Chang and Jackson (1957), but the main objective was to characterize inorganic P sources to the

water column of lakes. They emphasized the necessity of removing inorganic calcium with an

initial alcohol NH4Cl extract, prior to extraction with NaOH, to avoid precipitation of Ca-P

compounds. Hieltjes and Lijklema also fractionated inorganic P compounds with a soil to

solution ratio of 1:1000. They extracted P from the inorganic P compounds iron hydroxy-










phosphate and aluminum orthophosphate, as well as from calcium phosphates. They also

demonstrated the importance of the sequence of extractions, finding that when performed alone,

the 0.5M HCI extracted 99% of the P bound in the Fe and Al minerals, as well as the P in

calcium phosphates. The authors pointed out "...the characterization of phosphates is defined by

the analytical procedure."

The method proposed by Hedley et al. (1982) was also similar to Chang and Jackson

(1957), but included an attempt to measure microbial P. Both techniques employed a preliminary

exchange extraction, followed by extraction with base, and then acid. Hedley at al. (1982)

wanted to examine changes in P sequestration as a function of tillage practices, so they

investigated organic P fractions. The authors realized that identification of particular organic P

compounds was beyond their capabilities, and stated "At present, this is an impossible task as

much of soil P remains unidentified." Their study soils were low in organic carbon (approx. 3%)

compared to typical wetland soils. The authors classified organic P (Po) into four categories,

bicarbonate-extractable Po, microbial biomass Po, alkaline extractable Po, and residual Po. The

bulk of the changes in soil P resulting from the two tillage practices occurred in the residual

(non-extracted) P fraction. The difference between the two tillage practices indicated that

refractory P had declined as a result of crop harvesting. Thus, the crop had depleted labile pools,

and caused the replenishment of this pool by increasing mineralization of refractory pools.

Bowman (1989) proposed a method to determine total organic P in soils. The method was

a sequential technique, first using a strong acid extractant (H2SO4), followed by a base (0.5 M

NaOH). The acid extraction involved the generation of considerable heat, as water was added to

a dried soil sample containing concentrated acid. For both extraction steps, Po was determined by

the difference of inorganic P (Pi) in the extractant of a digested and undigested aliquot of the









same sample. The acid extraction step extracted 79% of the Po present in the soil sample, without

causing significant hydrolysis of organic P. Any hydrolysis of Po that did occur, the author

believed occurred in the highly labile organic P sources. Since this pool is a very small fraction

of TPo, its addition to the Pi pool was not a significant error. Bowman (1980) also points out that

the proposed extraction technique does not "...identify the nature of the Po extracted from the

unamended soil."

Ivanoff et al. (1998) focused their attempts at soil P fractionation on organic soils. Their

obj ective was to establish stability of P in the Everglades. Their technique used a sequential

extraction with bicarbonate, HC1, and NaOH. No attempt was made to distinguish between

inorganic P pools; all Pi was removed with a IM HCI extract immediately following a 0.5M

NaHCO3 extract. They used standard organic P compounds of known composition and liability to

determine how much Po was hydrolyzed to Pi as a consequence (or artifact) of the extractant.

They found that, except for the highly labile organic P material they used, acid pretreatment

hydrolyzed only 1 2 % of the Po present. Also, 0.5 M NaOH caused little hydrolysis of Po, with

the exception of glucose-6-phosphate (41% hydrolyzed). This indicates the potential for this

extractant to hydrolyze organic P, causing possible overestimation of Pi, at the expense of Po.

They extracted soils from the Everglades Agricultural Area and found that the acid extract

removed approximately 12% of the Po pool without causing significant hydrolysis of this pool.

The acid extractable pool probably includes amino-sugars, amino acids and other acid-soluble

organo-P compounds, as well as Ca-P compounds (Stevenson 1996). Ivanoff et al. (1998) were

also able to demonstrate that the Murphy-Riley reagent used in the P determination caused

minimal (<1.5%) hydrolysis of the organic P contained in glycerophosphate, glucose-6-

phosphate, and phytic acid. Of the total Po in their soils, approximately 40% was recovered in the









NaOH extract, 30% was attributed to microbial biomass, and 17% was non-extractable. The

alkaline extractable fraction was further divided into two fractions after extraction by acidifying

the extract, causing precipitation of humic materials. This allowed for separate determination of

the P content of humic and fulvic acids. Kuo (1996) described an ignition technique for TPo

determination, where ignited (550oC) and un-ignited samples are compared and TPo is

determined by difference. The rationale was that organic compounds are completely oxidized at

high temperature and the P associated with these compounds is liberated.

Characterization of the organic P pool is especially important in wetland soils, since this

tends to be the dominant fraction (Wetzel 1999, Reddy and DeLaune 2007). Reddy et al. (1998)

found that organic P represented from 56-70% of total P for the peat soils of the Everglades

Water Conservation Areas. Its mineralization rate can be an important regulator of primary

production, especially in P-limited wetlands and other aquatic ecosystems. This rate is

determined primarily by the recalcitrance of the organic P pool. Stability of Po has been shown to

increase with soil depth in wetlands. Reddy et al. (1998) has shown that in WCA-2A in the

Everglades, refractory P increased from 33% of total P in surface soils, to 70% in deeper strata.

Similar trends in refractory P as those observed in the northern Everglades were found in soils

from BCMCA. Olila and Reddy (1995) showed generally increasing residual (or recalcitrant) P

content, increasing from 20% (of TP) in surface horizons to approximately 80% at a depth of 35

cm. Determination of organic-P stability can be a useful metric of the extent to which an aquatic

ecosystem may be P-limited. Chemical extraction of Po from soils provides little information on

the stability of organic phosphorus compounds beyond classification into several operationally

defined groups. New techniques are needed to provide further resolution of the extent to which

Po has become immobilized. The obj ective of this study was to characterize the stability of










organic P in peat wetland soils using three techniques; two new thermal-based and one

enzymatic extraction. The main hypothesis was that as organic P becomes increasingly

recalcitrant, more thermal energy is required to liberate it from the organic compounds with

which it is associated.

Methods

Site Description

Field sampling was conducted in two subtropical wetlands, Water Conservation Area 2A

(WCA-2A) in the northern Florida Everglades and Blue Cypress Marsh Conservation Area

(BCMCA) in east central Florida, USA. The peat soils are derived from sawgrass (Cladium

jamnaicense Crantz) (Davis 1943). They are approximately 2 m in thickness and the basal

deposits were formed approximately 5,000 years before present (Gleason and Stone 1994). They

thus present a vertical profile that progresses from recently deposited plant litter that can be

presumed to be relatively labile, to deeper organic deposits that are recalcitrant due to millennia

of biogeochemical weathering. Two sampling locations were selected in WCA-2A, long-term

South Florida Water Management District research stations El and E5 (Fig. 1-2). Station El is

located in a nutrient-impacted region, approximately 1.4 km south of South Florida Water

Management District' s (SFWMD) inflow structure S-10C. Station E5 is located approximately

10 km south of this structure, in the un-impacted central marsh. Vegetation at station El was

predominantly cattail (Typha domingensis Pers.), and at station ES, sawgrass (Cladum

jamnaicense Crantz).

Soils were also collected from Blue Cypress Marsh Conservation Area (BCMCA). The

BCMCA is a 1 16 km2 Shallow marsh in the headwaters of the St. Johns River. It is also underlain

by peat soils that vary from 1.5 m thick in the eastern marsh to greater than 4 m thick near the

center. The soils of BCMCA appear to be both higher in P than those of the WCA-2A (Grunwald









et al. 2006) and less P-limited (Prenger and Reddy 2004). The vegetation is a mosaic of open

slough communities (Nymphaea odorata, Eleocharis elongata, Utricularia spp.), Maidencane

flats (Panicum hemitomon), broad expanses of sawgrass, and tree islands (Acer rubrum,

Ta-xodium distichum, M~yrica cerifera). Agricultural discharges have created a region of elevated

soil P in the northeastern area of the marsh, though those discharges ceased in 1994. One

sampling location was selected from this region of the marsh, Station C1, and one station was

selected from the interior unimpacted part of the marsh, Station B4 (Fig. 1-3)



Sampling and Analysis

Soil samples were collected in triplicate from all stations in both marshes on July 17 and

18, 2001. Soil cores were obtained with a 2-m long x 7.3-cm-diameter stainless steel soil corer.

Soil samples were extruded in the field into ZiploCTM balgs at 10-cm intervals. The samples were

immediately chilled to 4oC and were kept on ice for transport to the laboratory. Laboratory

processing was performed at the University of Florida' s Wetland Biogeochemistry Laboratory.

Wet weight was recorded for each sample and samples were transferred to rigid polyethylene

containers and thoroughly mixed. Three P extraction techniques were used to characterize soil

organic P.

Method 1: autoclave extractable P. The samples used in this experiment were sub-

samples of those described in Chapter 2. The technique used in this extraction is similar to that

described by Kenney et al. (2000). Twenty-five ml of deionized water was added to 0.5 g of

oven-dried equivalent wet soil. The soils were shaken at room temperature for 3 hours,

centrifuged, and the supernatant water was removed and vacuum filtered through 0.45 CIM

poresize polyethersulfone filters. The residual soil sample was then placed into an autoclave for









90 minutes at 128oC and 1.7 atmospheres (25 lbs in-2). The sample was removed from the

autoclave, 20 ml of deionized water was added, and the sample was equilibrated for a one-hour,

end-to-end shaking period. The sample was then vacuum filtered through 0.45 CIM filters.

Extracts were analyzed for dissolved reactive P (DRP) and total P. Triplicate samples of model

organic P compounds glycerophosphate and phytic acid were also extracted using the above

procedure to determine the extent of organic phosphorus mineralization in model compounds.

The difference between the initial deionized water extraction and the post-autoclave extract is

termed hot water extractable P (HEP).

Method 2: pyrolytic extraction of organic P. The soil samples used in this experiment

were subsamples of those used in the previous extraction. Pyrolytic extraction of organic P was

performed in a furnace that was constantly purged of 02 with N2. This method is similar to a

thermal gravimetric method used to characterize the degree of decomposition in peat (Levesque

and Dinel 1978). The oven consisted of a 7.6-cm (3-in) diameter galvanized steel pipe with two

threaded end caps. A 120V heating element was spiral-wound around the pipe and connected to a

temperature controller. The end caps were drilled and threaded to accept a gas inlet and outlet

port, and the entire assembly was placed into an insulated metal box. One end cap was removed

and a batch of samples was placed into the furnace. Samples were first extracted at room

temperature with IM HCI on a reciprocating shaker for three hours to obtain an estimate of

inorganic P. Oven dried samples were weighed in triplicate onto 5 cm squares of aluminum foil.

The foil squares were placed onto a tray, the tray was inserted into the oven, and the oven was

sealed. The chamber was purged for approximately 10 min and was then increased to the set

point temperature. It remained at this temperature for one hour for each treatment temperature.

Three soil depths (0 -10, 40 50, and 90 100 cm) were used for this experiment. Pyrolysis









temperatures were 160, 200, 260, 300, 360, and 550oC. The samples were removed from the

oven and transferred to 43 ml centrifuge tubes. They were then extracted with IM HCI at a 1:50

ratio for three hours on a reciprocating shaker, then filtered through 0.45 C1M fi1ters. The P that

was extracted at room temperature with IM HCI was subtracted from each value to yield a net

pyrolyzed P.

Method 3: enzymatic hydrolysis of soil organic phosphorus. Organic P that is

biologically available via extracellular enzymes was determined using phosphatase enzymes.

Soils from only the unimpacted sites in WCA-2A and BCMCA were used in this study. A 0.5 g

oven-dried equivalent fresh soil was weighed into 50 ml centrifuge tubes. Each enzymatically

treated tube was amended with 5 ml of modified universal buffer (MUB) containing 25 enzyme

units of akaline phosphatase monoesterase (Sigma-Aldrich Co., St Louis, MO, catalog no. P-

8361, lot no. 062Kl359). Stock MUB consisted of the following: 12. 1 g tris hydroxymethyl

amino methane (THAM), 11.6 g maleic acid, 14 g citric acid, 6.3 g boric acid, and 488 ml 1N

NaOH diluted in 1 L of deionized water (Tabatabai 1982). Two hundred ml of the stock MUB

was diluted to 500 ml, and this was pH adjusted with 0. 1N NaOH to pH 10, the optimum for this

enzyme. The enzyme concentration in the stock MUB was high, and represents enzyme non-

limiting conditions. The activity of the MUB + enzyme solution was qualitatively examined on a

fluorometer and high activity was confirmed. One set of samples was treated with the MUB +

enzyme, and a duplicate set was treated with MUB only. Eight samples were incubated in

triplicate to examine the reproducibility of the technique. All samples were incubated at 30oC for

4 hours and then extracted with 0.5M NaHCO3 (pH = 8.24) for 30 min on a reciprocating

shaker. Samples were then directly filtered, without centrifuging, through 0.45 CIM









polyethersulfone filters. Enzymatically extractable P was determined by difference between the

MUB and MUB + enzyme samples.

Statistical comparisons between pyrolysis treatments were performed with the General

Linear Model procedure in MiniTab (rel. 14.2) with fixed effects (temperature and soil depth)

and crossed factors. Differences among individual treatments were determined using Tukey's

multiple post-hoc comparison, and judged to be significant at p<0.05. Regression analysis (HEP

experiment) was performed using the Microsoft Excel (ver. 2003) Data Analysis ToolPak. Slope

was judged significantly different from zero for p<0.05.

Results and Discussion

Autoclave Extractable Phosphorus

Both Everglades WCA-2A stations showed much lower HEP than the BCMCA stations

(Fig. 3-1). HEP accounted for 10 -50% of total P in surficial (0-10 cm) soils, declining to 5-10%

of total P at a depth of 60-cm in all cores. All stations showed declines in HEP with depth,

indicating increasing resistance to thermal breakdown with depth, or soil age (Table 3-2). HEP of

WCA-2A soils was considerably lower than BCMCA soils, ranging from 8% to 15% of total P at

the soil surface, declining to 2-5% of total P at a depth of 60-cm, compared to 36-47 % of total P

in BCMCA surface soils and 1 1-19% at 60-cm. In general, the proportion of total P that could be

extracted with hot water declined at a rate of 0.04 to 0.22 % cm-l at the two Everglades stations,

and approximately 0.4% cm-l to 0.6% cm-l at the BCMCA stations. This decline suggests that

deeper soils were more resistant to thermal decomposition than surface soils.

The dramatic difference between the HEP in the two wetlands may be due to differences in

calcium concentration. WCA-2A is a hard water system, with high calcium particularly near

inflow sources, whereas BCMCA is a low pH, soft-water system. Since calcium carbonate

solubility declines with increasing temperature, precipitation of calcite in Everglades samples









may have occurred during autoclaving, and this new carbonate may have sorbed any P released

during the extraction. If this was the case, it is not surprising that both BCMCA stations showed

higher levels of HEP than even the most impacted station, E5. Modifying the procedure such that

the final, post-autoclave extraction was performed with IM HCI would re-dissolve any calcite-P

compounds that may have been formed, thus preventing this potential artifact.

Hot water extractable P explained 69 percent of the variability in microbial biomass P (Fig.

3-2). The slope of the linear fit of the HEP to microbial biomass P data implies approximately 48

percent of the P extracted in the HEP extract came from the microbial biomass pool. This is

consistent with other researchers who attribute most HEP to bacterial storage of polyphosphates

(Kenney et al. 2000). The relationship between HEP and water soluble P (WSP) was not as

robust, probably due to removal of WSP in the initial deionized water extract (Fig. 3-2).

Approximately 60% of both glycerophosphate and phytic acid was mineralized at the

experimental temperature and pressure (Fig. 3-3). This indicates that even at the relatively low

temperature used here, there was considerable breakdown of recalcitrant organic P that was

incorporated into phytin. This further suggests that hot water extractions may liberate more than

microbial and algal stores of polyphosphates, as has been postulated by other researchers

(Kenney et al. 2000).

Pyrolytic Extractable Phosphorus

When the results from all depths and both wetlands are combined, increasing temperature

resulted in greater mineralization of organic P, with the exception of the 550oC treatment (Fig. 3-

4). At this temperature, P recovery declined dramatically. This decline was confirmed by

repeating the experiment with only the last two temperatures 360oC and 550oC and the results

were the same; paradoxically reduced recovery of P. One possible explanation for this is that

mineralized P in the reducing environment of the pyrolysis chamber was converted to phosphine









gas (PH3) and lost from the system. Loss of phosphine gas from wetland soils has been suggested

as a significant biogeochemical pathway for P export from wetlands (Devai and DeLaune 1995).

However, when the P from the post-pyrolysis extract plus the total residue P was summed for the

two temperatures, they were found to be equivalent, indicating no losses from the system. It is

not clear what caused the large reduction in IM HCI extractable P in the highest temperature

treatment, however one plausible explanation is the formation of a material with strongly

hydrophobic properties. For example, one method of treating wood for decay resistance, called

torrefaction, is to subj ect it to a high temperature, inert atmosphere (Felfli et al. 2005). This

treatment apparently eliminates some chemically-bound water, and the final product is markedly

more hydrophobic. The hydrophobicity of the soil after the 550oC treatment may have prevented

the dilute acid access to the soil matrix. The results from this and the previous 550oC treatment

indicate that this temperature may be too high to be useful for characterizing the recalcitrance of

organic P. If the 550oC treatment is removed from the analysis, increasing temperature

significantly increased mineralization of organic P (Fig. 3-5, Table 3-3). The significant

(p<0.001) interaction term can be interpreted to mean that some temperatures used in the

extraction were not effective at distinguishing between the three soil depths used in this

experiment. For example, the 160oC treatment extracted significantly more P from the shallow

(0-10 cm) soil sample, but was incapable of distinguishing between the mid (40-50 cm) and deep

(90-100 cm) sample. The temperature that most effectively distinguished between the shallow,

mid, and deep sample was 360oC. When stations from both wetlands are combined for the 360oC

treatment, there were significant differences among depths in mean extractable P for all shallow,

mid, and deep samples: 64%, 80%, and 84% of P extracted, respectively (p = 0.013) (Table 3-4).

There is therefore evidence to support the hypothesis that the amount of thermal energy needed









to mineralize organic P is dependent on its recalcitrance. Further subdividing the dataset by

wetland and nutrient impact status reduces respective sample sizes, and statistically significant

differences among soil depths decline accordingly.

If the analysis is repeated on the surficial peat, more P is liberated at lower temperature, for

instance 75% at 300oC. This suggests that the less decomposed peat is not as resistant to thermal

decomposition. The lower temperatures extracted very little P in the deep peat, approximately

2% at 160oC and 10% at 200oC, compared to 22% and 38% in the surface peat.

As each sample was subj ected to progressively higher heat, weight loss of the sample

increased, as would be expected. This weight loss was likely due to loss of gaseous carbon

thermal decomposition products. Bartkowiak and Zakrzewski (2004) subj ected extracted lignin

to pyrolytic thermal gravimetry and found that the lignin had a declining hydrogen content,

relative to carbon and oxygen, over the range of temperatures they used. They attributed this to

destruction and loss of methoxyl groups associated with the lignin molecule. At 360oC, the

BCMCA samples had lost as much as 40% of their initial dry weight. There was a relationship

between the sample dry weight loss and the amount of P that could be extracted with the dilute

acid, with 73% of the variance in extractable P accounted for by weight lost during the pyrolysis

procedure. For every 10% increase in weight loss there was a 24% increase in extractable P. This

suggests that the increase in extractable P was due to destruction of soil organic matter and

subsequent liberation of P from organic compounds.

Enzymatic Hydrolysis

The amount of P that was enzymatically extractable was very low. There are several

possible explanations for this. The microbial communities at both stations are functioning in a P-

limited environment, though less so at BCMCA station B4. This would have the effect of

increasing enzyme production in those organisms capable of producing it. The small pool of










organic P that could be hydrolyzed by the added phosphatase may thus have already been

exposed to the native phosphatase, thus reducing the pool that was susceptible to this enzyme.

Another possible explanation is that the dominant form of organic P in these wetlands was

diester-P, and therefore was not hydrolyzable with the monoesterase used in this study.

Conclusions

The amount of P that could be extracted with hot water declined with depth at both stations

in BCMCA, and at the unimpacted station in WCA-2A. The variable results observed in WCA-

2A may be due to higher calcium levels in the soils of WCA-2A, particularly near inflow

sources. This procedure may better characterize stability of organic P if calcium is first removed

with a weak acid extract prior to autoclaving. For the low calcium soils of BCMCA, HEP ranged

from approximately 40 % of total P in BCMCA surface soils, declining to 15% at a depth of 60-

cm. The increasing resistance to extraction with hot water with soil depth suggests that HEP may

be a useful method to characterize the stability of the organic P pool.

Pyrolytic extraction of organic P yielded similar results. Since the final step with this

procedure was extraction with weak acid, potential underestimation of P mineralization, due to

calcium precipitation, was not observed. Increasing temperatures yielded progressively greater

extraction of Po. Also, at any given temperature, significantly more P was extracted from surface

soils than from deeper deposits. The most effective temperature for distinguishing the stability of

Po was 360oC. Addition of organic P hydrolyzing enzymes did not cause significant hydrolysis

of Po.

Results from this study suggest that 1) thermal stability of organic P seems to be related to

the recalcitrance of the compound that contains the organic P, and 2) thermal extraction of Po

from wetland soil may provide a useful diagnostic tool for investigating organic P liability.










The results from this study do not provide information on the mechanism for increasing

recalcitrance. Similarly, the results from sequential chemical fractionation (Chapter 2) provide

little compound-specific information on long-term changes that occur in Po in wetland soils. This

is where investigative techniques such as stable isotope analysis, C- and P-Nuclear Magnetic

Resonance, spectrophotometric, and fluorometric techniques may provide further insights. These

methods were used in the next chapter to better elucidate long-term changes in both soil organic

matter and soil Po.












Table 3-1. Basic physico-chemical properties of surficial (0-50 cm) soils
collected in July 2002 from WCA-2A and BCMCA.
BCMCA WCA-2A
Parameter B4 Cl El E5
TN, % 3.21 3.50 2.93 2.42
TC, % 47.4 48.3 45.1 45.7
Dry Weight, % 8.7 10.3 10.0 10.8
pH 5.70 5.88 7.59 7.30
Bulk Dens., g cm-3 0.081 0.101 0.095 0.111
Ash, % 5.8 8.8 7.3 11.1
TP, mg kg-1 327 367 546 197

Table 3-2. Regression statistics for autoclave extractable P verses depth in soil samples from
Blue Cypress Marsh Conservation Area and Water Conservation Area 2A.
Regression statistics based on the upper 60-cm of all cores.
Parameter BCMCA BCMCA WCA-2A WCA-2A
(Impacted) (Unimpated) (Impacted) (Unimpacted)
r2 0.84 0.95 0.43 0.88
p 0.004 <0.001 0.112 0.002
Slope -0.366 -0.594 -0.039 -0.224
Intercept 41.6 43.4 9.3 14.6


Table 3-3. ANOVA General Linear Model results of pyrolysis temperature and soil depth
effects. 550oC treatment not included in analysis.
Source DF Seq. SS Adj. SS Adj. MS F p
Depth 2 8581 7066 3533 32.2 0.000
Temperature 4 89640 90382 22596 205.7 0.000
Depth X Temp. 8 4521 4521 565 5.14 0.000
Error 144 15820 15820 110
Total 158 118564


Table 3-4. ANOVA results of differences between peat sample depths for
360oC treatment. All stations combined for analysis. Two samples excluded
from analysis: Station El deep sample (sand) and B4 mid depth (incomplete
pyrolysis).
Source DF SS MS F p
Depth 2 2228 1114 5.1 0.013
Error 27 5893 218
Total 29 8121





































0 20 40 60 80 100


20 40 60 80 100


HEP, % of TP




Figure 3-1. Hot water extractable P (HEP) of Blue Cypress March Conservation Area and Water
Conservation Area 2A soils. Panel A data from stations that have been impacted by
nutrient-rich inflows. Panel B data from stations that have not been impacted by
nutrient inflows. Error bars represent one standard deviation.
















200 -1r2 = 0.691

cr 150 + *



** *
50 +1 *


05

+ y = 0.049x + 0.387
30
r2 = 0.373
,25
20 +2

S15

10
5 ** *


0 50 100 150 200 250 300 350 400
HEP, mg/kg
Figure 3-2. Hot water extractable phosphorus (HEP) of Blue Cypress March Conservation Area
and Water Conservation Area 2A vs. water soluble (WSP) and microbial biomass
phosphorus (MBP).
















































I I I I


80 E


mNo Heat
0 Heat


Figure 3-3. Hot water extractable P (HEP) of two model compounds, glycerophosphate (GP) and
phytic acid (PA). Error bars represent one standard deviation.








































I BCMCA Impacted
SBCMCA Unimpacted





160 200 260 300 360


0-10 cm










40-50 cm


-


120
100
80
60
40

20

120
100
80
60
40
20
0
120
100
80
60
40
20
0


~IWCA2A
WClrA2A


Impacted
Unimpacted


90-100 cm


Temperature, oC





Figure 3-4. Percent of total phosphorus extracted from dried, ground peat using pyrolytic
extraction under nitrogen atmosphere.











WCA2A l




Be df




ab ab
aa
ab

dms d ms d ms d ms dms
160 200 260 300 360

BCMCA



f9
he -h




ba baii~


Depth
Temperature


20

0
Depth
Temperature


160


200


260


300


360


Figure 3-5. Thermal extraction of organic P. Bars with same letter not significantly different
(p>0.05). Each bar represents six samples. Letters beneath x-axis refer to sample
depth interval: d (deep) = 90-100 cm, m (middle) = 40 50 cm, and s (shallow) =
10 cm. The 550oC treatment is not shown.









CHAPTER 4
CHARACTERIZATION OF LONG-TERM CHANGES INT ORGANIC PHOSPHORUS INT A
SUBTROPICAL WETLAND

Introduction

Organic phosphorus (Po) is often the principal pool of phosphorus in aquatic ecosystems

(Wetzel 1999, Reddy et al. 1998). It's mineralization rate, chemical structure, and pool sizes are

thus important determinants of primary production in aquatic ecosystems (Wetzel 1999). Less

than 50% of the compounds that comprise the Po pool have been identified to date (Stevenson

1994). Soil Po can be grouped into several main categories: inositol phosphates, nucleic acids,

phospholipids, with phosphoproteins and metabolic phosphates making up a minor contribution

to soil Po (Stevenson 1994). Phosphoesters can be further subdivided into monoesters, diesters,

and triesters depending on the number of covalent ester linkages to ortho-P molecules.

Monoesters include inositol hexaphosphate (phytate), a maj or component of plant seeds and

ATP; this group accounts for approximately half of the total soil organic P pool. Diester P is a

structural component of nucleic acids and cell membranes. This group is rapidly mineralized in

upland soils, although Turner and Newman (2005) found that phosphate diesters dominated the

Po fraction in a study of soils from the Florida Everglades.

Chemical fractionation of Po with sequential acid and base extractions provides insight into

respective inorganic and organic pool sizes, as well as some information on operationally defined

fractions. However, more compound-specific techniques are needed if better resolution of the

many compounds that comprise the Po pool is desired. Chromatographic separation has been

used to identify specific Po compounds (McKelvie 2005). Gel filtration chromatography is useful

for separating Po according to molecular weight. In this technique, water samples or soil extracts

can be passed through non-ionic columns, the eluted fraction collected, and analyzed for total P

and molybdate-reactive P. Ion exchange columns have also been used to separate ionizable Po









compounds. Numerous researchers have used 31P Nuclear Magnetic Resonance (31P-NMR) to

identify broad classes of Po compounds in soil extracts (Turner et al. 2003b, Cade-Menun 2005).

This technique involves extracting soil with NaOH, concentrating the sample, and placing the

sample into a strong magnetic field. The sample is then subj ected to pulsed radio frequency

energy, energy is absorbed, and then re-emitted by the sample. The energy required for

resonance of the 31P nuclei emission is determined in part by the amount of electron shielding of

the Po molecule and this property is used to classify the Po. Mass spectrometry has also been used

for characterizing Po, though only in surface waters (Cooper et al. 2005). This involves

production of gas-phase atoms, ionization, separation based on mass to charge ratio, and

detection. The difficulty in non-destructively- generating ionized, gas-phase molecules from large

Po molecules and distinguishing between Po and the background of other organic compounds in

environmental samples has slowed the development of this method. New techniques such as

electrospray ionization and matrix assisted laser desorption ionization (MALDI) appear to have

overcome some of these limitations (Cooper at al. 2005). Established methods based on organic

matter spectral absorbance (Chen et al. 1977) and fluorescence (Cox et al. 2000) may also be

useful for characterizing Po, since fate of organic matter and Po are closely coupled (Stewart and

Tiessen 1987).

In this study, several techniques were used to characterize the biogeochemical

transformations that have occurred in the soil profile. The optical properties of extracted soil

organic matter have been shown to be related to average molecular weight and the extent of

humification (Cox et al. 2000, Chen et al. 1977). Natural dissolved organic matter and extracted

soil organic matter contain molecules that are capable of absorbing light energy and promoting

an electron to an excited state. Upon returning to the ground state, the molecule emits light, i.e.,









it fluoresces. The pattern of fluorescence depends on the nature of the organic molecule. It has

been shown that as organic matter ages, the fluoresced light is longer in wavelength, or red-

shifted. By taking the ratio of the intensities of the higher fluorescence wavelengths to the lower

wavelengths, the extent of humification can be approximated (Cox et al. 2000). This index may

be useful as a measure of the extent of soil diagenesis.

The ratio of absorbance of light at 465 and 665 nm, or E4:E6 ratio, has been shown to be

related to viscosity, which in turn is related to the average molecular weight of compounds in

solution (Chen et al. 1977). Increasing molecular weight has been indicated as a consequence of

soil aging and humification (Stevenson 1994).

Changes in the depth distribution of broad groups of soil organic carbon compounds were

investigated using solid state 13C nuclear magnetic resonance (13C-NMR). Solution 31P-NMR

was used in a similar way to examine diagenetic changes in soil organic phosphorus. The

preservation of the principal plant polymers, lignin, cellulose, and hemicellulose was determined,

as well as the 13C and 15N content of this material. Lignin, cellulose, hemicellulose and soluble

carbon are the principal carbonaceous plant polymers deposited at the soil surface after plant

senescence. These polymers form the basis for the development of the peat deposit. They also

contain an isotopic record of changes in plant communities. Isotopic analysis of the fibric portion

of the peat deposit is superior to analysis of bulk peat, as it does not contain inorganic carbon,

algal remains, or percolated humic and fulvic compounds. This was used as an indication of

changes in floral community composition. Lastly, intact plant macrofossils (i.e. seed husks,

stems) were dated using 14C to establish the chronology of biogeochemical changes.









Methods


Site Description

The study location was BCMCA, a 12,000 ha peat marsh (Fig. 1-2). The eastern extent of

the marsh is characterized by open-water sloughs and tree islands. Typical vegetation in the

sloughs is a Utricularia sp. and Nymphaea odorata community, while the tree islands are

characterized by Ta-xodium distichum/Acer rubrum/Salix caroliniana. community. The western

region of the marsh i s predominately a Cladium jamnaicense/Panicum hemitomon prairie, with

some Cephalan2thus occidentalis, Pontederia cordata, and Sagittaria~tt~t~tttt~t~tt lan2cifolia. The soil at the

site consists of well-decomposed peat (Terra Ceia series, Euic, hyperthermic Typic

Haplosaprists) ranging from 2 m thick in the eastern region to over 5 m thick in the marsh's

center. The peat is underlain by coarse sand on the east, and sandy clay in the center. Surface

water in the marsh is characterized as soft and slightly acidic, with pH of approximately 6.5.

Nutrients are low, though not as low as the Everglades, with median surface water TP

concentration in the central region of the marsh of 50 Clg L 1. The hydrology of the marsh is

highly constrained by a network of water control structures. Inflows are primarily from

rainwater, a small stream on the western side (Padgett Branch), and three water control structures

on the southern side. Outflows occur through the water control structure, S-96C, on the northern

end of the marsh. The marsh surface elevation is approximately 7 m above mean sea level

(MSL). Water level in the marsh typically varies from 7 m (dry) to 9 m above MSL.

Soil Collection

Samples were collected from BCMCA at stations C1 and B4. These are the same

locations used in studies described in Chapters 2 and 3. Samples were collected on March 16,

2004. A peat corer was used to retrieve 2 7/8-inch ID soil cores (Clark 2002). This corer is a










piston-type corer with a sharpened steel leading edge. Handles on each side allow the device to

be rotated, which facilitates cutting through roots and peat fibers. The corer was constructed of

steel and employs a thin wall, semi-rigid, polybutyrate core liner. The liner containing the soil

sample was taken out of the core barrel after the core was removed from the ground. The

butyrate tubing was sufficiently thin that it was easily cut with a razor knife. Four cores were

collected from each station; three for fractionation and spectral properties of organic matter

extracts, and one for lignin fractionation and 14C analysis. Total core length ranged from 105 cm

to 150 cm. All cores consisted of peat or muck and were stored at 4oC until analyzed.

Organic Extract Fluorescence

Soils were sequentially fractionated using IM HCI extraction to remove inorganic C. This

was followed by a 0.5 M NaOH extraction for 17 hours on a reciprocating shaker at room

temperature. Twenty-five ml of extractant solution was added to approximately 0.5 g of oven-

dried soil, for a 1:50 extraction ratio. This solution was vacuum filtered through 0.45Cpm

polyethersulfone (PES) filters, diluted 1000X, and pH adjusted to 7 using a 0.05 M NaHCO3

solution that had been pH adjusted to 6.7. This dilution resulted in an optical density at 254 nm

of approximately 0. 15 (1-cm path length). This optical density has been shown to reduce both

inner filtering of fluoresced light, and to diminish absorption by the color of the extract (Ohno,

2002). A subsample of the NaOH extract was acidified to pH<2 to precipitate humic materials,

leaving only the acid soluble fraction, or fulvic acid in solution. This was done by adding 5 drops

of concentrated sulfuric acid to five ml of NaOH extract. The precipitated humic acid was

separated from the supernatant by centrifuging. Fluorescence properties of both samples were

determined. Fluorescence was determined with a Shimadzu model RFl501 (Shimadzu Corp.,

Tokyo, Japan) scanning fluorometer. Fluorometric scans were conducted with an excitation









wavelength of 254 nm and a fluorescence emission scan of 260 through 650 nm. The sample

vessel consisted of a 1-cm quartz cuvette. The humification index (HIX) was calculated as

480
C F/
HIX = A=435 (4-1)
345
C F/
A=300


where FI = fluorescence intensity (Cox et al. 2000).

Spectrophotometric Absorbance

Spectrophotometric properties were determined with a Shimadzu model UV160

spectrophotometer (Shimadzu Corp., Tokyo, Japan) on April 9, 2004. Absorbance was

determined at 465 and 665 nm in a 1-cm path length, flow-through cell. Analyses were

performed on the same 0.5M NaOH extract as used for the fluorometric analyses, though these

samples were diluted 40X and adjusted to pH 7 with dilute HC1. This dilution resulted in average

absorbance at 465 nm and 665 nm of 0.25 (f0. 18 SD) and 0.04 (f0.03 SD), well within the

measurement range of the instrument. The ratio of the absorbance at these two wavelengths is

reported here; i.e. the E4:E6 ratio. This ratio has been shown to be related to average molecular

weight of dissolved compounds (Chen et al 1977).

13C Nuclear Magnetic Resonance ( 3C-NMR)

Samples were collected February 11, 2003 from station B4 only. Oven-dried (700C)

samples were analyzed at a spectrometer frequency of 75.5 Mhz, contact time = 1 msec, pulse

delay = 2 sec, acquisition time = 2 sec, and accumulation of 4000 scans. The spectrum was

divided into four principal regions: carboxyls (benzene, carboxylic acid, amides and ethers),

aromatics phenolss), O-alkyls (plant polymers such as lignin and cellulose), and alkyls aliphaticc,

alkanes, fatty acids, and waxes). The fractions correspond to the spectra peak shift regions of 46-









110 ppm, 110-162 ppm, 59-92 ppm, and 0-46 ppm, respectively. The total area under each

region was determined and its proportion in the whole sample was calculated by dividing by the

total area of the spectra. It was assumed that the total area under the curve represented the total

carbon of the sample. This value was multiplied by the soil carbon content to determine the

actual mass of each fraction in the original soil sample.

3P Nuclear Magnetic Resonance (31P-NMR)

Samples were collected July 18, 2002 from station B4 only. Oven-dried (700C) and ground

samples were extracted by shaking the soil: solution ratio of 5:100 with a solution containing

0.25 M NaOH and 0.05 M EDTA (ethylenediaminetetraacetic acid) for 4 h at 200C (Cade-

Menun and Preston 1996). Each sample was extracted individually and centrifuged at 10,000 x g

for 30 min. Equal volumes of the extracts were then frozen immediately at -800C, lyophilized,

and ground to a fine powder. For solution 31P NMR spectroscopy, each freeze-dried extract

(~100 mg) was re-dissolved in 0. 1 mL of deuterium oxide and 0.9 mL of a solution containing 1

M NaOH and 0.1 M EDTA, then transferred to a 5-mm NMR tube. The deuterium oxide

provided an NMR signal lock and the NaOH raised the pH to >13 to ensure consistent chemical

shifts and optimum spectral resolution. Inclusion of EDTA in the NMR tube reduces line

broadening by chelating free Fe in solution (Turner 2004). Solution 31P NMR spectra were

obtained using a Bruker Avance DRX 500 MHz spectrometer operating at 202.456 MHz for 31P

and 500.134 MHz for 1H. Samples were analyzed using a 6 Cls pulse (450), a delay time of 1.0 s,

and an acquisition time of 0.8 s. The delay time used here ensured sufficient spin-lattice

relaxation between scans for P nuclei (Cade-Menun et al. 2002). Between 48,000 and 69,000

scans were acquired depending on the P concentration of the lyophilized extract, and broadband

proton decoupling was used for all samples. Chemical shifts of signals were determined in parts

per million (ppm) relative to an external standard of 85% H3PO4. Signals were assigned to










individual P compounds or functional groups based on literature reports (Turner et al. 2003) and

signal areas calculated by integration. Spectra were plotted with a line broadening of 8 Hz,

although additional spectra were plotted with a line broadening of 1 Hz to examine signals in the

phosphate monoester region.

Isotopic Analysis of Plant Fibers

Samples for lignin fractionation were collected on March 16, 2004 from stations C1 and

B4 in BCMCA. A single core from each station was split length-wise and one half of each 10-cm

interval was used for fiber analysis. Analysis was performed using an Ankom Fiber Analyzer,

model 200/220 (Ankom Inc., Fairport NY) following the method of Van Soest (1970). Samples

were sequentially extracted with a neutral detergent, an acidic detergent, and concentrated

sulfuric acid. The mass of each fiber fraction was determined gravimetrically after each

extraction step. This technique was originally developed for crop residues, and was modified to

characterize plant-derived soil carbohydrates. Initial attempts to fractionate soil carbohydrates

indicated that much fine organic matter leaked from the 30-C1M glass fiber bags used in the

analyses. Since the procedure is based on weight loss, an alternate method was developed. The

modified procedure followed these steps:

1. Place each 10-cm bulk-soil interval into a No. 35, 500-micron brass sieve.

2. Immerse the sieve in 0.5% Liquinox solution to disperse fine soil organic matter from fibric
material.

3. Gently agitate it for approximately 5 minutes.

4. Wash fine particulate material from fibric material remaining in sieve with stream of
deionized water until it is clean.

5. Air-dry the remaining fiber.

6. Grind fiber through 20-mesh screen in a WileyTM mill.

7. Fractionate using Ankom methodology.










Therefore, the results reported here represent the properties of the fibric material in the

sample, not the bulk soil. A subsample of the peat fiber was analyzed for 13C and 15N on a

Thermo Finnigan DELTA Plus isotope ratio mass spectrometer at the University of Florida Soil

and Water Science Department. Purified N2 and CO2 WeTO USed as internal standards, with a

helium carrier gas. Results are expressed in delta notation, relative to the 13C COntent of PeeDee

Belemnite, and 15N content of atmospheric N2-

Soil Age

Soil dating was determined through 14C analysis of plant macrofossils. Ten centimeter

sections of peat to a depth of approximately 1.5 m were carefully examined for plant fragments

such as stems, seeds, and other woody fragments. Care was taken to avoid selection of root

fragments. These plant macrofossils were sequentially extracted with 0.5M HCI (0.5 hours),

0. 1M NaOH (2 hours), and again with 0.5M HCI (0.5 hours). This removed carbonates and any

dissolved organic materials such as humic or fulvic acids that may have percolated down through

the soil profile. The samples were analyzed at University of California, Irvine Keck-CCAMS

laboratory by accelerator mass spectrometer (NEC 0.5MV 1.5 SDH-2, National Electrostatics

Corporation). Carbon dates were corrected to calendar dates using the computer program

CALIB, version 5.0. 1(Stuiver and Reimer 1993).

Results and Discussion

Fluorescence and Spectrophotometric Properties

The HIX of the peat extracts showed a steady downcore increase throughout the soil

profile (Fig. 4-1). Extracted, live plant tissue had a HIX of approximately 2, detritus 4, whereas

peat samples at a depth of 140 cm were about 10. Both stations had very similar increases with

respect to depth, though station Cl had a reversal (declining HIX) at 80 cm. This reversal









coincided with a change in physical properties of the peat. At a depth of approximately 80 cm,

the texture of the peat at Cl changed from a very decomposed, black muck, to a reddish-brown

fibric peat immediately underneath. This more fibric, less decomposed peat likely caused the

decline in HIX at station C1.

There was no significant difference in average HIX between the two stations for the humic

+ fulvic samples. (One-way ANOVA; F = 1.45; p = 0.232; MiniTab). This was expected, since

both stations have been exposed to the same hydrologic regime, and must have accreted peat at a

similar rate. Fulvic acid extracted from the soil samples showed a very similar trend of

increasing humification with respect to greater depth. In general, the whole NaOH extract humicc

+ fulvic) had an average HIX value approximately one unit higher than the fulvic sample.

The sharp increase in HIX at station Cl (70-90 cm) coincided with a dramatic increase in

OM age. Station B4 also showed local maxima at approximately this same depth. One possible

reason for this increase is accelerated decomposition during a long dry period in the marsh. Both

stations showed increase in HIX at this depth, followed by a slight decline (less decomposed),

then a gradual increase to the base of the core.

The E4:E6 ratios observed in this experiment indicate that the soil extract was richer in

fulvic than humic acids at both stations for all depths (Fig. 4-2) The ratio E4:E6 ratio of fulvic

acid in most soils was typically in the range of 6.0 to 8.5, whereas humic acids tend to be less

than 5 (Stevenson 1994). Chen at al. (1977) found that this range corresponded to fulvic acids

with an average molecular weight of approximately 2000 atomic mass units (amu; or daltons).

For this study, an average weight of 2000 amu corresponds to the complete soil extract

containing both humic and fulvic materials.









The interior marsh station (B4) showed a slight decline in E4:E6 ratio with increasing

depth, beginning at approximately 12 in the 0-10 cm soil depth, declining to 7 at a depth of 80

cm, followed by very little change to the base of the core. This suggests a relatively constant rate

of diagenesis for the surficial 70 cm of peat, with an increase in molecular weight of the

extracted organic matter to 70 cm. The E4:E6 ratio at the nutrient impacted station showed a

slight decline to 50 cm, increased to a depth of 100 cm, then declined to the base of the core. The

E4:E6 profile at station B4 is characteristic of the diagenesis of a relatively constant source, or

quality, of organic matter. The ratio at the soil surface is high, reflecting lower molecular weight,

poorly humified organic matter, as would be expected from recently deposited plant materials.

The ratio suggests increasing molecular weight with respect to soil depth to the base of the core,

as would be expected from the progressive humification of a constant source of organic matter

input. The gradual decline therefore implies that this region has experienced a similar history of

vegetation over the time period that this soil profile represents. Humification index showed

increasing humic character with depth, however results from the sequential fractionation

(Chapter 2) did not indicate increasing humic-P with depth. This may mean that in wetland soils

long-term sequestering of soil P is not occurring in humic-type materials, but in some other

compounds. Stewart and Tiessen (1987) point out that highly charged phosphate groups of soil

Po may prevent it from interacting with humic materials.

The E4:E6 profile at station C1 does not show the same constant decline as station B4. The

E4:E6 ratios at station Cl are lower than station B4 for soil depths above 70 cm, and greater than

those of B4 for samples below 70 cm. Lower ratios in surface depths (than B4) may be a

consequence of two factors; vegetation history and root biomass. Soil organic matter with greater

molecular weight at station B4 for soil depths < 70 cm may reflect changes in the vegetation









community in this area of the marsh. The Cl region of the marsh is known to have experienced

nutrient impacts that have altered the vegetation in the northeastern corner of BCMCA. A result

of this impact is the conversion of slough to cattail (Typha latifolia) and willow (Salix

caroliniana) marsh. Aquatic plants that characterize slough vegetation include such plants as

Nymphaea odorata, Utricularia spp., phytoplankton, etc. These plants contain less structural

tissue, such as lignin, than the emergent vegetation and shrubs mentioned above. Lignin is one of

the principal refractory biogenic compounds and is thought to be a precursor of humic materials

(Stevenson 1994), and humic content is a maj or determinant of the E4:E6 ratio. The nutrient

impact may have, by altering the local plant community, led to alterations in the biochemical

characteristics of the recent peat in this area of the marsh. A second potential cause of greater

E4:E6 ratios at station Cl may be greater contribution of live root biomass to the carbon pool of

the soil extracts. Pan2icum hemitomon is the dominant plant at station B4. Many small roots were

noted in the samples during core sectioning, many more so than at station C1, which was

dominated by Typha latifolia and Salix caroliniana. This new carbon would have consisted of

lignin, cellulose, and other polysaccharides of much lower molecular weight than humic

materials and may have biased the E4:E6 ratio upwards in the surface soil horizons of B4 relative

to station C1.

At neither station was a dramatic increase in molecular weight with increasing depth

observed. In fact, at station C1, very little overall change with respect to peat age was seen. This

suggests that peat soil diagenesis does not involve dramatic changes in molecular weight of

extractable soil components.

13C Nuclear Magnetic Resonance ( 3C-NMR)

Total O-alkyls are associated with plant polymers, such as lignin and cellulose, and are

considered relatively labile. This group represented the maj or fraction of soil organic carbon









(SOC), comprising 30-40% of the total carbon pool, although they declined from 225 mg kg-l at

the surface, to a minimum of 133 mg kgl at a depth of 90 cm (Fig. 4-3). As a percentage of total

C, these value are slightly higher than reported by Sjoigersten et al. (2003) for forest and tundra

soils. Concentrations declined with greater depth in soil profile, with the exception of a slight

increase for the bottom two depth intervals. An inflection at 90 cm coincided with a textural

change to peat characterized by much much finer organic matter. The most likely explanation for

this change is a period of greater water depths, favoring soft-tissued plants such as phytoplankton

or floating macrophytes. These plants would have contained less refractory carbonaceous tissue

(such as lignin), and thus would have been more rapidly mineralized. The more fibric materials

underlying this strata visually appeared less decomposed, and this was reflected in an increase in

O-alkyl concentration below 90 cm.

Alkyl-C compounds consist of fatty acids, waxes, and alkanes and include compounds

such as acetate, butyrate, and other fermentation end-products, as well as methane. Alkyls

constituted the second greatest fraction, comprising 25-30% of the soil organic carbon pool. This

fraction increased from 91 mg kg-l at the soil surface, to a maximum of 191 mg kgl at a depth of

90 cm, or an increase from 20% to 38% of SOC. This is slightly greater than observed by

Sjoigersten et al. (2003), probably due to increased production and storage of fermentation

products in the flooded organic soils of this study, as compared to their tundra soils.

Total aromatic carbon increased linearly with respect to depth, showing a very similar

concentration profile as alkyl-C, suggesting their accumulation at similar rates. Total aromatics

accounted for 20-30% of SOC. Baldock et al. (1997) reviewed the C-MNR results from organic

and mineral soils, forest litter, and composts and found that the proportion of aromatic carbon

compounds was not a particularly good surrogate for extent of decomposition. However, their









review did not include flooded soils. Organic wetland soils lack the fungi responsible for

production of the extracellular enzymes that are capable of hydrolyzing aromatic compounds.

This is likely the reason that the soils from BCMCA increase in aromaticity with depth, in

contrast to results from Baldock et al. (1997).

The rate of increase in alkyl-C and total aromatic compounds with respect to depth in the

soil profile was inversely related to the O-alkyl content, suggesting the biogeochemical

transformation of plant polymers to microbial metabolites and resistant organic compounds. The

ratio of alkyl to O-alkyl functional groups (A:O-A) has been shown to be a better indicator of the

extent of decomposition, especially for organic soils (Baldock et al. 1997). The ratio increases

steadily through the soil profile, with a dramatic increase in the region of silty organic matter.

The ratio begins at 0.4 at the soil surface, with a maximum value of 1.44 at 80-90 cm. Baldock et

al. (1997) reviewed C-NMR data from numerous studies on mineralization of forest, wetland,

and agricultural soils and found a range of A:O-A values, from 0. 1 to 1.5, with the ratio

depending mostly on degree of decomposition and secondarily on plant community and soil

mineral content. The ratios found in this study are on the high end of this range, indicating

advanced decomposition, especially at the lowest depths.

The steady increase in phenolic and aromatic compounds was very similar to the increase

seen in the humification index (HIX) reported earlier. Both techniques indicate a progressive

transformation of plant components to more humified, aromatic, and presumably stable soil

organic matter. This is a clear indication that plant litter that constitutes the bulk of peat soil

undergoes radical biochemical changes after senescence. Thus, inferences concerning ecosystem

history that are based on stratigraphic changes in peat chemistry need to consider that a

significant proportion of observed change may not be due to changes in ecosystem history (i.e.,










plant community, nutrient loading, etc.), but are due to biogeochemical transformations in soil

orgamic matter.

3P Nuclear Magnetic Resonance (31P-NMR)

Organic phosphorus in soil extracts, as a proportion of total P, increased throughout the

soil profile, representing approximately 20% of total P at the soil surface, and increasing to 40%

at a depth of 120 cm (Fig. 4-4). Within the organic P pool there was no significant difference

between phosphate monoesters and diesters (Student' s paired t-test; p = 0.321; df = 7). However,

it should be noted that the EDTA-NaOH extraction is known to degrade phosphate diesters such

as phospholipids and RNA to their phosphate monoester constituents, so the actual proportion of

phosphate diesters is likely higher and thus probably constitutes the dominant form of organic P

in these soils. Notably absent was myo-inositol hexaphosphate (phytic acid), the dominant

phosphate monoester in most upland soils (Turner et al. 2002) (Fig. 4-5). This form of P is of

plant origin, with maj or concentration in seeds. Its high charge density causes it to adhere to

soils, making it less available to microbial attack than other organic phosphates, and therefore

more stable. Few studies have documented the long-term fate of these two functional classes of

organic P in soils, particularly in wetlands. It appears from this study that both phosphate

monoesters and diesters increased slightly with soil depth, and maintained approximately equal

proportions. This represents an unusually high proportion of phosphate diesters. In a study by

Turner et al. (2003) the ratio of mono- to diester P ranged from approximately 5 to 26 for 29

temperate pastures, compared to 1:1 in this study. One of the principal sources of

phosphodiester is microbial tissue. The soils in this study appear to be much higher in microbial

biomass, or alternatively diester P is much better preserved under the anaerobic conditions of

these flooded soils. Turner and Newman (2005) also found elevated levels of diester P in the

Florida Everglades, which they attributed to slowed hydrolysis of plant and microbial inputs of










phosphate diesters. The slight increase in organic forms of P appeared to be at the expense of

orthophosphate, which declined throughout the profie. A strong signal from DNA is apparent

down to a depth of 80-cm. Since DNA is not well preserved in upland soils, this suggests a

prominent microbial community at these depths. However, preservation of DNA under flooded

conditions in a pond near Titusville, FL was quite good (Doran 2002). The Windover site is

approximately 90-km northeast of BCMCA and has gained international notoriety as having

some of the best-preserved human remains and cultural artifacts (particularly woven fiber) ever

recovered. In that study, 8000-year old human brain tissue was recovered from remains found in

a shallow pond. The DNA from this tissue was intact enough to partially sequence, though the

overall DNA yield was low; approximately 1% of that from fresh tissue. Thus, phosphodiesters

such as DNA may be better preserved than previously thought and thus may play a more

prominent role in long-term phosphorus stabilization.

Isotopic Analysis of Plant Fiber

The amount of plant fiber recovered using the wet-sieving technique varied from 5 to 15

grams dry weight, per 10-cm interval. Note that each core was split length-wise, so the actual

recoverable fibric material is twice this value. There was no apparent trend down core, though

plant fiber declined at depths where Eine organic matter increased. The relative abundance of

each plant polymer was lignin > cellulose > hemicellulose.

Lignin content increased gradually downcore at both stations, from approximately 40% of

the material dry weight in the surface soil, to 80% at the base of the core (Fig. 4-6). Lignin was

the most abundant plant polymer, from three to five times as abundant as cellulose and

hemicellulose. Station Cl showed a pronounced decline in lignin for the 50 to 60 cm depth,

coinciding with an increase in fine organic matter, and a sudden increase in organic matter age.

This may reflect either a deepening of the marsh, favoring soft-tissued macrophytes and










phytoplankton that possess less lignin, or alternatively that a long-term drought reduced much of

the lignin to smaller plant polymers. Evidence for a circum-Caribbean dry period that occurred

approximately 1000 years before present favors the second hypothesis (Buck et al. 2005, Hodell

et al. 2005).

Cellulose showed a gradual decline throughout the soil profile, declining from

approximately 20% at the surface, to 5-10% at 150 cm. Downcore lignin and cellulose content

was very similar to that found by DeBusk and Reddy (1998) for soil samples taken from WCA-

2A in the northern Everglades. They found that the lignin cellulose index (LCI; lignin/(lignin +

cellulose)) varied from 0.26 in standing dead plant litter, to 0.81 in peat taken from a depth of

30-cm. The LCI of the samples in this study varied from 0.73 at the soil surface, to greater than

0.90 at the base of the core, indicating a greater degree of decomposition at the lower soil depth.

The greater lignin content, as well as greater LCI in this study reflects the greater soil depths

investigated in BCMCA. DeBusk and Reddy (1998) examined peat samples from a maximum of

30 cm, whereas in this study peat from a depth of 150 cm was analyzed. Hemicellulose declined

somewhat erratically with respect to depth, from approximately 10% at the surface, to 3-5% at

the base of the core.

Soil lignin content increased with greater depth at both locations, while cellulose and

hemicellulose declined. Enzymes responsible for lignin decomposition, such as phenol oxidase

and lignin peroxidase, are secreted by white rot fungi such as Phanerochaete chrysosporium.

These fungi are not active in flooded soil, hence the buildup of lignin. Some of the increase in

lignin content with depth is undoubtedly due to the compaction and concentration of lignin as

more peat is accreted. However, the predominant mechanism is likely the loss of more labile

plant polymers from the separated soil fibric material. Overall, it is clear that as this fibric









material ages, it takes on a more purely lignin character. This data seems to conflict with the 13C

NMR results described earlier. The NMR data showed declining lignin throughout the soil

profile. However, those results were from a bulk soil sample (not sieved). It is likely that even

though lignin degrades slowly under flooded conditions, it was being converted to humic

materials and thus it represented a declining fraction of the bulk soil. In fact, this is one of the

principal theories of humic matter formation, that lignin is a precursor to humic matter

(Stevenson 1994). The sieved material contained only the fibric material of macrophytes. That

the lignin content of this material increased with depth reflects the loss of other plant

polysaccharides from the fiber, such as cellulose, hemicellulose, and sugars, thus leaving only

lignin. In summary, even though the lignin content of the soil declined with respect to depth, the

plant fiber that remained became enriched in lignin.

Live plant biomass at both stations was enriched in both 13C and 15N, relative to the

surficial detrital plant litter (Fig. 4-7). The method of analysis used in this study (i.e. analysis of

intact plant fiber) has the advantage of revealing the isotopic signature of only the principal

vascular plants that colonized the site at that depth interval. Thus, extraneous carbon from

carbonates and percolating organic acids from other depths do not contribute to the signature.

Biochemical transformations of carbon and nitrogen tend to discriminate against the heavier 13C

isotope, or fractionate carbon isotopes (Peterson and Fry 1987). Thus, plant biomass is depleted

in 13C relative to atmospheric carbon dioxide, which has a 613C Of approximately -8%o. The

isotopic signatures of aboveground plant biomass collected in this study are typical of results

observed in other studies of wetland emergent plants (Meyers 1994, Keough et al. 1996). Above-

ground samples of Typha collected from the nutrient impacted station averaged -29.1%o (10.2

SD), whereas Cladium samples collected from the unimpacted station averaged -26.7%o (10.3









SD). The 613C COntent of soil at impacted station Cl increased from -29%o in surface soil to -

26%o at a depth of 50 cm, with little additional change to the base of the core. Increases at the

unimpacted station were more modest, with 813C increasing only 1%o in the upper 50 cm of soil.

The two stations showed distinctly different isotopic signatures to a depth of approximately 50

cm, for 13C and 80 cm for 15N. This probably reflects a long-term difference in plant

communities at the two stations. Currently, the community at station B4 is mostly an emergent

marsh PanicuntPPP~~~~PPP~~~PPP hentitonzon Cladium jantaicense community whereas the Cl station is a deeper

slough community dominated by Utricularia spp., Typha latifolia, and Nynaphaea odorata.

Enrichment in the heavier carbon isotope with increasing depth may be due to several factors.

First, methanogenesis produces methane that tends to be depleted in 13C, Or light. Gu and

Schelske (2004) attributed extreme enrichment in 13C in lake phytoplankton to conversion of

methane to CO2 in the water column of a hypereutophic lake. Over time, this would tend to

enrich the remaining organic matter in the heavier isotope. Both chemical and biological

reactions fractionate isotopes such that the substrate becomes progressively enriched in the heavy

isotope, while products are enriched in the light isotope. Another potential cause of increasing

downcore 613C iS a gradual shallowing of the marsh that promoted a shift from submerged

macrophytes, to the emergent community prevalent today. However, other results from this study

favor the first hypothesis, that of gradual processing and loss of the light isotope. For example,

lignin was observed to increase with increasing depth and this is not consistent with a greater

presence of submerged vegetation. Both C:P and C:N also increased downcore and this is also

inconsistent with a historical community that had more submerged vegetation. Lastly, physical

examination of fresh soils revealed the presence of fibric peat to the base of the core and this is

further evidence of a relatively stable community (Appendix A). The difference in composition









between plant and litter was most dramatic for 15N, with plant litter containing essentially an

atmospheric 1N content (6 "N = 0). This may reflect extensive colonization of the detrital

material by N2-fixing cyanobacteria. Plant litter was submerged at the time of sampling and

therefore had probably become colonized with epiphytic algae and enriched in microbial

biomass.

There was a slight enrichment in 13C at the Typha-dominated station and a decline in 1N

approximately mid-core, compared to the Cladium-dominated station, perhaps reflecting a shift

in plant communities at respective stations. Carbon-14 dates at both stations indicate that this

change occurred approximately 1000AD (Fig. 4-8, Table 4-2). As mentioned previously, this

period was characterized by a long-term tropical drought that may have led to changes in marsh

plant communities.

An extended dry period, perhaps several hundred years, would have led to a dramatic loss

of soil organic matter and a lowered marsh surface elevation. Upon reflooding, the shallow water

adapted community would have been replaced by deeper water adapted plants, such as

Nymphaea and others. An abrupt increase in age of plant macrofossils is apparent at a depth of

between 70 and 90cm. This sudden increase in carbon age has been observed at other locations

in BCMCA (M. Fisher, unpublished data). There are two plausible explanations for this sudden

increase in the age of BCMCA peat: little or no accretion for almost a millennium, or (more

likely) loss of carbon due to an extended dry period. This marsh dryout would have been

accompanied by greatly accelerated aerobic peat decomposition, a peat fire, or both. In any case,

once normal rainfall patterns returned, a much deeper marsh community would have developed

and this change seems to be suggested by the isotopic evidence.









Conclusions

Results from this study provide information on the biogeochemical transformations that

organic matter undergoes after it is deposited on the soil surface. Peat cores taken to a depth of

approximately 1.5-m covered a soil age range that dated to approximately 3000 years before

present.

There is ample evidence that substantial changes occur both in organic matter composition,

and organic P. The fluorescence and spectrophotometric procedures indicated a progression to

more humified material, but not necessarily a concomitant change in molecular weight of

extractable organic matter. Increasing humic character with increasing depth observed in the

fluorescence experiments and the lack of increasing humic-extractable P (Chapter 2) would seem

to indicate that P is not closely associated with formation of humic compounds. Nuclear

magnetic resonance analysis of carbon shows a decline in plant carbohydrates with respect to

depth and an increase in aromatic compounds. Nuclear magnetic resonance of phosphorus

showed stability of diester-P compounds, and an absence of inositol hexaphosphate. This

contrasts sharply with upland soil 32P-NMR work, where diester-P is has been shown to quickly

degrade, and inositol phosphates are the dominant P-storage compound. Organic forms of

phosphorus increased with respect to depth.

Lignin fractionation showed that the lignin content of soil fibric material increases with

age, while cellulose and hemicellulose decline. Stable isotopic composition seems to indicate a

relatively stable plant community with a community shift approximately 1000 years ago, perhaps

due to an extended drought. These events would have had important consequences for the flora

and fauna of the marsh, the organic remains of these organisms, and thus the peat deposit itself.

Results from this study suggest that P does not behave as a conservative compound. It is

subject to the same biogeochemical processes that dramatically alter soil organic matter. These










processes act to transform the original floral and faunal material into inorganic and organic

compounds that are progressively less and less similar to the original material. These processes

lead to loss of labile forms of P, and an increasing proportion of stable organic phosphorus

compounds. Comparative studies that contrast surficial peat soils to their deeper counterparts

need to consider the effects that soil forming processes have had on soil properties, particularly

in the case of organic soils.









Table 4-1. Results of 14C AMS dating. Values in parentheses represent standard error of
analytical dating procedure. "YBP" years before present. Radiocarbon ages
have been adjusted for varying atmospheric 14C COncentration using calibration
program CALIB 5.01.
Station Soil Depth Sample Description Age, YBP 613C
B4 -35 Seed >1800 AD -21.2
B4 -65 Woody husk 1550 AD (f100) -24.2
B4 -85 Panicum root 1440 AD (f50) -22.9
B4 -115 Stem fragment 95 AD (f20) -26.4
B4 -125 Pine needle 1150 BC (f125) -24.1
B4 -145 Seed husk 950 BC (f60) -22.6
Cl -35 Grass blade >1800 AD -21.7
Cl -55 Seed pod case 1415 AD (f10) -24.7
Cl -75 Grass blade 985 AD (f60) -25.5
Cl -85 Twig 780 BC (f10) -26.4
Cl -95 Nuphar rhizome 825 BC (f25) -22.0
Cl -125 Nuphar rhizome 1290 BC (f30) -24.5
Cl -135 Wood fragment 1150 BC (f100) -19.4
Cl -145 Wood fragment 1650 BC (f30) -22.9












Humic + Fulvic Fulvic Acid
20
Plant tissue

0 Detritus


-20


-40


0 -60


t5 -80


-1 00


-1 20
SImpacted
SUnimpacted
-1 40


0 2 4 6 8 10 12 0 2 4 6 8 10 12

HIX




Figure 4-1. The humification index (HIX) ofNaOH-extracted peat samples. Error bars represent
the standard deviation of three soil extracts of the same sample. Panel A. reflects the
HIX of the whole extract, whereas panel B. has had humic materials removed by
acidic precipitation, leaving only fulvic compounds. Station B4 = nutrient
unimpacted, Station Cl = nutrient impacted.













0


-20


-40


-60


-80


-100


-120


-140


0 2 4 6 8 10 12 14

E4:E6


+MW

"' I fulvic

humicc


Figure 4-2. Ratio of absorbance at 465 and 665 nm (E4:E6) of NaOH peat extracts at stations Cl
and B4 in BCMCA. Station B4 = nutrient unimpacted, Station Cl = nutrient
impacted.

















0e

-20-



-40 -1




E-60-



o -80 -1 \



-1 00 -



-$ Total O-alkyl (lignin, cellulose)
-1 20-
-- -Alkyl-C (fatty acids, waxes, alkanes)
M-- Total Aromatic

-1 40
0 50 100 150 200 250

Carbon, g/kg


Figure 4-3. Results of solid-state 13C-NMR spectroscopy of dried, ground peat samples collected
from station B4 on February 11, 2003. Depicted is the soil carbon concentration
associated with each carbon group, not soil concentration of respective compounds.









109





MV


VM






MS V


-20



-40


-60



-80


Phosphate
Monoester-P
Diester-P


-1 00


-1 20 -1




0 10


Phosphorus fraction, % of TP

Figure 4-4. Results of solution 31P-NMR spectroscopy ofNaOH-extracted samples collected
from station B4 (unimpacted) on February 11, 2003. Note approximately equal
concentration of mono and diester P.


























-- n~L......~... _.. 10-20




20-30


30-40


40-50



~ 50-60

a..,....70-80

80-90


,..,. 90-1 00


L~c*rr*~.~nrluLrruln~l~lrrrm~WYLhlL~~hn
~~


0, 2 Soil Depth, cm
C O

0-10


10 8 6 4 2 0 -2-
Chemical shift (ppm)


4 -6 -8


100-110


Figure 4-5. The 3P-NMR spectra of 0.5M NaOH extracts of soil samples from Blue Cypress
Marsh Conservation Area, station B4 (unimpacted).


































0 5 10 15 20 25
Cellulose, %

O
OO
O





O



O *
O *



OO

O


0 20 40 60 80 100
Soluble carbon, %


0*
O
*O
O *



SOO
O
*O


B4 (unimpact) C
C1 (Impact) Q g
O *
O *


O
09

O



C)

O

0

OO
OS
()


0)


120 i


0 20 40 60 80
Lignin, %


100


OO


O *


O



O



O
0-


OO

O


0 10 20 30 40 50
Hemicellulose, %


Figure 4-6. Results of lignin fractionation of plant fiber. Fiber was separated from bulk peat
samples that were collected from BCMCA stations B4 and C1. Note differing
horizontal scales. Percentages reflect proportion of total organic matter present in peat
fibric materials.


















Typha Cladium


20


-20


-40




-*- Unimpace
-80 _--o- impacted


-1 00


-1 20


-1 40



-30 -28 -26 -24 -22 -2 0 2 4

813(' 615N


Figure 4-7. Stable isotopic composition of plant fiber that was separated from bulk peat samples
collected from BCMCA stations B4 (unimpacted) and Cl (impacted).




























-20-

-* Unimpacted
-401 ~- Impacted


-60-






-100-


-120-


-140-



0 1000 2000 3000 4000

1C Age, YBP





Figure 4-8. Uncalibrated 14C age of plant macrofossils from selected depths at stations B4 and
Cl in BCMCA (YBP = years before present).









CHAPTER 5
CONCEPTUAL MODEL OF PHOSPHORUS BURIAL INT WETLANDS, CONCLUSIONS,
AND FUTURE RESEARCH NEEDS

Conceptual Model of Soil Phosphorus Transformations.

Conceptual models are useful in summarizing, communicating, and gaining consensus on

significant inputs that are thought to influence a given process. For wetlands, processes leading

to long-term storage of phosphorus in wetland soils can be depicted as a system of stocks and

flows, or storage and transfers (Fig. 5-1). In this model, most of the flows among storage are

hypothesized to be bidirectional, reflecting a dynamic equilibrium among fractions. The

exception to this is the proportion of C in the residual organic fraction ("recalcitrant P"), which is

shown here to only increase with time. This was the only fraction in the sequential chemical

fractionation that was shown to consistently increase with time. Principal above and below

ground storage and transfers are given in Table 5-1. Primary production and microbial P

represent the only living biomass pools and correspond to living plant tissue (roots and shoots)

and microbial biomass P. Relatively labile Po consists of sugar phosphates, nucleic acids (DNA,

RNA), nucleotides (ATP), sugar phosphates, polyphosphates, and other immediate by-products

of cellular decomposition. These compounds are rapidly available to the microbial community

after exocellular enzymatic hydrolysis, and also through direct incorporation back into cellular

biomass. Through reactions that are mostly unknown, but thought to be abiotic, some proportion

of this pool is transferred into humic- and fulvic-like compounds and are illustrated as unknown

reactions in the conceptual model. These pools are hypothesized to be in a state of dynamic

equilibrium, therefore exchanges are first-order processes. Ultimately, a proportion of the humic

pool takes on increasing phenolic characteristics and a molecular geometry such that it is largely

inaccessible to enzymatic attack, especially under flooded conditions. Inorganic pools consist of

the ortho-P monomer, amorphous P compounds such as P associated with iron (Fe) and










aluminum (Al) hydroxides, and mineral P forms. These pools are regulated by the reaction

kinetics of sorption and crystallization. Soluble inorganic P is available for plant uptake,

incorporation into microbial biomass, diffusion and loss from the system, or sorption to Fe and

Al hydrolxides. This conceptual model proposed is a simplified depiction of principal pools and

transfers of P in wetland soils. Exchanges among pools are hypothesized only, since there was no

attempt made in this dissertation to measure reaction kinetics or to document transfers.

Conclusions

Wetland soils are typically long-term sinks for carbon and nutrients. Results from this

study provide information on the biogeochemical transformations that organic matter undergoes

after it is deposited on the soil surface. Peat cores taken to a depth of approximately 1.5 m

spanned a soil age that dated to approximately 5000 years before present. There is ample

evidence that substantial changes occur both in organic matter composition, and organic P. The

fluorescence and spectrophotometric experiments indicated a progression to more humified

material, but not necessarily a concomitant change in molecular weight of extractable organic

matter. Nuclear magnetic resonance analysis of carbon shows a decline in plant carbohydrates

with respect to depth and an increase in aromatic compounds. Nuclear magnetic resonance of

phosphorus showed stability of diester-P compounds, and an absence of inositol hexaphosphate.

This contrasts sharply with upland soil 31P-NMR work, where diester-P has been shown to

quickly degrade, and inositol phosphates are the dominant P-storage compound. The proportion

of organic to inorganic P increased with respect to depth at two stations, remained the same at

one, and declined at the last station.

Lignin fractionation showed that the lignin content of soil fibric material increases with

age, while cellulose and hemicellulose decline. Stable isotopic composition seems to indicate a









relatively stable plant community with a community change approximately 1000 years ago,

perhaps due to an extended drought.

Results from this study suggest that P does not behave as a conservative compound. It is

subject to the same biogeochemical processes that dramatically alter soil organic matter. These

processes act to transform the original floral and faunal material into inorganic and organic

compounds that are progressively less similar to the original material and lead to loss of labile

forms of P, with an increasing proportion of stable organic phosphorus compounds. This

progression is consistent with the pedogenic model proposed by Walker and Syers (1976). Their

model, based on upland soil chronosequences, showed long-term depletion of P, asymptotically

approaching minimum values for respective P fractions. They also found that with weathering,

Po became an increasingly larger proportion of soil total P. Their findings were supported for

two of the wetland sites in this study, though the ratio of Pi to Po increased at one site and

remained essentially unchanged at another. The most likely reason for this minor discrepancy

between the two studies reflects the principal source of P in uplands, verses wetlands. Natural

(non-agricultural) uplands begin the pedogenic process with the principal pool of P in primary

soil minerals. Weathering reduces this pool, while increasing storage of Po. Wetlands are situated

at bottom of the landscape, and thus receive nutrient subsidies from higher areas in the

topographic gradient. Organic matter produced mostly in situ accumulates due to slowed

decomposition and appears to be the principal form of stored P at all stages of the pedogenic

process in peat dominated wetlands. It is clear from the mathematical model results presented in

this study that the two principal long-term storage pools in peat-dominated wetland soils are the

P associated with the unextractable fraction and humic P (Fig. 5-2).









There are three main mechanisms that could account for the soil phosphorus profiles

observed in this study. Declining soil P with depth could be a result of ecosystem succession, to

plants that were capable of retaining P in more resistant compounds. Second, ecosystem loading

rates may have gradually increased over the millenia as a result of cultural eutrophication. Last,

weathering of organic P and organic P mineralization, with enrichment in refractory P (and

refractory carbon), accompanied by subsequent diffusion to the soil surface and advection via

surface transport from the wetland may have led to the observed declines. Results from this

dissertation suggest that this last mechanism is the principal reason for the observed declines in

soil P with respect to increasing depth.

Future Research

This research yielded little insight into potential transfers among P pools. For example, P

(and carbon) diagenesis is thought to involve progressive soil enrichment in higher molecular

weight, more humified substances. However, whether this enrichment involves preservation of a

constant initial mass of humic-like materials, or transformation of biological molecules into

humic substances is not known. This same question holds for other fractions as well. It is

possible that there exists a dynamic equilibrium among P fractions, and that P pools constantly

adjust to maintain this shifting equilibrium.

Little is known about the chemical characteristics of the residual fraction remaining behind

after sequential fractionation. The phosphorus and carbon associated with this material is thought

to be highly refractory, though this is speculation based on the premise that resistance to

chemical extraction is analogous to environmental stability. Advances in solid-state 31P-NMR

may be helpful in characterizing this material. Further investigation is needed to determine why

the thermal techniques investigated in this dissertation were able to distinguish between labile

and recalcitrant soil organic P. The hypothesis that thermal stability was related to environmental









recalcitrance seemed to hold, yet the nature of the compounds that confer this stability is

unknown. Lastly, ecological studies that contrast surficial peat soils to their deeper counterparts

need to consider the effects that soil forming processes have had on soil properties, particularly

in the case of organic soils.

















































V _I


Table 5-1. Components of conceptual model depicting phosphorus pools and transfers in
peat soils.
Model Element Description
Pool
Primary production Plant biomass P
Microbial P Cell walls, nucleic acids, ATP/ADP
Soluble Po Labile organic P
Fulvic P assoc. with Fulvic acid
Humic P assoc. with Humic acid
Recalcitrant P in long-term storage compounds. Stable.
PO4 IHOrganic P in solution
Amorphous P Non crystalline inorganic, iron, calcium, and aluminum P
Mineral Crystalline P


Transfer
Uptake
External transfer
Sorption
Crystal
PiMin

PoMin


UnRxl

UnRx2

UnRx3

Burial


Plant uptake
Transfer of P into/out of system
Sorption of P to inorganic compounds
Crystallization
Mineralization of organic matter by microbial community to
PO4
Mineralization of labile organic P organic matter by microbial
community to labile organic compounds, by microbial
community (or uptake of Po by them)
Unknown biotic or abiotic reaction leading to exchanges
between labile organic and fulvic compounds
Unknown biotic or abiotic reaction leading to exchanges
between of humic and fulvic compounds
Unknown biotic or abiotic reaction leading to exchanges
between humic and labile organic compounds
Long-term transfer to very recalcitrant P












External
Transfers


Primary
Production


Amorphous P






Shallow soil (rapid rates)

Deep soil (slow rates)


Microbial P


Soluble Po


Humic


UnRx2


Burial


Mineral


Recalcitrant P


Figure 5-1. Conceptual model of dynamic exchanges among soil phosphorus pools in wetland
soils.


























* Mbicroia


+ Humic


c~F~-
~~t~~tctt-t+
___________________ ___t~~ttt~ ~


XWwyvvvvwyvvvywyvvvywyvvvywyvvvvywyvvvYywyvvy


200


150 -


100 -


50 -


x Labile Pi
- Residual


0 500


1000 1500 2000 2500 3000


Time, yrs


Figure 5-2. Results of mathematical modeling of long-term declines of five phosphorus fractions
in Blue Cypress March Conservation Area.


1 *+++


X









APPENDIX
PHYSICAL SOIL DESCRIPTION

Table 1. Physical description of soil samples collected in July, 2002 from Water Conservation
Area 2A and Blue Cypress Marsh Conservation Area.
Station Depth Description
Water Conservation Area 2A

El Rep I 0-10 Very fibrous, spongy peat.
Many plant fragments; brown
10-20 Slightly drier peat; fibrous,
brown
20-30 Similar to 10-20, but more
compacted
30-40 Fibers becoming shorter,
compacted peat
40-50 Fibrous brown peat
50-60 Fibrous brown peat, slightly
darker, increase in fine
organic matter (OM)
60-70 Fibrous brown peat, slightly
darker, increase in fine
organic matter (OM)
70-80 Fibrous brown peat, slightly
darker, increase in fine
organic matter (OM)
80-85 Approx. 2.5 cm brown coarse
sand, overlain by 3.5 cm peat
El Rep II 0-10 Very fibrous, wet, many plant
parts
10-20 Sim. to above, many poorly
decomposed plant parts. Large
live Typha root; drier
20-30 More decomposed, still many
identifiable plant parts,
becoming fibrous peat
30-40 Slightly decomposed, dark
brown fibrous peat
40-50 Dry, fibrous brown compacted
peat





I


Depth
50-60


Table A-1. Continued
Station
El Rep II


Description
Very compacted dry brown
fibrous peat
Darker, more decomposed
peat
Similar to above, slightly
more watery
Compacted dry slightly sandy
peat
Compacted dry slightly sandy
peat
Very wet, poorly decomposed
peat, many identifiable plant
parts
Fibrous, drier peat
Fibrous, somewhat dry peat
Fibrous, somewhat dry peat
(dark brown; all sections)
fibrous, wetter peat. This was
likely the break between
successive core drives
Compacted dry fibrous peat.
Slightly lighter brown
Fibrous peat
Fibrous peat, slightly darker
Slightly sandy dry peat. Many
Eine roots, even at this depth
Fibrous rooty almost black
peat
Fibrous rooty almost black
peat. Less moisture
Fibrous, many Eine roots,
almost black
Fibrous, many Eine roots,
almost black
Similar to above, less roots
More decomposed,
considerably higher moisture
content
Well decomposed, still
somewhat fibrous, high
moiture content
Very well decomposed muck,
black, increased moisture.


60-70

70-80

80-90

90-97

0-10


10-20
20-30
30-40

40-50


50-60

60-70
70-80
80-90

0-10

10-20

20-30

30-40

40-50
50-60


60-70


70-80


El Rep III


E5 Rep I






















































10-20

20-30
30-40
40-50
50-60
60-70

70-80


Table A-1. Continued
Station
E5 Rep I


Depth
80-90


Description
Muck, slight increase in
amount of Eine roots and fibers
Muck, high moisture content
Decrease in moisture, fibrous
muck
Slightly sandy peat, color
change to light brown
Very sandy peat
Silty sand, high OM, some
identifiable ancient plant
fragments
Dark very fibrous peat
Dark very fibrous peat.
Slightly drier
Dark very fibrous peat.
Slightly drier
Dark very fibrous peat.
Moisture content same as
above
Some very large, very
decomposed woody material +
peat
Piece of wood 2" diam. and
peat
Well decomposed, almost
muck, somewhat dry
Mucky, fibrous peat
Sandy peat, slightly red in
color
Peat intermixed with bands of
sand

peaty sand
Very rooty peat, many plant
parts
Rooty peat, drier, cladimm
roots
Less rooty, drier, fibrous peat
Fibrous peat, slightly mucky
Mucky peat
Mucky peat
Mucky peat, increased water
content
Fibrous, mucky peat


E5 Rep II


90-100
100-110

110-120

120-130
130-140


0-10
10-20

20-30

30-40


40-50


50-60

60-70

70-80
80-90

90-100

100-110
110-120
0-10


E5 Rep III



















































10-20

20-30
30-40
40-50

50-60
60-70
70-80
80-90
90-98


Table A-1. Continued
Station


Depth
80-90


Description
Fibrous, slightly mucky, drier
peat
Orange, fibrous peat
Sandy peat
Peaty sand

Much detrital matter, high
water content
Fibrous peat, much drier
Fibrous peat, much drier
Drier still, compacted fibrous
peat
Fibrous peat
Mucky fibrous peat, increases
in water content
Well-compacted fibrous peat,
decrease in water content,
lighter brown
Well-compacted fibrous peat,
decrease in water content,
lighter brown, low water
content for peat
Dry light brown peat, many
plant fragments
Darker mucky peat, increase
in water content
Fibrous brown peat
Fibrous brown peat
Fibrous brown peat
Fibrous brown peat
Fibrous, poorly compacted
peat. Many large plant
fragments
Much drier, slightly
compacted peat
Brown fibrous peat
Brown fibrous peat
Brown fibrous peat, slightly
darker
Brown fibrous peat
Brown fibrous peat
Compacted brown fibrous peat
Compacted brown fibrous peat
Compacted brown fibrous peat


90-100
100-110
110-120
Blue Cypress Marsh Conservation Area
ep I 0-10

10-20
20-30
30-40

40-50
50-60

60-70


70-80



80-90

90-100

100-110
110-120
120-130
130-135
ep II 0-10


Cl Re


Cl Re









Table A-1. Continued
Station Depth Description
Cl Rep III 0-10 Fibrous poorly compacted
peat. Mostly detrital
10-20 Much drier, more compacted.
Peat decreasing in plant fibers
20-30 Compacted peat, smaller
fibers
30-40 Darker, muckier peat
40-50 Increasing water content, not
as compacted
50-60 Drier, dense mucky peat
60-70 Wet poorly consolidated peat,
many large fragments
70-80 Similar to above, many large
fragments
80-90 Brown, dry compacted peat
90-100 Brown, dry compacted peat.
Sandy silty peat at base of
core
B4 Rep I 0-10 Very rooty consolidated peat,
no detrital material
10-20 Very rooty consolidated peat,
no detrital material. Fibrous
consolidated peat
20-30 Rooty fibrous peat
30-40 Rooty fibrous peat
40-50 Rooty fibrous peat
50-60 Slightly wetter consolidated
peat
60-70 Fibrous rooty peat
70-80 Rooty peat, fibrous; increased
muck content
80-90 Mucky peat
90-100 Mucky peat
100-110 Mucky peat
B4 Rep II 0-10 Very rooty somewhat
consolidated peat little or no
detritus
10-20 Rooty fibrous consolidated
peat
20-30 Rooty fibrous consolidated
peat
30-40 Rooty fibrous consolidated
peat





Table A-1. Continued
Station


Depth
40-50


Description
Rooty fibrous consolidated
peat, increasing muck content
Rooty fibrous consolidated
peat, slightly mucky
Rooty fibrous consolidated
peat, slightly mucky
Mucky peat, slightly fibrous
Mucky peat, slightly fibrous,
very mucky, well decomposed
Mucky peat, slightly fibrous,
very mucky, well decomposed
Mucky peat, slightly fibrous,
very mucky, well decomposed

Very rooty (Panicum)
consolidated peat
Rooty consolidated peat
Fibrous peat
Fibrous peat
Rooty fibrous peat
Slightly mucky fibrous peat
Mucky fibrous peat
Mucky fibrous peat
Mucky fibrous peat
Mucky fibrous peat


50-60

60-70

70-80
80-90

90-100

100-110


B4 Rep III


0-10

10-20
20-30
30-40
40-50
50-60
60-70
70-80
80-90
90-98




























90-100
100-110
110-120
120-130
0-10

10-20

20-30

30-40
40-50
50-60

60-70

70-80

80-90
90-100
100-110
110-120
120-130
120-140
140-150


:h 2004 from Blue Cypress


Table A-2. Physical description of soil samples collected in Marc
Marsh Conservation Area.
Station Depth
B4 0-10


Description
Loosely consolidated detrital
peat
Fibric, slightly more
consolidated peat
Rooty, very fibric peat
Rooty, very fibric peat
Rooty, very fibric peat
Rooty, fibric, small increase in
OM
Rooty, fibric peat
Rooty, fibric peat
Noticeable increase in very
fine OM
Rooty, fibric peat
Rooty, fibric peat
Rooty, fibric peat
Rooty, fibric peat
Loosely consolidated detrital
peat; rooty (live)
Slightly more consolidated
peat
Slightly compacted detrital
peat
Fibric peat
Fibric peat
Big increase in fine black OM
in matrix of peat
Similar to above, even more
fine OM
Decline in fine OM; base of
slough?
Fibric peat
Lighter brown fibric peat
Reddish brown fibric peat
Reddish brown fibric peat
Reddish brown fibric peat
Reddish brown fibric peat
Similar to above, some
charcoal


10-20

20-30
30-40
40-50
50-60

60-70
70-80
80-90










LIST OF REFERENCES


Anderson, J. M., 1976. An ignition method for determination of total phosphorus in lake
sediments. Water Research 10:329-331.

Baldock, J.A., J.M. Oades, P.N. Nelson, T.M. Skene, A. Golchin, and P. Clarke. 1997. Assessing
the extent of decomposition of natural organic materials using solis-state 13C NMR
spectroscopy. Australian J. Soil Research. 35:1061-1083.

Bartkowiak, M., and R. Zakrzewski. 2004 Thermal degradation of lignins isolated from wood. J.
Thermal Analysis and Calorimetry. 77:295-304.

Binford, M.W., E.S. Deevey, and T.L. Crisman. 1983. Paleolimnology: an historical perspective
on lacustrine ecosystems. Ann. Rev. Ecol. Syst. 14: 255-286.

Brenner, M., C.L. Schelske, and L.W. Keenan. 2001. Historical rates of sediment and nutrient
accumulation in marshes of the Upper St. Johns River Basin, Florida, USA. J. Paleolimnol.
26:241-247.

Brezonik, P.L., and D.R. Engstrom. 1998. Modern and historic accumulation rates of phosphorus
in Lake Okeechobee, Florida. J. Paleolimnol. 20:31-46.

Brookes, P.C., D.S. Powlson, and D.S. Jenkinson. 1982. Measurement of microbial biomass
phosphorus in soil. Soil Biol. Biochem. 14:319-329.

Buck, D. G., M. Brenner, D.A. Hodell, and J.H. Curtis. 2005. A late Holocene record of climate
variability from hypersaline Lago Enriquillo, Dominican Republic. Presentation at 2005
American Society of Limnology and Oceanography meetings, Salt Lake City, Utah.

Burrough, P.A. 1996. Opportunities and limitations of GIS based modeling of solute transport at
regional scale. p 19-38. In: Applications of GIS to the Modeling of Non-point Source
Pollutants in the Vadose Zone. SSSA Special Pub. No. 48. SSSA. Madison, WI.

Cade-Menun, B.J., and C.M. Preston. 1996. A comparison of soil extraction procedures for 31P
NMR spectroscopy. Soil Science. 161: 770-785.

Cade-Menun, B.J., C.W. Liu, R. Nunlist, and J. G. McColl. 2002. Soil and litter 31PNMR
spectroscopy: extractants, metals and phosphorus relaxation times. J. Eny. Qual. 31:457-
465.

Cade-Menun. 2005. Using phosphorus-31 nuclear magnetic resonance spectroscopy to
characterize organic phosphorus in environmental samples. In: Organic Phosphorus in the
Environment. pp. 21-44. B.L. Turner, E. Frossard, and D.S. Baldwin (eds.). CABI
Publishing. Cambridge, MA.

Carignan, R., and R.J. Flett. 1981. Postdepositional mobility of phosphorus in lake sediments.
Limnol. Oceanogr. 26:361-366










Celi, L. and E. Barberis. 2005. Abiotic stabilization of organic phosphorus in the environment.
In: Organic Phosphorus in the Environment. pp. 113-132. B.L. Turner, E. Frossard, and
D.S. Baldwin (eds.). CABIPublishing. Cambridge, MA.

Chang, S.C., and M.L. Jackson. 1957. Fractionation of soil phosphorus. Soil Sci. 84: 133-144.

Chen, Y., N. Senesi, and M. Schnitzer. 1977. Information provided on humic substances by
E4/E6 ratios. Soil Sci. Soc. Am. J. 41:352-358.

Clark, M.W. 2000. Biophysical characterization of floating wetlands (flotant) and its influence
on vegetative succession of a warm-temperature aquatic ecosystem. Doctoral Dissertation.
University of Florida.

Cooper, W.T., J.M. Llewelyn, G.L. Bennett, A.C. Stenson, and V.J. Salters. 2005. Organic
phosphorus speciation in natural waters by mass spectrometry. In: Organic Phosphorus in
the Environment. pp. 45-74. B.L. Turner, E. Frossard, and D.S. Baldwin (eds.). CABI
Publishing. Cambridge, MA.

Cox, L., R. Celis, M.C. Hermosin, J. Comnejo, A. Zsolnay, and K. Zeller. 2000. Effect of organic
amendments on herbicide sorption as related to the nature of dissolved organic matter.
Environ. Sci. Technol. 34: 4600-4605.

Craft, C.B., and C.J. Richardson. 1993. Peat accretion and phosphorus accumulation along a
eutrophication gradient in the northern Everglades. Biogeochem. 22:133-156.

Craft, C.B., and C.J. Richardson. 1998. Recent and long-term organic soil accretion and nutrient
accumulation in the Everglades. Soil Sci. Soc. Am. J. 62:834-843.

Crews, T.E., K. Kitayama, J.H. Fownes, R.H. Riley, D.A. Herbert, D. Mueller-Dombois, P. M.
Vitousek. 1995. Changes in soil phosphorus fractions and ecosystem dynamics across a
long chronosequence in Hawaii. Ecology. 76: 1407-1424.

Crisman, T.L. 1978. Interpretations of past environments from lacustrine animal remains. pp.69-
101. In: Biology and Quartemnary Environments. 264 pp. D. Walter (ed.). Australian
Academy of Sciences. Canberra, Australia.

Davis, J.H. 1943. The natural features of southern Florida, especially the vegetation and the
Everglades. Geol Bull., 25, Florida Dept. of Conservation, FL Geol. Sury. Tallahassee.
311pp.

Davis, S.M. 1991. Growth, decomposition, and nutrient retention of Cladium jamaicense Crantz
and Typha domingensis Pers. in the Florida Everglades. Aquat. Bot. 40: 203-224.

DeBusk, W.F. and K.R. Reddy. 1998. Turnover of detrital organic carbon in a nutrient impacted
Everglades Marsh. Soil Sci. Soc. Am. J. 62: 1460-1498.










DeBusk, W.F., K.R. Reddy, M.S. Koch and Y. Wang. 1994. Spatial distribution of soil nutrients
in a northern Everglades marsh: Water Conservation Area 2A. Soil Sci. Soc. Am J.
58:543-552.

Devai, I. and R. D. DeLaune. 1995. Evidence for phosphine production and emission from
Louisiana and Florida marsh soils. Org. Geochem. 23:277-279.

Dixit, S.S., J.P. Smol, J.C. Kingston, and D.F. Charles. 1992 Diatoms: powerful indicators of
environmental change. Environ. Sci. and Technol. 26:23-33.

Doran, G. 2002. Windover : multidisciplinary investigations of an early Archaic Florida
cemetery. University Press of Florida, Gainesville FL.

Engstrom, D.R., S.P. Schottler, P.R. Leavitt, and K.E. Havens. 2006. A reevaluation of cultural
eutrophication of Lake Okeechobee using multiproxy sediment records. Ecological
Applications. 16:1194-1206.

Felfli, F.F. A.A. Luengo, J. Suarez, and P. Beaten. 2005. Wood briquette torrefaction. Energy for
Sustain. Develop. 9:19-22.

Fertilizer Institute. 2007. http://www.tfi. org/factsandstats/stati stics.cfm. The Fertilizer Institute.
820 First St. N.E. Suite 430, Washington, DC 20002.

Fisher, M. M., K. R. Reddy, and R. T. James. 2005. Internal nutrient loads from sediments in a
shallow, subtropical lake. Lake and Reserv. Manage. 21:338-349.

Florida Department of Environmental Protection. 2004. Integrated Water Quality Assessment for
Florida: 2004 305(b) Report and 303(d) List Update.

Freeman, C., G. Liska, N.J. Ostle, M.A. Lock, B. Reynolds, and J. Hudson. 1996. Microbial
activity and enzymic decomposition processes following peatland water table drawdown.
Plant and Soil. 180:121-127.

Frey, D.G. 1969. The rationale for paleolimnology. Mitt. Internat. Verein. Limnol. 17:7-18.

Gleason, P. J., and P. Stone. 1994. Age, origin, and landscape evolution of the Everglades
peatland. pp. 149-197 hz S. M. Davis and J. C. Ogden (eds.) Everglades The ecosystem
and its restoration. St. Lucie Press, Delray Beach, FL.

Gleason. P.J. 1974. Chemical quality of water in Water Conservation Area 2A and associated
canals. Technical Report 74-1. South Florida Water Management District. West Palm
Beach, Florida.

Grunwald, S., K.R. Reddy, J.P. Prenger, and M.M. Fisher. 2007. Modeling the spatial variability
of biogeochemical soil properties in a freshwater ecosystem. Ecological Modelling.
201:521-535.










Hedin, L.O. P.M. Vitousek, P.A. Matson. 2003. Nutrient losses over four million years of
tropical forest development. Ecology. 84:223 1-225 5.

Hieltj es, A.H., and L. Lijklema. 1980. Fractionation of inorganic phosphates in calcareous lake
sediments. J. Environ. Qual. 9:405-407.

Hillel, D.A. 1986. Modeling in soil physics: A critical review. pp 35-42. In: Future
Developments in Soil Science Research. SSSA Misc. Publ. SSSA, Madison WI.

Hodell, D.A., M. Brenner, and J.H. Curtis. 2005. Terminal Classic drought in the northern Maya
lowlands inferred from multiple sediment cores in Lake Chichancanab (Mexico).
Quaternary Sci. Rev. 24:1413-1427.

Hupfer, M., B. Rube, P. Schmieder. 2004. Origin and diagenesis of polyphosphate in lake
sediments: a 31P-NMR study. Limnol. Oceanogr. 49:1-10.

Ivanoff, D.B., K.R. Reddy, and S. Robinson. 1998. Chemical fractionation of organic
phosphorus in selected Histosols. Soil Science. 163:36-45.

Jenny, H. 1941. Factors of soil formation. Mcgraw Hill, New York, NY.

Kenney, W.F., C.L. Schelske, and M. Brenner. 2000. Changes in polyphosphate sedimentation: a
response to excessive phosphorus enrichment in a hypereutrophic lake. Can. J. Fish. Aquat.
Sci. 58: 879-887.

Keough, J.R., M.E. Sierszen, and C.A, Hagley. 1996. Analysis of a Lake Superior food web with
stable isotope techniques. Limnol. Oceanogr. 41:136-146.

Levesque, M. and H. Dinel. 1978. Applicability of thermal methods for characterization of peats
and plants. Geoderma. 20: 201-213.

Lowe, E.F. and L.W. Keenan. 1997. Managing phosphorus-based cultural eutrophication in
wetlands: a conceptual approach. Ecol. Eng. 9: 109-118.

McLatchey, G.P. and K.R. Reddy. 1998. Regulation of organic matter decomposition and
nutrient release in a wetland soil. J. Environ Qual. 27:1268-1274.

Meyers, P.A. 1994. Preservation of elemental and isotopic source identification of sedimentary
organic matter. Chem. Geol. 114:289-302.

Mitsch, W.J. and J.G. Gosselink. 1993. Wetlands. Second edition. Van Nostrand Reinhold. New
York, NY.

Moore, P. A. K. R. Reddy and M. M. Fisher. 1998. Phosphorus flux between sediment and
overlying water in Lake Okeechobee, Florida: spatial and temporal variations. J. Environ.
Qual. 27:1428-1439.










Murphy, J. and J.P. Riley. 1962. A modified single solution method for the determination of
phosphate in natural waters. Anal. Chim. Acta. 27:31-36.

Ohno, T. 2002. Fluorescence inner filtering correction for determining the humification index of
dissolved organic matter. Environ. Sci. Technol. 36: 742-746.

Olila O. G., K.R. Reddy and L. Keenan. 1995. Labile and non-labile pools of phosphorus in
surface waters and soils in the Upper St. Johns River Basin. Final report to the St. Johns
River Water Management District. Contract No. 94D 179.

Page, A.L. 1983. In: Methods of Soil Analysis, Part 2: Chemical and Microbiological Properties.
2nd ed. A.L. Page (ed.). Amer. Society of Agronomy.

Penn, M.R., M.T. Auer, E.L. Van Orman, and J.J. Korienek. 1995. Phosphorus diagenesis in lake
sediments using fractionation techniques. Marine Freshwater Res. 46:89-99.

Peterson, B.J. and B. Fry. 1987. Stable isotopes in ecosystem studies. Ann. Rev. Ecol. Syst.
18:292-320.

Ponnamperuma, F.N. 1972. The chemistry of submerged soils. Adv. Agron. 24:29-96.

Prenger, J.P. and K.R. Reddy. 2004. Microbial enzyme activities in a freshwater marsh after
cessation of nutrient loading. Soil Sci. Soc. Am. J. 68:1796-1804.

Reddy, K. R., and W. H. Patrick, Jr. 1983. Effect of aeration on reactivity and mobility of soil
constituents.1In: Chemical Mobility and Reactivity in Soil Systems (eds. D. W. Nelson, D.
E. Elrick, and K. K. Tanji). SSSA Special Publ. No. 11:11-33. Soil Sci. Soc. Am. J.
Madison, WI.

Reddy, K.R. and R. DeLaune. In press. Chapter 9: Phosphorus In: Biogeochemistry of
Wetlands: Science and Applications.

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

Reddy, K.R., Y Wang, W.F. DeBusk, M.M. Fisher, and S. Newman. 1998. Forms of soil
phosphorus in selected hydrologic units of the Florida Everglades. Soil Sci. Soc. Am. J.
62:1134-1147.

Redfield, A.C., B.H. Ketchum and F.A. Richards. 1963. The influence of organisms on the
composition of seawater. pp. 26-77. In The Sea, Volume 2. M.N. Hill (ed.). Interscience
Publ., New York, NY.

Richardson, C.J. 1985. Mechanisms controlling phosphorus retention capacity in freshwater
wetlands. Science. 228:1424-1427.










Richardson, C.J., S. Qian, C.B. Craft, and R.G. Qualls. 1997. Predictive models for phosphorus
retention in wetlands. Wetlands Ecology and Management. 4: 159-175.

Richter, D.D. H. Lee Allen, J. Li, D. Markewitz, J. Raikes. 2006. Bioavailability of slowly
cycling phosphorus: maj or restructuring of soil P fractions over four decades in an
aggrading forest. Oecologia. 150: 259-271.

Rowell, D.L. 1981. Chapter 7: Oxidation and Reduction. Dr: The Chemistry of Soil Processes.
pp. 401-461. D.J. Greenland and M.H.B. Hayes, eds. John Wiley and Sons.

Rydin, E. 2000. Potentially mobile phosphorus in Lake Erken sediment. Water Res. 34:2037-
2042.

Sharpley, A.N. 2000. Global issues of phosphorus in terrestrial ecosystems. pp. 15-46. Dr
Phosphorus biogeochemistry in subtropical ecosystems. K.R. Reddy, G.A. O'Connor, and
C.L. Schelske (eds.). Lewis Publishers, Boca Raton, FL.

Shumate, B.C., C.L. Schelske, T.L. Crisman, and W.F. Kenney. 2002. Response of the
cladoceran community to trophic state changes in Lake Apopka, Florida. J. Paleolimnol.
27:71-77.

Sjoigersten, Sofie, B.L. Turner, N. Mahieu, L.M. Condron, and P. A. Wookey. 2003. Global
Change and Biol. 9: 759-772.

Smol, J.P. 1992. Paleolimnology: an important tool for effective ecosystem management. J.
Aquatic Ecosys. Health. 1:49-58.

Stevenson, F.J. 1994. Humus Chemistry: Genesis, Composition, and Reactions. 2nd Edition.
John Wiley & Sons.

Steward, K.K., and W.H. Ornes. 1975. The autecology of sawgrass in the Florida Everglades.
Ecology 56: 162-171.

Stewart, J.W.B and H. Tiessen. 1987. Dynamics of soil organic phosphorus. Biogeochemistry.
4:41-60.

Stuiver, M. and P.J. Reimer. 1993. Extended 14C database and revised CALB 3.0 14C age
calibration program. Radiocarbon. 35: 215-230.

Tabatabai, M.A. 1982. Soil Enzymes. p. 903-947. h?: Methods of Soil Analysis. Part 2 Chemical
and Microbiological Methods. A.L. Page et al. (eds.). 2nd ed. Agronomy Monogr. 9. ASA
and SSSA. Madison, WI.

Turner, B.L. 2004. Optimizing phosphorus characterization in animal manures by solution
Phosphorus-31 Nuclear Magnetic Resonance spectroscopy. J. Environ. Qual. 33: 757-766.

Turner, B.L., and S. Newman. 2005. Phosphorus cycling in wetland soils: the importance of
phosphate diesters. J. Environ. Qual. 34:1921-1929.










Turner, B.L., M.J. Paphazy, P.M. Haygarth, and I.D. McKelvie. 2002. Inositol phosphates in the
environment. Phil. Trans. R. Soc. Lond. 357:449-469.

Turner, B.L., N. Mahieu, and L.M. Condron. 2003a. 31P MNR spectral assignments of
phosphorus compounds in soil NaOH-EDTA extracts. Soil Sci. Soc. Am. J. 67: 497-510.

Turner, B.L., N. Mahieu, and L.M. Condron. 2003b. The phosphorus composition of temperate
pasture soils determined by NaOH-EDTA extraction and solution 31P MNR spectroscopy.
Organic Geochemistry. 34: 1199-1210.

United States Environmental Protection Agency. 1996. Environmental indicators of water quality
in the United States. EPA-841-R-96-002. USEPA Office of Water (4503f). US
Government Printing Office, Washington DC. 25 pp.

United States Environmental Protection Agency. 1998. 303d List of Impaired Waters.

United States Environmental Protection Agency. 2002. National Water Quality Inventory: 2000
Report. EPA-841-R-02-001. USEPA Office of Water (4503f). US Govemnment Printing
Office, Washington DC.

Urban, N.H., S.M. Davis, and N.G. Aumen. 1993. Fluctuations in sawgrass and cattail densities
in Everglades Water Conservation Area 2A under varying nutrient, hydrologic and fire
regimes. Aquat. Bot. 46: 203-223.

Van Soest, J.P. 1970. Forage and Fiber Analysis Handbook No. 379. USDA. Washington, D.C.

Waksman, S.A. 1936. Humus. Williams and Wilkins Co., Baltimore, MD.

Walker, T.W. and J.K. Syers. 1976. The fate of phosphorus during diagenesis. Geoderma. 15:1-
19.

Wetzel, R.G. 1999. Organic phosphorus mineralization in organic soils. pp. 225-245. In
Phosphorus biogeochemistry in subtropical ecosystems. K.R. Reddy, G.A. O'Connor, and
C.L. Schelske (eds.). Lewis Publishers, Boca Raton, FL..

Whitmore, T.J. 1989. Florida diatom assemblages as indicators of trophic state and pH. Limnol.
Oceanogr. 34: 882-895.

Wright, A.L. and K.R. Reddy. 2001. Phosphorus loading effects on extracellular enzyme activity
in Everglades wetland soils. Soil Sci. Soc. Am. J. 65:588-595.









BIOGRAPHICAL SKETCH

Millard Mathias Fisher, III was born in Hattiesburg Mississippi in 1958 and

moved to Florida at the age of two years. He grew up in St. Augustine, Florida, USA,

where he attended Catholic parochial and high school. He holds a bachelor' s degree in

engineering technology from the University of Central Florida, a bachelor' s degree in

environmental engineering from the University of Florida, and a master' s degree in soil

and water science, also from the University of Florida. In 2007, he completed his doctoral

degree in soil and water science at the University of Florida. Fishing, skiing, kayaking,

and flying are his favorite hobbies.





PAGE 1

1 BIOGEOCHEMICAL TRANSFORMATIONS OF PHOSPHORUS IN WETLAND SOILS By MILLARD M. FISHER, III 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 2007

PAGE 2

2 2007 Millard M. Fisher, III

PAGE 3

3 For Santa A. Fisher

PAGE 4

4 ACKNOWLEDGMENTS I am indebted to many who made this di ssertation possible. Without the support and encouragement of Dr. Lawrence Keenan, this disse rtation would not have been possible. I am very appreciative of the guidan ce and patience of my supervisory committee; Dr. Mark Brenner, Dr. Bill DeBusk, Dr. Ping Hsieh, Dr. Jim Jawitz Dr. Lawrence Keenan and Dr. Ramesh Reddy. I owe a great debt to Mrs. Yu Wang, who tirelessl y assisted me in many phases of this study. Dr. Ben Turner was essential fo r assisting with both carbon a nd phosphorus nuclear magnetic resonance analysis and interpretation of resu lts. Special thanks go to Dr. Sue Newman for arranging transportation into Water Conservati on Area 2A. The research described in this dissertation was funded by the St. Johns River Wate r Management District. Lastly, most of my professional development was under the mentorsh ip of Dr. Ramesh Reddy. I can never repay him for the guidance, patience, and trust he ha s shown me during this dissertation and over the last twenty-five years.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........8 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 INTRODUCTION..................................................................................................................12 Background..................................................................................................................... ........12 Phosphorus in Natural Ecosystems.................................................................................12 Estimating Historical Ecosystem Loading Rates............................................................13 Characterizing Changes in Soil Phosphorus....................................................................16 Overview of the Problem........................................................................................................18 Hypotheses and Objectives.....................................................................................................20 Dissertation Format............................................................................................................ ....20 2 LONG-TERM CHANGES IN SOIL PHOSPHORUS FORMS IN TWO SUBTROPICAL WETLANDS..............................................................................................25 Introduction................................................................................................................... ..........25 Methods........................................................................................................................ ..........28 Site Description...............................................................................................................28 Sampling and Analysis....................................................................................................29 Sequential Organic Phosphorus Fractionation................................................................30 Results and Discussion......................................................................................................... ..32 Soil Properties................................................................................................................ .32 Soil Fractionation Results................................................................................................34 Long-Term Declines of Total Phosphorus in BCMCA...................................................38 Mathematical Model of Cha nges in Phosphorus Fractions.............................................40 Historical Accumulation of Phosphorus..........................................................................43 Conclusions.................................................................................................................... .........45 3 ESTIMATING STABILITY OF ORGANIC PHOSPHORUS IN WETLAND SOILS........64 Introduction................................................................................................................... ..........64 Methods........................................................................................................................ ..........68 Site Description...............................................................................................................68 Sampling and Analysis....................................................................................................69 Results and Discussion......................................................................................................... ..72 Autoclave Extractable Phosphorus..................................................................................72

PAGE 6

6 Pyrolytic Extractable Phosphorus....................................................................................73 Enzymatic Hydrolysis.....................................................................................................75 Conclusions.................................................................................................................... .........76 4 CHARACTERIZATION OF LONG-TERM CHANGES IN ORGANIC PHOSPHORUS IN A SU BTROPICAL WETLAND............................................................84 Introduction................................................................................................................... ..........84 Methods........................................................................................................................ ..........87 Organic Extract Fluorescence..........................................................................................88 Spectrophotometric Absorbance.....................................................................................89 13C Nuclear Magnetic Resonance (13C-NMR)................................................................89 31P Nuclear Magnetic Resonance (31P-NMR).................................................................90 Isotopic Analysis of Plant Fibers.....................................................................................91 Soil Age....................................................................................................................... ....92 Results and Discussion......................................................................................................... ..92 Fluorescence and Spectrophotometric Properties...........................................................92 13C Nuclear Magnetic Resonance (13C-NMR)................................................................95 31P Nuclear Magnetic Resonance (31P-NMR).................................................................98 Isotopic Analysis of Plant Fiber......................................................................................99 Conclusions.................................................................................................................... .......104 5 CONCEPTUAL MODEL OF PHOSPH ORUS BURIAL IN WETLANDS, CONCLUSIONS, AND FUTURE RESEARCH NEEDS...................................................115 Conceptual Model of Soil Phosphorus Transformations......................................................115 Conclusions.................................................................................................................... .......116 Future Research................................................................................................................ ....118 APPENDIX PHYSICAL SOIL DESCRIPTION.......................................................................123 LIST OF REFERENCES.............................................................................................................130 BIOGRAPHICAL SKETCH.......................................................................................................137

PAGE 7

7 LIST OF TABLES Table page 2-1 Basic physico-chemical properties of surf icial (0 cm) soils collected in July 2002 from WCA-2A and BCMCA.............................................................................................48 2-2 Long-term depth, mass, and phosphorus accre tion at two locations in Blue Cypress Marsh Conservation Area..................................................................................................48 2-3 Model parameters for long-term soil accretion and phosphorus accretion at two stations in Blue Cypress Marsh Conservation Area..........................................................48 2-4 Results of phosphorus fraction model calib ration and validation at two stations in Blue Cypress Marsh Conservation Area............................................................................49 2-5 Rate constants, equations and initial va lues used in Stella model of long-term P declines at nutrient unimpacted station in Blue Cypress Marsh Conservation Area.........49 3-1 Basic physico-chemical properties of surf icial (0 cm) soils collected in July 2002 from WCA-2A and BCMCA.............................................................................................78 3-2 Regression statistics for autoclave extrac table P verses depth in soil samples from Blue Cypress Marsh Conservation Ar ea and Water Conservation Area 2A.....................78 3-3 ANOVA General Linear Model results of pyrolysis temperature and soil depth effects........................................................................................................................ .........78 4-1 Results of 14C AMS dating..............................................................................................106 5-1 Components of conceptual model depict ing phosphorus pools and transfers in peat soils.......................................................................................................................... ........120 A-1 Physical description of soil samples co llected in July, 2002 from Water Conservation Area 2A and Blue Cypress Marsh Conservation Area....................................................123 A-2 Physical description of soil samples co llected in March 2004 fr om Blue Cypress Marsh Conservation Area................................................................................................129

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Conceptual diagram of phosphorus cycle in wetland soils (adapted from Stewart and Tiessen 1987).................................................................................................................. ...22 1-2 Study location in Water Conservation Area 2A, Florida, USA.........................................23 1-3 Study location in Blue Cypress Ma rsh Conservation Area, Florida, USA........................24 2-1 Sequential fractionation tec hnique used in first extraction, adapted from Ivanoff et al. 1998........................................................................................................................... .........50 2-2 Soil bulk density (g cm-3) at nutrient impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Water Conservation Area 2A..............................51 2-3 Soil loss on ignition (%) at nutrient imp acted and unimpacted stations in Blue Cypress Marsh Conservation Area and Water Conservation Area 2A..............................52 2-4 Soil total phosphorus at nut rient impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Wa ter Conservation Area 2A............................................53 2-5 Soil carbon to nitrogen ratio at nutrien t impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Water Conservation Area 2A..............................54 2-6 Soil carbon to phosphorus ratio at nutrient impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Water Conservation Area 2A..............................55 2-7 Ratio of inorganic phosphorus to orga nic phosphorus at nutrient impacted and unimpacted stations in Blue Cypre ss Marsh Conservation Area and Water Conservation Area 2A........................................................................................................56 2-8 Results of sequential chemical fractio nation. Values are % of total P extracted...............57 2-9 Ratio of humic to fulvic aci d in sequentially extracted soils.............................................58 2-10 Rate of increase in the % residual (unextractable) phosphor us fraction with respect to time in Blue Cypress Marsh Conservation Area................................................................59 2-11 Relationship between soil depth and age in Blue Cypress Mars h Conservation Area. ....60 2-12 Relationship between soil age and tota l phosphorus concentration in Blue Cypress Marsh Conservation Area..................................................................................................61 2-13 Dynamic model of storages and tran sfers from major phosphorus fractions....................62 2-14 Results of model calibration and validation.......................................................................63

PAGE 9

9 3-1 Hot water extractable P (HEP) of Blue Cypress March Conservation Area and Water Conservation Area 2A soils...............................................................................................79 3-2 Hot water extractable phosphorus (HEP) of Blue Cypress March Conservation Area and Water Conservation Area 2A vs. wate r soluble (WSP) and microbial biomass phosphorus (MBP).............................................................................................................80 3-3 Hot water extractable P (HEP) of tw o model compounds, glycerophosphate (GP) and phytic acid (PA)............................................................................................................... ..81 3-4 Percent of total phosphorus extracte d from dried, ground peat using pyrolytic extraction under nitrogen atmosphere................................................................................82 3-5 Thermal extraction of organic P........................................................................................83 4-1 The humification index (HIX) of NaOH-extracted peat samples....................................107 4-2 Ratio of absorbance at 465 and 665 nm (E4: E6) of NaOH peat extracts at stations C1 and B4 in BCMCA...........................................................................................................108 4-3 Results of solid-state 13C-NMR spectroscopy of dried, ground peat samples collected from station B4 on February 11, 2003.............................................................................109 4-4 Results of solution 31P-NMR spectroscopy of NaOH-extracted samples collected from station B4 (unimpac ted) on February 11, 2003.......................................................110 4-5 The 31P-NMR spectra of 0.5M NaOH extracts of soil samples from Blue Cypress Marsh Conservation Area, st ation B4 (unimpacted)........................................................111 4-6 Results of lignin fractionation of plant fiber. Fiber was separated from bulk peat samples that were collected from BCMCA stations B4 and C1......................................112 4-7 Stable isotopic composition of plant fibe r that was separated from bulk peat samples collected from BCMCA stations B4 (unimpacted) and C1 (impacted)...........................113 4-8 Uncalibrated 14C age of plant macrofossils from selected depths at stations B4 and C1 in BCMCA.................................................................................................................114 5-1 Conceptual model of dynamic exch anges among soil phosphorus pools in wetland soils.......................................................................................................................... ........121 5-2 Results of mathematical modeling of l ong-term declines of five phosphorus fractions in Blue Cypress March Conservation Area.....................................................................122

PAGE 10

10 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 BIOGEOCHEMICAL TRANSFORMATIONS OF PHOSPHORUS IN WETLAND SOILS By Millard M. Fisher, III December 2007 Chair: K. Ramesh Reddy Major: Soil and Water Science Phosphorus (P) is an essential el ement required by all organisms and is a principal nutrient influencing agricultural productivity worldwide. Surface runoff and subsurface leaching of P to surface waters has led to nutrient enrichment of aquatic ecosystems. Nutrient enrichment is the third largest cause of impairment of surface waters in the U.S. Excess nutrients, particularly P, have been linked to changes in flora and fauna, such as encroachment of invasive species and nuisance algal blooms. Eutrophication of historic ally low-nutrient lake s and wetlands alters trophic structure, favoring species that are more adapted to higher nutrient conditions. Determining pre-disturbance nutrient loadi ng regimes is a major challenge for aquatic ecosystem researchers. Wetland soils and lake sedi ments that accrete materials preserve a partial record of ecosystem history that can be us ed to determine some pre-impact ecosystem characteristics. However, biogeochemical processe s continually alter soil properties. Describing the extent of these changes was the central goa l of this dissertation. Results from both P and carbon (C) characterization suggest movement of P and C into more recalcitrant organic compounds with time. I examined changes in soil phosphorus pools over a 5000-year period in two subtropical wetlands using chemical fractionation. Results indi cate an increase in the relative proportion of

PAGE 11

11 recalcitrant P over time, with minimal change in humicand fulvic-P fractions. Long-term wetland P accretion was estimated to be approximately 2 mg m2 yr-1, a rate much lower than has been reported for wetland soils. Two thermal methods were developed to better resolve the extent of P recalcitrance. Autoclave-extractabl e P declined throughout the soil profile from 40% of total P in surface soils to 10% in soils 1 m deep. Soils thermally extracted under N2 atmosphere showed both increasing extractable P with increasing thermal energy, and declining P recovery with depth for any gi ven temperature. Both techniques suggest that susceptibility to thermal degradation may be a useful method to characterize organi c P recalcitrance. Spectrophotometric investigati on of organic matter extracts showed little change in average molecular weight with soil depth, though fluorescence and 13C-NMR analyses indicate increasing humic and phenolic character with in creasing soil depth. The lignin content of soil fiber separated from peat increased from 35% in surface soil, to 75% at a depth of 1 m. Isotopic characterization of soil fiber seemed to indica te rapid ecological changes in the marsh that occurred at approximately 1000AD.

PAGE 12

12 CHAPTER 1 INTRODUCTION Background Phosphorus in Natural Ecosystems Phosphorus (P) is one the pr incipal macronutrients influe ncing agricultu ral production worldwide. It is an essential elemen t required by all organisms and is the 12th most abundant element in the lithosphere. It constitutes 0.1% of the earths crust (R eddy and DeLaune 2007) as well as 0.1% of plant and animal tissue, a nd tends to maintain a proportion to other macronutrients of approximately 106:16:16:1, C: N: Si: P in biomass (Redfield 1963). Phosphorus export from anthropogenic sources is the third largest cause of impairment of surface waters in the U.S. (USEPA 1998), and in Florid a approximately 35% of large lakes do not meet their designated use due to excess nutrients (F DEP 2004). Excess nutrients, particularly P, have been linked to changes in flora and fauna, such as encroachment of invasive species and nuisance algal blooms. A major cause of nutrient impacts to surface waters is increase d fertilizer use in the United States (U.S.) (Sharpley 2000). Application of phosphorus fertilizers increased in the U.S. from 2.6 million tons in 1960, to a maximum of 5 .6 million tons in 1977, and has declined since to 4.6 million tons in 2005 (Fertilizer Institute 20 07). Advanced treatment of point sources of P, as well as removal of P from ma ny detergents has increased the relative contribu tion of P from agriculture. The result is that agriculture now constitutes the leading source of P exports to surface waters (USEPA 1996). Additionally, long-ter m applications of P in excess of crop and livestock removal have led to build up in soil P st orages as well as regional imbalances of P in surface soils, particularly in developed countries (Sharpley 2000). Lakes and wetlands often occupy the lowest position on the landscape. They are thus closely linked with activities on adjacent uplan ds, and tend to accumulate nutrients and other

PAGE 13

13 materials associated with surface runoff. Eutrop hication of historically low-nutrient lakes and wetlands alters trophic structure, favoring specie s that are better adap ted to higher nutrient conditions. In particular, phosphorus (P) has often been shown to be the principal nutrient influencing ecosystem composition (Steward and Ornes 1975, Davis 1991, Urban et al. 1993). It is thus critical to the preserva tion of wetland trophic structure to both estimate and maintain the nutrient loading regimes that favored the establ ishment of the regions historic community. Estimating Historical Ecosystem Loading Rates The determination of pre-disturbance nutrient lo ading regimes is a major challenge for aquatic ecosystem managers and researchers. Water an d soil monitoring programs for most lakes and wetlands in the U.S. were initiate d only within the last four to five decades; perhaps none predate large-scale agricultural and urban development. Wetland soils and lake sediments that accrete materials preserve a partial record of ecosystem history. Proxy variables th at have been used to indirectly estimate predevelopment conditions incl ude investigations of pollen, diatoms, algal pigments, zooplankton exoskeletons, isotopes, and soil nutrient conten t. Estimates of predevelopmental water quality in lakes have been determined through analysis of invertebrate and diatom remains (Crisman 1978, Dixit et al. 1992, Shumate et al. 2002, Whitmore 1989). These fossil assemblages can be well-preserved in sedi ment and reflect water conditions at the time they were deposited. Changes in sediment chemistry have also been used in lakes to infer previous water column conditions, as well as historical nutri ent loading rates (Br ezonik and Engstrom 1998, Engstrom et al. 2006, Brenner et al. 2001, Craf t and Richardson 1998). Craft and Richardson (1998) estimated an eight-fold increase in P lo ading to the Everglades, while Brezonik and Engstrom (1998) estimated a four-fold increase in P loading to Lake Okeechobee, Florida based on increases in down-core sediment total P. Brenner et al. (2001) estimated recent P

PAGE 14

14 accumulation rates to be from two to seventeen times greater than accumulation rates in 1920 for a marsh in east central Florida. Wetland nutrien t assimilation models ha ve been adapted to calculate sustainable nutrient loading rates (L owe and Keenan 1997). Typically, such models rely on an empirically determined settling coeffi cient that accounts for loss to the sediments. Estimation of this term is critical to nutrient loading applications, especially when used to determine acceptable thresholds of nut rient loading to natural wetlands. Temporal Changes in Soil Phosphorus One assumption of using temporal (calculated) changes in nutrient accumulation to infer past water column conditions is that the estimated net accumulation rate for a particular time in the past truly reflects the nutrient loading ra te that prevailed at that time. However, biogeochemical processes progressively alter bo th organic and inorgani c P compounds, perhaps confounding accurate estimates of past P accumula tion. Transformations of soil inorganic and organic P compounds are highly dynamic and are influenced by complex transfers among biological and mineral pools (Ste wart and Tiessen 1987). For exampl e, once labile inorganic P is depleted, the microbial and plant communities ar e capable of producing extracellular enzymes that are responsible for hydrolysis of monoa nd diester phosphorus in organic matter (Freeman et al. 1996, Wright and Reddy 2001) t hus mineralizing soil organic P. It is also suspected that predominantly abiotic reactions alter soil orga nic matter (and thus organic P), resulting in increasing humic character with time (Stevenson 1994) (Fig 1-1). Factors associated with the formation of soils, such as climate, landscape position, time, biological pr ocesses, and parent material all influence the rate and extent of th ese transformations (Jenny, 1943). It is therefore likely that the floral and fauna l precursors of the soil organic matter pool are transformed into compounds that are progressively less similar to their parent compone nts. In a P-limited

PAGE 15

15 environment, such as many pristine aquatic ecosy stems, it might also be expected that these processes would alter materials at the expense of P, leading to a decline in soil P concentration with greater depth, or time. It is difficult to estimate the extent of these alterations if total P is measured. Phosphorus fractionation schemes are capable of distinguishing be tween inorganic and organic pools of P, and labile and recalcitrant P. Phosphorus fractionation has been used to describe changes in P chemistr y across horizontal surf icial soil chronosequences. For example, Walker and Syers (1976) chemically fractiona ted soil P across a landscape that spanned approximately one hundred thousand years of so il pedogenesis. They demonstrated that P fractions were not static, and th at soil-forming factors continuous ly acted on mineral and organic forms of P. They postulated that the distributi on of soil P fractions might be a useful pedogenic tool for dating soils. Their results were duplicated by Crews et al. (1995) for a soil chronosequence spanning four million years in Hawa ii. In both studies, apatite P in surface soil horizons declined due to weatheri ng, while organic P initially increa sed, then slowly declined in older landscapes. Hedin et al. ( 2003) examined the same Hawaiian soils and determined that as soils aged, export of soluble re active P declined while the pr oportion exported as dissolved organic P increased. This slow leakage of P was found to be sufficient to maintain older forests in a state of constant P limitation. Thus fo r upland soils, soil P has been shown to undergo continual change due to soil weathering processe s. In organic wetland soils, redistribution of both total and soil P fractions could also be expe cted to occur, though so il weathering processes differ. In submerged soils, oxidative processe s are much slower due to limited diffusion of oxygen and anaerobic conditions (Reddy and Patric k 1975). Energy yield of alternate terminal electron acceptors is low relativ e to oxygen, thus slowing the mi neralization of organic matter

PAGE 16

16 (McLatchey and Reddy 1998, DeBusk and Reddy 1998). This is the primary reason for accretion of organic soils in peat soil wetlands. Walker and Syers (1976) compared phosphorus fractionation results from surface soils taken acro ss a landscape of progre ssively older soils to estimate temporal changes in P pools. The same approach is useful for describing stratigraphic changes in P fractions for progre ssively deeper peat samples. Characterizing Changes in Soil Phosphorus Characterization of the availability of soil and se diment P has been a principal goal of soil and environmental scientists. Chemical extraction schemes are used to classify soil P into operationally defined pools (Chang and Jackson 1957, Hedley et al. 1982, Hieljtes and Lijklema 1980, Ivanoff et al. 1998). Mild acids extract inor ganic P that is associated with calcium, magnesium, iron, and aluminum, while bases such as sodium hydroxide can be used to extract a portion of organically-bound P. Th e un-extractable residue that remains after fractionation is considered to be recalcitrant, though little is known about its chemical composition. The resultant P fractions can be furthe r subdivided into varying levels of labile and refractory P with the goal being to estimate the short and long-te rm availability of P to biota. Organic P fractionation derives in part from attempts to characterize soil organic matter. Late in the 18th century, scientists extracted peat with alkaline ex tractants and obtained a da rk fluid that formed a precipitate upon acidification, which was termed humic acid (Stevenson, 1994). They observed that more humic acid could be extracted fr om deeper, older peat. By 1900, it was widely recognized that humus was a complex mixture of organic substances that were mostly colloidal in nature and had weakly acidic prope rties. It was also rec ognized that soil organic matter was comprised of over 40 compounds in the broad categories of organic acids, hydrocarbons, fats, sterols, aldehydes, and carboh ydrates. Waksman (1936) argued that the terms humic and fulvic acids should be abandoned, because these labels designatecertain

PAGE 17

17 preparations which have been obt ained by specific procedures, ra ther than identifying specific compounds. Stevenson (1994) points out reference to the acid-insoluble material as humic acid is considerably less cumbersome than repeated reference to alkalisoluble, acid-insoluble fraction. Soil humus contains most, if not all, of the biochemical co mpounds synthesized by living organisms and many additional compounds that are largely the product of abiotic secondary synthesis reactions that occur in the soil environment. These compounds are thus dissimilar to the biopoly mers of microorganisms and higher plants. Few procedures have been developed to ev aluate compounds at the opposite end of the availability continuum, i.e. P-containing co mpounds that have become resistant to mineralization. Organic P in soils (especially peat) is present in chemical compounds of varying environmental stability. Reddy et al. (1998) have shown that in Water Conservation Area 2A in the Florida Everglades, refractor y P increased from 33% of total P in surface soils, to 70% in deeper strata. Stratigraphic trends in refractor y P, similar to those obs erved in the northern Everglades, were found in soils from marshes of the upper St. Johns River, Florida. Olila and Reddy (1995) showed generally increasing residual (or recalcitrant) P cont ent, increasing from 20% (of TP) in surface horizons to approximate ly 80% at a depth of 35 cm. DeBusk and Reddy (1998) showed that the proportion of refractory carbon in soil pr ofiles (from sta nding dead litter to deeper soil strata) increases significantly wi th depth. In their study, lignin content increased from approximately 12% of total dry weight for st anding dead plant litter, to 50% for peat at a depth of 30 cm. There are three principal mechanisms that could lead to en richment in refractory compounds in the soil profile. Firs t, the plant community that forms the peat soil may have changed to a community that is less enriched in refractory compounds. This could be expected if a shrub community was succeeded by a slough co mmunity dominated by herbaceous plants.

PAGE 18

18 Second, changes in external load ing may have resulted in enrich ment of labile compounds in surficial peat. Lastly, over long periods of time, there may have been mobilization and loss of labile fractions and enrichment of refractory co mpounds, such as lignin in deeper deposits. The compounds associated with the more labile fr actions are thus subject to mineralization, mobilization, and subsequent loss from the soil profile. Pristine wetland regions that have not been impacted by nutrient loading may therefor e serve as experimental controls in that chronological (depth) changes in P fractions would be expected to reflect primarily the effects of organic matter decomposition. Further insight into long-term changes in soil organic P may be gained by characterizing both organic P and soil organic matter as a functi on of age or depth. Mole cular level techniques may provide insight into longterm alterations in soil Po and organic matter. Techniques such as 31P-Nuclear Magnetic Resonance (NMR) and 13C-NMR provide this information for broad classes of both organic P and organic C compounds (Turner et al. 2003). Fluorescence and absorbance properties of soil organic matter extracts have also been shown to be related to extent of soil humification (Cox et al. 2000, Chen et al 1977). Thermal techniques have been used to characterize polyphosphates in microb ial and algal biomass (Kenney et al. 2000) as well as the extent of peat decomposition (Levesque and Dinel 1978). It is also possible that thermal stability of organic P is related to its recalcitrance. Overview of the Problem Accelerated P loading to freshwater wetlands often causes changes in flora and fauna, and these community changes are generally regarded as undesirable. All aqua tic ecosystems receive baseline P loads that vary depending on la ndscape position, region, and other factors. Determination of this historical rate of P input is a major goal of lake and wetland scientists. One approach to determining a sustai nable P loading rate is to examine the soil or sediment record

PAGE 19

19 and characterize vertical changes in soil P c ontent, and in combination with the chronostratigraphy, calculate hi storic P loading. A potential shortcom ing of this approach is that diagenetic changes may have altered soil P pool s such that calculated P accumulation rates do not accurately estimate the original nutrient loadi ng rates. The main focus of this dissertation is the characterization of those changes. The experiments described in this dissertati on were performed in two subtropical wetlands, Blue Cypress Marsh Conservation Area (BCMCA ), in east central Florida, and Water Conservation Area 2A (WCA-2A), in the northern Florida Everglades (Fi g. 1-2). Both wetlands are underlain by peat soils, and both have regions that have been impacted by nutrient rich agricultural discharges, yet still contain relatively pristine area s. This region has a subtropical climate and receives approximately 1.5 m of rainfall annually. WCA2A and other surrounding WCAs are used for flood control, recreation, wild life habitat, and water storage. The flora and fauna of the Everglades are P-limited (Ste ward and Ornes 1975), though anthropogenic nutrient loading from the Everglades Agricultural Ar ea (EAA) has reduced some of this limitation, particularly in regions adjacent to main canal s (Reddy et al. 1998, DeBusk et al. 1994, Craft and Richardson 1993). The peat soils are derived from sawgrass ( Cladium jamaicense Crantz) (Davis 1943). They are approximately 2-m thic k and the basal deposits were formed about 5,000 years before present (Gleason and Stone 1994). They thus present a vertical profile that grades from recently deposited plant litte r that can be presumed to be enriched in relatively labile compounds, to deeper organic deposits that are recalcitrant after mille nnia of biogeochemical weathering. The BCMCA is a 116-km2 shallow marsh in the headwaters of the St. Johns River (Fig. 13). It is part of a joint St. Johns River Water Management Dist rict (SJRWMD) and United States

PAGE 20

20 Army Corps of Engineers floodplai n restoration project whose goal is to provide both ecological and flood protection benefits. BCMCA is also underl ain by peat soils that vary from 1.5 m in the eastern marsh to greater than 4 m near the ce nter. The vegetation is a mosaic of open slough communities ( Nymphaea odorata, Eleocharis elongata, Utricularia spp .), Maidencane flats ( Panicum hemitomon ), broad expanses of sawgrass, and tree islands ( Acer rubrum, Taxodium distichum, Myrica cerifera ). Agricultural discharges created a region of elevated soil P in the northeastern side of the marsh, though those discharges ceased in August 1994. Hypotheses and Objectives The central hypothesis of this dissert ation is that soil phosphorus undergoes biogeochemical processing that favors transforma tion into stable, recalcitrant pools. For wetland soils, the ultimate pool is hypothesized to be recalc itrant organic matter, with P as a part of the organic matrix. The following objectives were developed to investig ate this hypothesis: Characterize long-term changes in soil organi c P using a fractionation scheme that was developed for organic soils. Estimate the organic P is immobilization rate and develop techniques to distinguish extent of immobilization. Because organic P cycling and soil organi c matter (SOM) are intimately linked, characterize temporal changes in SOM pools. Dissertation Format I examined changes in phosphorus chemistry over a 5000-year soil record in two subtropical wetland soils using se quential chemical extraction. Chapter one examines the impacts of phosphorus on natural systems and methods us ed to characterize soil phosphorus. It also describes methods of determining historical eco system nutrient loading records by examining the soil stratigraphy. In chapte r two, the sequential chemical fracti onation of soils in cores from two subtropical wetlands were compared using a me thod of organic P fractio nation (Ivanoff et al.

PAGE 21

21 1998). The fractionation data were ev aluated in light of carbon-14 so il dates to determine rate of P sequestration in wetland soil. A mathematical model of long-term phosphorus diagenesis is proposed. Chapter three focuses on development of two new thermal P extraction techniques: pyrolytic extraction of organi c P and an autoclave extracti on. An enzyme-based organic P extraction was also investigated. Chapter four cove rs characterization of al kaline extracts of soil organic matter (SOM), with the reasoning that in sights into changes in th is fraction will yield similar insight into cycling of organic P. Results from both 13C-NMR and 31P-NMR are described. Soil organic matter fluorescence and abso rbance techniques were used to characterize changes in SOM in two subtropical wetland soil s. The carbon and nitrogen isotopic signature of lignin separated from peat was analyzed to determine potential changes in plant communities. Chapter five provides a synthesis of the results of preceding chapters and proposes a conceptual model of organic P burial in wetland soils.

PAGE 22

22 Figure 1-1. Conceptual diagram of phosphorus cycle in wetland soils (adapted from Stewart and Tiessen 1987). Solution P Mineral and Amorphous P Plant Residues Bacterial, Fungal, Protozoan Processing Labile Po Increasingly Resistant Humic/Fulvic Po Plants

PAGE 23

23 Figure 1-2. Study location in Water C onservation Area 2A, Florida, USA. ( ( 800'0"W 800'0"W 8040'0"W 800'0"W 8030'0"W 800'0"W 8020'0"W 8020'0"W 2550'0"N 2550'0"N 26'0"N 26'0"N 2610'0"N 2610'0"N 2620'0"N 2620'0"N 2630'0"N 2630'0"N 2640'0"N 2640'0"N Water Conservation Area 2A 02550 12.5Km Map AreaE5 E1Water Conservation Area 1 Water Conservation Area 3

PAGE 24

24 Figure 1-3. Study location in Blue Cypress Marsh Conservation Area, Florida, USA. ( ( 800'0"W 8050'0"W 8045'0"W 8045'0"W 8040'0"W 8040'0"W 2740'0"N 270'0 2745'0"N 275'0 2750'0"N 270'0 Blue Cypress Marsh Conservation Area Blue Cypress Lake 036 1.5Km Farm DischargeCitrus Farms Map AreaB4 C1

PAGE 25

25 CHAPTER 2 LONG-TERM CHANGES IN SOIL PHOSPHORUS FORMS IN TWO SUBTROPICAL WETLANDS Introduction Wetland soils and lake sediments tend to accumula te material, for two main reasons. First, wetlands occupy a low position in the landscape, typi cally at the ecotone between terrestrial and aquatic ecosystems (Mitsch and Gosselink 1993). Flow of water, suspended solids, nutrients and other constituents is therefore ge nerally towards them. This subs idy of water and nutrients leads to high primary production, of ten in excess of 1000 g m-2 year-1 (Mitsch and Gosselink 1993). Second, decomposition of organic matter is reduced due to slow diffusion of molecular oxygen into water and saturate d soil (Ponnamperuma 1972, Rowell 1981, Patrick and Reddy 1983, DeBusk and Reddy 1998). The end result of high primary production and slow decomposition is accumulation of organic matter, usua lly in the form of peat. Nutrient loading in excess of that to whic h aquatic ecosystems have been conditioned often results in undesirable changes in flora a nd fauna, such as algal blooms in lakes and encroachment of pioneer or high nutrient adapte d vegetation in wetlands. In the United States, nutrient enrichment is one of the leading causes of impairment to wetlands (USEPA 2002). Eutrophication of historically lo w-nutrient lakes and wetlands a lters trophic structure, favoring species that are more adapted to higher nutrient conditions. In particular, phosphorus (P) has been shown to be the principal nutrient infl uencing wetland ecosystem composition (Steward and Ornes 1975, Davis 1991, Urban et al. 1993). It is thus critical to the preservation of wetland trophic structure to both estimate and maintain the nutrient loading regimes that favored the establishment of the regions historic community. Unlike surface water samples, which reflect instantaneous conditions, soils and sediments integrate conditions over long time scales (Frey 1969, Smol 1992). For example, estimates of

PAGE 26

26 pre-developmental water quality in lakes have be en determined through analysis of invertebrate and diatom remains (Shumate et al. 2002, Whitm ore 1989). These fossil assemblages can be well preserved in sediment and reflect water conditio ns at the time they were deposited. Wetland soils and lake sediments may also preser ve a direct record of ecosystem nutrient loading that provides information on predisturbance conditions. This re cord can then be used to help determine acceptable lake or wetland nutrient loading rates (Brezonik and E ngstrom 1998, Craft and Richardson 1998, Brenner et al. 2001, Engstrom et al. 2006). This can be done by associating dates with vertical changes in soil P content to derive P accumulation rates, then reasoning that changes in P accumulation are mostly due to changes in ecosystem deposition, or loading rates. The ages of specific soil horizons are typically ob tained through radioisotopes such as cesium137, lead-210, or carbon-14. For example, Craft and Richardson (1998) estimated an eight-fold increase in P accumulation since 1960 in nutrient impacted areas of the northern Everglades. Brezonik and Engstrom (1998) estimated a fou r-fold increase in P accumulation since 1910 to Lake Okeechobee, Florida based on sediment P c ontent and overall sediment accumulation rate. Brenner et al. (2001) estimated a 2.3 to 17 fold increase in P accumulati on when comparing post1970 to pre-1920 rates for a marsh in east central Fl orida, and concluded th at the increases were due to increased external nutrient loading. Ho wever, dynamic biogeochemical forces act in concert to alter the organic and inorganic materi als deposited at the so il surface (C arignan and Flett 1981, Walker and Syers 1976, Bi nford et al. 1983). For example, once labile inorganic P is depleted, the microbial and plant communities ar e capable of producing extracellular enzymes that can hydrolyze monoand diester phosphorus in detrital material (Freeman et al. 1996, Wright and Reddy 2001) thus mineraliz ing soil organic P. It is also suspected that predominantly abiotic reactions alter soil organic matter (and thus organi c P) resulting in increasing

PAGE 27

27 humification with time (Stevenson 1994). It is th erefore likely that th e floral and faunal precursors of the soil organic matter pool are tran sformed into compounds that are progressively different from their precursor compounds. In a P-limited environment, such as many pristine aquatic ecosystems, it might also be expected th at these processes alter these materials at the expense of P, causing a decline in soil P concentration with dept h, or time. If inferences about past nutrient loading rates are to be extrapolated from soil nutrient content, some knowledge of the rate of P diagenesis is needed. Sequential chemical fractionation techniques ha ve been developed to distinguish between forms of P in agricultural soil s (Chang and Jackson 1957), for lake sediments (Hieltjes and Liklema 1980) and organic wetland soils (Ivanoff et al. 1998). They are ca pable of distinguishing between inorganic and organic pools of P, as we ll as extent of recalcitrance. Many researchers have documented changes in soil P fractions along horizontal lands cape chronosequences (Walker and Syers 1976, Crews et al. 1995, Richter et al. 2006, Hedin et al. 2003). For example, Walker and Syers (1976) sequent ially fractionated soils along a horizontal chronosequence and showed that long-term surface soil weatheri ng caused a reduction in calcium-bound P and an initial increase in organic P, followed by long-term declines. The rate and extent of these ch anges, particularly with respec t to soil organic P, have not been well characterized. Deep peat soils provide an opportunity to investigate the nature of longterm alterations in soil phosphorus fractions These soils contain a record of changing environmental conditions that in many cases spans thousands of years. Combining P fractionation results with soil isot opic dating methods allows for estimation of rates of change in major soil fractions, as well as for the calculation of historic soil P accretion rates. The two main

PAGE 28

28 objectives of this study were to 1) characterize long-term changes in soil organic and inorganic pools, and 2) determine rate of transfer into long-term recalcitrant P fractions. Methods Site Description Field sampling was conducted in two subtropi cal wetlands, Water C onservation Area 2A (WCA-2A) in the northern Florida Everglades and Blue Cypress Ma rsh Conservation Area (BCMCA) in east central Florida, USA. This region has a subtropical climate and receives approximately 1.5 m of rainfall annually. The WCA-2A is a 447-km2 wetland underlain by peat soils. WCA-2A and other surround ing WCAs are used for flood control, recreation, wildlife habitat, and water storage. The flora and fauna of the Everglades are P-limited (Steward and Ornes 1975), though anthropogenic nut rient loading from the Ever glades Agricultural Area (EAA) has reduced some of this limitation, par ticularly in regions adjacent to main canals (Reddy et al. 1998, DeBusk et al. 1994, Craft and Richardson 1993). The peat soils are derived from sawgrass ( Cladium jamaicense Crantz) (Davis 1943). They are approximately 2 m in thickness and the basal deposits we re formed approximately 5,000 years before present (Gleason and Stone 1994). They thus present a vertical pr ofile that grades from recently deposited plant litter that can be presumed to be relatively labile to deeper organic depos its that are recalcitrant due to millennia of biogeochemical weathering. Two sampling locations were selected in WCA2A, long-term research stations E1 and E5 (Fig. 1-2). Station E1 is located in a nutrient-impacted region, approximately 1.4-km south of South Fl orida Water Management Districts (SFWMD) inflow structure S-10C. Station E5 is located approximately 10 km s outh of this structure, in the unimpacted central marsh. Vegetation at station E1 was predominantly cattail ( Typha domingensis Pers.), and at station E5, sawgrass.

PAGE 29

29 The BCMCA is a 116-km2 shallow marsh in the headwaters of the St. Johns River (Fig. 13). It is part of a joint St. Johns River Water Management Dist rict (SJRWMD) and United States Army Corps of Engineers floodplai n restoration project whose goal is to provide both ecological and flood protection benefits. BCMCA is also underl ain by peat soils that vary from 1.5 m in the eastern marsh to greater than 4 m near the ce nter. The vegetation is a mosaic of open slough communities ( Nymphaea odorata, Eleocharis elongata, Utricularia spp .), Maidencane flats ( Panicum hemitomon ), broad expanses of sawgrass, and tree islands ( Acer rubrum, Taxodium distichum, Myrica cerifera ). Agricultural discharges have creat ed a region of elevated soil P in the northeastern area of the marsh, though thos e discharges ceased in August 1994. One sampling location was selected from this region of the marsh, station C1, and one station was selected from the interior unimpact ed region of the marsh, station B4. Sampling and Analysis Soil samples were collected in triplicate from both marshes on July 17 and 18, 2001. Soil cores were obtained with a 2-m long x 7.3-cm diamet er stainless steel soil corer. The extent to which the soil was compacted during sampling was noted, and tended to be less than 15 cm for a 1.5 m core sample. Soil samples were extruded in th e field into Ziploc bags at 10 cm intervals. The samples were immediately chilled to 4oC and were kept on ice for transport to the laboratory. Laboratory processi ng was performed at the Univ ersity of Floridas Wetland Biogeochemistry Laboratory. Wet weight was re corded for each sample and samples were transferred to rigid polyethylene containe rs and thoroughly mixed by hand. A 10 20 g subsample was removed for pH determination. For samp les that were insufficiently moist to make a pH determination, deionized water was added until the sample was sufficiently wet. A separate 50 100 g sub-sample was dried at 80oC to constant weight for percent moisture determination.

PAGE 30

30 Sequential Organic Phosphorus Fractionation The soil P extraction technique closely followed the technique of Ivanoff et al. (1998) (Fig. 2-1). The analysis was performed on a field-moist sample (0.5 gram dry weight equivalent), with a soil to solution ratio of 1:50. The procedure wa s initiated by adding a field-moist sample to a 43 ml polyethylene centrifuge t ube, then adding 2 ml of CHCl3 to the tube. The tubes were left uncapped overnight and extracted th e following day with 0.5 M NaHCO3. Twenty-five ml of each extractant was added to the tube and it was end-to-end shaken on a reciprocating shaker for 16 hours. The samples were centrifuged at 6000 rp m, the supernatant ex tractant was removed and vacuum filtered through 0.45m polyethersulfone filters, and the tube was re-weighed. Twenty-five milliliters of the next extractant wa s then added to the tube and the process was repeated. This procedure was continued in such a manner that with each successive step, less P remained in the soil sample. After the final step, only resistant P remained. A separate, non-sequentially extracted wet s ub-sample was extracted with 0.5 M NaHCO3, but with no chloroform added to the sample. The difference between the P liberated in the chloroformed and the P liberated in the non-chloroformed NaHCO3 extraction is believed to represent the P content of the soil microbial biom ass. The reported values were not adjusted to account for incomplete extraction of microbial bi omass with chloroform (Brookes et al. 1982). In these soils, chloroform treatment may result in overestimation of biomass P due to extraction from labile non-microbial organic P pools (Reddy et al. 1998). The humic and fulvic acid content of the Na OH extract was operationally determined in the following manner. After extracting with 1M NaOH and centrifuging, 7 ml of the supernatant fluid was withdrawn from each tube. Seven drops of concentrated H2SO4 were added to each subsample to condense the acid-insoluble humic acid fraction. The sample was again centrifuged

PAGE 31

31 at 6000 rpm for five minutes, and an aliquot of the supernatant was withdrawn. This aliquot was analyzed for total P by adding 1 ml of 11N H2SO4 and 0.3 g K2S2O8 and digesting for 3 hours at 380oC. The P remaining in solution after the acidification step was considered to represent P associated with fulvic acid. The total P conten t of an un-acidified NaOH subsample was also determined. Humic acid-bound P was determined as the difference between the TP content of the un-acidified sample and the fulvic acid P. The to tal P content of the residu al soil remaining after the sequential extraction was determined by combustion of residue at 550oC, followed by dissolution of ashes in boiling 6N HCL (Ande rsen 1976). All sequential extractions were conducted at room temperature, and all extractions were completed within thirty days of sample collection. Phosphorus was determined colorimetr ically using the method of Murphy and Riley (1962), and analysis was performed on a Tec hnicon AutoAnalyzer (Tarrytown, NY) (U.S. Environmental Protection Agency, 1983, Method 365.1). Soil dating (BCMCA only) was determined through 14C analysis of plant macrofossils. Ten centimeter sections of peat to a depth of appr oximately 1.5 m were carefully examined for plant fragments such as stems, seeds, and other woody fragments. Care was taken to avoid selection of root fragments. These plant macrofossils were sequentially extracted with 0.5M HCl (0.5 hours), 0.1M NaOH (2 hours), and again with 0.5M HC l (0.5 hours). This removed carbonates and any dissolved organic materials such as humic or fulv ic acids that may have percolated down through the soil profile. The samples were analyzed at University of Califor nia, Irvine Keck-CCAMS laboratory by accelerator mass spectrometer (N EC 0.5MV 1.5SDH-2, National Electrostatics Corporation). Carbon dates were corrected to calendar dates using the computer program CALIB, version 5.0.1(Stuiver and Reimer 1993).

PAGE 32

32 Results and Discussion Soil Properties The soils of WCA-2A and BCMCA were simila r with respect to most physical properties (Table 2-1). The pH of soil from BCMCA was cons iderably more acidic at almost 2 units lower than WCA-2A soils. Bulk density was similar at both locations, increasing from approximately 0.090 g (dry) cm-3 (wet) at the soil surface, to 0.110 g (dry) cm-3 (wet) at a depth of 80 cm (Fig. 2-2). Below 100 cm, bulk density increases drama tically at the two WCA2A stations due to increasing sand content of the p eat. Total nitrogen was slightly greater in BCMCA than WCA2A. Total P at the nutrient impacted station in WC A-2A was greater than the nutrient impacted site in BCMCA, though TP at the unimpacted site in WCA-2A was much lower than the unimpacted site in BCMCA (Fig. 2-4). Total P was highest in surface soils (0-50 cm) of WCA2A station E1 (546 mg kg-1), and declined in the order E1>C 1>B4>E5. As expected, the nutrient impacted stations showed higher total P content in the 0 50 cm soil layer than the unimpacted stations. However, the difference in total P be tween the impacted and unimpacted stations was not nearly as dramatic at the BCMCA stations, 367 vs. 327 mg kg-1. This comparison is based on a large depth interval, so any differences in P lo ading between the two sta tions would have been obscured by inclusion of relativel y deep soil in the sample. Also, total P at the unimpacted station in BCMCA was 66% greater than the unimp acted Everglades stati on. This suggests that BCMCA was not historically under the same degr ee of P limitation as the Everglades, and that nutrient enrichment was more pronounced at the impacted WCA-2A station. Other studies have shown that nutrient impacted su rface soils of BCMCA appear to be higher in P than WCA-2A (Grunwald et al. 2006), and that BCMCA is less P-limited (Prenger and Reddy 2004). This difference may be principally due to differences in landscape pos ition of the two wetlands than

PAGE 33

33 to differences in anthropogenic nutrient loading. For example, th e headwaters of the two river systems that are connected to respective wetlands ar e only 50 km apart, separated by the narrow Holopaw sand ridge. The difference is that BCMCA is located in the southernmost headwaters of the St. Johns River, whereas WCA-2A is locate d at the downstream end of the Kissimmee River Lake Okeechobee Everglades ecosystem comp lex. Thus there are ample opportunities for P to become sequestered by the lake-river-wet land complex upstream of WCA-2A, compared to the more upstream position of BCMCA. The carbon to nitrogen ratio (w eight:weight) in BCMCA increased downcore, from 15:1 in surface soils at both stations, to 19:1 at a depth of 90 cm (Fig 2-5). For WCA-2A, the ratio was relatively constant at the nutrient impacted station, 15.7:1 ( 1 SD). The ratio initially increased at the unimpacted station, from 17.3:1 ( 0.7 SD) at the surface, to 22.3:1 ( 1.8 SD) at a depth of 50 cm. Thereafter, the ra tio declined to 16.3:1 ( 0.6 SD) near the base of the core. Carbon to P ratio (C:P) increased consisten tly downcore at both BCMCA stations, from 600:1 to 900:1 in surface soil, to approximately 6000:1 at a de pth of 120 cm. Though the nutrient impacted BCMCA core showed initially lower C:P, due to elevated surface soil P, C:P increased more rapidly with increasing depth than at the interi or unimpacted marsh station. One interpretation of this is that the station that is now impacted by farm nutrient discharges may have historically been a low nutrient region of the marsh. Both WC A-2A cores showed initially increasing C:P, to a depth of 60 to 80 cm, then decreases to the base of the core. This decline is principally due to increasing inorganic matter conten t at a depth of approximately 70 cm in WCA-2A, as reflected in downcore increases in loss on ignition at that depth, particularly at the interior unimpacted station (Fig. 2-3).

PAGE 34

34 Organic P (TPo) constituted the majority of total P at both wetlands. The proportion of total inorganic P (TPi) ranged from an average of 0.06 at the unimpacted station in BCMCA, to 0.43 at the nutrient impacted station in WCA-2A (Fig. 2-7). Surface so ils at the unimpacted station in BCMCA were enriched in TPi, which declined with depth to approximately 50-cm, with little change thereafter. TPi:TPo declined slightly with greater depth in the soil profile at the impacted BCMCA station. Greater TPi:TPo in WCA-2A soils reflects the more calcareous soils of WCA2A and the influence that calcium has on P biogeoc hemistry in these soils. Highest ratios were observed in surface soils at the im pacted WCA-2A station, with TPi approaching 50 percent of TPo. The trend in TPi: TPo with respect to soil depth was different between the two WCA-2A stations, with the ratio increasing with respect to depth at the unimpacted station and declining at the impacted station. One possible explanation is that inflows to WCA-2A are high in both P and calcium (Gleason 1974). However, these inflows have resulted in a nor th-south gradient in surface soil P, but not in soil calcium (Re ddy et al. 1998, Craft and Richardson 1993). It may also be that interior water column concentration at the WCA-2A site is so low that adsorption of P on precipitated calc ite is inhibited. Soil Fractionation Results Labile phosphorus (NaHCO3 extractable Pi). Labile Pi accounted for <1 to 12% of soil P (Fig. 2-8). Labile Pi was greatest in surface so ils of BCMCA, accounting fo r as much as 12% of TP at the unimpacted station, de clining to approximatel y 4% at one meter. The reverse was true of the impacted station in WCA2A, with low values (3 % of TP) in surface soils, increasing to 13% near the base of the peat deposit. At th e impacted station in BCMCA and the unimpacted station in WCA-2A there was no clear trend in this pool, with labile Pi accounting for <1 to 7% of TP.

PAGE 35

35 Labile organic phosphorus (NaHCO3 extractable Po). Labile Po was a minor fraction of total P for both wetlands. Labile Po was frequently below the limit of detection, especially in WCA-2A. This fraction was highest in surface soils for both wetla nds, representing 2-3 % of TP in BCMCA and 0-2% in WCA-2A. Non-labile phosphorus (HCL extractable P). One molar HCl extracted inorganic P bound to Ca and Mg compounds. This fraction is a re latively stable storage of soil P. Except for the unimpacted BCMCA station, this fraction declined with depth. There were large differences between BCMCA and WCA-2A. Non-labile Pi accounted for 2 50% of the TP pool in WCA2A, yet < 10 % in BCMCA. This reflects the infl uence that the calcareous geology of WCA-2A has on P cycling, compared to the relatively softer water of BCMCA. Microbial phosphorus. Microbial biomass P (MBP) typically declined slightly with depth. On average, it represented 17% ( SD) of TP. MBP in surface (0-10 cm) soil averaged 19 % ( SD) of TP, and declined at most stations to an averag e of 13% ( SD) at a depth of approximately 1 m. Microbial biomass P increased s lightly with depth at the impacted station in WCA-2A, from 14% (.6 SD) at the surface, to 17% ( SD) at a depth of 0.8 m. The MBP was significantly greater in surface (0-20 cm) soils at the two unimpacted stations (Students t; P = 0.02). Fulvic and humic P Fulvic P accounted for 20% ( SD) of TP. With the exception of the impacted station in WCA-2A, fulvic P remained relatively constant thro ughout the soil profile. At the impacted WCA-2A station fulvic P was slightly higher in surface soils (0-20 cm), averaging 27% of TP ( SD), declining to 15% of TP ( SD) at a depth of 70 cm. Humic P accounted for 15 % of TP ( SD) and generally declined with greater depth. On average, humic P represented from 6 % of TP ( SD) at the impacted WCA-2A station, to 23 % of TP

PAGE 36

36 ( SD) at the unimpacted station in BCMCA. The ratio of humic to fulvic P increased from approximately 0.5 in surface soils, to a maximu m of 2.8 at the unimpacted station in BCMCA (Fig. 2-9). The ratio was consiste ntly greater at both unimpacted stations. The ratio began to decline at a depth of 90 cm in BCMCA and 60 cm in WCA-2A. An increasing ratio of humic to fulvic P through most of the soil profile probably reflects prevalen ce of more labile plant-derived fulvic compounds in the upper regions of the so il profile, and increasi ng importance of more recalcitrant humic compounds with time. Humic P is determined by difference between the total P content of the NaOH extract and the total P content of a subsample that has had humic materials precipitated from it. Hupfer et al. (2004) has shown that for lake sediments, up to 49% of this fraction can consist of inorganic polyphosphates. It is therefore possi ble that a significant proportion of this fraction is not due to humic P, and thus humic P may be overestimated. Residual phosphorus. Residual P, or unextractable P, s howed the clearest change in the soil profile. Residual P increased with respect to soil depth at all stat ions, representing an average of 20 % of TP ( SD) in surface soils, to 68 % of TP ( SD) at the base of the peat deposit. Both the proportion of P in this fraction and the rate of increase with depth were very similar for the upper 60-cm of soil for both wetlands and for both impacted and unimpacted stations. Below 60-cm, residual P increased more dramatically at the WCA-2A nutrient impacted station (E1) than the other stations, to a maximu m of 88% of TP at a depth of 90 cm. Residual P was the only fraction that showed consistent incr eases in pool size with increasing depth, or soil age. It increased mostly linearl y with respect to time (Figure 2-9). The proportion of P in this fraction increased by 0.012% per year and repres ented approximately 27% of total P in surface soils of BCMCA. This either reflects transfer fr om other fractions into this pool, or that this

PAGE 37

37 fraction maintained a relatively constant soil c oncentration, while others declined. Results from studies on the diagenesis of lake sediment provide evidence for the second mechanism. Penn et al. (1995) fractionated sediments from Ononda ga Lake, NY and found increasing refractory P with depth. They chemically fractionated approx imately 50 yr of deposition and assigned P to two groups, a fast and a slow d ecay pool. At the sediment surface, there was a 1:1 ratio between these pools, however, the proportion of P in the slow decay pool increased with depth. They estimated decay constants of 0.1 yr-1 and 4.8 yr-1 for the slow and fast pools, respectively, or a 50-fold decline in decay ra te for recalcitrant P. Total P declined dramatically with depth at a ll stations, while there was an increase in the unextractable, or recalcitrant fraction. The microbi al P pool generally declined with depth, as did the fraction associated with humic materials. Be cause total P declined sharply with depth and there were only modest changes in most other fractions, soil P pools were apparently maintain the same downcore proportions as those in surface soils. The maintenance of relatively stable proportions of P fractions indi cates the stability of respect ive fractions i.e., a dynamic equilibrium is maintained. Increases in the residual fractions were mostly at the expense of hu mic and microbial P. The principal difference in phosphorus fractions between the two wetlands was the proportion of P in the HCl-extractable fraction. The HCl-extr actable P in BCMCA was approximately 10%, compared to 27% in WCA-2A. This is probably due to the fact that BCMCA is a slightly acidic, soft-water wetland, underlain by sand, whereas WCA-2A is a slightly alkaline, hardwater wetland, underlain by sand and marl, and therefore ri ch in calcium. That humic P declined with increasing depth was not expected, since soil hum ification is thought to involve a progression of soil organic matter to more phenolic, less bioa vailable organic compounds. This suggests that

PAGE 38

38 either the compounds that comprise soil humic matte r are not closely associated with organic P, or that the alkaline extraction was not completely effective at removing humic P from the residual soil material. It is possible that accu mulation of organic P in soils does not occur through incorporation into humicand fulvic-type compounds. Celi and Barberis (2005) describe mechanisms that are largely ab iotic and involve stabilization of organic P through sorption of organic P compounds to iron, aluminum, and clay minerals. Though the proportion of P in the fulvic pool did not decline with depth as did humic P, it remained as a mostly constant percentage of total P. Long-Term Declines of Total Phosphorus in BCMCA The carbon-dated plant macrofossils in BCMCA allowed for the association of soil depth and soil age (Fig. 2-11). A first-order model was fitted to the age and depth data (Equation 2-1). Soil age (years before present) = Yo e (k*Z) (2-1) where: Yo = soil age at time = 0, years before present, k = first order decay constant, years-1, Z = soil depth, cm For the impacted and unimpacted stations, Yo equals 54 and 128 years, respectively. Rate constants at both stations were 0.03 yr-1. The parameter Yo is the soil age at a depth of zero centimeters. The older surface soil age for the nut rient impacted station is counterintuitive as nutrient impacts tend promote rapid plant growth and therefore faster so il accretion rates. In actuality, the nutrient impacted station likely does accrete mate rial faster. However, the 14C dating method is not particularly accurate for rece nt surficial material, as it is influenced by atomic bomb testing, plant roots, and burrowing of invertebrates. Therefor e, the surficial soil age difference of 84 years between the two stations is probably not signi ficant. Both stations appear to accrete material over long time intervals at the same rate. The rate constant (0.03 yr-1) implies

PAGE 39

39 that soil age increases at a rate of 3% for ever y unit increase depth. Equa tion 2-1 suggests that soil age is exponentially proportional to soil dept h, for example each unit increase of soil depth at lower depths represents a greater change in so il date than surface horizons. This could be due to either changing rates of accr etion over time, or that subsurface horizons are both physically compressed by the overburden, or that lower so il horizons have lost mass (and volume) over time, or both. Since there were only modest incr eases in bulk density, changing accretion and/or mass loss seem more likely (Fig. 2-2). The relationship between soil age and depth can be used to express sample phosphorus content as a function of age, ra ther than depth. A first order m odel can be fit to this data: Total Pt = Pf + Po exp(-kt) (2-2) where: Pf = long-term asymptotic soil P concentration, mg kg-1; Pf + Po = initial soil P concentration at time = 0, mg kg-1; k = first order rate constant, year-1. Fitting this model to the total P concentration at the two BCMCA statio ns yields a loss rate constant of 0.006 yr-1, or a decline of 0.6% per year, a nd a soil minimum of 118 and 80 mg kg-1 for the unimpacted and impacted stations, respect ively (Fig. 2-12; Tabl e 2-4). Soils at both WCA-2A and BCMCA asymptotically approached th is value, and for both marshes, this lower soil horizon dates to approximately 5,000 years befo re present (YBP). The exponential decline in soil P could be due to a combina tion of factors, such as mobili zation and loss of P from the soil profile, or a gradual increase in P loading to the marsh. Given that the decline was very similar for both impacted and unimpacted stations, the ch ange is likely principa lly due to decomposition of the original organic material mineralization, and loss of P. More support for this explan ation can be found from lakes in catchments with very little development. Lake Erken is a 24 km2 mesotrophic lake in east central Sweden that has

PAGE 40

40 undergone very little change in the catchment for the past cen tury (Rydin 2000). The drainage basin consists of approximately 70% forest and 10% farmland. Similar to theWCA-2A and BCMCA cores, Rydin (2000) found st eep declines in sediment tota l P with respect to depth, and attributed the difference in total P between surface sediment and deeper strata to P that had been mobilized and lost during diagenes is. Other studies have demonstrat ed that this diffusive loss to overlying water of mineralized P can be a majo r component of total ecosystem nutrient budgets (Moore et al. 1998, Fisher et al. 2005). Carignan and Flett (1981) ex amined the influence of postdepositional processes on porew ater concentration gradients in lake sediments. They homogenized lake sediments and placed them into aquaria, and within five weeks steep concentration gradients had beco me established in their microc osms. They concluded that postdepositional processes may obscure any relati onship between P concentration in surficial sediments and in the original sedimented material Mineralization and flux of P from the soil to water column may therefore accoun t for much of the observed dec line in P with depth in lake sediments and wetland soils. Mathematical Model of Changes in Phosphorus Fractions Temporal changes in soil phosphor us fractions can be depicted as a system of storages and transfers by coupling the data from the phosphorus sequential fractionation with soil age data. However, whether changes occurred through adv ective losses from the wetland, or by transfers into other P fractions cannot be ascertained from this research. In the model proposed here, all changes (typically declines) in respective fractions result in direct loss from the system (Fig. 213). Each fraction was modeled as a first-orde r process, with conc entration declines proportional to the concentration at the beginni ng of the time step. The unimpacted station in BCMCA was used to calibrate the model. Dec lines in soil P fractions were modeled as

PAGE 41

41 P(t) = Pi exp(-k*t) + P* (2-3) where Pi = P at t = 0 mg kg-1, k = first order rate constant, yrs-1, t = time, years, and P* = longterm asymptotic P value. For model calibration an d validation, the time verses depth model was adjusted such that the surface (0-10 cm) sample represented time zero for both stations. This was done by subtracting the model-determ ined age of the surface interval from each of the lower soil intervals for respective sa mpling stations. For both BCMCA and WCA-2A, phosphorus concentration asymptotically d eclined with age to a background concentration, then dropped little thereafter. The P* model parameter prevents the P of each fraction from declining below these minimum long-term levels. This model was fit to the P and time data for each fraction using two objective functions: minimizing the sum of the differences between the model prediction and the data (SSR), and by minimizi ng the root mean square error (RMSE). Both objective functions were minimized using the So lver utility in Microsoft Excel. For most fractions, the RMSE objective func tion resulted in the better correlation coefficient between model and data (Table 2-4). Solving for model parameters by minimizing RMSE for humic P did not result in convergence on a solu tion, therefore minimization of SSR was used to determine k and Pi. Whether RSME or SSR was minimized made little difference in solving for k; the average difference in k between the two methods was approximately 8%. Rate of decline in respective fractions, as evidenced by ra te constants was of the order NaHCO3 Pi > microbial biaomass P > HCl Pi > fulvic P > residual P > NaHCO3 Po > Humic P. The apparently slow decline of the NaHCO3 Po fraction is partially explained by th e very low initial concentration of this fraction, typically only 10 mg kg-1 in surface soils. Also, mode l fit to this fraction was relatively poor compared to other fractions, with the model explaining only 72% of variability in the data at the unimpacted station. Hu mic P declined at a rate of 0.09% yr-1 and represented the

PAGE 42

42 lowest loss rate. Humic P increased with increas ing depth, relative to other fractions, at the unimpacted station in BCMCA (Fig. 2-8). The lowe r loss rate of humic P, compared to other fractions, led to enrichment in this fraction at this station, though this enrichment was less apparent at the other three stati ons. The fit of the exponential m odel to the humic P data was the least robust (r2 = 0.48) and this data is probably better described by a zero-order linear process. Bicarbonate extractable inorganic Pi declined fastest, at a rate of 4.6% yr-1, a rate over 50-fold greater than humic P. The model was validated using the nutrien t impacted BCMCA station, with the Pi value of that station substituted into the model (Fig. 2-14). Goodness of model fit to th e actual data at the impacted station was judged by comparing th e RMSE of the station that the model was calibrated to, against the RMSE of the other marsh station. Th is comparison is the RMSE ratio given in Table 2-4. For example, the RMSE ratio of microbial biomass P is 2.2. This implies that the error in prediction of long-te rm changes in microbial biomass P at the impacted station are 2.2 times greater than at the station at which the model was developed. This estimate of longterm behavior is based only on the surficial (0 -10 cm) concentration at the new, unmeasured station, and the rate constant and long-term asymptotic conc entration determined at the calibration station. Model predicti on was worst for humic P, with an RSME ratio of 45. The poor prediction for this fraction is due to the combin ation of high initial va lue of humic P at the unimpacted station and the very lo w rate constant. In essence, this ensures that the modeled humic P always lags actual soil concentration, th us generating a high residual sum of squares for the BCMCA validation station. This disparity indicates some sensit ivity of the model to initial conditions (Pi). It is also possible that, since the m odel was developed for a low nutrient region of the marsh, it may not characterize P dynamics for a region that has been impacted by nutrient

PAGE 43

43 runoff. Despite this potential s hortcoming, the model seems to de scribe long-term declines in most fractions reasonably well. The rate constants (k) an d long-term asymptotes (P*) were incorporated into a systems dynamic model (Stella version 9, isee sytems Inc., Le banon NH). The model provides a simultaneous description of decay kinetics for the seven P fractions (Fig. 5-3). The total P mineralized is summed in a component termed M ineralized. The loss from each pool reflects a first order process in that the amount of P lost for each time step is proportional to the mass contained in respective pools at the beginning of the time step. The transfer, or flow term for each P fraction is simply: P(ti+1)= (P(ti) P*) k (2-4) Where P(ti+1) is the P content at the end of the time step, P(ti) is the P content at the beginning of the time step, P* is the long-term asymptotic P content, and k is the first order rate constant. The model allows for comparisons to be made of relative rates of change among P fractions (Fig. 5-4). The model components consist of a system of stocks and flows, and this system of equations and initial values can be readily imported into other sy stems dynamics software (Table 2-5). To apply this model to a new site, the surf icial soil total P concen tration and the proportion of P in each fraction would be specified as initi al concentrations for respective fractions. For subtropical peat wetlands, rate constants and long-term minimum c oncentrations are likely to be similar to those found here, yet s hould be independently verified. Historical Accumulation of Phosphorus Soil age increased exponentially with depth. However, prio r to 500 YBP, soil age (and therefore accretion) increase was mostly linear (Fig. 2-11). If a linear model is assumed for ages greater than 500 YBP, long-te rm soil accretion in BCMCA ha s been approximately 2 mm yr-1

PAGE 44

44 (Table 2-2). As noted previ ously, at both the Everglades and Blue Cypress Marsh (both impacted and unimpacted stations) soil P levels declined asymptotically to approximately 100 mg kg-1, and declined very little thereafter. The n early steady-state historical soil accumulation rate and total P concentration can be combined to yield an estimate of long-term P accretion of 2.28 and 1.84 mg m-2 yr-1 for stations B4 (unimpacted) a nd C1 (impacted) in BCMCA (Table 22). This value represents the mass transfer rate to very long-term P sinks. This estimate is very low compared to previous accretion estimates for BCMCA and other wetlands. The calculation of wetland P assimilation is highly influenced by the depth interval over which the accretion is calculated, particularly during early diagenesis. For example, Brenner et al. (2001) estimated P accretion in BCMCA prior to 1920 to be 20 mg m-2 day-1, or 10-fold higher than this study. Craft and Richardson (1998) estimated P accumulation rate s in the upper 30-cm soil layer for a station in WCA-2A to be 60 mg P m-2 yr-1, or 30-fold greater than thes e results. For this study, the interval over which rate of accr etion and P concentration were relatively constant was chosen. This period spanned approximately 2,000 years a nd represented a depth interval of 80 150 cm (Table 2-2). This long-term assimilation rate can be contrasted with estimates of overall wetland nutrient loading rates. Richardson et al. (1997) examined in flow and outflow data from over 100 North American treatment wetlands, as well as a site in WCA-2A. Infl ow loading rates ranged from 0.01 to over 1000 g P m-2 yr-1. Analysis of the inflow-outfl ow relationship suggested that wetland loading rates in excess of 1 g P m-2 yr-1 appeared to overwhelm the capacity of the wetland to move P into long-term sinks. For BCMCA the historical rate of movement into longterm pools is much slower than this and coul d not represent the prin cipal mechanism for P removal below this threshold loading rate. Richardson (1985) compared P accumulation in freshwater wetlands and uplands an d concluded that in general, wetlands have a relatively low

PAGE 45

45 capacity to absorb P. He pointed out that even though peat wetlands store massive amounts of P, they do so over very long time intervals. Penn et al. (1995) described th e kinetic changes in P fractions for the sediments of Onondaga Lake, Ne w York. They suggested that for those organicrich sediments, below a depth of 30 cm, almost all of the P was highly resistant to further decomposition due to losses of labile fractions. Th is fraction was constant with increasing depth and therefore represents the long-term pool of accreting P. Lowe and Keenan (1997) developed a conceptual model of nutrient as similation in natural wetlands that predicts areal exte nt of impact on wetlands due to anthropogenic nutrient loading. The model can be thought of as a thin lea ky sponge, with the thickness of the sponge proportional to short-term sinks such as plant biomass and other relatively labile pools. The leakage rate represents movement to long-term pools, with the rate dependent on size of the labile pool. The leakage rate is empirically determ ined from inflow-outflow concentration data in treatment wetlands, thou gh calculation using actual P burial ra tes such as found in this study may provide better protectio n to natural wetlands. Conclusions The peat soils of BCMCA and WCA-2A were si milar in terms of physical properties such as texture, loss on ignition and total N and P content. The unimpacted station in BCMCA was greater in total P than the unimpacted station in WCA-2A, perhaps refl ecting differences in landscape position, whereas the nutrient impacted station in WCA-2A was greater in surface total P compared to the impacted station in BCMCA Over very long time periods, soil total P in both wetlands declined to approximately 100 mg kg-1, and this represents an exponential annual reduction for both wetlands of approximately 0.6% per year. At a depth of approximately 75 cm, further reductions in soil P content are minimal. Soil P accretion at depths lower than this was

PAGE 46

46 approximately 2 mg m-2 yr-1 for BCMCA soils. This is a very low accretion rate compared to other estimates in BCMCA and other wetlands. Soil fractionation results showed that WCA-2A soils are much higher in inorganic P, probably due to the calcareous nature of these so ils. The most dramatic change in P pools with depth was the enrichment of the residual, recal citrant pool. Residual P increased from 20% of total P in surface soil, to 68% at the base of the peat deposit in both wetlands. Changes in P associated with humic and fulvic compounds did not change in a consiste nt way among stations and wetlands. In general, humic P declined with increasing soil depth, while fulvic P was mostly constant. My results cannot disprove conclusively the claim that chronological changes observed in soil P profiles were caused by changing environm ental conditions such as vegetation succession, hydrology, and nutrient loading. However, two findi ngs from this study suggest that observed vertical changes in P are at least in part due to diagenesis. First, marsh stations far removed from sources of anthropogenic nutrient show declines in soil P with depth, though it could be argued that even locations remote from nutri ent inflows are affected by increased 20th century P use. Second, there is a constant downcore increase in re sidual P, as well as a widening in C to P ratios at all stations. It seems unlikely that these soils were enriched in residual P at the time they were accreted. The operationally-defined extraction used in this study distinguished between five organic P pools: microbial, labile Po, fulvic, humic, and residual Po. The greatest changes in the sequential fractionation were observed in the resi dual, unextractable frac tion. This fraction is believed to consist of refractory P compounds, and for peat soils, these are predominately in an organic form. Little is known concerning the ch emical composition of this fraction, though it is

PAGE 47

47 assumed to be recalcitrant due to its resistance to solubilization with the chemicals used. These five fractions constitute a categorical method of characterizing soil organic P stability. Noncategorical (i.e., continuous) methods may be more useful at distinguishi ng the extent to which organic P has been immobili zed. More resolution of Po immobilization is needed to better predict the nutrient release potential of wetland soils, when they are drained for agricultural or urban development; reflooded, as a consequence of restoration projects; constructed with the inte ntion of nutrient removal; In all of these cases, a common concern is how nutrient dynamics will change under the new conditions. For organic soils, a principal dete rminant of this is the stability of Po. Thermal techniques have been used to char acterize the extent of decomposition of peat and may also prove useful in ch aracterizing the continuum of Po stability (Levesque and Dinel 1978). Also, enzymatic attack of wetland soil or ganic matter is the principal means by which Po is broken down to more labile compounds. Thus extraction of peat with phosphoesterases may provide insights into the degr adability of organic wetland so ils. These techniques will be investigated in Chapter 3.

PAGE 48

48 Table 2-1. Basic physico-chemical prope rties of surficia l (0 cm) soils collected in July 2002 from WCA-2A and BCMCA. BCMCA WCA-2A Parameter B4 C1 E1 E5 TN, % 3.213.502.932.42 TC, % 47.448.345.145.7 Dry Weight, % 8.710.310.010.8 pH 5.705.887.597.30 Bulk Density, g cm-3 0.0810.1010.0950.111 Loss on ignition, % 5.88.87.311.1 TP, mg kg-1 327367546197 Table 2-2. Long-term depth, mass, and phosphorus accretion at two locations in Blue Cypress Marsh Conservation Area. Parameter B4 C1 Units Depth interval 85-14575-145cm Age interval 510-2900965-3600YBP Accretion 0.0190.025cm yr-1 Mean total P 11880mg kg-1 Mean bulk density 0.1070.092g cm-3 P accumulation rate 2.281.84mg m-2 yr-1 Table 2-3. Model parameters for long-term soil accretion and phosphorus accretion at two stati ons in Blue Cypress Marsh Conservation Area. Parameter B4 C1 Units Soil accretion: Yo e (k*Z) Accretion rate constant, k 0.0290.026yr-1 Surficial age, Yo 54128YBP Model fit, r2 0.960.79units Phosphorus accretion: C + Co exp(-kt) Ct=0 389729mg kg-1 Ct= 11880mg kg-1 Rate constant -0.006-0.006yr-1 Model fit, r2 0.990.99units

PAGE 49

49 Table 2-4. Results of phosphorus fraction model calibration and validation at two stations in Blue Cypress Marsh Conservation Area. A ll values, except for RSME ratio, from model calibration only. Fraction Objective Function r2 Rate Constant Pi P* RSME Ratio yr-1 mg kg-1 mg kg-1 Microbial RMSE 0.97330.0235128.710 2.2 SS 0.97290.0223130.0 NaHCO3 Pi RMSE 0.99460.046267.83 0.8 SS 0.99400.043269.0 NaHCO3 Po RMSE 0.71550.003311.70 2.3 SS 0.71190.003910.8 Inorganic RMSE 0.99110.021946.62 3.5 SS 0.98900.019547.9 Fulvic RMSE 0.95160.0103102.710 3.4 SS 0.94470.0079106.8 Humic RMSE No Convergence SS 0.48010.000938.610 45aResidual RMSE 0.96830.003886.730 8.8 SS 0.96790.004085.0 Note: aRatio determined from SSR, not RSME. Table 2-5. Rate constants, equations and initia l values used in Stella model of long-term P declines at nutrient unimpacted station in Blue Cypress Marsh Conservation Area. Fraction Fraction (t) Initial Value Transfer Term Fulvic-P Fulv(t dt)+(-Loss_Fulv)*dt 0.226* TPinitial (Fulv-10)*0.0103 Humic-P Hum(t dt)+(Loss_Hum)*dt 0.106* TPinitial (Hum-10)*0.0009 HCl-P Inorg(t dt)+(Loss_Inorg)*dt 0.087* TPinitial (Inorg-2)*0.0219 NaHCO3-Pi Lab_Pi(t dt)+(LossLabPi)*dt 0.123* TPinitial (Lab_Pi-3)*(0.0462) NaHCO3-Po Lab_Po(t dt)+(-Loss_Lab_Po)*dt 0.008* TPinitial Lab_Po*0.0033 Microbial P MBP(t dt)+(Loss_MBP)*dt 0.248* TPinitial (MBP-10)*(0.0235)

PAGE 50

50 Figure 2-1. Sequential fractionation t echnique used in firs t extraction, adapted from Ivanoff et al. 1998. Soil Extractant, Extraction Time, and Fraction Definition A. 0.5M NaHCO3+ CHCl316 hrs Microbial P (TP A. TP B. ) B 0.5M NaHCO316 hrs Labile Pi(SRP) Labile Po(TP-SRP) 1M HCl 3 hrs Non-labile Pi(SRP) 0.5M NaOH 16 hrs C. TP Wet soil residue Recalcitrant Po(TP) Wet Soil Sample Wet Soil Sample Remove aliquot Acidify pH 2; centrifuge FulvicP (TP) HumicP ( C. FulvicP )

PAGE 51

51 Figure 2-2. Soil bulk density (g cm-3) at nutrient impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Wa ter Conservation Area 2A. Error bars represent one standard deviation. Note di fferent x-axis scale for WCA-2A cores. BCMCA Unimpacted -140 -120 -100 -80 -60 -40 -20 0 BCMCA ImpactedDepth, cmBulk Density, g dry cm-3wet0.10.20.10.2 WCA-2A Unimpacted -140 -120 -100 -80 -60 -40 -20 0 0.00.51.01.5 WCA-2A Impacted 0.00.51.01.5 BCMCA Unimpacted -140 -120 -100 -80 -60 -40 -20 0 BCMCA ImpactedDepth, cmBulk Density, g dry cm-3wet0.10.20.10.2 WCA-2A Unimpacted -140 -120 -100 -80 -60 -40 -20 0 0.00.51.01.5 WCA-2A Impacted 0.00.51.01.5

PAGE 52

52 Figure 2-3. Soil loss on ignition ( %) at nutrient impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Wa ter Conservation Area 2A. Error bars represent one standard deviation. Note di fferent x-axis scale for WCA-2A cores -140 -120 -100 -80 -60 -40 -20 0 -140 -120 -100 -80 -60 -40 -20 0 2060100 2060100BCMCA Unimpacted BCMCA UnimpactedLoss on Ignition, %Soil Depth, cm

PAGE 53

53 Figure 2-4. Soil total phosphorus at nutrient impacted and unimpact ed stations in Blue Cypress Marsh Conservation Area and Water Conserva tion Area 2A. Error bars represent one standard deviation. -140 -120 -100 -80 -60 -40 -20 0 -140 -120 -100 -80 -60 -40 -20 0 050010001500 050010001500BCMCA Unimpacted BCMCA Impacted WCA-2A Unimpacted WCA-2A ImpactedDepth, cmTotal P, mg kg-1 -140 -120 -100 -80 -60 -40 -20 0 -140 -120 -100 -80 -60 -40 -20 0 050010001500 050010001500BCMCA Unimpacted BCMCA Impacted WCA-2A Unimpacted WCA-2A ImpactedDepth, cmTotal P, mg kg-1

PAGE 54

54 Figure 2-5. Soil carbon to nitrogen ratio at nutri ent impacted and unimpacted stations in Blue Cypress Marsh Conservation Area and Wa ter Conservation Area 2A. Error bars represent one standard deviation. -140 -120 -100 -80 -60 -40 -20 0 BCMCA Impacted 102030WCA-2A Impacted BCMCA Unimpacted -140 -120 -100 -80 -60 -40 -20 0 102030WCA-2A Unimpacted -140 -120 -100 -80 -60 -40 -20 0 BCMCA Impacted 102030WCA-2A Impacted 102030WCA-2A Impacted BCMCA Unimpacted -140 -120 -100 -80 -60 -40 -20 0 102030WCA-2A Unimpacted -140 -120 -100 -80 -60 -40 -20 0 102030WCA-2A UnimpactedDepth, cmCarbon : Nitrogen

PAGE 55

55 Figure 2-6. Soil carbon to phosphorus ratio at nutrient impacted a nd unimpacted stations in Blue Cypress Marsh Conservation Area and Wa ter Conservation Area 2A. Error bars represent one standard deviation. -140 -120 -100 -80 -60 -40 -20 0 -140 -120 -100 -80 -60 -40 -20 0 0500010000 0500010000Depth, cmCarbon : PhosphorusBCMCA Unimpacted BCMCA Impacted WCA-2A Unimpacted WCA-2A Impacted -140 -120 -100 -80 -60 -40 -20 0 -140 -120 -100 -80 -60 -40 -20 0 0500010000 0500010000Depth, cmCarbon : PhosphorusBCMCA Unimpacted BCMCA Impacted WCA-2A Unimpacted WCA-2A Impacted

PAGE 56

56 Figure 2-7. Ratio of inorganic phosphorus to or ganic phosphorus at nutrient impacted and unimpacted stations in Blue Cypre ss Marsh Conservation Area and Water Conservation Area 2A. Error bars represent one standard deviation. BCMCAUnimpacted-120 -100 -80 -60 -40 -20 0 BCMCA Impacted WCA2AUnimpacted-120 -100 -80 -60 -40 -20 0 0.00.20.4 WCA2A Impacted0.00.20.4Depth, cmTPi:TPo

PAGE 57

57 Figure 2-8. Results of sequential chemical frac tionation. Values are % of total P extracted. 050100 0 10 20 30 40 50 60 70 80 100 110 120 130WCA2A Impacted 90 050100 WCA2A Unimpacted 050100BCMCA Impacted 050100 BCMCA UnimpactedLabile Pi Non-labile Pi Labile Po MBP FAP HAP Residual P 050100 0 10 20 30 40 50 60 70 80 100 110 120 130WCA2A Impacted 90 050100 WCA2A Unimpacted 050100BCMCA Impacted 050100 BCMCA UnimpactedLabile Pi Non-labile Pi Labile Po MBP FAP HAP Residual P Labile Pi Non-labile Pi Labile Po MBP FAP HAP Residual P

PAGE 58

58 Figure 2-9. Ratio of humic to fulvic acid in sequentially extracted soils. 0123 -160 -140 -120 -100 -80 -60 -40 -20 0 0123Depth, cmHumic: FulvicImpacted Unimpacted Impacted UnimpactedBCMCAWCA-2A 0123 -160 -140 -120 -100 -80 -60 -40 -20 0 0123Depth, cmHumic: FulvicImpacted Unimpacted Impacted UnimpactedBCMCAWCA-2A

PAGE 59

59 Figure 2-10. Rate of increase in the % residual (unextractable) phosphorus fraction with respect to time in Blue Cypress Marsh Conservation Area. Time, YBP 0100020003000 Residual P, % of TP 10 20 30 40 50 Time, YBP 0100020003000 ResP = 29 + 0.012*t r2 = 0.39 p = 0.038 ResP = 25 + 0.012*t r2 = 0.73 p < 0.001 UnimpactedImpacted

PAGE 60

60 Figure 2-11. Relationship between soil depth and age in Blue Cypress Marsh Conservation Area. Error bars represent the analytical 14C counting error of one standard deviation. Age = 54*e(0.03*depth)r2= 0.96 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 050100150 Depth, cm Age = 128*e(0.03*depth)r2= 0.79Age, YBPUnimpacted Impacted Age = 54*e(0.03*depth)r2= 0.96 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 050100150 Depth, cm Age = 128*e(0.03*depth)r2= 0.79Age, YBPUnimpacted Impacted

PAGE 61

61 Figure 2-12. Relationship between soil age and to tal phosphorus concentration in Blue Cypress Marsh Conservation Area. Error bars repr esent one standard deviation of three replicate measures of soil total phosphorus. Soil Age, YBP 0200400600800100012001400 Total Phosphorus, mg kg-1 0 200 400 600 800 1000 Total Phosphorus, mg kg-1 0 200 400 600 800 1000 TP = 80 + 729*exp(-0.006*t) r2 = 0.99 BCMCA Unimpacted TP = 118 + 389*exp(-0.006*t) r2 = 0.99 BCMCA Impacted

PAGE 62

62 Figure 2-13. Dynamic model of storages and transfers from major phosphorus fractions. Lab Pi Lab Po Inorg Fulv Hum Mineralized LossLabPi Loss Lab Po LossInorg LossFulv Loss Hum MBP Loss MBP Residual Loss Res Total P initial Lab Pi Lab Po Inorg Fulv Hum Mineralized LossLabPi Loss Lab Po LossInorg LossFulv Loss Hum MBP Loss MBP Residual Loss Res Total P initial

PAGE 63

63 Figure 2-14. Results of model ca libration and validation. Calibrati on and validation performed at unimpacted and impacted BCMCA stati ons, respectively. So lid line indicates modeled P fractions at impacted station. Fulvic Validation Calibration Model Phosphorus, mg kg-1Microbial Humic Validation Calibration Model 50 100 150 200 250 Validation Calibration Model Residual 0 50 100 150 200 250 Validation Calibration Model 0 NaHCO3PiNaHCO3PoHClPi 0 25 50 75 100 Validation Calibration Model 110100100010000 Time, years Validation Calibration Model 0 25 50 75 100 110100100010000 Time, years Validation Calibration Model Fulvic Validation Calibration Model Phosphorus, mg kg-1Microbial Humic Validation Calibration Model 50 100 150 200 250 Validation Calibration Model Residual 0 50 100 150 200 250 Validation Calibration Model 0 NaHCO3PiNaHCO3PoHClPi 0 25 50 75 100 Validation Calibration Model 110100100010000 Time, years Validation Calibration Model 0 25 50 75 100 110100100010000 Time, years Validation Calibration Model

PAGE 64

64 CHAPTER 3 ESTIMATING STABILITY OF ORGANIC PHOSPHORUS IN WETLAND SOILS Introduction Fractionation of soil P into organic, inorganic, labile and recalcitra nt pools is useful for estimating effects of crop manage ment practices in agricultural settings, and to determine soil nutrient status in natural ecosystems. Numerous pr ocedures have been deve loped to divide soil P into respective pools. Chang and Jacks on (1957) developed a sequential inorganic P fractionation procedure that used sequential extraction with NH4Cl, NaOH, and H2SO4 to characterize P according to its a ssociation with soil minerals. Sommers et al. (1972) used a sequential extraction to characte rize organic P in lake sediment s, followed by anion exchange chromatography of the extracts. Th eir sequential extraction consisted of a 1N HCl extract, a cold 0.3N NaOH, and finally a 0.3 N hot (90oC) NaOH extract. The extractant solutions were eluted through an anion exchange column to identify P fractions based on their interactions with the column material. The second (hot) NaOH extract removed an additional 30% of the organic P present in one of the highly organic lake se diments in their study. They believed the hot NaOH fraction consisted of biologically resistant organic P forms. Their overall objective was to characterize the type of organic P present to gain insight into the processes relating to mineralization and immobilization. Hieltjes and Lijklema (1980) developed a similar scheme for the fractionation of inorganic P in calcareous lake sediments. Their technique was similar to Chang and Jackson (1957), but the main objective was to characterize inorganic P sources to the water column of lakes. They emphasized the ne cessity of removing inorganic calcium with an initial alcohol NH4Cl extract, prior to extraction with NaOH, to avoid precipitation of Ca-P compounds. Hieltjes and Lijklema also fracti onated inorganic P com pounds with a soil to solution ratio of 1:1000. They extracted P from the inorganic P compounds iron hydroxy-

PAGE 65

65 phosphate and aluminum orthophosphate, as well as from calcium phosphates. They also demonstrated the importance of the sequence of extractions, finding that when performed alone, the 0.5M HCl extracted 99% of the P bound in the Fe and Al minerals, as well as the P in calcium phosphates. The authors pointed out th e characterization of p hosphates is defined by the analytical procedure. The method proposed by Hedley et al. (1982) was also similar to Chang and Jackson (1957), but included an attempt to measure microb ial P. Both techniques employed a preliminary exchange extraction, followed by extraction with base, and then acid. Hedley at al. (1982) wanted to examine changes in P sequestration as a function of tillage practices, so they investigated organic P fractions. The authors realized that identif ication of particular organic P compounds was beyond their capabilities, and stated A t present, this is an impossible task as much of soil P remains unidentified. Their study soils were low in organic carbon (approx. 3%) compared to typical wetland soils. The authors classified organic P (Po) into four categories, bicarbonate-extractable Po, microbial biomass Po, alkaline extractable Po, and residual Po. The bulk of the changes in soil P resulting from the two tillage practices occurred in the residual (non-extracted) P fraction. The difference betwee n the two tillage practices indicated that refractory P had declined as a re sult of crop harvesting. Thus, the crop had depleted labile pools, and caused the replenishment of this pool by in creasing mineralization of refractory pools. Bowman (1989) proposed a method to determine total organic P in soils. The method was a sequential technique, first us ing a strong acid extractant (H2SO4), followed by a base (0.5 M NaOH). The acid extraction involved the generation of considerable heat, as water was added to a dried soil sample containing concentrat ed acid. For both extraction steps, Po was determined by the difference of inorganic P (Pi) in the extractant of a digested and undigested aliquot of the

PAGE 66

66 same sample. The acid extraction step extracted 79% of the Po present in the soil sample, without causing significant hydrolysis of organic P. Any hydrolysis of Po that did occur, the author believed occurred in the highly labi le organic P sources. Since this pool is a very small fraction of TPo, its addition to the Pi pool was not a significant error. Bo wman (1980) also points out that the proposed extraction technique does not identify the nature of the Po extracted from the unamended soil. Ivanoff et al. (1998) focused their attempts at soil P fractionation on organic soils. Their objective was to establish stability of P in th e Everglades. Their tec hnique used a sequential extraction with bicarbonate, HCl, and NaOH. No attempt was made to distinguish between inorganic P pools; all Pi was removed with a 1M HCl ex tract immediately following a 0.5M NaHCO3 extract. They used standard organic P compounds of known compos ition and lability to determine how much Po was hydrolyzed to Pi as a consequence (or artifact) of the extractant. They found that, except for the hi ghly labile organic P material they used, acid pretreatment hydrolyzed only 1 2 % of the Po present. Also, 0.5 M NaOH caused little hydrolysis of Po, with the exception of glucose-6-phosphate (41% hydrolyzed). This indi cates the potential for this extractant to hydrolyze organic P, causing possible overestimation of Pi, at the expense of Po. They extracted soils from the Everglades Ag ricultural Area and found th at the acid extract removed approximately 12% of the Po pool without causing significan t hydrolysis of this pool. The acid extractable pool probably includes ami no-sugars, amino acids and other acid-soluble organo-P compounds, as well as Ca-P compounds (S tevenson 1996). Ivanoff et al. (1998) were also able to demonstrate that the Murphy-Riley reagent used in the P determination caused minimal (<1.5%) hydrolysis of the organic P contained in gly cerophosphate, glucose-6phosphate, and phytic acid. Of the total Po in their soils, approximately 40% was recovered in the

PAGE 67

67 NaOH extract, 30% was attributed to microbial biomass, and 17% was non-extractable. The alkaline extractable fraction was further divided into two fracti ons after extrac tion by acidifying the extract, causing precipitation of humic materials. This allowe d for separate determination of the P content of humic and fulvic acids. Kuo (1996) described an ignition technique for TPo determination, where ignited (550oC) and un-ignited samples are compared and TPo is determined by difference. The rationale was that organic compounds are co mpletely oxidized at high temperature and the P associated with these compounds is liberated. Characterization of the organic P pool is esp ecially important in wetland soils, since this tends to be the dominant fraction (Wetzel 1999, Reddy and DeLaune 2007) Reddy et al. (1998) found that organic P represented from 56-70% of total P for the peat soils of the Everglades Water Conservation Areas. Its mineralization ra te can be an important regulator of primary production, especially in P-limite d wetlands and other aquatic ecosystems. This rate is determined primarily by the recalcitrance of the organic P po ol. Stability of Po has been shown to increase with soil depth in we tlands. Reddy et al. (1998) has s hown that in WCA-2A in the Everglades, refractory P increased from 33% of total P in surface soils, to 70% in deeper strata. Similar trends in refractory P as those observed in the northern Everglad es were found in soils from BCMCA. Olila and Reddy (1995) showed gene rally increasing residual (or recalcitrant) P content, increasing from 20% (of TP) in surface ho rizons to approximately 80% at a depth of 35 cm. Determination of organic-P stability can be a us eful metric of the extent to which an aquatic ecosystem may be P-limited. Chemical extraction of Po from soils provides little information on the stability of organic phosphorus compounds beyond classification into several operationally defined groups. New techniques are needed to provi de further resolution of the extent to which Po has become immobilized. The ob jective of this study was to characterize the stability of

PAGE 68

68 organic P in peat wetland so ils using three techniques; two new thermal-based and one enzymatic extraction. The main hypothesis was that as organic P becomes increasingly recalcitrant, more thermal energy is required to liberate it from the organic compounds with which it is associated. Methods Site Description Field sampling was conducted in two subtropi cal wetlands, Water C onservation Area 2A (WCA-2A) in the northern Florida Everglades and Blue Cypress Ma rsh Conservation Area (BCMCA) in east central Florida, USA. Th e peat soils are derived from sawgrass ( Cladium jamaicense Crantz) (Davis 1943). They are approximately 2 m in thickness and the basal deposits were formed approximately 5,000 years be fore present (Gleason and Stone 1994). They thus present a vertical profile that progresses from recently deposited plant litter that can be presumed to be relatively labile, to deeper organic deposits that are recalcitrant due to millennia of biogeochemical weathering. Two sampling loca tions were selected in WCA-2A, long-term South Florida Water Management Di strict research stations E1 a nd E5 (Fig. 1-2). Station E1 is located in a nutrient-impacted region, approximately 1.4 km south of South Florida Water Management Districts (SFWMD) inflow structure S-10C. Station E5 is located approximately 10 km south of this structure, in the un-impact ed central marsh. Vegetation at station E1 was predominantly cattail ( Typha domingensis Pers.), and at station E5, sawgrass ( Cladium jamaicense Crantz). Soils were also collected from Blue Cypr ess Marsh Conservation Area (BCMCA). The BCMCA is a 116 km2 shallow marsh in the headwaters of th e St. Johns River. It is also underlain by peat soils that vary from 1.5 m thick in the eastern marsh to gr eater than 4 m thick near the center. The soils of BCMCA appear to be both higher in P than those of the WCA-2A (Grunwald

PAGE 69

69 et al. 2006) and less P-limited (Prenger and Re ddy 2004). The vegetation is a mosaic of open slough communities ( Nymphaea odorata, Eleocharis elongata, Utricularia spp .), Maidencane flats ( Panicum hemitomon ), broad expanses of sa wgrass, and tree islands ( Acer rubrum, Taxodium distichum, Myrica cerifera ). Agricultural discharges have created a region of elevated soil P in the northeastern ar ea of the marsh, though those di scharges ceased in 1994. One sampling location was selected from this region of the marsh, Station C1, and one station was selected from the interior unimpacted pa rt of the marsh, Station B4 (Fig. 1-3) Sampling and Analysis Soil samples were collected in triplicate from all stations in both marshes on July 17 and 18, 2001. Soil cores were obtained with a 2-m long x 7.3-cm-diameter stainless steel soil corer. Soil samples were extruded in the field into Zipl oc bags at 10-cm intervals. The samples were immediately chilled to 4oC and were kept on ice for trans port to the laboratory. Laboratory processing was performed at the University of Floridas Wetland Biogeochemistry Laboratory. Wet weight was recorded for each sample and sa mples were transferred to rigid polyethylene containers and thoroughly mixed. Three P extrac tion techniques were used to characterize soil organic P. Method 1: autoclave extractable P. The samples used in this experiment were subsamples of those described in Chapter 2. The techni que used in this extraction is similar to that described by Kenney et al. (2000). Twenty-five ml of deionized water was added to 0.5 g of oven-dried equivalent wet soil. The soils were shaken at room temperature for 3 hours, centrifuged, and the supernatant water wa s removed and vacuum filtered through 0.45 M poresize polyethersulfone filters. The residual soil sample was then placed into an autoclave for

PAGE 70

70 90 minutes at 128oC and 1.7 atmospheres (25 lbs in-2). The sample was removed from the autoclave, 20 ml of deionized water was adde d, and the sample was equilibrated for a one-hour, end-to-end shaking period. The sample was then vacuum filtered through 0.45 M filters. Extracts were analyzed for dissolved reactive P (DRP) and total P. Triplicate samples of model organic P compounds glycerophosphate and phytic ac id were also extrac ted using the above procedure to determine the extent of organi c phosphorus mineralization in model compounds. The difference between the initial deionized wate r extraction and the post-autoclave extract is termed hot water extractable P (HEP). Method 2: pyrolytic extraction of organic P. The soil samples used in this experiment were subsamples of those used in the previous extraction. Pyrolytic extr action of organic P was performed in a furnace that was constantly purged of O2 with N2. This method is similar to a thermal gravimetric method used to characterize the degree of decomposition in peat (Levesque and Dinel 1978). The oven consisted of a 7.6-cm (3-in) diameter ga lvanized steel pipe with two threaded end caps. A 120V heating element was spiral-wound around the pipe and connected to a temperature controller. The end caps were drille d and threaded to accept a gas inlet and outlet port, and the entire assembly was placed into an insulated metal box. One end cap was removed and a batch of samples was placed into the furn ace. Samples were first extracted at room temperature with 1M HCl on a re ciprocating shaker for three hours to obtain an estimate of inorganic P. Oven dried samples were weighed in triplicate onto 5 cm squares of aluminum foil. The foil squares were placed onto a tray, the tray was inserted into the oven, and the oven was sealed. The chamber was purged for approximately 10 min and was then increased to the set point temperature. It remained at this temperature for one hour for each treatment temperature. Three soil depths (0 -10, 40 50, and 90 100 cm ) were used for this experiment. Pyrolysis

PAGE 71

71 temperatures were 160, 200, 260, 300, 360, and 550oC. The samples were removed from the oven and transferred to 43 ml centrifuge tubes. Th ey were then extracted with 1M HCl at a 1:50 ratio for three hours on a reciproca ting shaker, then filtered through 0.45 M filters. The P that was extracted at room temperature with 1M HCl was subtracted from each value to yield a net pyrolyzed P. Method 3: enzymatic hydrolysis of soil organic phosphorus. Organic P that is biologically available via extr acellular enzymes was determin ed using phosphatase enzymes. Soils from only the unimpacted sites in WCA-2A and BCMCA we re used in this study. A 0.5 g oven-dried equivalent fresh soil was weighed into 50 ml centrifuge tubes. Each enzymatically treated tube was amended with 5 ml of modi fied universal buffer (M UB) containing 25 enzyme units of akaline phosphatase monoesterase (Sig ma-Aldrich Co., St Louis, MO, catalog no. P8361, lot no. 062K1359). Stock MUB consisted of the following: 12.1 g tris hydroxymethyl amino methane (THAM), 11.6 g maleic acid, 14 g citric acid, 6.3 g boric acid, and 488 ml 1N NaOH diluted in 1 L of deionized water (Tab atabai 1982). Two hundred ml of the stock MUB was diluted to 500 ml, and this was pH adjusted with 0.1N NaOH to pH 10, the optimum for this enzyme. The enzyme concentration in the st ock MUB was high, and represents enzyme nonlimiting conditions. The activity of the MUB + enzyme solution was qualitatively examined on a fluorometer and high activity wa s confirmed. One set of samples was treated with the MUB + enzyme, and a duplicate set was treated with MUB only. Eight samples were incubated in triplicate to examine the reproduc ibility of the technique. All samples were incubated at 30oC for 4 hours and then extracted with 0.5M NaHCO3 (pH = 8.24) for 30 min on a reciprocating shaker. Samples were then directly filtered, without centrifuging, through 0.45 M

PAGE 72

72 polyethersulfone filters. Enzymatically extractable P was determined by difference between the MUB and MUB + enzyme samples. Statistical comparisons between pyrolysis tr eaments were perfomed with the General Linear Model procedure in MiniTab (rel. 14.2) with fixed effects (temperatu re and soil depth) and crossed factors. Differences among individual treatements were determined using Tukeys multiple post-hoc comparison, and judged to be significant at p<0.05. Regression analysis (HEP experiment) was performed using the Microsoft Ex cel (ver. 2003) Data Analysis ToolPak. Slope was judged significantly diffe rent from zero for p<0.05. Results and Discussion Autoclave Extractable Phosphorus Both Everglades WCA-2A stations showed much lower HEP than the BCMCA stations (Fig. 3-1). HEP accounted for 10 -50% of total P in surficial (0-10 cm) soils, declining to 5-10% of total P at a depth of 60-cm in all cores. All stations showed dec lines in HEP with depth, indicating increasing resistance to thermal breakdown with dept h, or soil age (Table 3-2). HEP of WCA-2A soils was considerably lower than BCMC A soils, ranging from 8% to 15% of total P at the soil surface, declining to 2-5% of total P at a depth of 60-cm, compared to 36-47 % of total P in BCMCA surface soils and 11-19% at 60-cm. In ge neral, the proportion of total P that could be extracted with hot water declin ed at a rate of 0.04 to 0.22 % cm-1 at the two Everglades stations, and approximately 0.4% cm-1 to 0.6% cm-1 at the BCMCA stations. This decline suggests that deeper soils were more resistant to thermal decomposition than surface soils. The dramatic difference between the HEP in th e two wetlands may be due to differences in calcium concentration. WCA-2A is a hard water system, with high calcium particularly near inflow sources, whereas BCMCA is a low pH soft-water system. Since calcium carbonate solubility declines with increasing temperature, precipitation of calcite in Everglades samples

PAGE 73

73 may have occurred during autoclaving, and this new carbonate may have sorbed any P released during the extraction. If this was the case, it is not surprising th at both BCMCA stations showed higher levels of HEP than even the most impacted station, E5. Modifying th e procedure such that the final, post-autoclave extrac tion was performed with 1M HCl would re-dissolv e any calcite-P compounds that may have been formed, t hus preventing this potential artifact. Hot water extractable P explained 69 percent of the variability in microbial biomass P (Fig. 3-2). The slope of the linear f it of the HEP to microbial biomass P data implies approximately 48 percent of the P extracted in the HEP extract came from the microbial biomass pool. This is consistent with other researchers who attribut e most HEP to bacterial storage of polyphosphates (Kenney et al. 2000). The relationship between HEP and water soluble P (WSP) was not as robust, probably due to removal of WSP in th e initial deionized wate r extract (Fig. 3-2). Approximately 60% of both glycerophosphate and phytic acid was mineralized at the experimental temperature and pressu re (Fig. 3-3). This indicates th at even at the relatively low temperature used here, there was considerable breakdown of recalcitrant organic P that was incorporated into phytin. This fu rther suggests that hot water extr actions may liberate more than microbial and algal stores of polyphosphates, as has been postulated by other researchers (Kenney et al. 2000). Pyrolytic Extractable Phosphorus When the results from all depths and both wetlands are combined, increasing temperature resulted in greater mineralization of or ganic P, with the exception of the 550oC treatment (Fig. 34). At this temperature, P recovery declined dramatically. This d ecline was confirmed by repeating the experiment with onl y the last two temperatures 360oC and 550oC and the results were the same; paradoxically reduce d recovery of P. One possible explanation for this is that mineralized P in the reducing environment of th e pyrolysis chamber was converted to phosphine

PAGE 74

74 gas (PH3) and lost from the system. Loss of phosphine gas from wetland soils has been suggested as a significant biogeochemical pathway for P export from wetlands (Devai and DeLaune 1995). However, when the P from the post-pyrolysis extr act plus the total residue P was summed for the two temperatures, they were found to be equivalent, indicating no losses from the system. It is not clear what caused the large reduction in 1M HCl extractabl e P in the highest temperature treatment, however one plausible explanation is the formation of a material with strongly hydrophobic properties. For example, one method of treating wood for decay resistance, called torrefaction, is to subject it to a high temperatur e, inert atmosphere (Fel fli et al. 2005). This treatment apparently eliminates some chemica lly-bound water, and the fi nal product is markedly more hydrophobic. The hydrophobicity of the soil after the 550oC treatment may have prevented the dilute acid access to the soil matrix. The results from this and the previous 550oC treatment indicate that this temperature ma y be too high to be useful for ch aracterizing the r ecalcitrance of organic P. If the 550oC treatment is removed from the analysis, increasing temperature significantly increased mineraliz ation of organic P (Fig. 3-5, Table 3-3). The significant (p<0.001) interaction term can be interpreted to mean that some temperatures used in the extraction were not effective at distinguishing between the thre e soil depths used in this experiment. For example, the 160oC treatment extracted significantly more P from the shallow (0-10 cm) soil sample, but was incapable of dis tinguishing between the mid (40-50 cm) and deep (90-100 cm) sample. The temperature that most e ffectively distinguished between the shallow, mid, and deep sample was 360oC. When stations from both we tlands are combined for the 360oC treatment, there were significant differences amo ng depths in mean extrac table P for all shallow, mid, and deep samples: 64%, 80%, and 84% of P extracted, respectively (p = 0.013) (Table 3-4). There is therefore evidence to support the hypothesis th at the amount of thermal energy needed

PAGE 75

75 to mineralize organic P is dependent on its r ecalcitrance. Further subdividing the dataset by wetland and nutrient impact status reduces respective sample sizes and statistically significant differences among soil depths decline accordingly. If the analysis is repeated on the surficial peat, more P is liberated at lower temperature, for instance 75% at 300oC. This suggests that the less decomposed peat is not as resistant to thermal decomposition. The lower temperatures extracted ve ry little P in the de ep peat, approximately 2% at 160oC and 10% at 200oC, compared to 22% and 38% in the surface peat. As each sample was subjected to progressive ly higher heat, weight loss of the sample increased, as would be expected. This weight loss was likely due to loss of gaseous carbon thermal decomposition products. Bartkowiak and Zakrzewski (2004) subjected extracted lignin to pyrolytic thermal gravimet ry and found that the lignin ha d a declining hydrogen content, relative to carbon and oxygen, over th e range of temperatures they us ed. They attributed this to destruction and loss of methoxyl groups asso ciated with the lignin molecule. At 360oC, the BCMCA samples had lost as much as 40% of th eir initial dry weight. There was a relationship between the sample dry weight loss and the amount of P that could be extr acted with the dilute acid, with 73% of the variance in extractable P accounted for by weight lo st during the pyrolysis procedure. For every 10% increase in weight loss there was a 24% increase in extractable P. This suggests that the increase in extractable P was due to destruction of soil organic matter and subsequent liberation of P from organic compounds. Enzymatic Hydrolysis The amount of P that was enzymatically extr actable was very low. There are several possible explanations for this. The microbial co mmunities at both stations are functioning in a Plimited environment, though less so at BCMCA st ation B4. This would have the effect of increasing enzyme production in those organism s capable of producing it. The small pool of

PAGE 76

76 organic P that could be hydrol yzed by the added phosphatase ma y thus have already been exposed to the native phosphatase thus reducing the pool that wa s susceptible to this enzyme. Another possible explanation is that the dominant form of orga nic P in these wetlands was diester-P, and therefore was not hydrolyzable with the monoest erase used in this study. Conclusions The amount of P that could be extracted with hot water declined with depth at both stations in BCMCA, and at the unimpacted station in WCA-2A. The variable re sults observed in WCA2A may be due to higher calcium levels in the soils of WCA2A, particularly near inflow sources. This procedure may better characterize stability of organic P if calcium is first removed with a weak acid extrac t prior to autoclaving. For the low calcium soils of BCMCA, HEP ranged from approximately 40 % of total P in BCMCA su rface soils, declining to 15% at a depth of 60cm. The increasing resistance to extraction with hot water with soil depth suggests that HEP may be a useful method to characterize the stability of the organic P pool. Pyrolytic extraction of organic P yielded similar results. Since the final step with this procedure was extraction with weak acid, potentia l underestimation of P mineralization, due to calcium precipitation, was not observed. Increasing temperatures yielded progressively greater extraction of Po. Also, at any given temper ature, significantly more P was extracted from surface soils than from deeper deposits. The most effectiv e temperature for distingui shing the stability of Po was 360oC. Addition of organic P hydrolyzing enzy mes did not cause significant hydrolysis of Po. Results from this study suggest that 1) thermal stability of organic P seems to be related to the recalcitrance of the compound that contains th e organic P, and 2) thermal extraction of Po from wetland soil may provide a useful diagnostic tool for investigati ng organic P lability.

PAGE 77

77 The results from this study do not provide information on the mechanism for increasing recalcitrance. Similarly, the results from sequent ial chemical fractionatio n (Chapter 2) provide little compound-specific information on long-term changes that occur in Po in wetland soils. This is where investigative techniques such as stab le isotope analysis, Cand P-Nuclear Magnetic Resonance, spectrophotometric, an d fluorometric techniques may pr ovide further insights. These methods were used in the next chapter to bette r elucidate long-term changes in both soil organic matter and soil Po.

PAGE 78

78 Table 3-1. Basic physico-ch emical properties of su rficial (0 cm) soils collected in July 2002 from WCA-2A and BCMCA. BCMCA WCA-2A Parameter B4 C1 E1 E5 TN, % 3.213.502.932.42 TC, % 47.448.345.145.7 Dry Weight, % 8.710.310.010.8 pH 5.705.887.597.30 Bulk Dens., g cm-3 0.0810.1010.0950.111 Ash, % 5.88.87.311.1 TP, mg kg-1 327367546197 Table 3-2. Regression statistics for autoclave ex tractable P verses depth in soil samples from Blue Cypress Marsh Conservation Ar ea and Water Conservation Area 2A. Regression statistics based on th e upper 60-cm of all cores. Parameter BCMCA (Impacted) BCMCA (Unimpacted) WCA-2A (Impacted) WCA-2A (Unimpacted) r2 0.840.950.43 0.88 p 0.004<0.0010.112 0.002 Slope -0.366-0.594-0.039 -0.224 Intercept 41.643.49.3 14.6 Table 3-3. ANOVA General Linear Model results of pyrolysis temperature and soil depth effects. 550oC treatment not included in analysis. Source DF Seq. SS Adj. SS Adj. MS F p Depth 2 858170663533 32.2 0.000 Temperature 4 896409038222596 205.7 0.000 Depth X Temp. 8 45214521565 5.14 0.000 Error 144 1582015820110 Total 158 118564 Table 3-4. ANOVA results of differences between peat sample depths for 360oC treatment. All stations combined for analysis. Two samples excluded from analysis: Station E1 deep sample (sand) and B4 mid depth (incomplete pyrolysis). Source DF SS MS F p Depth 2 222811145.10.013 Error 27 5893218 Total 29 8121

PAGE 79

79 Figure 3-1. Hot water extractable P (HEP) of Bl ue Cypress March Conservation Area and Water Conservation Area 2A soils. Panel A data fr om stations that ha ve been impacted by nutrient-rich inflows. Panel B data from stations that have not been impacted by nutrient inflows. Error bars repr esent one standard deviation. 020406080100 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 BCMCA WCA2A HEP, % of TP 020406080100 BCMCA WCA2A A. B.

PAGE 80

80 Figure 3-2. Hot water extractable phosphorus (HEP ) of Blue Cypress March Conservation Area and Water Conservation Area 2A vs. wate r soluble (WSP) and microbial biomass phosphorus (MBP). y = 0.48x + 12.9 r2= 0.691 0 50 100 150 200 250MBP, mg/kg y = 0.049x + 0.387 r2 = 0.373 0 5 10 15 20 25 30 35 050100150200250300350400 HEP, mg/kgWSP, mg/kg y = 0.48x + 12.9 r2= 0.691 0 50 100 150 200 250MBP, mg/kg y = 0.049x + 0.387 r2 = 0.373 0 5 10 15 20 25 30 35 050100150200250300350400 HEP, mg/kgWSP, mg/kg

PAGE 81

81 Figure 3-3. Hot water extractable P (HEP) of two model compounds, glycerophosphate (GP) and phytic acid (PA). Error bars repr esent one standard deviation. GPPA % of Total P Extracted 0 20 40 60 80 100 No Heat Heat GPPA % of Total P Extracted 0 20 40 60 80 100 No Heat Heat

PAGE 82

82 Figure 3-4. Percent of total phosphorus extrac ted from dried, ground peat using pyrolytic extraction under nitrogen atmosphere. 0 20 40 60 80 100 120 WCA2A Impacted WCA2A Unimpacted BCMCA Impacted BCMCA Unimpacted % Phosphorus Extracted 0 20 40 60 80 100 120 Temperature, o C 160200260300360550 0 20 40 60 80 100 120 0-10 cm 90-100 cm 40-50 cm

PAGE 83

83 Figure 3-5. Thermal extraction of organic P. Bars with same letter not significantly different (p>0.05). Each bar represents six samples. Letters beneath x-axis refer to sample depth interval: d (deep) = 90-100 cm, m (middle) = 40 50 cm, and s (shallow) = 010 cm. The 550oC treatment is not shown. Temperature Depth 360 300 260 200 160 s m d s m d s m d s m d s m d 120 100 80 60 40 20 0 WCA2Aa b a b c b a a b d c d c d c d cef c d e f e g d f f% Total P Extracted Temperature Depth 360 300 260 200 160 s m d s m d s m d s m d s m d 120 100 80 60 40 20 0 BCMCAa b c a b a b a c e hcec e h e d g d f d f d h f f g% Total P Extracted Temperature Depth 360 300 260 200 160 s m d s m d s m d s m d s m d 120 100 80 60 40 20 0 WCA2Aa b a b c b a a b d c d c d c d cef c d e f e g d f f% Total P Extracted Temperature Depth 360 300 260 200 160 s m d s m d s m d s m d s m d 120 100 80 60 40 20 0 BCMCAa b c a b a b a c e hcec e h e d g d f d f d h f f g% Total P Extracted

PAGE 84

84 CHAPTER 4 CHARACTERIZATION OF LONG-TERM CHAN GES IN ORGANIC PHOSPHORUS IN A SUBTROPICAL WETLAND Introduction Organic phosphorus (Po) is often the principal pool of phosphorus in aquatic ecosystems (Wetzel 1999, Reddy et al. 1998). Its mineralization rate, chemical structure, and pool sizes are thus important determinants of primary producti on in aquatic ecosystems (Wetzel 1999). Less than 50% of the compounds that comprise the Po pool have been identifi ed to date (Stevenson 1994). Soil Po can be grouped into several main categorie s: inositol phosphates, nucleic acids, phospholipids, with phosphoproteins and metabo lic phosphates making up a minor contribution to soil Po (Stevenson 1994). Phosphoesters can be furthe r subdivided into monoesters, diesters, and triesters depending on the number of cova lent ester linkages to ortho-P molecules. Monoesters include inositol hexa phosphate (phytate), a major component of plant seeds and ATP; this group accounts for approximately half of the total soil organic P pool. Diester P is a structural component of nucleic acids and cell me mbranes. This group is rapidly mineralized in upland soils, although Turner and Newman (2005) found that phosphate diesters dominated the Po fraction in a study of soils from the Florida Everglades. Chemical fractionation of Po with sequential acid and base extractions provides insight into respective inorganic and organic pool sizes, as well as some information on operationally defined fractions. However, more compound-specific techni ques are needed if bette r resolution of the many compounds that comprise the Po pool is desired. Chromatographic separation has been used to identify specific Po compounds (McKelvie 2005). Gel filtration chromatography is useful for separating Po according to molecular weight. In this technique, water samples or soil extracts can be passed through non-ionic columns, the elut ed fraction collected, a nd analyzed for total P and molybdate-reactive P. Ion exchange columns ha ve also been used to separate ionizable Po

PAGE 85

85 compounds. Numerous researchers have used 31P Nuclear Magnetic Resonance (31P-NMR) to identify broad classes of Po compounds in soil extracts (Tur ner et al. 2003b, Cade-Menun 2005). This technique involves extracting soil with Na OH, concentrating the sample, and placing the sample into a strong magnetic field. The sample is then subjected to pulsed radio frequency energy, energy is absorbed, and then re-emitted by the sample. The energy required for resonance of the 31P nuclei emission is determined in part by the amount of el ectron shielding of the Po molecule and this property is used to classify the Po. Mass spectrometry has also been used for characterizing Po, though only in surface waters (C ooper et al. 2005). This involves production of gas-phase atoms, ionization, separation based on mass to charge ratio, and detection. The difficulty in non-destructively gene rating ionized, gas-phase molecules from large Po molecules and distinguishing between Po and the background of ot her organic compounds in environmental samples has slowed the development of this method. New techniques such as electrospray ionization a nd matrix assisted laser desorption ionization (MALDI) appear to have overcome some of these limitations (Cooper at al. 2005). Established methods based on organic matter spectral absorbance (Chen et al. 1977) and fluorescence (C ox et al. 2000) may also be useful for characterizing Po, since fate of organic matter and Po are closely coupled (Stewart and Tiessen 1987). In this study, several techniques were us ed to characterize the biogeochemical transformations that have occurred in the soil profile. The optical prope rties of extracted soil organic matter have been shown to be related to average molecular wei ght and the extent of humification (Cox et al. 2000, Chen et al. 1977). Natural dissolve d organic matter and extracted soil organic matter contain molecules that are cap able of absorbing light energy and promoting an electron to an excited state. Upon returning to the ground state, the molecule emits light, i.e.,

PAGE 86

86 it fluoresces. The pattern of fluorescence depends on the nature of the organic molecule. It has been shown that as organic matter ages, the fl uoresced light is longer in wavelength, or redshifted. By taking the ratio of the intensities of the higher fluorescence wavelengths to the lower wavelengths, the extent of humi fication can be approximated (C ox et al. 2000). This index may be useful as a measure of the extent of soil diagenesis. The ratio of absorbance of light at 465 and 665 nm, or E4:E6 ratio, has been shown to be related to viscosity, which in tu rn is related to th e average molecular weight of compounds in solution (Chen et al. 1977). Increas ing molecular weight has been indicated as a consequence of soil aging and humification (Stevenson 1994). Changes in the depth distribution of broad groups of soil organic carbon compounds were investigated using solid state 13C nuclear magnetic resonance (13C-NMR). Solution 31P-NMR was used in a similar way to examine diagen etic changes in soil organic phosphorus. The preservation of the principal plant polymers, li gnin, cellulose, and hemicellulose was determined, as well as the 13C and 15N content of this material. Lignin, ce llulose, hemicellulose and soluble carbon are the principal carbonaceous plant polyme rs deposited at the soil surface after plant senescence. These polymers form the basis for th e development of the p eat deposit. They also contain an isotopic record of changes in plant co mmunities. Isotopic analysis of the fibric portion of the peat deposit is superior to analysis of bulk peat, as it does not contain inorganic carbon, algal remains, or percolated humic and fulvic compounds. This was used as an indication of changes in floral community composition. Lastly, intact plant macrofoss ils (i.e. seed husks, stems) were dated using 14C to establish the chronology of biogeochemical changes.

PAGE 87

87 Methods Site Description The study location was BCMCA, a 12,000 ha peat ma rsh (Fig. 1-2). The eastern extent of the marsh is characterized by open-water sloughs and tree islands. Typical vegetation in the sloughs is a Utricularia sp. and Nymphaea odorata community, while the tree islands are characterized by Taxodium distichum / Acer rubrum / Salix caroliniana community. The western region of the marsh is predominately a Cladium jamaicense / Panicum hemitomon prairie, with some Cephalanthus occidentalis Pontederia cordata and Sagittaria lancifolia The soil at the site consists of well-decom posed peat (Terra Ceia seri es, Euic, hyperthermic Typic Haplosaprists) ranging from 2 m thick in the eastern region to over 5 m thick in the marshs center. The peat is underlain by coarse sand on the east, and sandy clay in the center. Surface water in the marsh is characteri zed as soft and slightly acidic with pH of approximately 6.5. Nutrients are low, though not as low as th e Everglades, with median surface water TP concentration in the central region of the marsh of 50 g L-1. The hydrology of the marsh is highly constrained by a network of water control structures. Inflows are primarily from rainwater, a small stream on the western side (Pa dgett Branch), and three water control structures on the southern side. Outflows occur through the water control structure, S-96C, on the northern end of the marsh. The marsh surface elevation is approximately 7 m above mean sea level (MSL). Water level in the marsh typically varies from 7 m (dry) to 9 m above MSL. Soil Collection Samples were collected from BCMCA at st ations C1 and B4. These are the same locations used in studies described in Chapte rs 2 and 3. Samples were collected on March 16, 2004. A peat corer was used to retrieve 2 7/8-in ch ID soil cores (Clark 2002). This corer is a

PAGE 88

88 piston-type corer with a sharpened steel leading edge. Handles on each side allow the device to be rotated, which facilitates cutt ing through roots and peat fibers The corer was constructed of steel and employs a thin wall, semi-rigid, polybutyrate core liner. The liner containing the soil sample was taken out of the core barrel af ter the core was removed from the ground. The butyrate tubing was sufficiently thin that it was easily cut with a razor knife. Four cores were collected from each station; three for fracti onation and spectral propert ies of organic matter extracts, and one for lignin fractionation and 14C analysis. Total core length ranged from 105 cm to 150 cm. All cores consis ted of peat or muck and were stored at 4oC until analyzed. Organic Extract Fluorescence Soils were sequentially fractionated using 1M HCl extraction to remove inorganic C. This was followed by a 0.5 M NaOH extraction for 17 hours on a reciprocating shaker at room temperature. Twenty-five ml of extractant solution was added to approximately 0.5 g of ovendried soil, for a 1:50 extraction ratio. This solution was vacuum filtered through 0.45 m polyethersulfone (PES) filters, diluted 1000X, and pH adjusted to 7 using a 0.05 M NaHCO3 solution that had been pH adjusted to 6.7. This d ilution resulted in an optical density at 254 nm of approximately 0.15 (1-cm path length). This op tical density has been shown to reduce both inner filtering of fluoresced light, and to dimini sh absorption by the color of the extract (Ohno, 2002). A subsample of the NaOH extract was acidifie d to pH<2 to precipitate humic materials, leaving only the acid soluble fr action, or fulvic acid in solution. This was done by adding 5 drops of concentrated sulfuric acid to five ml of NaOH extract. The precipitated humic acid was separated from the supernatant by centrifuging. Fluorescence properties of both samples were determined. Fluorescence was determined w ith a Shimadzu model RF1501 (Shimadzu Corp., Tokyo, Japan) scanning fluorometer. Fluorometri c scans were conducted with an excitation

PAGE 89

89 wavelength of 254 nm and a fluorescence emissi on scan of 260 through 650 nm. The sample vessel consisted of a 1-cm quartz cuvette. Th e humification index (H IX) was calculated as 345 300 480 435 FI FI HIX (4-1) where FI = fluorescence inte nsity (Cox et al. 2000). Spectrophotometric Absorbance Spectrophotometric properti es were determined with a Shimadzu model UV160 spectrophotometer (Shimadzu Corp., Tokyo, Japan) on April 9, 2004. Absorbance was determined at 465 and 665 nm in a 1-cm path length, flow-through cell. Analyses were performed on the same 0.5M NaOH extract as us ed for the fluorometric analyses, though these samples were diluted 40X and adjusted to pH 7 with dilute HCl. This dilu tion resulted in average absorbance at 465 nm and 665 nm of 0.25 ( 0.18 SD) and 0.04 ( 0.03 SD), well within the measurement range of the instrument. The ratio of the absorbance at these two wavelengths is reported here; i.e. the E4:E6 rati o. This ratio has been shown to be related to average molecular weight of dissolved compounds (Chen et al 1977). 13C Nuclear Magnetic Resonance (13C-NMR) Samples were collected February 11, 2003 from station B4 only. Oven-dried (70oC) samples were analyzed at a spectrometer freq uency of 75.5 Mhz, contact time = 1 msec, pulse delay = 2 sec, acquisition time = 2 sec, and accumulation of 4000 scans. The spectrum was divided into four principal re gions: carboxyls (benzene, carboxy lic acid, amides and ethers), aromatics (phenols), O-alkyls (plant polymers such as lignin and cel lulose), and alkyls (aliphatic, alkanes, fatty acids, and waxes). The fractions correspond to the sp ectra peak shift regions of 46-

PAGE 90

90 110 ppm, 110-162 ppm, 59-92 ppm, and 0-46 ppm, respectively. The total area under each region was determined and its proportion in the whole sample was calculated by dividing by the total area of the spectra. It was assumed that th e total area under the curve represented the total carbon of the sample. This value was multiplied by the soil carbon content to determine the actual mass of each fraction in the original soil sample. 31P Nuclear Magnetic Resonance (31P-NMR) Samples were collected July 18, 2002 from station B4 only. Oven-dried (70oC) and ground samples were extracted by shaking the soil: so lution ratio of 5:100 with a solution containing 0.25 M NaOH and 0.05 M EDTA (ethylenediaminete traacetic acid) for 4 h at 20C (CadeMenun and Preston 1996). Each sample was ex tracted individually a nd centrifuged at 10,000 x g for 30 min. Equal volumes of the extracts were then frozen immediately at C, lyophilized, and ground to a fine powder. For solution 31P NMR spectroscopy, each freeze-dried extract (~100 mg) was re-dissolved in 0.1 mL of deuter ium oxide and 0.9 mL of a solution containing 1 M NaOH and 0.1 M EDTA, then transferred to a 5-mm NMR tube. The deuterium oxide provided an NMR signal lock and the NaOH raised th e pH to >13 to ensure consistent chemical shifts and optimum spectral resolution. Inclus ion of EDTA in the NMR tube reduces line broadening by chelating free Fe in solution (Turner 2004). Solution 31P NMR spectra were obtained using a Bruker Avance DRX 500 MH z spectrometer operating at 202.456 MHz for 31P and 500.134 MHz for 1H. Samples were analyzed using a 6 s pulse (45), a delay time of 1.0 s, and an acquisition time of 0.8 s. The delay time used here ensured sufficient spin-lattice relaxation between scans for P nuclei (Cad e-Menun et al. 2002). Between 48,000 and 69,000 scans were acquired depending on the P concentra tion of the lyophilized extract, and broadband proton decoupling was used for all samples. Chemi cal shifts of signals were determined in parts per million (ppm) relative to an external standard of 85% H3PO4. Signals were assigned to

PAGE 91

91 individual P compounds or functi onal groups based on literature re ports (Turner et al. 2003) and signal areas calculated by integration. Spectra were plotted with a line broadening of 8 Hz, although additional spectra were plotted with a line broadening of 1 Hz to examine signals in the phosphate monoester region. Isotopic Analysis of Plant Fibers Samples for lignin fractionation were collect ed on March 16, 2004 from stations C1 and B4 in BCMCA. A single core from each station wa s split length-wise and one half of each 10-cm interval was used for fiber analysis. Analysis was performed using an Ankom Fiber Analyzer, model 200/220 (Ankom Inc., Fair port NY) following the method of Van Soest (1970). Samples were sequentially extracted with a neutral detergent, an acidic detergent, and concentrated sulfuric acid. The mass of each fiber fracti on was determined gravimetrically after each extraction step. This technique wa s originally developed for crop residues, and was modified to characterize plant-derived soil ca rbohydrates. Initial at tempts to fractionate soil carbohydrates indicated that much fine or ganic matter leaked from the 30M glass fiber bags used in the analyses. Since the procedure is based on weight loss, an alternate method was developed. The modified procedure followed these steps: 1. Place each 10-cm bulk-soil interval into a No. 35, 500-micron brass sieve. 2. Immerse the sieve in 0.5% Liqui nox solution to disperse fine so il organic matter from fibric material. 3. Gently agitate it for approximately 5 minutes. 4. Wash fine particulate material from fibric material remaining in sieve with stream of deionized water until it is clean. 5. Air-dry the remaining fiber. 6. Grind fiber through 20-mesh screen in a Wiley mill. 7. Fractionate using Ankom methodology.

PAGE 92

92 Therefore, the results reported here represent the properties of the fi bric material in the sample, not the bulk soil. A subsample of the peat fiber was analyzed for 13C and 15N on a Thermo Finnigan DELTA Plus isotope ratio mass sp ectrometer at the University of Florida Soil and Water Science Department. Purified N2 and CO2 were used as internal standards, with a helium carrier gas. Results are expre ssed in delta notati on, relative to the 13C content of PeeDee Belemnite, and 15N content of atmospheric N2. Soil Age Soil dating was determined through 14C analysis of plant macr ofossils. Ten centimeter sections of peat to a depth of approximately 1.5 m were carefully examined for plant fragments such as stems, seeds, and other woody fragment s. Care was taken to avoid selection of root fragments. These plant macrof ossils were sequentially extr acted with 0.5M HCl (0.5 hours), 0.1M NaOH (2 hours), and again with 0.5M HC l (0.5 hours). This removed carbonates and any dissolved organic materials such as humic or fulv ic acids that may have percolated down through the soil profile. The samples were analyzed at University of Califor nia, Irvine Keck-CCAMS laboratory by accelerator mass spectrometer (N EC 0.5MV 1.5SDH-2, National Electrostatics Corporation). Carbon dates were corrected to calendar dates using the computer program CALIB, version 5.0.1(Stuiver and Reimer 1993). Results and Discussion Fluorescence and Spectrophotometric Properties The HIX of the peat extracts showed a st eady downcore increas e throughout the soil profile (Fig. 4-1). Extracted, liv e plant tissue had a HIX of appr oximately 2, detritus 4, whereas peat samples at a depth of 140 cm were about 10. Both stations had very similar increases with respect to depth, though station C1 had a reversal (declining HI X) at 80 cm. This reversal

PAGE 93

93 coincided with a change in physical properties of the peat. At a depth of approximately 80 cm, the texture of the peat at C1 changed from a very decomposed, black muck, to a reddish-brown fibric peat immediately underneat h. This more fibric, less decom posed peat likely caused the decline in HIX at station C1. There was no significant difference in average HIX between the two stations for the humic + fulvic samples. (One-way ANOVA; F = 1.45; p = 0.232; MiniTab). This was expected, since both stations have been exposed to the same hydrol ogic regime, and must have accreted peat at a similar rate. Fulvic acid extracted from the soil samples showed a very similar trend of increasing humification with respect to greater de pth. In general, the whole NaOH extract (humic + fulvic) had an average HIX value approximate ly one unit higher than the fulvic sample. The sharp increase in HIX at st ation C1 (70-90 cm) coincided with a dramatic increase in OM age. Station B4 also showed local maxima at approximately this same depth. One possible reason for this increase is accelerated decompos ition during a long dry period in the marsh. Both stations showed increase in HIX at this dept h, followed by a slight decl ine (less decomposed), then a gradual increase to the base of the core. The E4:E6 ratios observed in this experiment indicate that the soil extract was richer in fulvic than humic acids at both stations for all de pths (Fig. 4-2) The ratio E4:E6 ratio of fulvic acid in most soils was typically in the range of 6.0 to 8.5, wher eas humic acids tend to be less than 5 (Stevenson 1994). Chen at al. (1977) found th at this range corresp onded to fulvic acids with an average molecular weight of approxima tely 2000 atomic mass units (amu; or daltons). For this study, an average weight of 2000 am u corresponds to the complete soil extract containing both humic and fulvic materials.

PAGE 94

94 The interior marsh station (B4) showed a slight decline in E4:E6 ratio with increasing depth, beginning at approximately 12 in the 0-10 cm soil depth, declining to 7 at a depth of 80 cm, followed by very little change to the base of the core. This s uggests a relatively constant rate of diagenesis for the surficial 70 cm of peat, with an increase in molecular weight of the extracted organic matter to 70 cm. The E4:E6 rati o at the nutrient impacted station showed a slight decline to 50 cm, increased to a depth of 100 cm, then declined to the base of the core. The E4:E6 profile at station B4 is ch aracteristic of the diagenesis of a relatively constant source, or quality, of organic matter. The ra tio at the soil surface is high, re flecting lower molecular weight, poorly humified organic matter, as would be expe cted from recently deposited plant materials. The ratio suggests increasing molecular weight with respect to soil depth to the base of the core, as would be expected from the progressive humif ication of a constant source of organic matter input. The gradual decline therefore implies that this region has experienced a similar history of vegetation over the time period that this soil profile represents. Humi fication index showed increasing humic character with depth, however results from the sequential fractionation (Chapter 2) did not indicate incr easing humic-P with depth. This may mean that in wetland soils long-term sequestering of soil P is not occurring in humic-type materials, but in some other compounds. Stewart and Tiessen (1987) point out that highly charged phosphate groups of soil Po may prevent it from interacting with humic materials. The E4:E6 profile at station C1 does not show the same constant decline as station B4. The E4:E6 ratios at station C1 are lower than stati on B4 for soil depths above 70 cm, and greater than thos of B4 for samples below 70 cm. Lower ratios in surface depths (than B4) may be a consequence of two factors; vege tation history and root biomass. Soil organic matter with greater molecular weight at station B4 for soil depths < 70 cm may reflect changes in the vegetation

PAGE 95

95 community in this area of the marsh. The C1 re gion of the marsh is known to have experienced nutrient impacts that have altered the vegetati on in the northeastern corner of BCMCA. A result of this impact is the conversion of slough to cattail ( Typha latifolia ) and willow ( Salix caroliniana ) marsh. Aquatic plants that characterize slough vegetation include such plants as Nymphaea odorata, Utricularia spp. phytoplankton, etc. These plants contain less structural tissue, such as lignin, than the emergent vegetati on and shrubs mentioned above. Lignin is one of the principal refractory biogenic compounds and is thought to be a precursor of humic materials (Stevenson 1994), and humic content is a major de terminant of the E4:E6 ratio. The nutrient impact may have, by altering the local plant community, led to a lterations in the biochemical characteristics of the recent peat in this area of the marsh. A second potential cause of greater E4:E6 ratios at station C1 may be greater contri bution of live root biomass to the carbon pool of the soil extracts. Panicum hemitomon is the dominant plant at station B4. Many small roots were noted in the samples during core sectioning, ma ny more so than at station C1, which was dominated by Typha latifolia and Salix caroliniana This new carbon would have consisted of lignin, cellulose, and other polysaccharides of much lower molecular weight than humic materials and may have biased the E4:E6 ratio upwa rds in the surface soil horizons of B4 relative to station C1. At neither station was a dramatic increase in molecular weight with increasing depth observed. In fact, at station C1, ve ry little overall change with resp ect to peat age was seen. This suggests that peat soil diagenesis does not involve dramatic ch anges in molecular weight of extractable soil components. 13C Nuclear Magnetic Resonance (13C-NMR) Total O-alkyls are associated with plant polym ers, such as lignin and cellulose, and are considered relatively labile. This group repr esented the major fraction of soil organic carbon

PAGE 96

96 (SOC), comprising 30-40% of th e total carbon pool, although th ey declined from 225 mg kg-1 at the surface, to a minimum of 133 mg kg-1 at a depth of 90 cm (Fig. 4-3). As a percentage of total C, these value are slightly higher than reported by Sjgersten et al. (2003) for forest and tundra soils. Concentrations declined wi th greater depth in soil profile, with the exception of a slight increase for the bottom two depth intervals. An inflection at 90 cm coin cided with a textural change to peat characterized by much much fine r organic matter. The most likely explanation for this change is a period of greater water depths, favoring soft-tissued plants such as phytoplankton or floating macrophytes. These pl ants would have contained le ss refractory carbonaceous tissue (such as lignin), and thus would have been more rapidly mineralized. The more fibric materials underlying this strata visually appeared less decom posed, and this was reflected in an increase in O-alkyl concentration below 90 cm. Alkyl-C compounds consist of fatty acids, wa xes, and alkanes and include compounds such as acetate, butyrate, and other fermen tation end-products, as well as methane. Alkyls constituted the second greatest fraction, comprisi ng 25-30% of the soil organic carbon pool. This fraction increased from 91 mg kg-1 at the soil surface, to a maximum of 191 mg kg-1 at a depth of 90 cm, or an increase from 20% to 38% of SO C. This is slightly greater than observed by Sjgersten et al. (2003), probabl y due to increased production and storage of fermentation products in the flooded organic soils of this study, as compared to their tundra soils. Total aromatic carbon increased linearly with respect to depth, showing a very similar concentration profile as alkyl-C suggesting their accumulation at similar rates. Total aromatics accounted for 20-30% of SOC. Baldock et al. (1 997) reviewed the C-MNR results from organic and mineral soils, forest litter, and composts and found that the proporti on of aromatic carbon compounds was not a particularly good surrogate for extent of decomposition. However, their

PAGE 97

97 review did not include flooded soils. Organic we tland soils lack the fungi responsible for production of the extracellular enzymes that ar e capable of hydrolyz ing aromatic compounds. This is likely the reason that the soils from BCMCA increase in aromaticity with depth, in contrast to results from Baldock et al. (1997). The rate of increase in alkyl-C and total arom atic compounds with respect to depth in the soil profile was inversely related to the Oalkyl content, suggesting the biogeochemical transformation of plant polymers to microbial metabolites and resistan t organic compounds. The ratio of alkyl to O-alkyl functiona l groups (A:O-A) has been shown to be a better indicator of the extent of decomposition, especially for organic soils (Baldock et al. 1997). The ratio increases steadily through the soil profile, with a dramatic increase in the region of silty organic matter. The ratio begins at 0.4 at the soil surface, with a maximum value of 1.44 at 80-90 cm. Baldock et al. (1997) reviewed C-NMR data from numerous studies on mi neralization of forest, wetland, and agricultural soils and f ound a range of A:O-A values, from 0.1 to 1.5, with the ratio depending mostly on degree of decomposition and secondarily on plant community and soil mineral content. The ratios found in this st udy are on the high end of this range, indicating advanced decomposition, especia lly at the lowest depths. The steady increase in phenolic and aromatic compounds was very similar to the increase seen in the humification index (HIX) reported earlier. Both te chniques indicate a progressive transformation of plant components to more hum ified, aromatic, and presumably stable soil organic matter. This is a clear indication that pl ant litter that constitutes the bulk of peat soil undergoes radical biochemical changes after senes cence. Thus, inferences concerning ecosystem history that are based on stra tigraphic changes in peat chemis try need to consider that a significant proportion of observed ch ange may not be due to change s in ecosystem history (i.e.,

PAGE 98

98 plant community, nutrient loading, etc.), but are due to biogeochemical transformations in soil organic matter. 31P Nuclear Magnetic Resonance (31P-NMR) Organic phosphorus in soil extracts, as a pr oportion of total P, increased throughout the soil profile, representing approximately 20% of total P at the soil surface, and increasing to 40% at a depth of 120 cm (Fig. 4-4). Within the organic P pool there was no significant difference between phosphate monoesters and diesters (Student s paired t-test; p = 0. 321; df = 7). However, it should be noted that the EDTA-NaOH extracti on is known to degrade phosphate diesters such as phospholipids and RNA to their phosphate monoest er constituents, so th e actual proportion of phosphate diesters is likely higher and thus proba bly constitutes the dominant form of organic P in these soils. Notably absent was myoinositol hexaphosphate (phytic acid), the dominant phosphate monoester in most upland soils (Turner et al. 2002) (Fig. 4-5). This form of P is of plant origin, with major concentration in seeds. Its high charge density causes it to adhere to soils, making it less available to microbial attack than other organic phos phates, and therefore more stable. Few studies have documented the longterm fate of these tw o functional classes of organic P in soils, particularly in wetlands. It appears from this study that both phosphate monoesters and diesters increase d slightly with soil depth, and maintained approximately equal proportions. This represents an unusually high pr oportion of phosphate diesters. In a study by Turner et al. (2003) the ratio of monoto dies ter P ranged from approximately 5 to 26 for 29 temperate pastures, compared to 1:1 in this study. One of the principal sources of phosphodiester is microbial tissue. The soils in this study appear to be much higher in microbial biomass, or alternatively diester P is much be tter preserved under the anaerobic conditions of these flooded soils. Turner and Newman (2005) also found elevated levels of diester P in the Florida Everglades, which they attributed to sl owed hydrolysis of plant and microbial inputs of

PAGE 99

99 phosphate diesters. The slight increase in organic forms of P appeared to be at the expense of orthophosphate, which declined throughout the pr ofile. A strong signal from DNA is apparent down to a depth of 80-cm. Since DNA is not we ll preserved in upland soils, this suggests a prominent microbial community at these depths However, preservati on of DNA under flooded conditions in a pond near Titusville, FL wa s quite good (Doran 2002). The Windover site is approximately 90-km northeast of BCMCA and ha s gained international notoriety as having some of the best-preserved human remains and cultural artifacts (particularly woven fiber) ever recovered. In that study, 8000-year old human br ain tissue was recovered from remains found in a shallow pond. The DNA from this tissue was in tact enough to partially sequence, though the overall DNA yield was low; approx imately 1% of that from fre sh tissue. Thus, phosphodiesters such as DNA may be better preserved than prev iously thought and thus may play a more prominent role in long-term phosphorus stabilization. Isotopic Analysis of Plant Fiber The amount of plant fiber recovered using the wet-sieving technique varied from 5 to 15 grams dry weight, per 10-cm interval. Note that each core was split length-wise, so the actual recoverable fibric material is twice this value. There was no apparent trend down core, though plant fiber declined at depths where fine orga nic matter increased. The relative abundance of each plant polymer was lignin > cellulose > hemicellulose. Lignin content increased gradually downcore at both stations, from approximately 40% of the material dry weight in the surface soil, to 80 % at the base of the core (Fig. 4-6). Lignin was the most abundant plant polymer, from three to five times as abundant as cellulose and hemicellulose. Station C1 showed a pronounced decline in lignin for the 50 to 60 cm depth, coinciding with an increase in fine organic matt er, and a sudden increase in organic matter age. This may reflect either a deepening of th e marsh, favoring soft-tissued macrophytes and

PAGE 100

100 phytoplankton that possess less lignin, or alternativ ely that a long-term dr ought reduced much of the lignin to smaller plant polymers. Evidence fo r a circum-Caribbean dry period that occurred approximately 1000 years before present favors the second hypothesis (B uck et al. 2005, Hodell et al. 2005). Cellulose showed a gradua l decline throughout the soil profile, declining from approximately 20% at the surface, to 5-10% at 150 cm. Downcore lignin and cellulose content was very similar to that found by DeBusk a nd Reddy (1998) for soil samples taken from WCA2A in the northern Everglades. They found that the lignin cellulo se index (LCI; lignin/(lignin + cellulose)) varied from 0.26 in standing dead plant litter, to 0.81 in peat taken from a depth of 30-cm. The LCI of the samples in this study varied from 0.73 at th e soil surface, to greater than 0.90 at the base of the core, indicating a greate r degree of decomposition at the lower soil depth. The greater lignin content, as well as greater LC I in this study reflects the greater soil depths investigated in BCMCA. DeBusk and Reddy (1998) examined peat samples from a maximum of 30 cm, whereas in this study peat from a depth of 150 cm was analyzed. Hemicellulose declined somewhat erratically with respect to depth, from approximately 10% at the surface, to 3-5% at the base of the core. Soil lignin content increased with greater de pth at both locations, while cellulose and hemicellulose declined. Enzymes responsible fo r lignin decomposition, such as phenol oxidase and lignin peroxidase, are secreted by white rot fungi such as Phanerochaete chrysosporium These fungi are not active in fl ooded soil, hence the buildup of li gnin. Some of the increase in lignin content with depth is undoubtedly due to the compaction and concentration of lignin as more peat is accreted. However, the predominan t mechanism is likely th e loss of more labile plant polymers from the separated soil fibric mate rial. Overall, it is clear that as this fibric

PAGE 101

101 material ages, it takes on a more purely lignin ch aracter. This data seems to conflict with the 13CNMR results described earlier. The NMR data showed declining lignin throughout the soil profile. However, those results were from a bulk so il sample (not sieved). It is likely that even though lignin degrades slowly under flooded c onditions, it was being converted to humic materials and thus it represented a declining fraction of th e bulk soil. In fact, this is one of the principal theories of humic matter formation, that lignin is a precu rsor to humic matter (Stevenson 1994). The sieved material contained only the fibric material of macrophytes. That the lignin content of this material increased with depth reflects the loss of other plant polysaccharides from the fiber, such as cellulo se, hemicellulose, and sugars, thus leaving only lignin. In summary, even though the lignin content of the soil declin ed with respect to depth, the plant fiber that remained became enriched in lignin. Live plant biomass at both stations was enriched in both 13C and 15N, relative to the surficial detrital plan t litter (Fig. 4-7). The method of analys is used in this study (i.e. analysis of intact plant fiber) has the advantage of reveali ng the isotopic signature of only the principal vascular plants that colonized the site at th at depth interval. Thus, extraneous carbon from carbonates and percolating organic acids from othe r depths do not contribute to the signature. Biochemical transformations of carbon and nitrog en tend to discriminate against the heavier 13C isotope, or fractionate carbon is otopes (Peterson and Fry 1987). T hus, plant biomass is depleted in 13C relative to atmospheric carbon dioxide, which has a 13C of approximately -8. The isotopic signatures of aboveground plant biomass co llected in this study ar e typical of results observed in other studies of wetland emergent plants (Meyers 1994, Keough et al. 1996). Aboveground samples of Typha collected from the nutrient imp acted station averaged -29.1 (.2 SD), whereas Cladium samples collected from the unimpacted station averaged -26.7 (.3

PAGE 102

102 SD). The 13C content of soil at impacted station C1 increased from -29 in surface soil to 26 at a depth of 50 cm, with little additional change to the base of the core. Increases at the unimpacted station were more modest, with 13C increasing only 1 in the upper 50 cm of soil. The two stations showed distinct ly different isotopic signatures to a depth of approximately 50 cm, for 13C and 80 cm for 15N. This probably reflects a l ong-term difference in plant communities at the two stations. Currently, the comm unity at station B4 is mostly an emergent marsh Panicum hemitomon/Cladium jamaicense community whereas the C1 station is a deeper slough community dominated by Utricularia spp. Typha latifolia and Nymphaea odorata Enrichment in the heavier carbon is otope with increasing depth may be due to several factors. First, methanogenesis produces methane that tends to be depleted in 13C, or light. Gu and Schelske (2004) attributed extreme enrichment in 13C in lake phytoplankton to conversion of methane to C02 in the water column of a hypereutophic lake. Over time, this would tend to enrich the remaining organic matter in the h eavier isotope. Both chemical and biological reactions fractionate isotopes such that the substrate becomes progressively enriched in the heavy isotope, while products are enriched in the light isotope. Another poten tial cause of increasing downcore 13C is a gradual shallowing of the marsh that promoted a shift from submerged macrophytes, to the emergent community prevalent today. However, other results from this study favor the first hypothesis, that of gradual proces sing and loss of the light isotope. For example, lignin was observed to increase with increasing de pth and this is not cons istent with a greater presence of submerged vegetation. Both C:P and C: N also increased downcore and this is also inconsistent with a historical community that had more subm erged vegetation. Lastly, physical examination of fresh soils revealed the presence of fi bric peat to the base of the core and this is further evidence of a relatively stable commun ity (Appendix A). The difference in composition

PAGE 103

103 between plant and litter was most dramatic for 15N, with plant litter c ontaining essentially an atmospheric 15N content ( 15N = 0). This may reflect extensiv e colonization of the detrital material by N2-fixing cyanobacteria. Plant litter was submerged at the time of sampling and therefore had probably become colonized with epiphytic algae and enriched in microbial biomass. There was a slight enrichment in 13C at the Typha -dominated station and a decline in 15N approximately mid-core, compared to the Cladium -dominated station, perh aps reflecting a shift in plant communities at respective stations. Carbon -14 dates at both stations indicate that this change occurred approximately 1000AD (Fig. 4-8, Table 4-2). As mentioned previously, this period was characterized by a longterm tropical drought that may ha ve led to changes in marsh plant communities. An extended dry period, perhaps several hundred years, would have led to a dramatic loss of soil organic matter and a lowered marsh surf ace elevation. Upon refl ooding, the shallow water adapted community would have been replaced by deeper water adapted plants, such as Nymphaea and others. An abrupt increase in age of pl ant macrofossils is apparent at a depth of between 70 and 90cm. This sudden increase in ca rbon age has been observed at other locations in BCMCA (M. Fisher, unpublished data). There ar e two plausible explanations for this sudden increase in the age of BCMCA peat: little or no accretion for almost a millennium, or (more likely) loss of carbon due to an extended dr y period. This marsh dryout would have been accompanied by greatly accelerated aerobic peat decomposition, a peat fire, or both. In any case, once normal rainfall patterns returned, a much deeper marsh community would have developed and this change seems to be s uggested by the isotopic evidence.

PAGE 104

104 Conclusions Results from this study provide informati on on the biogeochemical transformations that organic matter undergoes after it is deposited on the soil surface. Peat core s taken to a depth of approximately 1.5-m covered a soil age range th at dated to approximately 3000 years before present. There is ample evidence that substantial changes occur bot h in organic matter composition, and organic P. The fluorescence and spectrophotometric procedur es indicated a progression to more humified material, but not necessarily a concomitant change in molecular weight of extractable organic matter. Increasing humic ch aracter with increasing depth observed in the fluorescence experiments and the lack of increasi ng humic-extractable P (C hapter 2) would seem to indicate that P is not cl osely associated with forma tion of humic compounds. Nuclear magnetic resonance analysis of carbon shows a decline in plant carbohydrates with respect to depth and an increase in aromatic compounds Nuclear magnetic resonance of phosphorus showed stability of diester-P compounds, a nd an absence of inos itol hexaphosphate. This contrasts sharply with upland soil 32P-NMR work, where diester-P is has been shown to quickly degrade, and inositol phosphates are the dom inant P-storage compound. Organic forms of phosphorus increased with respect to depth. Lignin fractionation showed that the lignin content of soil fibr ic material increases with age, while cellulose and hemicellulose decline. Stable isotopic composition seems to indicate a relatively stable plant community with a community shift appr oximately 1000 years ago, perhaps due to an extended drought. These events would have had important consequences for the flora and fauna of the marsh, the organic remains of thes e organisms, and thus the peat deposit itself. Results from this study suggest that P does not behave as a conservative compound. It is subject to the same biogeochemical processes that dramatically a lter soil organic matter. These

PAGE 105

105 processes act to transform the original floral and faunal material into inorganic and organic compounds that are progressively le ss and less similar to the orig inal material. These processes lead to loss of labile forms of P, and an in creasing proportion of st able organic phosphorus compounds. Comparative studies that contrast surficial peat soils to their deeper counterparts need to consider the effects that soil forming pr ocesses have had on soil properties, particularly in the case of organic soils.

PAGE 106

106 Table 4-1. Results of 14C AMS dating. Values in parenthese s represent standard error of analytical dating procedure. YBP = y ears before present. Radiocarbon ages have been adjusted for varying atmospheric 14C concentration using calibration program CALIB 5.01. Station Soil Depth Sample Description Age, YBP 13C B4 -35 Seed >1800 AD -21.2 B4 -65 Woody husk 1550 AD ( 100) -24.2 B4 -85 Panicum root 1440 AD ( 50) -22.9 B4 -115 Stem fragment 95 AD ( 20) -26.4 B4 -125 Pine needle 1150 BC ( 125) -24.1 B4 -145 Seed husk 950 BC ( 60) -22.6 C1 -35 Grass blade >1800 AD -21.7 C1 -55 Seed pod case 1415 AD ( 10) -24.7 C1 -75 Grass blade 985 AD ( 60) -25.5 C1 -85 Twig 780 BC ( 10) -26.4 C1 -95 Nuphar rhizome 825 BC ( 25) -22.0 C1 -125 Nuphar rhizome 1290 BC ( 30) -24.5 C1 -135 Wood fragment 1150 BC ( 100) -19.4 C1 -145 Wood fragment 1650 BC ( 30) -22.9

PAGE 107

107 Figure 4-1. The humification index (HIX) of NaOH-extracted peat samples. Error bars represent the standard deviation of thr ee soil extracts of the same sample. Panel A. reflects the HIX of the whole extract, whereas pane l B. has had humic materials removed by acidic precipitation, leaving only fulv ic compounds. Station B4 = nutrient unimpacted, Station C1 = nutrient impacted. HIX 024681012 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 20 Impacted Unimpacted Plant tissue DetritusHumic+Fulvic FulvicAcid 024681012 HIX 024681012 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 20 Impacted Unimpacted Plant tissue DetritusHumic+Fulvic FulvicAcid 024681012

PAGE 108

108 Figure 4-2. Ratio of absorbance at 465 and 665 nm (E 4:E6) of NaOH peat extracts at stations C1 and B4 in BCMCA. Station B4 = nutri ent unimpacted, Station C1 = nutrient impacted. E4:E6 02468101214 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 B4 C1 +humic +MW +fulvic E4:E6 02468101214 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 B4 C1 +humic +MW +fulvic

PAGE 109

109 Figure 4-3. Results of solid-state 13C-NMR spectroscopy of dried, ground peat samples collected from station B4 on February 11, 2003. Depi cted is the soil carbon concentration associated with each carbon group, not so il concentration of respective compounds. Carbon, g/kg 050100150200250 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 Total O-alkyl (lignin, cellulose) Alkyl-C (fatty acids, waxes, alkanes) Total Aromatic

PAGE 110

110 Figure 4-4. Results of solution 31P-NMR spect roscopy of NaOH-extracted samples collected from station B4 (unimpacted) on Febr uary 11, 2003. Note approximately equal concentration of mono and diester P. Phosphorus fraction, % of TP 01020304050 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 Phosphate Monoester-P Diester-P Phosphorus fraction, % of TP 01020304050 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 Phosphate Monoester-P Diester-P

PAGE 111

111 Figure 4-5. The 31P-NMR spectra of 0.5M NaOH extracts of soil samples from Blue Cypress Marsh Conservation Area, station B4 (unimpacted). Chemical shift (ppm) -8 -6 -4 -2 0 2 4 6 8 10 0-10 10-20 20-30 30-40 40-50 50-60 70-80 80-90 90-100 100-110 Soil Depth, cm Chemical shift (ppm) -8 -6 -4 -2 0 2 4 6 8 10 0-10 10-20 20-30 30-40 40-50 50-60 70-80 80-90 90-100 100-110 Soil Depth, cm PO4-P Diester-P Monoester-P Polyphosphate

PAGE 112

112 Figure 4-6. Results of lignin fract ionation of plant fiber. Fiber was separated from bulk peat samples that were collected from BCMCA stations B4 and C1. Note differing horizontal scales. Percentages reflect proportion of total orga nic matter present in peat fibric materials. Cellulose, % 0510152025 Lignin, % 020406080100 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 B4 (unimpact.) C1 (impact.) Hemicellulose, % 01020304050 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 Soluble carbon, % 020406080100

PAGE 113

113 Figure 4-7. Stable isotopic compos ition of plant fiber that was se parated from bulk peat samples collected from BCMCA stations B4 (unimpacted) and C1 (impacted). 13C -30-28-26-24-22 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 20 Unimpacted Impacted -2024 Typha Cladium Detritus15N 13C -30-28-26-24-22 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 20 Unimpacted Impacted -2024 Typha Cladium Detritus15N

PAGE 114

114 Figure 4-8. Uncalibrated 14C age of plant macrofossils from selected depths at stations B4 and C1 in BCMCA (YBP = years before present). 14 C Age, YBP 01000200030004000 Depth, cm -140 -120 -100 -80 -60 -40 -20 0 Unimpacted Impacted

PAGE 115

115 CHAPTER 5 CONCEPTUAL MODEL OF PHOSPHORUS BURIAL IN WETLANDS, CONCLUSIONS, AND FUTURE RESEARCH NEEDS Conceptual Model of Soil Phosphorus Transformations. Conceptual models are useful in summar izing, communicating, and gaining consensus on significant inputs that are thought to influence a given process. For wetlands, processes leading to long-term storage of phosphorus in wetland soils can be depicted as a system of stocks and flows, or storages and transfers (Fig. 5-1). In this model, most of the flows among storages are hypothesized to be bidirectional, reflecti ng a dynamic equilibrium among fractions. The exception to this is the proportion of C in the resi dual organic fraction (recalcitrant P), which is shown here to only increase with time. This was the only fraction in the sequential chemical fractionation that was shown to consistently in crease with time. Principal above and below ground storages and transfers are given in Table 5-1. Primary produc tion and microbial P represent the only living bioma ss pools and correspond to living plant tissue (roots and shoots) and microbial biomass P. Relatively labile Po consists of sugar phosphates, nucleic acids (DNA, RNA), nucleotides (ATP), sugar phosphates, polyphosphates, and othe r immediate by-products of cellular decomposition. These compounds are rapidly available to the microbial community after exocellular enzymatic hydrolysis, and also through direct incorporation back into cellular biomass. Through reactions that are mostly unknown, but thought to be abiotic, some proportion of this pool is transferred into humicand fu lvic-like compounds and are illustrated as unknown reactions in the conceptual m odel. These pools are hypothesized to be in a state of dynamic equilibrium, therefore exchanges are first-order processes. Ulti mately, a proportion of the humic pool takes on increasing phenolic ch aracteristics and a molecular geom etry such that it is largely inaccessible to enzymatic attack, especially unde r flooded conditions. Inorganic pools consist of the ortho-P monomer, amorphous P compounds su ch as P associated with iron (Fe) and

PAGE 116

116 aluminum (Al) hydroxides, and mineral P forms. These pools are regulated by the reaction kinetics of sorption and crysta llization. Soluble inorganic P is available for plant uptake, incorporation into microbial biomass, diffusion a nd loss from the system, or sorption to Fe and Al hydrolxides. This conceptual model proposed is a simplified depiction of principal pools and transfers of P in wetland soils Exchanges among pools are hypothe sized only, since there was no attempt made in this dissertation to measure reaction kinetics or to document transfers. Conclusions Wetland soils are typically long-term sinks for carbon and nutrients. Results from this study provide information on the biogeochemical transformations that organic matter undergoes after it is deposited on the soil surface. Peat cores taken to a depth of approximately 1.5 m spanned a soil age that dated to approximately 5000 years before pres ent. There is ample evidence that substantial changes occur both in organic matter composition, and organic P. The fluorescence and spectrophotometri c experiments indicated a pr ogression to more humified material, but not necessarily a concomitant change in molecular weight of extractable organic matter. Nuclear magnetic resonance analysis of carbon shows a decline in plant carbohydrates with respect to depth and an increase in ar omatic compounds. Nuclear magnetic resonance of phosphorus showed stability of diester-P compo unds, and an absence of inositol hexaphosphate. This contrasts sharply with upland soil 31P-NMR work, where diester-P has been shown to quickly degrade, and inositol phosphates are the dominant P-storag e compound. The proportion of organic to inorganic P increase d with respect to depth at two stations, remained the same at one, and declined at the last station. Lignin fractionation showed that the lignin content of soil fibr ic material increases with age, while cellulose and hemicellulose decline. Stable isotopic composition seems to indicate a

PAGE 117

117 relatively stable plant commun ity with a community change approximately 1000 years ago, perhaps due to an extended drought. Results from this study suggest that P does not behave as a conservative compound. It is subject to the same biogeochemical processes that dramatically a lter soil organic matter. These processes act to transform the original floral and faunal material into inorganic and organic compounds that are progressively less similar to the original materi al and lead to loss of labile forms of P, with an increasing proportion of stable organic phosphorus compounds. This progression is consistent with the pedogenic mo del proposed by Walker and Syers (1976). Their model, based on upland soil chronos equences, showed long-term depl etion of P, asymptotically approaching minimum values for respective P fr actions. They also found that with weathering, Po became an increasingly larger proportion of soil total P. Their findings were supported for two of the wetland sites in th is study, though the ratio of Pi to Po increased at one site and remained essentially unchanged at another. Th e most likely reason for this minor discrepancy between the two studies reflects the principal so urce of P in uplands, verses wetlands. Natural (non-agricultural) uplands begin the pedogenic process with the pr incipal pool of P in primary soil minerals. Weathering reduces this pool, while increasing storage of Po. Wetlands are situated at bottom of the landscape, and thus receiv e nutrient subsidies from higher areas in the topographic gradient. Organic matter produced mostly in situ accumulates due to slowed decomposition and appears to be the principal form of stored P at all stages of the pedogenic process in peat dominated wetlands. It is clear from the mathematical model results presented in this study that the two principa l long-term storage pools in peat -dominated wetland soils are the P associated with the unextractabl e fraction and humic P (Fig. 5-2).

PAGE 118

118 There are three main mechanisms that could account for the soil phosphorus profiles observed in this study. Declining soil P with dept h could be a result of ecosystem succession, to plants that were capable of retaining P in more resistant compounds. Second, ecosystem loading rates may have gradually increased over the mille nia as a result of cultural eutrophication. Last, weathering of organic P and or ganic P mineralization, with en richment in refractory P (and refractory carbon), accompanied by subsequent di ffusion to the soil surface and advection via surface transport from the wetland may have led to the observed declines. Results from this dissertation suggest that this last mechanism is the principal reason for th e observed declines in soil P with respect to increasing depth. Future Research This research yielded little insight into pot ential transfers among P pools. For example, P (and carbon) diagenesis is thought to involve progressi ve soil enrichment in higher molecular weight, more humified substances. However, whet her this enrichment involves preservation of a constant initial mass of humic-like materials, or transformation of biological molecules into humic substances is not known. This same questi on holds for other fractions as well. It is possible that there exists a dynamic equilibrium among P fractions, and that P pools constantly adjust to maintain this shifting equilibrium. Little is known about the chemical characteri stics of the residual fraction remaining behind after sequential fractiona tion. The phosphorus and carbon associated with this material is thought to be highly refractory, though this is specula tion based on the premise that resistance to chemical extraction is analogous to environmental stabilit y. Advances in solid-state 31P-NMR may be helpful in characterizing this material. Fu rther investigation is needed to determine why the thermal techniques investigated in this dissert ation were able to distinguish between labile and recalcitrant soil organic P. The hypothesis that thermal stabil ity was related to environmental

PAGE 119

119 recalcitrance seemed to hold, yet the nature of the compounds that conf er this stability is unknown. Lastly, ecological studies that contrast surficial peat soil s to their deeper counterparts need to consider the effects that soil forming pr ocesses have had on soil properties, particularly in the case of organic soils.

PAGE 120

120 Table 5-1. Components of conceptual model depicting phosphorus pools and transfers in peat soils. Model Element Description Pool Primary production Plant biomass P Microbial P Cell walls, nucleic acids, ATP/ADP Soluble Po Labile organic P Fulvic P assoc. with Fulvic acid Humic P assoc. with Humic acid Recalcitrant P in long-term storage compounds. Stable. PO4 Inorganic P in solution Amorphous P Non crystalline inorgani c, iron, calcium, and aluminum P Mineral Crystalline P Transfer Uptake Plant uptake External transfer Transfer of P into/out of system Sorption Sorption of P to inorganic compounds Crystal Crystallization PiMin Mineralization of organic ma tter by microbial community to PO4 PoMin Mineralization of labile orga nic P organic matter by microbial community to labile organic compounds, by microbial community (or uptake of Po by them) UnRx1 Unknown biotic or abiotic reaction leading to exchanges between labile organic and fulvic compounds UnRx2 Unknown biotic or abiotic reaction leading to exchanges between of humic and fulvic compounds UnRx3 Unknown biotic or abiotic reaction leading to exchanges between humic and labile organic compounds Burial Long-term transfer to very recalcitrant P

PAGE 121

121 Figure 5-1. Conceptual model of dynamic ex changes among soil phosphorus pools in wetland soils. UnRx3 Recalcitrant P Fulvic Mineral Burial Humic Crystal UnRx1 UnRx2 Uptake Decomposition PoMin Primary Production Amorphous P External Transfers Sorption PiMin Microbial P Soluble Po Deep soil (slow rates) Shallow soil (rapid rates) PO4

PAGE 122

122 Figure 5-2. Results of mathemati cal modeling of long-term declin es of five phosphorus fractions in Blue Cypress March Conservation Area. 0 50 100 150 200 050010001500200025003000 Microbial Fulvic Humic Labile Pi Residual Time, yrsPhosphorus, mg kg-1 0 50 100 150 200 050010001500200025003000 Microbial Fulvic Humic Labile Pi Residual Time, yrsPhosphorus, mg kg-1

PAGE 123

123 APPENDIX PHYSICAL SOIL DESCRIPTION Table 1. Physical description of soil samples collected in Ju ly, 2002 from Water Conservation Area 2A and Blue Cypress Marsh Conservation Area. Station Depth Description Water Conservation Area 2A E1 Rep I 0-10 Very fibrous, spongy peat. Many plant fragments; brown 10-20 Slightly drier peat; fibrous, brown 20-30 Similar to 10-20, but more compacted 30-40 Fibers becoming shorter, compacted peat 40-50 Fibrous brown peat 50-60 Fibrous brown peat, slightly darker, increase in fine organic matter (OM) 60-70 Fibrous brown peat, slightly darker, increase in fine organic matter (OM) 70-80 Fibrous brown peat, slightly darker, increase in fine organic matter (OM) 80-85 Approx. 2.5 cm brown coarse sand, overlain by 3.5 cm peat E1 Rep II 0-10 Very fibrous, wet, many plant parts 10-20 Sim. to above, many poorly decomposed plant parts. Large live Typha root; drier 20-30 More decomposed, still many identifiable plant parts, becoming fibrous peat 30-40 Slightly decomposed, dark brown fibrous peat 40-50 Dry, fibrous brown compacted peat

PAGE 124

124 Table A-1. Continued Station Depth Description E1 Rep II 50-60 Very compacted dry brown fibrous peat 60-70 Darker, more decomposed peat 70-80 Similar to above, slightly more watery 80-90 Compacted dry slightly sandy peat 90-97 Compacted dry slightly sandy peat E1 Rep III 0-10 Very wet, poorly decomposed peat, many identifiable plant parts 10-20 Fibrous, drier peat 20-30 Fibrous, somewhat dry peat 30-40 Fibrous, somewhat dry peat (dark brown; all sections) 40-50 fibrous, wetter peat. This was likely the break between successive core drives 50-60 Compacted dry fibrous peat. Slightly lighter brown 60-70 Fibrous peat 70-80 Fibrous peat, slightly darker 80-90 Slightly sandy dry peat. Many fine roots, even at this depth E5 Rep I 0-10 Fibrous rooty almost black peat 10-20 Fibrous rooty almost black peat. Less moisture 20-30 Fibrous, many fine roots, almost black 30-40 Fibrous, many fine roots, almost black 40-50 Similar to above, less roots 50-60 More decomposed, considerably higher moisture content 60-70 Well decomposed, still somewhat fibrous, high moiture content 70-80 Very well decomposed muck, black, increased moisture.

PAGE 125

125 Table A-1. Continued Station Depth Description E5 Rep I 80-90 Muck, slight increase in amount of fine roots and fibers 90-100 Muck, high moisture content 100-110 Decrease in moisture, fibrous muck 110-120 Slightly sandy peat, color change to light brown 120-130 Very sandy peat 130-140 Silty sand, high OM, some identifiable ancient plant fragments E5 Rep II 0-10 Dark very fibrous peat 10-20 Dark very fibrous peat. Slightly drier 20-30 Dark very fibrous peat. Slightly drier 30-40 Dark very fibrous peat. Moisture content same as above 40-50 Some very large, very decomposed woody material + peat 50-60 Piece of wood 2 diam. and peat 60-70 Well decomposed, almost muck, somewhat dry 70-80 Mucky, fibrous peat 80-90 Sandy peat, slightly red in color 90-100 Peat intermixed with bands of sand 100-110 110-120 peaty sand E5 Rep III 0-10 Very rooty peat, many plant parts 10-20 Rooty peat, drier, cladium roots 20-30 Less rooty, drier, fibrous peat 30-40 Fibrous peat, slightly mucky 40-50 Mucky peat 50-60 Mucky peat 60-70 Mucky peat, increased water content 70-80 Fibrous, mucky peat

PAGE 126

126 Table A-1. Continued Station Depth Description 80-90 Fibrous, slightly mucky, drier peat 90-100 Orange, fibrous peat 100-110 Sandy peat 110-120 Peaty sand Blue Cypress Marsh Conservation Area C1 Rep I 0-10 Much detrital matter, high water content 10-20 Fibrous peat, much drier 20-30 Fibrous peat, much drier 30-40 Drier still, compacted fibrous peat 40-50 Fibrous peat 50-60 Mucky fibrous peat, increases in water content 60-70 Well-compacted fibrous peat, decrease in water content, lighter brown 70-80 Well-compacted fibrous peat, decrease in water content, lighter brown, low water content for peat 80-90 Dry light brown peat, many plant fragments 90-100 Darker mucky peat, increase in water content 100-110 Fibrous brown peat 110-120 Fibrous brown peat 120-130 Fibrous brown peat 130-135 Fibrous brown peat C1 Rep II 0-10 Fibrous, poorly compacted peat. Many large plant fragments 10-20 Much drier, slightly compacted peat 20-30 Brown fibrous peat 30-40 Brown fibrous peat 40-50 Brown fibrous peat, slightly darker 50-60 Brown fibrous peat 60-70 Brown fibrous peat 70-80 Compacted brown fibrous peat 80-90 Compacted brown fibrous peat 90-98 Compacted brown fibrous peat

PAGE 127

127 Table A-1. Continued Station Depth Description C1 Rep III 0-10 Fibrous poorly compacted peat. Mostly detrital 10-20 Much drier, more compacted. Peat decreasing in plant fibers 20-30 Compacted peat, smaller fibers 30-40 Darker, muckier peat 40-50 Increasing water content, not as compacted 50-60 Drier, dense mucky peat 60-70 Wet poorly consolidated peat, many large fragments 70-80 Similar to above, many large fragments 80-90 Brown, dry compacted peat 90-100 Brown, dry compacted peat. Sandy silty peat at base of core B4 Rep I 0-10 Very rooty consolidated peat, no detrital material 10-20 Very rooty consolidated peat, no detrital material. Fibrous consolidated peat 20-30 Rooty fibrous peat 30-40 Rooty fibrous peat 40-50 Rooty fibrous peat 50-60 Slightly wetter consolidated peat 60-70 Fibrous rooty peat 70-80 Rooty peat, fibrous; increased muck content 80-90 Mucky peat 90-100 Mucky peat 100-110 Mucky peat B4 Rep II 0-10 Very rooty somewhat consolidated peat little or no detritus 10-20 Rooty fibrous consolidated peat 20-30 Rooty fibrous consolidated peat 30-40 Rooty fibrous consolidated peat

PAGE 128

128 Table A-1. Continued Station Depth Description 40-50 Rooty fibrous consolidated peat, increasing muck content 50-60 Rooty fibrous consolidated peat, slightly mucky 60-70 Rooty fibrous consolidated peat, slightly mucky 70-80 Mucky peat, slightly fibrous 80-90 Mucky peat, slightly fibrous, very mucky, well decomposed 90-100 Mucky peat, slightly fibrous, very mucky, well decomposed 100-110 Mucky peat, slightly fibrous, very mucky, well decomposed B4 Rep III 0-10 Very rooty ( Panicum ) consolidated peat 10-20 Rooty consolidated peat 20-30 Fibrous peat 30-40 Fibrous peat 40-50 Rooty fibrous peat 50-60 Slightly mucky fibrous peat 60-70 Mucky fibrous peat 70-80 Mucky fibrous peat 80-90 Mucky fibrous peat 90-98 Mucky fibrous peat

PAGE 129

129 Table A-2. Physical descripti on of soil samples collected in March 2004 from Blue Cypress Marsh Conservation Area. Station Depth Description B4 0-10 Loosely consolidated detrital peat 10-20 Fibric, slightly more consolidated peat 20-30 Rooty, very fibric peat 30-40 Rooty, very fibric peat 40-50 Rooty, very fibric peat 50-60 Rooty, fibric, small increase in OM 60-70 Rooty, fibric peat 70-80 Rooty, fibric peat 80-90 Noticeable increase in very fine OM 90-100 Rooty, fibric peat 100-110 Rooty, fibric peat 110-120 Rooty, fibric peat 120-130 Rooty, fibric peat C1 0-10 Loosely consolidated detrital peat; rooty (live) 10-20 Slightly more consolidated peat 20-30 Slightly compacted detrital peat 30-40 Fibric peat 40-50 Fibric peat 50-60 Big increase in fine black OM in matrix of peat 60-70 Similar to above, even more fine OM 70-80 Decline in fine OM; base of slough? 80-90 Fibric peat 90-100 Lighter brown fibric peat 100-110 Reddish brown fibric peat 110-120 Reddish brown fibric peat 120-130 Reddish brown fibric peat 120-140 Reddish brown fibric peat 140-150 Similar to above, some charcoal

PAGE 130

130 LIST OF REFERENCES Anderson, J. M., 1976. An ignition method for determination of total phosphorus in lake sediments. Water Research 10:329-331. Baldock, J.A., J.M. Oades, P.N. Nelson, T.M. Sk ene, A. Golchin, and P. Clarke. 1997. Assessing the extent of decomposition of natural or ganic materials using solis-state 13C NMR spectroscopy. Australian J. Soil Research. 35:1061-1083. Bartkowiak, M., and R. Zakrzewski. 2004 Thermal de gradation of lignins isolated from wood. J. Thermal Analysis and Calorimetry. 77:295. Binford, M.W., E.S. Deevey, and T.L. Crisman. 1983. Paleolimnology: an historical perspective on lacustrine ecosystems. A nn. Rev. Ecol. Syst. 14: 255-286. Brenner, M., C.L. Schelske, and L.W. Keenan. 2001. Historical rates of sediment and nutrient accumulation in marshes of the Upper St. Johns River Basin, Florida, USA. J. Paleolimnol. 26:241-247. Brezonik, P.L., and D.R. Engstrom. 1998. Modern and historic accumulation rates of phosphorus in Lake Okeechobee, Florida. J. Paleolimnol. 20:31-46. Brookes, P.C., D.S. Powlson, and D.S. Jenki nson. 1982. Measurement of microbial biomass phosphorus in soil. Soil Biol. Biochem. 14:319-329. Buck, D. G., M. Brenner, D.A. Hodell, and J.H. Curtis. 2005. A late Holocene record of climate variability from hypersaline Lago Enriquillo, Dominican Re public. Presentation at 2005 American Society of Limnology and Ocea nography meetings, Salt Lake City, Utah. Burrough, P.A. 1996. Opportunities and limitations of GIS based modeling of solute transport at regional scale. p 19-38. In: Applications of GIS to the Modeling of Non-point Source Pollutants in the Vadose Zone. SSSA Speci al Pub. No. 48. SSSA. Madison, WI. Cade-Menun, B.J., and C.M. Preston. 1996. A co mparison of soil extraction procedures for 31P NMR spectroscopy. Soil Science. 161: 770-785. Cade-Menun, B.J., C.W. Liu, R. Nunlist, and J. G. McColl. 2002. Soil and litter 31PNMR spectroscopy: extractants, metals and phos phorus relaxation times. J. Env. Qual. 31:457465. Cade-Menun. 2005. Using phosphorus-31 nuclear magnetic resonance spectroscopy to characterize organic phosphorus in environmental samples. In: Organic Phosphorus in the Environment. pp. 21-44. B.L. Turner, E. Fr ossard, and D.S. Baldwin (eds.). CABI Publishing. Cambridge, MA. Carignan, R., and R.J. Flett. 1981. Postdepositional mobility of phosphorus in lake sediments. Limnol. Oceanogr. 26:361-366

PAGE 131

131 Celi, L. and E. Barberis. 2005. Ab iotic stabilization of organic phosphorus in the environment. In: Organic Phosphorus in the Environment. pp. 113-132. B.L. Turner, E. Frossard, and D.S. Baldwin (eds.). CABI Publishing. Cambridge, MA. Chang, S.C., and M.L. Jackson. 1957. Fractionati on of soil phosphorus. Soil Sci. 84:133-144. Chen, Y., N. Senesi, and M. Schnitzer. 1977. Information provided on humic substances by E4/E6 ratios. Soil Sci. Soc. Am. J. 41:352-358. Clark, M.W. 2000. Biophysical charact erization of floating wetlands (flotant) and its influence on vegetative succession of a warm-temperature aquatic ecosystem. Doctoral Dissertation. University of Florida. Cooper, W.T., J.M. Llewelyn, G.L. Bennett, A. C. Stenson, and V.J. Salters. 2005. Organic phosphorus speciation in natura l waters by mass spectrometry. In: Organic Phosphorus in the Environment. pp. 45-74. B.L. Turner, E. Frossard, and D.S. Baldwin (eds.). CABI Publishing. Cambridge, MA. Cox, L., R. Celis, M.C. Hermosin, J. Cornejo, A. Zsolnay, and K. Zeller. 2000. Effect of organic amendments on herbicide sorpti on as related to the nature of dissolved organic matter. Environ. Sci. Technol. 34: 4600-4605. Craft, C.B., and C.J. Richardson. 1993. Peat accretion and phosphorus accumulation along a eutrophication gradient in the north ern Everglades. Biogeochem. 22:133-156. Craft, C.B., and C.J. Richardson. 1998. Recent a nd long-term organic soil accretion and nutrient accumulation in the Everglades. Soil Sci. Soc. Am. J. 62:834-843. Crews, T.E., K. Kitayama, J.H. Fownes, R.H. R iley, D.A. Herbert, D. Mueller-Dombois, P. M. Vitousek. 1995. Changes in soil phosphorus fractions and ecosystem dynamics across a long chronosequence in Ha waii. Ecology. 76: 1407-1424. Crisman, T.L. 1978. Interpretations of past environments from lacustrine animal remains. pp.69101. In: Biology and Quarternary Environments 264 pp. D. Walter (ed.). Australian Academy of Sciences. Canberra, Australia. Davis, J.H.1943. The natural features of southern Florida, especially the vegetation and the Everglades. Geol Bull., 25, Flor ida Dept. of Conservation, FL Geol. Surv. Tallahassee. 311pp. Davis, S.M. 1991. Growth, decompos ition, and nutrient retention of Cladium jamaicense Crantz and Typha domingensis Pers. in the Flor ida Everglades. Aquat. Bot. 40: 203-224. DeBusk, W.F. and K.R. Reddy. 1998. Turnover of detr ital organic carbon in a nutrient impacted Everglades Marsh. Soil Sc i. Soc. Am. J. 62: 1460-1498.

PAGE 132

132 DeBusk, W.F., K.R. Reddy, M.S. Koch and Y. Wang. 1994. Spatial distribution of soil nutrients in a northern Everglades marsh: Water C onservation Area 2A. Soil Sci. Soc. Am J. 58:543-552. Devai, I. and R. D. DeLaune. 1995. Eviden ce for phosphine production and emission from Louisiana and Florida marsh soils. Org. Geochem. 23:277-279. Dixit, S.S., J.P. Smol, J.C. Kingston, and D.F. Charles. 1992 Diatoms: powerful indicators of environmental change. Environ. Sci. and Technol. 26:23-33. Doran, G. 2002. Windover : multidisciplinary inves tigations of an early Archaic Florida cemetery. University Press of Florida, Gainesville FL. Engstrom, D.R., S.P. Schottler, P.R. Leavitt, and K.E. Havens. 2006. A reevaluation of cultural eutrophication of Lake Okeechobee using multiproxy sediment records. Ecological Applications. 16:1194-1206. Felfli, F.F. A.A. Luengo, J. Suarez, and P. Beaton. 2005. Wood briquette torrefaction. Energy for Sustain. Develop. 9:19-22. Fertilizer Institute. 2007. http://www.tfi.org/factsa ndstats/statistics.cfm The Fertilizer Institute. 820 First St. N.E. Suite 430, Washington, DC 20002. Fisher, M. M., K. R. Reddy, and R. T. James. 2 005. Internal nutrient loads from sediments in a shallow, subtropical lake. Lake and Reserv. Manage. 21:338-349. Florida Department of Environmental Protecti on. 2004. Integrated Water Quality Assessment for Florida: 2004 305(b) Report and 303(d) List Update. Freeman, C., G. Liska, N.J. Ostle, M.A. Lo ck, B. Reynolds, and J. Hudson. 1996. Microbial activity and enzymic decomposition processe s following peatland water table drawdown. Plant and Soil. 180:121-127. Frey, D.G. 1969. The rationale for paleolimnol ogy. Mitt. Internat. Ve rein. Limnol. 17:7-18. Gleason, P. J., and P. Stone. 1994. Age, origin, and landscape evolution of the Everglades peatland. pp. 149-197 In S. M. Davis and J. C. Ogden (eds.) Everglades The ecosystem and its restoration. St. Luci e Press, Delray Beach, FL. Gleason. P.J. 1974. Chemical quality of water in Water Conservation Area 2A and associated canals. Technical Report 74-1. South Florid a Water Management District. West Palm Beach, Florida. Grunwald, S., K.R. Reddy, J.P. Prenger, and M.M. Fisher. 2007. Modeling the spatial variability of biogeochemical soil properties in a fr eshwater ecosystem. Ecological Modelling. 201:521-535.

PAGE 133

133 Hedin, L.O. P.M. Vitousek, P.A. Matson. 2003. Nutrient losses over four million years of tropical forest developm ent. Ecology. 84:2231-2255. Hieltjes, A.H., and L. Lijklema 1980. Fractionation of inorganic phosphates in calcareous lake sediments. J. Environ. Qual. 9:405-407. Hillel, D.A. 1986. Modeling in soil physics: A critical review. pp 35-42. In: Future Developments in Soil Science Resear ch. SSSA Misc. Publ. SSSA, Madison WI. Hodell, D.A., M. Brenner, and J.H. Curtis. 2005. Terminal Classic drought in the northern Maya lowlands inferred from multiple sediment cores in Lake Chichancanab (Mexico). Quaternary Sci. Rev. 24:1413-1427. Hupfer, M., B. Rube, P. Schmieder. 2004. Origin and diagenesis of polyphosphate in lake sediments: a 31P-NMR study. Limnol. Oceanogr. 49:1-10. Ivanoff, D.B., K.R. Reddy, and S. Robins on. 1998. Chemical frac tionation of organic phosphorus in selected Histos ols. Soil Science. 163:36-45. Jenny, H. 1941. Factors of soil formation. Mcgraw Hill, New York, NY. Kenney, W.F., C.L. Schelske, and M. Brenner. 2000. Changes in polyphosphate sedimentation: a response to excessive phosphorus enrichment in a hypereutrophic lake. Can. J. Fish. Aquat. Sci. 58: 879-887. Keough, J.R., M.E. Sierszen, and C.A, Hagley. 1996. An alysis of a Lake Superior food web with stable isotope techniques. Limnol. Oceanogr. 41:136-146. Levesque, M. and H. Dinel. 1978. Applicability of thermal methods for characterization of peats and plants. Geoderma. 20: 201-213. Lowe, E.F. and L.W. Keenan. 1997. Managing p hosphorus-based cultural eutrophication in wetlands: a conceptual appr oach. Ecol. Eng. 9:109-118. McLatchey, G.P. and K.R. Reddy. 1998. Regulation of organic matter decomposition and nutrient release in a wetland so il. J. Environ Qual. 27:1268-1274. Meyers, P.A. 1994. Preservation of elemental and isotopic source identifi cation of sedimentary organic matter. Chem. Geol. 114:289-302. Mitsch, W.J. and J.G. Gosselink. 1993. Wetlands Second edition. Van Nostrand Reinhold. New York, NY. Moore, P. A. K. R. Reddy and M. M. Fi sher. 1998. Phosphorus flux between sediment and overlying water in Lake Okeechobe e, Florida: spatial and tem poral variations. J. Environ. Qual. 27:1428-1439.

PAGE 134

134 Murphy, J. and J.P. Riley. 1962. A modified sing le solution method for the determination of phosphate in natural waters. Anal. Chim. Acta. 27:31-36. Ohno, T. 2002. Fluorescence inner filtering correction fo r determining the humification index of dissolved organic matter. Envi ron. Sci. Technol. 36: 742-746. Olila O. G., K.R. Reddy and L. Keenan. 1995. Labile and non-labile pools of phosphorus in surface waters and soils in the Upper St. Johns River Basin. Final report to the St. Johns River Water Management Di strict. Contract No. 94D179. Page, A.L. 1983. In: Methods of Soil Analysis, Part 2: Ch emical and Microbiological Properties. 2nd ed. A.L. Page (ed.). Am er. Society of Agronomy. Penn, M.R., M.T. Auer, E.L. Van Orman, and J.J. Korienek. 1995. Phosphorus diagenesis in lake sediments using fractionation techniques. Marine Freshwater Res. 46:89-99. Peterson, B.J. and B. Fry. 1987. Stable isotopes in ecosystem studies. Ann. Rev. Ecol. Syst. 18:292-320. Ponnamperuma, F.N. 1972. The chemistry of submerged soils. Adv. Agron. 24:29-96. Prenger, J.P. and K.R. Reddy. 2004. Microbial enzy me activities in a freshwater marsh after cessation of nutrient loading. So il Sci. Soc. Am. J. 68:1796-1804. Reddy, K. R., and W. H. Patrick, Jr. 1983. Effect of aeration on reactivity and mobility of soil constituents. In : Chemical Mobility and Reactivity in Soil Systems (eds. D. W. Nelson, D. E. Elrick, and K. K. Tanji). SSSA Special Publ. No. 11:11-33. Soil Sci. Soc. Am. J. Madison, WI. Reddy, K.R. and R. DeLaune. In press. Chapter 9: Phosphorus In : Biogeochemistry of Wetlands: Science and Applications. Reddy, K.R. and W.H. Patrick. 1975. Effect of alternate aerobic and anaerobic conditions on redox potential, organic matter decomposition a nd nitrogen loss in a flooded soil. Soil Biol. Biochem. 7:87-94. Reddy, K.R., Y Wang, W.F. DeBu sk, M.M. Fisher, and S. Newman. 1998. Forms of soil phosphorus in selected hydrologi c units of the Florida Everglades. Soil Sci. Soc. Am. J. 62:1134-1147. Redfield, A.C., B.H. Ketchum and F.A. Rich ards. 1963. The influence of organisms on the composition of seawater. pp. 26-77. In The Sea, Volume 2. M.N. Hill (ed.). Interscience Publ., New York, NY. Richardson, C.J. 1985. Mechanisms controlling phos phorus retention capacity in freshwater wetlands. Science. 228:1424-1427.

PAGE 135

135 Richardson, C.J., S. Qian, C.B. Craft, and R.G. Qualls. 1997. Predictive models for phosphorus retention in wetlands. Wetlands Ecology and Management. 4: 159-175. Richter, D.D. H. Lee Allen, J. Li, D. Markewit z, J. Raikes. 2006. Bioavailability of slowly cycling phosphorus: major restructuring of so il P fractions over four decades in an aggrading forest. O ecologia. 150: 259-271. Rowell, D.L. 1981. Chapter 7: Oxidation and Reduction. In : The Chemistry of Soil Processes. pp. 401-461. D.J. Greenland and M.H.B. Ha yes, eds. John Wiley and Sons. Rydin, E. 2000. Potentially mobile phosphorus in Lake Erken sediment Water Res. 34:20372042. Sharpley, A.N. 2000. Global issues of phosphor us in terrestrial ecosystems. pp. 15-46. In Phosphorus biogeochemistry in subtropical ecosystems. K. R. Reddy, G.A. OConnor, and C.L. Schelske (eds.). Lewis Publishers, Boca Raton, FL. Shumate, B.C., C.L. Schelske, T.L. Crisman, and W.F. Kenney. 2002. Response of the cladoceran community to trophic state change s in Lake Apopka, Florida. J. Paleolimnol. 27:71-77. Sjgersten, Sofie, B.L. Turner, N. Mahie u, L.M. Condron, and P. A. Wookey. 2003. Global Change and Biol. 9: 759-772. Smol, J.P. 1992. Paleolimnology: an important t ool for effective ecosystem management. J. Aquatic Ecosys. Health. 1:49-58. Stevenson, F.J. 1994. Humus Chemistry: Gene sis, Composition, and Reactions. 2nd Edition. John Wiley & Sons. Steward, K.K., and W.H. Ornes. 1975. The auteco logy of sawgrass in th e Florida Everglades. Ecology 56: 162-171. Stewart, J.W.B and H. Tiessen. 1987. Dynamics of soil organic phosphorus. Biogeochemistry. 4:41-60. Stuiver, M. and P.J. Reimer.1993. Extended 14C database and revised CALIB 3.0 14C age calibration program. Radiocarbon. 35: 215-230. Tabatabai, M.A. 1982. Soil Enzymes. p. 903-947. In : Methods of Soil Analysis. Part 2 Chemical and Microbiological Methods. A.L. Page et al. (eds.). 2nd ed. Ag ronomy Monogr. 9. ASA and SSSA. Madison, WI. Turner, B.L. 2004. Optimizing phosphorus charact erization in animal manures by solution Phosphorus-31 Nuclear Magnetic Resonance spectroscopy. J. Environ. Qual. 33: 757-766. Turner, B.L., and S. Newman. 2005. Phosphorus cycling in wetland soil s: the importance of phosphate diesters. J. Environ. Qual. 34:1921-1929.

PAGE 136

136 Turner, B.L., M.J. Paphazy, P.M. Haygarth, and I.D. McKelvie. 2002. Inositol phosphates in the environment. Phil. Trans. R. Soc. Lond. 357:449-469. Turner, B.L., N. Mahieu, and L.M. Condron. 2003a. 31P MNR spectral assignments of phosphorus compounds in soil NaOH-EDTA extrac ts. Soil Sci. Soc. Am. J. 67: 497-510. Turner, B.L., N. Mahieu, and L.M. Condron. 2003b. The phosphorus composition of temperate pasture soils determined by Na OH-EDTA extraction and solution 31P MNR spectroscopy. Organic Geochemistry. 34: 1199-1210. United States Environmental Protection Agency. 1996. Environmental indica tors of water quality in the United States. EPA-841-R-96-002. USEPA Office of Water (4503f). US Government Printing Office, Washington DC. 25 pp. United States Environmental Protection Ag ency. 1998. 303d List of Impaired Waters. United States Environmental Protection Agen cy. 2002. National Water Quality Inventory: 2000 Report. EPA-841-R-02-001. USEP A Office of Water (4503f). US Government Printing Office, Washington DC. Urban, N.H., S.M. Davis, and N.G. Aumen. 1993. Fluc tuations in sawgrass and cattail densities in Everglades Water Conservation Area 2A under varying nutrient, hydrologic and fire regimes. Aquat. Bot. 46: 203-223. Van Soest, J.P. 1970. Forage and Fiber An alysis Handbook No. 379. US DA. Washington, D.C. Waksman, S.A. 1936. Humus. Williams and Wilkins Co., Baltimore, MD. Walker, T.W. and J.K. Syers. 1976. The fate of phosphorus during diagenesis. Geoderma. 15:119. Wetzel, R.G. 1999. Organic phosphorus mine ralization in organic soils. pp. 225-245. In Phosphorus biogeochemistry in subtropical ecosystems. K. R. Reddy, G.A. OConnor, and C.L. Schelske (eds.). Lewis Publishers, Boca Raton, FL.. Whitmore, T.J. 1989. Florida diatom assemblages as indicators of trophic state and pH. Limnol. Oceanogr. 34: 882-895. Wright, A.L. and K.R. Reddy. 2001. Phosphorus load ing effects on extrace llular enzyme activity in Everglades wetland soils. Soil Sci. Soc. Am. J. 65:588-595.

PAGE 137

137 BIOGRAPHICAL SKETCH Millard Mathias Fisher, III was born in Hattiesburg Mississippi in 1958 and moved to Florida at the age of two years. He grew up in St. Augustine, Florida, USA, where he attended Catholic parochial and hi gh school. He holds a b achelors degree in engineering technology from the University of Central Florida, a bachelors degree in environmental engineering from the University of Florida, and a masters degree in soil and water science, also from the University of Florida. In 2007, he completed his doctoral degree in soil and water science at the Univ ersity of Florida. Fi shing, skiing, kayaking, and flying are his favorite hobbies.