Phytoplankton Productivity and Dynamics in the Caloosahatchee Estuary, Florida, USA

MISSING IMAGE

Material Information

Title:
Phytoplankton Productivity and Dynamics in the Caloosahatchee Estuary, Florida, USA
Physical Description:
1 online resource (215 p.)
Language:
english
Creator:
Mathews, Ashley Loren
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Fisheries and Aquatic Sciences, Forest Resources and Conservation
Committee Chair:
Phlips, Edward J
Committee Members:
Parkyn, Daryl Charles
Havens, Karl
Baker, Shirley M
Montague, Clay L

Subjects

Subjects / Keywords:
bzpi0 -- caloosahatchee -- estuary -- florida -- model -- phytoplankton -- productivity -- subtropical
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre:
Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Urban and agricultural development in and around the Caloosahatchee Estuary on the southwest coast of Florida in the USA has altered the flow and quality of water in the system since the late 1800s.  Increasing algal blooms have brought attention to water quality and processes affecting phytoplankton production and biomass accumulation there.  The primary objectives of this project were to (1) measure phytoplankton productivity to test a previously developed empirical model based on simple measures of phytoplankton biomass and light availability in the photic zone, (2) define the abundance and composition of the phytoplankton community to identify patterns of succession, (3) assess changes in water quality for its direct and indirect effect on the phytoplankton community, primary production, and the model relationship, and (4) apply the phytoplankton productivity model to analyze long-term changes in the estuary’s trophic status.  Experiments and analyses were conducted using integrated water samples collected monthly between February 2009 and February 2010 at four sites (one in each region of the estuary and bay).  Primary production rates, in terms of oxygen evolution, were measured using simulated in situ light:dark bottle incubations in a flow-through raceway.  Measurements of daily gross primary productivity (GPPd) ranged from 90 to 3121 mg•C•m-2•d-1 with the overall annual mean estimated at 346 g•C•m-2•yr-1.  When the estimates from all four sites were pooled there was a strong linear relationship between GPPd and the ‘light•biomass’ model predictor (r2 = 0.84, p less than 0.001).  The model tended to overestimate productivity during the dry period when dinoflagellates were dominant and underestimate productivity during the wet period when diatoms were dominant.  These deviations in the model were examined as indicators of secondary controls on phytoplankton production in the Caloosahatchee Estuary.  The model relationship was then applied over a twenty-five year period using a water quality data set compiled there between January 1986 and December 2010.  Estimates of annual gross primary productivity (GPPy) varied spatially and temporally from oligotrophic (less than 100 g•C•m-2•yr-1) to hypertrophic levels (greater than 500 g•C•m-2•yr-1), given the influence of both natural and anthropogenic drivers on the productivity potential of the phytoplankton community.
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 Ashley Loren Mathews.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Phlips, Edward J.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-05-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045307:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 PHYTOPLANKTON PRODUCTIVITY AND DYNAMICS IN T HE CAL OOSAHATCHEE ESTUARY, FLORIDA, USA By ASHLEY LOREN MATHEW S A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIR EMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Ashley Loren Mathews

PAGE 3

3 To Fisher

PAGE 4

4 ACKNOWLEDGMENTS This research was primarily funded by the South Florida Water Managemen t District, West Palm Beach. Dr. Peter Doering coordinated an introductory tour of the Caloosahatchee Estuary and provided valuable insight on sampling methods and historical conditions in the system. Drs. Chenxia Qiu and Miao Li Chang offered helpful fe edback on experimental methods. Kathy Haunert provided access to technical data and documents and facilitated contract communication. Additional funding and support was provided by the Dr. Edward J. Phlips Laboratory and the University of Florida, Gaine sville. Field assistance was provided by and colleague Dr. Nico le Dix Pangle. I thank them for being agreeable and diligent despi te the early mornings, long day s, unpredictab le weather conditions, and physically demanding tasks. Laboratory assistance was primarily provided by Dorota Roth and Joey Chait, but I would like to recognize the entire staff of the Dr. Edward J. Phlips Laboratory for their support and encouragement du ring this project. Phytoplankton enumeration and identification was conducted by Susan Badylak. Technical assistance with the design and construction of the experimental equipment was provided by Dr. Lance Riley. A number of people/groups in the Fort M yers community deserve recognition for helping this project run so smoothly. The Peppertree Community Marina and Lee County Parks and Recreation Department provided boat ramp access and experimental set up space near our sampling sites. I would also like to thank the staff of Legacy Harbor Marina of Fort Myers, particularly Marc Sullivan Sheila Turregano, Dave

PAGE 5

5 Eric Ravenschlag Lana Carlin, and Marisa Zavala for their hospitality and generous accommodations. This project would not have been possible without the guidance and direction of Professor Ed P hlips, my advisor and committee chair, who encouraged me to apply myself, pursue my PhD, and lead this research effort. I would also like to thank Drs. Shirley Baker, Karl Havens, Clay Montague, and Daryl Parkyn for serving on my supervisory committee and for providing feedback and support throughout this process. Some of the data analyses used in this project were conducted by Dr. Chuck Jacoby, who offered hours of statistical support to a stud ent he had never met before. Additional data approaches and considerat ions were offered by Dr. Dana Bigham and Mark Hoyer Lastly, I would like to thank my family and friends for having faith in me and encouraging me throughout this journey. I am fo rever grateful to my husband Matt, my parents Dan and Karen Stephens, and my parents in law Joe and Beth Mathews, for helping me find the time, strength, and motivation to finish this project. My baby boy Fisher brought such joy to my life during this exp erience, and I thank him for making me feel like I have already accomplished the greatest achievement in my life

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIG URES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 18 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 20 Anthropogenic Influences in Estuaries and Coastal Ecosystems ........................... 20 Primary Production in Estuaries and Coastal Ecosystems ................................ ..... 21 Study Overview and Objectives ................................ ................................ .............. 23 2 SYSTEM BACKGROUND AND LITERATURE REVIEW ................................ ........ 26 History of th e Caloosahatchee Estuary ................................ ................................ ... 26 Critical Issues in the Caloosahatchee Estuary ................................ ........................ 28 Management and Restoration Efforts in the Caloosahatchee Estuary .................... 32 3 MODELING PHYTOPLANKTON PRODUCTIVITY IN A SHALLOW, MICROTIDAL, EXTENSIVELY MODIFIED, SUBTROPICAL ESTUARY IN SOUTHWEST FLORIDA ................................ ................................ ........................ 36 Concepts and Applications ................................ ................................ ..................... 36 Methods ................................ ................................ ................................ .................. 40 Study Area ................................ ................................ ................................ ........ 40 Water Sampling ................................ ................................ ................................ 41 Field Measurements ................................ ................................ ......................... 42 Meteorological and Hydrological Data ................................ .............................. 43 Chemical Analyses ................................ ................................ ........................... 44 Phytoplankton Analyses ................................ ................................ ................... 45 Primary Productivity Experiments ................................ ................................ ..... 46 Primary Productivity Model Test ................................ ................................ ....... 49 Primary Productivity Model Application ................................ ............................ 51 Results ................................ ................................ ................................ .................... 51 Physical Chemical Conditions ................................ ................................ ....... 51 Phytoplankton Abundance and Composition ................................ .................... 55 Measured Primary Productivity ................................ ................................ ......... 56 Primary Productivity Model Fit ................................ ................................ .......... 57

PAGE 7

7 Mod eled Primary Productivity ................................ ................................ ........... 58 Discussion ................................ ................................ ................................ .............. 59 Modeling Phytoplankton Productivity ................................ ................................ 59 Interpretation of Model Residuals ................................ ................................ ..... 63 Model overestimation ................................ ................................ ................. 66 Model underestimation ................................ ................................ ............... 69 Other considerations ................................ ................................ .................. 71 Comparative Rates of Production ................................ ................................ ..... 72 Summary ................................ ................................ ................................ ................ 76 4 NATURAL AND ANTHROPOGENIC INFLUENCES ON THE SPATIAL AND TEMPORAL PATTERNS OF PHYTOPLANKTON PRODUCTIVITY IN THE CALOOSAHATCHEE ESTUARY, FLORIDA ................................ .......................... 99 Concepts and Applications ................................ ................................ ..................... 99 Methods ................................ ................................ ................................ ................ 103 Study Area ................................ ................................ ................................ ...... 103 Water Quality Data Sources and Manipulations ................................ ............. 105 Water Sampling and Analyses ................................ ................................ ....... 109 Meteorological and Hydrol ogical Data Sources ................................ .............. 110 Primary Productivity Estimates ................................ ................................ ....... 113 Decomposition of Phytoplankton Biomass Variability ................................ ..... 114 Statistical Analyses for Identification of Environmental Drivers ...................... 115 Results ................................ ................................ ................................ .................. 118 Spatial and Temporal Similarity Groupings of Data ................................ ........ 118 Meteorological and Hydrological Observations ................................ .............. 119 Water Quality Variations ................................ ................................ ................. 123 Phytoplankton Biomass and Pro ductivity Patterns ................................ ......... 126 Spatial and Temporal Changes in Trophic Status ................................ .......... 129 Phytoplankton Biomass Variability at Different Scales ................................ ... 133 Environmental Drivers of Phytoplankton Productivity ................................ ..... 137 Discussion ................................ ................................ ................................ ............ 138 Relationships between Phytoplankton Product ivity and Environmental Variables ................................ ................................ ................................ ..... 138 Climate and weather ................................ ................................ ................ 138 Temperature ................................ ................................ ............................ 139 Rainfall ................................ ................................ ................................ ..... 140 Salinity ................................ ................................ ................................ ..... 142 Light ................................ ................................ ................................ ......... 145 Nutrients ................................ ................................ ................................ ... 146 Patterns of Phytoplankton Biomass Variability ................................ ............... 149 Annual variability ................................ ................................ ...................... 149 Seasonal variability ................................ ................................ .................. 152 Residual variability ................................ ................................ ................... 155 Variations in Trophic Status ................................ ................................ ............ 156 Summary ................................ ................................ ................................ .............. 159

PAGE 8

8 5 CONCLUSION ................................ ................................ ................................ ...... 197 LIST OF REFERENCES ................................ ................................ ............................. 203 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 215

PAGE 9

9 LIST OF TABLES Table page 3 1 Definitions of variables and their units of measure. ................................ ............ 79 3 2 m odel s. ................................ ................................ ................................ ............... 79 3 3 Phytoplankton biomass potential based on Redfield stoichiometric proportions of SRP, DIN, a nd SI ................................ ................................ ......... 92 3 4 Summary of measured and modeled d aily primary productivity ( GPP d in 2 1 ) from each of the four sites in the Caloosahatchee Estuary FL ... 96 3 5 Summar y of tested model relationships ................................ ............................. 96 3 6 Annual phytoplankton productivity estimates (system wide ranges or site specific means) of vari ous river dominated estuarie s. ................................ 98 4 1 Sampling sites used in water quality monitoring programs/research projects in the four regions of the Caloosahatchee Estuary, defined by their distance in kilometers from S 79, the Franklin Lock and Dam ................................ ....... 162 4 2 Median salinity values observed in the four regions of the Caloosahatchee Estuary from December 1985 to May 1989, November 1994 to August 1996, and April 1999 to June 2003 (Doering et al. 2006) ................................ ........... 162 4 3 Recommended (SFWMD et al. 2009) and observed (January 1986 to December 2010) frequency distribution of mean monthly inflows to the Caloosahatchee Estuary, FL from S 79. ................................ ........................... 169 4 4 Mean (with standard deviations) physical chemical, biomass, and productivity values associated with the fo ur regions (UE, ME, LE, and BY) ........................ 173 4 5 Mean (with standard deviations) physical chemical, biomass, and productivity values associated with the five salinity zones (freshwater, oligohaline, me sohaline, polyhaline, euhaline) ................................ ................................ .... 174 4 6 Standard deviations of the annual (SD y ), seasonal (SD m ), and residual (SD ) components of phytoplankton biomass variability in the Caloosahatchee Estuary, FL ................................ ................................ ................................ ...... 194 4 7 Best combinations of environmental drivers that explained patterns in daily gross primary productivity ( GPP d ) in the four regions of the Caloosahatchee Estuary, FL. ................................ ................................ ................................ ...... 196

PAGE 10

10 LIST OF FIGURES Figure page 3 1 Location of the Caloo sahatch ee Estuary, FL ................................ ..................... 78 3 2 Location of the four sampling sites in the Upper Estuary (UE), Middle Estuary (ME), Lower Estuary (LE), and San Carlos Bay (BY). ................................ ........ 78 3 3 Schematic showing how the specific light transmittance depth profile in the water column at each site (A) was emulated with various light treatment levels in the incubation raceway (B ). ................................ ................................ .. 80 3 4 Daily rainfall (cm) (A) and flow (m 3 1 ) (B) from February 2009 to February 2010 at the Franklin Lock and Dam (S 79), FL ................................ .................. 81 3 5 Surface and bottom sali nities (SAL in psu) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009). ................................ ................................ ..................... 82 3 6 Surface and bottom water temperature (TEMP_W in C) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009). ................................ ................................ ............ 82 3 7 Model parameters for site UE in t he Caloosahatchee Estuar y, FL. ................... 83 3 8 Model parameters for site ME in t he Caloosahatchee Estuary, FL .................... 84 3 9 Model parameters for site LE in t he Caloos ahatchee Estuary, FL ..................... 85 3 10 Model parameters of site BY in t he Caloosahatchee Estuary, FL ....................... 86 3 11 Model application for t he Ca loosahatchee Estuary, FL ................................ ..... 87 3 12 Estimates of light attenuation coefficients ( K T in m 1 ) and corresponding concentrations of (A) color dissolved organic matter (CDOM in pcu), (B) turbidity (T URB in ntu), (C) and uncorrected chlorophyll a (CHL A 3 ) ... 88 3 13 1 ) of (A) total soluble phosphorous (TSP), (B) total phosphorous (TP), (C) total soluble nitrogen (TSN), and (D) total nitrogen (TN) at each of the four sites in the Caloosahatchee Estuary, FL ...................... 89 3 14 1 ) of (A) dissolved inorganic nitrogen (DIN) and (B) soluble reactive phosphorous (SRP) and (C) mass ratios of DIN to SRP at each of the four sites in the Caloosahatchee Estuary, FL ................................ .. 90

PAGE 11

11 3 15 Concentrations 1 ) of silica (SI) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009). ................................ ................................ ..................... 91 3 16 Total phytoplankton biovolume concentration ( BV in 10 6 3 1 ) of dinoflagellates, diatoms, cyanobacteria, and other phytoplankton taxa at each of the four sites in the Caloosahatchee Estuary, FL ................................ .. 94 3 17 Percent contribution of A. sanguinea and other dinoflagellates, S. cf costatum and other diatoms, cyanobacteria, and other phytoplankton taxa to total phytoplankton biovolume concentration ( BV in 10 6 3 1 ) .......................... 95 3 18 Regression of daily gross primary productivity ( GPP d 2 1 ), against the composite parameter BZ p I 0 for 24 incubatio n experiments ............. 97 3 19 Estimated differences between predicted and measured primary productivity values for all four sites across six monthly incubation experiments. ................... 97 4 1 Location and connection of the Caloosahatchee Estuary, FL with respect to Lake Okeechobee via the Caloosahatchee River (C 43 Canal) and the Gulf of Mexico and Charlotte Harbor via San Carlos Bay. ................................ ....... 161 4 2 Four regions of the Caloosahatchee Estuary sampled in long term water quality monitoring programs/research projects. ................................ ................ 161 4 3 (A) Multivariate ENSO (El Ni o/Southern Oscillation) Index (MEI) between J anuary 1986 and December 201 0. ................................ ................................ 163 4 4 Average monthly air temperatures (C) (A) and departures from normal (C) (B) reco rded at the meteorological station in Fort Myers, FL (fro m January 1986 to December 2010 ................................ ................................ .................. 164 4 5 Actual water temperatures (TEMP_W in C), observed in the Caloosahatchee Estuary, FL between January 1986 and December 2010. ..... 165 4 6 M onthly rainfall totals (cm) (A) and departure s from normal (cm) (B) recorded at the meteorological station in Fort Myers, FL fro m January 1986 to December 2010 ................................ ................................ ............................... 166 4 7 Mean monthly flow (m 3 1 ) from S 79 into the Caloosahatchee Estuary, FL between January 1986 and December 2010. ................................ ................... 167 4 8 Daily flows (m 3 1 ) along the C aloosahatchee River between Lake Okeechobee and the Caloosahatchee Estuary at water control structures S 77 (A), S 78 (B), and S 79 (C). ................................ ................................ ..... 168

PAGE 12

12 4 9 Salinity (SAL in psu) ranges divided i nto five zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) observed in the four regions (UE, ME, LE, BY) of the Caloosahatchee Estuary, FL ................................ ........................... 170 4 10 Salinity (SAL in psu) observed in each of the four regions of the Caloosahatchee Estuary, FL between January 1986 and December 2010 with respect to mean monthly inflow (m 3 1 ) of freshwater at S 79. ................. 171 4 11 Predicted steady state salinity (SAL in psu) distribution along a spatial gradient in the Caloosahatchee Estuary, FL for a range of freshwater inflows (m 3 1 ) from S 79 ................................ ................................ ............................. 172 4 12 1 ) across five salinity (SAL) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of th e four regions (UE, ME, LE, BY) ................................ ................................ ... 175 4 13 1 ) across five salinity (SAL) zones (freshwater, oligohaline, m esohaline, polyhaline, euhaline) in each of th e four regions (UE, ME, LE, BY) ................................ ....................... 176 4 14 Monthly mean silica (S 1 ) across five salinity (SAL) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of th e four regions (UE, ME, LE, BY) ................................ ................................ ................ 177 4 15 Monthly mean photic depth ( Z p in m 1 ) across five salinity (SAL) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of th e four regions (UE, ME, LE, BY ) ................................ ................................ ................. 178 4 16 Monthly mean chlorophyll a (CHL A 3 ) across five salinity (SAL) zones (freshwater, oligohaline, mesohaline, polyhaline, eu haline) in each of th e four regions (UE, ME, LE, BY) ................................ ................................ .... 179 4 17 Monthly mean daily gross primary productivity ( G PP d 2 1 ) across five salinity (SAL) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of the four regions (UE, ME, LE, BY) ................................ ... 180 4 18 Close up view of phytoplankton biomass responses (in terms of CHL A 3 ) in variable salinities (F = freshwater, O = oligohaline, M = mesohaline, P = polyhaline, E = euhaline ) from April ...... 181 4 19 Close up view of phytoplank ton productivity responses (as GPP d estimates in 2 1 ) in variable salinities (F = freshwater, O = oligohaline, M = mesohaline, P = polyh aline, E = euhaline) from April ................................ 182 4 20 Close up view of phytoplankton biomass responses (in terms of CHL A 3 ) in variable salinities (F = freshwater, O = oligohaline, M = mesohaline, P = poly haline, E = euhaline) from June ...... 183

PAGE 13

13 4 2 1 Close up view of phytoplankton productivity responses (as GPP d estimates in 2 1 ) in variable salinities (F = freshwater, O = oligohaline, M=mesohaline, P=pol yhaline, E=euhaline) from June ................................ ..... 184 4 22 Modeled average annual gross primary prod uctivity ( GPP y 2 1 ) in the four regions (UE, ME, LE, and BY) of the Caloosahatchee Est uary, FL between 1986 and 2010 ................................ ................................ .................. 185 4 23 Modeled average annual gross primary productivity ( GPP y 2 1 ) by salinity zone (freshwater, oligohaline, mesohaline, polyhaline, and euh aline) in the Caloosahatchee Estu ary, FL between 1986 and 2010 .......................... 186 4 24 Modeled average annual gross primary productivity ( GPP y 2 1 ) in the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL between 1986 and 2010 with respect to season ................................ .............. 187 4 25 Modeled average annual gross primary productivity ( GPP y 2 1 ) by salinity zones (freshwater, oligohaline, mesohaline, polyhaline, and euhaline) of the Caloosahatchee Estuary, FL between 1986 and 2010 with respect to season (DRY = November through Ap ril, WET = May through October) ......... 188 4 26 Average expected annual gross primary productivity ( GPP y 2 1 ) for the Caloosahatchee Estuary, FL overall and for each region (UE, ME, LE, and BY) and salinity zone (freshwater, oligohaline, mesohali ne, polyhaline, and euhaline) ................................ ................................ ................................ ... 189 4 27 Region UE monthly m ean chlorophyll a concentrations (CHL A 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C), and seasonal, ( m ) (D) components of phytoplankton biomass variability ................ 190 4 28 Region ME monthly mean chlorophyll a concentrations (CHL A 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annu al ( y ) (C), and seasonal, ( m ) (D) components of ph ytoplankton biomass variability ................ 191 4 29 Region LE monthly mean chlorophyll a concentrations (CHL A 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C), and seasonal, ( m ) (D) components of phy toplankton biomass variability ................ 192 4 30 Region BY monthly mean chlorophyll a concentrations (CHL A 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C), and seasonal, ( m ) (D) components of phytoplankton biomass variability ................ 193 4 31 Bubble plot comparing patterns of chlorophyll a (CHL A ) variability across 84 sites sample d within 5 1 estuarine coastal ecosystems ................................ ... 195

PAGE 14

14 LIST OF ABBREVIATIONS B p hytoplankton biomass as chlorophyll a (CHL A ) concentration BV t otal phytoplankton biovolume concentration BY San Carlos Bay C carbon C long term CHL A mean C degrees celsius C 43 C aloosahatchee R iver portion of trans state canal CAL Caloosahatchee Estuary Program CCHM Coastal Charlotte Harbor Monitoring Program CDOM c olor dissolved organic matter CERP Comprehensive Everglades Restoration Plan CES Center for Environmental Studies program CESWQ Caloosahatchee Estuary Water Quality Monitoring Program CF confer or compare CHL A c hlorophyll a a measure of phytoplankton abundance C IJ CHL A concentration in year i and month j CWA Clean Water Act D day DIN dissolved inorganic nitrogen DIP dissolved inorganic phosphorous DO d isso lved oxygen residual component of phytoplankton biomass variability E 1 e folding time or water residence time, representing the time needed for a total water mass to reach 37% of its initial water mass ECP Everglades Construction Project

PAGE 15

15 E G for example ENSO El Ni o/Southern Oscillation ERD Environmental Research and Design Program ET AL and others FDEP Florida Department of Environmental Protection FEFA Florida Everglades Forever Act G gram GPP gross primary productivity GP P D daily gross p rimary productivity GP P Y annual g ross primary productivity HB Harbor Branch Project I E that is I 0 total daytime surface irradiance or the level of solar radiation just above the surface of the water column KM kilometer K T l ight attenuation coefficient L liter LBOD luminescent b iological oxygen demand LE Lower E stuary M meter M seasonal (monthly) component of phytoplankton biomass variability ME Middle Estuary MEI Multivariate ENSO Index MFL minimum flow and level MG milligram ML milliliter

PAGE 16

16 MOL mole N total number of observations NCP net community productivity NELAP National Environmental Laboratory Accreditation program NH 3 ammonia NO 2 nitrite NO 3 nitrate NPP net primary productivity NP P D n et daily primary productivity NPP Y n et annual primary productivity NTU nephelometric turbid ity units P level of significance for statistical hypothesis testing Spearman correlation coefficient adjusted for tied ranks PAR p hotosynthetic ally active radiation PCU platinum colbalt units PSU practical salinity unit PSU photosynthetic unit PUR p hoto synthetically usable radiation R 2 coefficient of determination S second S similarity coefficient S 77 Moore Haven Lock and Dam S 78 Ortona Lock and Dam S 79 Franklin Lock and Dam SAL s alinity

PAGE 17

17 SAV submerged aquatic vegetation SD standard deviation SI s ilica SRP s oluble reactive phosphorous SFWMD South Florida Water Management District TEMP_W w ater temperature TKN t otal K jelda h l nitrogen TMDL total maximum daily load TN t otal nitrogen TP t otal phosphorous TSN total soluble nitrogen TSP total soluble phosphor ous TSS t o tal suspended solids TURB t urbidity UE Upper Estuary USA United States of America USACE United States Army Corps of Engineers VEC v aluable ecosystem component W watt Y year Y annual component of phytoplankton biomass variability Z P p hoti c depth o r the depth where 1 % of the incident irradiance is available Z W w ater depth Z S s ecchi disk depth

PAGE 18

18 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Deg ree of Doctor of Philosophy PHYTOPLANKTON PRODUCTIVITY AND DYNAMICS IN THE C AL OOSAHATCHEE ESTUARY, FLORIDA, U SA By Ashley Loren Mathews May 2013 Chair: Edward Phlips Major: Fisheries and Aquatic Sciences Urban and agricultural development in and around the Caloosahatchee E stuary o n the southwest coast of Florida in the U SA has altered the flow and quality of water in the syste m since the late 1 800s. Increasing a lgal blooms have brought attention to water quality and processes affecting phytoplankton pr oduction and biomass accumulation there. The primary objectives of this project were to (1) measure phytoplankton productivity to test a previously developed empirical model that is based on simple measures of phytoplankton biomass and light availability in the photic zone, (2) define the abundance and composition of the phytoplankton community to ident ify patterns of succession, (3) assess changes in water quality for its direct and indirect effect on the phytoplankton community, primary production, and t he model relationship and (4) apply the phytoplankton productivity model to analyze long term changes in the Experiments and analyses were conducted using integrated water samples collected monthly between February 2009 and Febr uary 2010 at four sites (one in each region of the estuary and bay). P rimary production rates, in terms of oxygen evolution, were measured using simulated in situ light:dark bottle incubations in a flow through raceway. Measurements of d aily g ross p rimar y productivity ( GPP d )

PAGE 19

19 ranged from 9 0 to 3121 m 2 d 1 with the overall annual mean estimated at 3 46 m 2 yr 1 When the estimates from all four sites were pooled there was a strong linear relationship between GPP d and the model pred ictor ( r 2 = 0.84, p <0.001). The model tended to overes timate productivity during the dry period when dinoflagellates were domina nt and underes timate productivity during the wet period when diatoms were domina nt These deviations in the model were examine d as indicators of secondary controls on phytoplank ton production in th e Caloosahatchee Estuary. T he model relationship was then applied over a twenty five year period using a water quality data set compiled there between January 1986 and December 2010 Estimates of a nnual gross primary productivity ( GPP y ) varied spatially and temporally from m 2 yr 1 ) to hypertrop hic levels (greater than 500 m 2 yr 1 ), given the influence of both natural and anthropogenic drivers on th e productivity potential of the phytoplankton communit y.

PAGE 20

20 CHAPTER 1 INTRODUCTION Anthropogenic Influences in Estuaries and Coastal Ecosystems Estuaries and coastal ecosystems are both historical and modern hotspots for human development due to their provisi on of food, energy, transportation, recreation, aesthetics, and other valuable resources. The great civilizations of our time have been work within two hundred kilometers land surface (Hinrichsen 1999). Despite the constraints on space, coastal population growth is expected to continue in this century. In the United States alone, an increase from 153 million people in 200 3 to 165 million people in 2015 is expected within the limits of 673 contiguous coastal counties, constituting only 17% of the total land area of the United States, excluding Alaska (Crossett et al. 2004). This attraction to, and dependency on, our coastli nes does not come without consequences, and these impacts stem from human activities occurring both within and beyond the traditional, political boundaries of these areas. Increases in coastal population densities give rise to human domination and alterat ion of these ecosystems and their watersheds, making them susceptible to habitat degradation and loss, hydrological modification, pollution, overfishing, invasive species, and a number of other issues (Vitousek et al. 1997a, Crossett et al. 2004). Eutroph ication of estuaries and coastal ecosystems has also emerged as one of the most pressing and increasingly studied problems of the late 20th and early 21st centuries (Nixon 1995, Cloern 2001, Schindler 2006). The formation and accumulation of toxic, harmfu l, and/or nuisance

PAGE 21

21 algal blooms that are commonly associated with anthropogenic nutrient enrichment have been linked to a suite of ecological and economic impacts. Primary Production in Estuaries and Coastal Ecosystems The dominant influences of humanity, with eutrophication being a prime example, ultimately affect the structure and function of e stuaries and coastal ecosystems (Day et al. 1989, Vitousek et al. 1997 a Zingone et al. 2010). A better understanding of the structure and function of estuaries an d coasta l ecosystems is desired so that cross comparisons and sound management decisio ns can be made (Odum 1968, Boye r et al 1993, Sand Jensen 1997). One of the fundamental ways that scientists define and compare the structure and function of ecosystems is by their rates of production. The production of organic matter through the assimilation and transformation of solar energy serves as the basis for natural food webs and plays a major role in the global carbon cycle (MacFadyen 1948, Odum 1968, Sand Jens en 1997). Measures of production are therefore characterizations of the trophic status of ecosystems, indicating their ability to support the metabolic demands of organisms and the related processes going on within them relative to other places in the wor ld (Nixon 1995). For example, estuaries and coastal ecosystems having a supply of org anic carbon greater than 300 but 2 1 are said to be eutrophic (Nixon 1995). Measurements of primary production have thus become an important part o f characterizing both aquatic and terrestrial ecosystems, as well as defining the impacts of anthropogenic influences on their health (Barbour et al. 1987, Sorokin 1999). Phytoplankton are major contributors to primary production in estuaries and coastal ecosystems, making them a key component of the overall structure and function of these environments (Zingone et al. 2010). They are a ubiquitous and diverse

PAGE 22

22 assemblage of free floating, microscopic algae that produce food energy via photosynthesis (Day e t al. 1989, Paerl et al. 2010). Phytoplankton production typically varies spatially and temporally within and across different estuaries (Boynton et al. 1982, Day et al. 1989). This variability is due to morphological, physical, chemical, and biological factors that are unique to each ecosystem (Brylinsky and Mann 1973, Boynton et al. 1982, Day et al. 1989). These regulatory factors can influence phytoplankton production on various scales from the individual cell, through ecophysiological responses to en vironmental changes, to entire ecosystems by forcing species selection and succession (Day et al. 1989, Falkowski 1994). Phytoplankton are thus sensitive and important indicators for detecting ecological change in estuaries due to their fast growth rates a nd rapid responses to a wide range of environmental disturbances (Paerl et al. 2010). Efforts to directly measure rates of phytoplankton production are therefore of great importance, but the methods used often encou nter major logistical barriers to their application (Ryther 1956b, Tilzer 1989, Sand Jensen 1997). Attempts to model production using easily and broadly measured parameters have become popular for bridging gaps in methodology (Odum 1968, Vollenwider 1969, Brylinsky and Mann 1973, Boyer et al. 1 993, Scardi 1996). The development of phytoplankton productivity models has been based on the assumptions that (1) primary productivity is proportional to primary producer biomass, (2) phytoplankton biomass is closely correlated with chlorophyll a concent rations, and 3) photosynthesis is proportional to light availability (Ryther 1956a, Ryther 1956b, Ryther and Yentsch 1957, Brylinsky and Mann 1973, Geider and Osborne 1992). Cole and Cloern (1987) combined these principles in the form of a simple empirica l model that

PAGE 23

23 predicts daily phytoplankton productivity from the composite parameter BZ p I 0 which is the product of phytoplankton biomass ( B ) in terms of chlorophyll a the photic depth ( Z p ), and the total amount of solar radiation ( I 0 ). The relationship w as initially developed from field measurements taken in San Francisco Bay, California (Cole and Cloern 1984), but Cole and Cloern (1987) and other researchers have since tested and applied this roductivity in a number of estuaries, varying in location and classification (Keller 1988, Mallin et al. 1991, Boyer et al. 1993, Kelly and Doering 1997, Murrell et al. 2007, Bouman et al. 2010). Using easily and commonly measured variables from routine oceanographic investigations, the BZ p I 0 has provided an opportunity to eliminate burdensome, expensive, time consuming, and often unreliable primary production studies and acquire more ro bust productivity estimates over larger spat ial and temporal scales than the traditional approaches allow (Ryther and Yentsch 1957 Cole and Cloern 1987, Boyer et al. 1993 Kelly and Doering 1997 ). These productivity estimates can expand the usefulness of long time series by extracting the necessar y biomass and light availability data to create ecosystem baselines and track the effects of natural and anthropogenic changes (Zingone et al. 2010). The collec tion and interpretation of time series data is useful in the study of estuarine and coastal eco system structure and function since these environments limit the use of whole ecosystem experiments (Cloern 2001). Study Overview and Objectives The overall goal of this study was to describe the primary production of phytoplankton in the Caloosahatchee Estuary, Florida as a means of assessing its current conditions and long term changes in trophic status The Caloosahatchee

PAGE 24

24 Estuary is a shallow, microtidal, subtropical system and one that has been extensively modified and managed by humans since the lat e 1800s, creating a combination of influences. Due to its relatively shallow morphology, the photic zone of the Caloosahatchee Estuary often extends to the bottom, expanding the capacity for phytoplankton production throughout the water column. The microtidal range keeps the delivery and resuspension of sediments relatively low in the Caloosahatchee Estuary, allowing phytoplankton to respond positively to nutrient inputs wit hout being inhibited by turbidity. As a subtropical system, the Caloosahatchee Estuary experiences relatively high temperatures and solar radiation levels year round, supporting phytoplankton growth through all seasons, including the winter. The subtropi cal climate also contributes significant rainfall, particularly during the summer to fall wet season, increasing the supply of nutrients to support the productivity potential of the phytoplankton com munity. This study was implemented in the context of th e following obj ectives and corresponding hypotheses: Objective 1 The applicability of the BZ p I 0 in the Caloosahatchee Estuary was tested using measurements of phytoplankton biomass, photic depth, surface irradiance, and primary p roductivity. The complex and dynamic changes in water quality that either magnify or truncate the underlying natural, seasonal patterns. Objective 2 The phytoplankton communit y of the Caloosahatchee Estuary was defined in terms of species a bundance and composition to identify patterns of succession that could influence the predictive power of the model. The hypothesis was that any shifts in the phytoplankton community would co rrespond to shifts in water quality ( i.e., salinity, nutrient availability, light availability, etc.) and reflect the species specific preferences for and adaptations to variable environmental conditions thereby, altering the productivity potential of the

PAGE 25

25 phytoplankton community given the quantity of biomass and light that is present in the water column Objective 3 C hanges in water quality (i.e., salinity, nutrient availability, light availability, etc.) were assessed in the Caloosahatchee Estuary for their direct or indirect effect on the phytoplankton community, primary production, and the model relationship. The hypothesis was that water quality would vary with respect to natural and anthropogenic inputs of freshwater into the estuary, creating a co mbination of factors that could support or suppress phytoplankton productivity at different spatial and temporal scales, which may not be accounted Objective 4 Th e adapted primary productivity model was applied to an existing twenty five year discontinuous water quality data set from the Caloosahatchee Estuary as a means of examining the responses of the phytoplankton community to both natural and anthropogenic changes in t he system. The hypothesis was that phytopl ankton productivity would fluctuate between extremes in response to anthropogenically enhanced seasonal patterns, El Nio/Southern Oscillation (ENSO) cyc les, and episodic storm events, thereby creating spatial and temporal ic status that challenge current management practices and plan s

PAGE 26

26 CHAPTER 2 SYSTEM BACKGROUND AND LITERATURE REVI E W History of the Caloosahatchee Estuary Prior to the movement of settlers to and the development of the area, the Caloosahatchee Estuary and its watershed was an expansive mixture of sloughs, sawgrass marshes, and pine flatwoods (Flaig and Capece 1998, Knight and Steele 2005, Perry 2008). Like the rest of the Florida Everglades, the system was completely dependent on and susceptible to the nat ural flow of water that moved freely between Lake Okeechobee, the surrounding floodplains, the Caloosahatchee River, and out to the Gulf of Mexico. A small canoe trail used by Native Americans connected Lake Okeechobee to Lake Flirt (located just east of La Belle), which served as the headwaters to the small, meandering Caloosahatchee River that flowed to Beautiful Island and into the estuary (Flaig and Capece 1998, Knight and Steele 2005). tters were eager to settle in South Florida with the promise of cheap land that could be drained and used for farming (Foster and Wessel 200 7 Perry 2008). By 1881, the governor of Florida had encouraged Hamilton Disston, a Philadelphia toolmaker and deve loper, to purchase and drain the land around Lake Okeechobee, the heart of the Everglades, to allow farmers to work its surrounding rich, dark soils (Foster and Wessel 200 7 ). Disston began by dredging a n approximately fifteen meter wide canal along the ex isting Native American canoe trail between Lake Okeechobee and Lake Flirt to provide a navigable steamboat channel for the movement of people and goods throughout the area (Flaig and Capece 1998, Foster and Wessel 200 7 ). These development efforts were tes ted in the 1920s when two major hurricanes hit the area causing extensive flooding,

PAGE 27

27 devastation, and death. Despite the hard economic times of the Great Depression, President Hoover responded in 1930 with money appropriated by Congress to construct the He rbert Hoover Dike around Lake Okeechobee to prevent future flooding (Foster and Wessel 200 7 ). In the years to follow, southern Florida oscillated between too little and too much water due to periods of drought and hurricane flooding, respectively. The st ate and federal government responded again in 1948 to the water demands and devastation by authorizing the U.S. Army Corps of Engineers (USACE) to start a $208 million dollar flood control project that would eventually include over 2,800 kilometers of cana ls and levees and some 200 water control structures throughout central and southern Florida (Guest 2001, Perry 2008). The Central and Southern Florida Flood Control District, now known as the South Florida Water Management District (SFWMD), was subsequent ly created to serve as the local sponsor and manager of this federal project, which was referred to by the same name (Guest 2001, Kranzer 2003, Perry 2008). As part of the Central and Southern Florida Flood Control Project, the Caloosahatchee River was f urt her dredged, straightened, deepe ned, widened, and connected to the St. Lucie River on the opposi te side of Lake Okeechobee to create a tran s state shipping channel (named the C 43 Canal) that linked the Gulf of Mexico to the Atlantic Ocean. The dredgin g resulted in the construction of a series of other canals, locks, and pumping stations that were designed to remove excess water from the surrounding lands (Foster and Wessel 200 7 ). Two lock and dam structures were completed in the 1930s at the towns of Moore Haven (referred to as S 77) and Ortona (referred to as S 78) to control the river flow and discharge from Lake Okeechobee

PAGE 28

28 (Doering and Chamberlain 1999). The final structure, the W.P. Franklin Lock and Dam (referred to as S 79), was built in the 196 0s approximately forty two kilometers upstream from the Gulf of Mexico near Olga in east Lee County (Flaig and Capece 1998, Foster and Wessel 200 7 ). This structure was designed to assure a fresh water supply for the lower region of the watershed and to pr event salt water intrusion from the Caloosahatchee Estuary into the upstream aquifers (Flaig and Capece 1998, Foster and Wessel 200 7 Doering and Chamberlain 1999). Moreover, S 79 created two distinct river systems within the Caloosahatchee. Upstream and east of S 79, the river is characterized as a pooled freshwater system that is isolated from the tidal influences of the Gulf of Mexico, while the downstream and western part of the river remains an estuary where seawater from the Gulf of Mexico mixes wit h freshwater inflows from the surrounding watershed and eastern stretch of the river (Wessel and Capece 200 7 ). Today, the Caloosahatchee River, Estuary and Watershed represent a system that was originally used, subsequently alter ed, and currently maintai ned to meet the immediate needs and priorities of Florida residents. What started as an effort to make the lands more inhabitable and cultivatable has resulted in historic changes to the natural water flow patterns in the Everglades and across central and southern Florida. These historical changes and current demands have created a great challenge to understand and manage the balance between natural and anthropogenic processes going on in the entire Caloosahatchee River Watershed. Critical Issues in the C aloosahatchee Estuary The modifications to the physical shape and hydrology of the Caloosahatchee River, Estuary, and Watershed have subsequently altered the flow of water into and through out the entire system. These changes have caused large fluctuations in the

PAGE 29

29 quantity, quality, timing, and distribution of freshwater inflow from the river to the estuary, impacting the ecology and economy of the system (Chamberlain and Doering 1998a, Barnes 2005, SFWMD et al. 2009). Anthropogenic control over the deliver y of freshwater into the Caloosahatchee Estuary has been associated with unnatural variations in salinity and nutrient loads, which together are considered the major issues affecting downstream organisms and the overall health and value of the system (SFWM D et al. 2009). The delivery of freshwater into the Caloosahatchee Estuary varies naturally between periods of heightened and reduced rainfall and corresponding runoff occurring in the summer/fall wet season (May thro ugh October) and winter/spring dry se ason (November through April), respectively. However, the development of the area has caused rainfall runoff that was once retained within the watershed to reach the estuary more quickly and in higher volumes (Barnes 2005). In addition, these seasonal pat terns are exacerbated by regulatory releases of freshwater from Lake Okeechobee (through S 77) and the Upper Caloosahatchee River and Watershed (through S 78) into the Caloosahatchee Estuary (through S 79). Upstream discharges occur in order to maintain prescribed water levels in Lake Okeechobee, prevent flood ing, and flush algal blooms, salt water and other contaminants out of the Caloosahatchee River (Flaig and Capece 1998, Doering and Chamberlain 1999 Rand and Bachman 2008 ). Fr eshwater releases are withheld to meet the urban and agricultural demands on the public water supply. The combined effects of rainfall, runoff, and regulatory releases cause the estuary to receive excessive freshwater inflows during the wet season, while flow is reduced or sto pped completely during the dry season (SFWMD et al. 2009).

PAGE 30

30 Salinity in the Caloosahatchee Estuary varies widely in time and space in response to both natural and anthropogenic inputs of freshwater. However, the volume of freshwater entering the Caloosahat chee Estuary through S 79 overwhelms any other source (Chamberlain and Doering 1998a, Flaig and Capece 1998), so inflows from the Caloosahatchee River have the greatest implications for monitoring and control and, therefore, receive the highest attention w ith respect to their influence on the physiological tolerances of downstream organisms (Chamberlain and Doering 1998b). Oysters and submerged aquatic vegetation (SAV), including Vallisneria americana (tape grass) and the seagrasses Halodule wrightii (shoa l grass) and Thalassia testudinum (turtle grass) are considered key species or valuable ecosystem components (VECs) in the Caloosahatchee Estuary (SFWMD et al. 2009), so the limits of freshwater inflows that protect and enhance oyster and SAV productivity should lead to a healthy and diverse estuarine ecosystem (Chamberlain and Doering 1998a). A gradient of freshwater to euhaline conditions is present in the Caloosahatchee Estuary when monthly inflow averages between 14.2 and 28.3 m 3 1 (Bierman 1993), pr oviding desirable salinities for all organisms somewhere in the system (Chamberlain and Doering 1998a, Knight and Steele 2005). Flows below this range limit the distribution of tape grass and increase the susceptibility of oysters to disease and predation while flows above this range reduce salinities below the preferred tolerances of the seagrasses and oysters (Chamberlain and Doering 1998a, Chamberlain and Doering 1998b, SFWMD et al. 2009). Excessive mean monthly inflows greater than about 79.3 m 3 1 are detrimental to other estuarine biota and cause stress to the estuary (Chamberlain and Doering 1998b, SFWMD et al. 2009). A minimum mean monthly

PAGE 31

31 inflow of 8.5 m 3 1 from S 79 was established in 2001 as the specific minimum flow and level (MFL) require d to maintain sufficient salinities at the Ft. Myers monitoring station, although a minimum monthly mean target of 12.7 m 3 1 is necessary to avoid MFL violations during dry periods (Knight and Steele 2005, SFWMD 2008, SFWMD et al. 2009). Nutrient concent rations across the Caloosahatchee Estuary are the product of point and non point sources of natural and anthropogenic inputs, including those derived from river flow, rainfall runoff, wastewater disposal, sewage overflow, atmospheric deposition, and elemen tal recycling (Knight and Steele 2005). However, the loading of new nutrients into the Caloosahatchee Estuary is primarily dependent on both the quantity and quality of freshwater entering the system through S 79. Annual flow s through S 79 are large enou gh, on average, to fill the entire volume of the estuary over eight times per year (Doering and Chamberlain 1999). The majority of this freshwater comes in the form of rainfall runoff from the Caloosahatchee River Basin as opposed to direct discharges fro m Lake Okeechobee (Chamberlain and Doering 1998a, Doering and Chamberlain 1999), so its composition is a function of land use and anthropogenic activities in the upper watershed. Current land use in the upper (freshwater) portion of the river basin is dom inated by agriculture and ranching (Knight and Steele 2005). Runoff from these areas carries excess concentrations of nitrogen and phosphorous from over fertilization and animal defecation. Concentrations of phosphorous are already naturally high in the waters of the Caloosahatchee Estuary due to the leaching of phosphate rich rock formations underlying the watershed (Odum et al. 1998). As a result, the loading of nitrogen to the Caloosahatchee Estuary is

PAGE 32

32 perceived to be more critical with respect to lim iting primary productivity and managing eutrophication (McPherson and Miller 1990, McPherson et al. 1990, Montgomery et al. 1991, Doering et al. 2006, Heil et al. 2007, Bailey et al. 2009). Multiple studies in the Caloosahatchee Estuary have confirmed tha t increases in nitrogen loads are associated with increases in phytoplankton productivity, which in turn, are associated with peak concentrations of phytoplankton biomass (McPherson and Miller 1990, McPherson et al. 1990, Montgomery et al. 1991, Doering et al. 2006). The resulting algal blooms reduce water quality by decreasing light availability for SAV and depleting oxygen in the water column needed by fish and other aquatic organisms (Day et al. 1989). Additionally, certain species of algae produce tox ins that can directly or indirectly harm inve rtebrates, fish, birds, mammals, and humans (Cloern 2001, Brand and Compton 2007, SFWMD et al. 2009) Total maximum daily loads (TMDLs) of total nitrogen (TN) into the Caloosahatchee Estuary were recently imple mented by the Florida Department of Environmental Protection in order to reach target concentrations of chlorophyll a (annual average of 11 m m 3 ) in these impaired waters (Bailey et al. 2009). A maximum load of 4,121 metric tons of TN per year, represen ting a 23% load reduction, should restore the Caloosahatchee Estuary to pre development concentrations and provide suitable conditions for a healthy SAV community, and by default, the entire estuarine ecosystem (Bailey et al. 2009). Management and Restorat ion Efforts in the Caloosahatchee Estuary The extensive changes to freshwater inflows along with the associated salinity fluxes and nutrient loads have negatively impacted water quality in the Caloosahatchee Estuary for more than 130 years. These impacts have been most notably documented as declines in SAV beds and oyster reefs and increases in algal blooms in the area

PAGE 33

33 (Chamberlain and Doering 1998a, Volety 2008 Barnes 2005, Brand and Compton 2007, Lapointe and Bedford 2007, Perry 2008, SFWMD et al. 2009) These outcomes have been taken as indicators of poor estuarine health and eutrophication in the Caloosahatchee Estuary and have become the primary focus of management plans and restoration efforts in the system (SFWMD et al. 2009). Efforts to manage and restore the Caloosahatchee Estuary and the Florida Everglades have increased since the 1970s, starting with the passing of the Clean Water Act (CWA) in 1972 as an amendment to the Federal Water Pollution Control Act of 1948. The CWA established regulatio ns for discharging pollutants and setting water landowner rights into a form of a dministrative water law that brought all waters of the state under regulatory control through the establishment of five water management districts (Rand and Bachman 2008, Carriker and Borisova 2009). Around this time, the last of the major works of the 19 48 Central and Southern Flood Control Project was completed, and the Florida Legislature began to recognize the harm that the project had caused to South Florida (Guest 2001). The final report from the project concluded that ulture and drainage had begun to eutrophy Lake named SFWMD shifted its focus from strictly flood miti gation to more general water control in 1975, but it continued to ignore state mandates to regulate nutrient pollution from agricultural runoff and back pumping into the lake (Guest 2001). Just downstream,

PAGE 34

34 a waste load allocation study in the early 1980s confirmed that the Caloosahatchee Estuary had already reached its nutrient loading limits as indicated by elevated chlorophyll a and depressed dissolved oxygen concentrations, suggesting that increased nutrient inputs would cause chlorophyll a concentratio ns to reach dangerous levels (Degrove 1981, Doering et al. 2006). Water quality in these South Florida ecosystems received national attention in 1988 when the United States Federal Government filed a lawsuit against the SFWMD and the Florida Department o f Environmental Protection (FDEP) alleging that agricultural runoff and water discharges in the Everglades were in violation of state water quality standards (Guest 2001, Perry 2008). The resulting settlement essentially sparked the state of Florida to ta ke action to improve water quality, starting with the passing of the Florida Everglades Forever Act (FEFA) in 1994 (Guest 2001). Under the FEFA, the SFWMD initiated the Everglades Construction Project (ECP), whereby they were to purchase land, construct s tormwater treatment areas, and improve the quality of waters coming from the Everglades agricultural area (Guest 2001, Perry 2008). By the mid 1990s, it was more and more evident that the 1948 project that was intended to make the Everglades more conduci ve to development and agriculture was damaging Lake Okeechobee and the connected ecosystems (Guest 2001). A task force was formed under the direction of the USACE to review the original project and come up with a corrective plan, which became known as the modify the works of the Central and Southern Florida Flood C ontrol Project of 1948 to resto re natural hydrological patterns, while maintaining flood co ntrol and improving

PAGE 35

35 urban and agricultural water supplies (Guest 2001, Rand a nd Bachman 2008). The resultant plan was adopted by the State of Florida in 1999 and the United States Congress in 2000 as the Comprehensive Everglades Restoration Plan (CERP), w hich was to be implemented over a period of 40 years at an estimated cost of eight billion dollars (Guest 2001, Kranzer 2003, Rand and Bachman 2008). CERP ha s been described as the largest ecosystem restoration effort in history, encompassing an area of a pproximately 47,000 square kilometers that covers 16 counties (Kranzer 2003, Perry 2008). In conjunction with CERP, the Florida Legislature has passed a number of ecosystem r estoration in the greater Everglades. Much of these were focused on improving water quality in Lake Okeechobee, such as the 2000 Lake Okeechobee Protection Act and its subsequent program. It was not until 2007 that the state of Florida expanded this effo rt into the Northern Everglades and Estuaries Protection Program to recognize the connectivity of the watersheds surrounding Lake Okeechobee (SFWMD et al. 2008). This latest program guided the development of new comprehensive, systematic, and multi agency plans that specifically addressed the objectives and benefits of CERP and other projects in the rivers and estuaries connected to Lake Okeechobee. The Caloosahatchee River Watershed Protection Plan was adopted in 2009 and is still used today as the prefe rred plan that consolidates and outlines the best combination of construction projects, pollutant control programs, research, and water quality monitoring programs needed to improve the quantity, quality, timing, and distribution of water in the system (SF WMD et al. 2009).

PAGE 36

36 CHAPTER 3 MODELING PHYTOPLANKTON PRODUCTIVITY IN A SHALLOW, MICROTIDAL, EXTENSIVELY MODIFIED, SUBTROPICAL ESTUARY IN SOUTHWEST FLORIDA Concepts and Applications In many estuaries, phytoplankton are the largest contributors to primary pr oduction and therefore play a central role in estuarine ecosystem structure and function (Boynton et al.1982, Day et al. 1989). The ability to measure primary productivity, or rates of primary production, is therefore of great importance, but the methods used to make such determinations can be complicated, expensive, and inconsistent (Ryther 1956b, Ryther and Yentsch 1957, Tilzer 1989, Sand Jensen 1997). These challenges have led to the development of model relationships that estimate primary productivity using more easily determined and broadly measured parameters, such as light flux, photic depth, and phytoplankton biomass (most often in terms of chlorophyll a ). The utility of light availability and chlorophyll a in modeling phytoplankton productivity w as recognized by the mid twentieth century (Ryther 1956b, Ryther and Yentsch 1957). These early developments provided researchers a tool to estimate primary productivity using data routinely collected during oceanographic investigations (Ryther and Yentsc h 1957). More recently, Cole and Cloern (1987) combined these principles in the form of a simple empirically based model for predicting daily phytoplankton productivity, i.e. BZ p I 0 where B is phytoplankton biomass in terms of chlorophyll a Z p is photic depth, and I 0 i s daily incident irradiance. The BZ p I 0 model is essentially a measure of phytoplankton biomass multiplied by a term representing light availability in the wate r column (Boyer et al. 1993, Brush and Brawley 2009). The model was derived from its original form BI 0 K T 1 (Cole and Cloern 1984), where K T is the

PAGE 37

37 light attenuation coefficient and equated to 4.61 Z p 1 The more commonly used BZ p I 0 relationship was compar ed to field measurements of phytoplankton productivity for several estuaries, including San Francisco Bay in California, Puget Sound in Washington, Hudson River plume in New York, Delaware Bay in Delaware, and the Neuse and South Rivers in North Carolina ( Cole and Cloern 1987). A strong linear relationship ( r 2 = 0.82) was obtained between the observed and modeled values ( N = 211) of primary productivity (Cole and Cloern 1987). The model has been applied to a number of other estuaries, including Narraganse tt Bay in Rhode Island (Keller 1988), Neuse River Estuary in North Carolina (Mallin et al. 1991, Boyer et al. 1993), Massachusetts Bay and Boston Harbor in Massachusetts (Kelly and Doering 1997), Escambia Bay in Florida (Murrell et al. 2007), and Tokyo Bay in Japan (Bouman et al. 2010), reaffirming the adaptability of the model to a range of ecosystem types. Like other significant oceanographic investigations and advances made throughout history, little effort has been put forth to extend this concept t o regions beyond the temperate zone. Except for the study done in Escambia Bay, which lies at 30N latitude in a warm temperate/subtropical region of northwest Florida, all other previous model applications have been conducted in temperate estuaries locat ed at latitudes between 35 and 50N. This paper describes the first field test of the Cole and C loern (1987) model in a low latitude subtropical system, the Caloosahatchee Estuary, located at 26N latitude on the southwest coast of Florida. Despite being categorized into similar geographic and climatic regions, the Caloosahatchee Estuary differs from Escambia Bay by having water temperatures that typically do not fall below 20C in comparison to lows of 10C in the winter in Escambia Bay (Murrell et al. 2 007).

PAGE 38

38 Furthermore, the seasonal pattern of rainfall and river flow in Escambia Bay is more common to temperate systems with peaks occurring in the winter and spring (February to April) and valleys occurring in the summer and fall (June to November) (Murre ll et al. 2007). The Caloosahatchee Estuary has the opposite freshwater inflow pattern, which is more ch aracteristic of tropical systems, and one that is accentuated by anthropogenic flushing events. As a result, the Caloosahatchee Estuary experiences a scenario during the wet season (May through October) when freshwater inputs from the watershed are coupled with high light levels and warm water temperatures, creating a combination of presumably ideal cond itions for phytoplankton blooms. This high potential for phytoplankton production in the Caloosahatchee Estuary is further and microtidal range, which are shared characteristics of most river dominated Gulf of Mexico estuaries, including Escam bia Bay (Murrell et al. 2007). Further distinguishing itself from the previously tested systems, the Caloosahatchee Estuary has been significantly altered by human development since the late 1800s when it was artificially connected to Lake Okeechobee ( Doering et al. 2006). The system has been further modified from its original state by the addition of three water control structures along the Caloosahatchee River and years of dredging to maintain a trans state waterway. These and other changes to the p hysical shape and hydrology of the system have caused unnatural salinity fluctuations, excess nutrient loading, elevated dissolved color concentrations, increased sediment deposition, and the formation and accumulation of both autochthonous and allochthono us harmful algal blooms ( Chamberlain and Doering 1998 a Barnes 2005, Knight and Steele 2005,

PAGE 39

39 SFWMD et al. 2009, RECOVER 2011). The responses of the phytoplankton community to changes in water quality, in particular the formation and accumulation of harmfu l algal blooms, have become a growing management concern (McPherson and Miller 1990, Doering et al. 2006, Brand and Compton 2007, Perry 2008). Considering the characteristics and issues that are unique to this shallow, microtidal, extensively modified, su btropical system, the usefulness of a simple model for estimating phytoplankton productivity in the Caloosahatchee Estuary was questioned. The purpose of this study was primarily to test the applicability of the BZ p I 0 model in the Caloosahatchee Estuary u sing field measurements of phytoplankton productivity collected in a manner similar to those described by Cole and Cloern (1987). Modifications to the study methods (using O 2 evolution instead of 14 C uptake productivity measurements) and model parameters, particularly the biomass ( B ) and photic depth ( Z p ) terms, were done to investigate their influence on the overall relationship. Model deviations and productivity patterns were described in terms of spatial and temporal variations in freshwater inflow ( i. e., rainfall and river flow), water quality ( i.e., salinity, light, nutrients, etc.), and phytoplankton community structure ( i.e., abundance and (in terms of its statistical significance and strength) (in terms of its appropriateness and usefulness) BZ p I 0 model adapted to the Caloosahatchee Estuary was used to determine annual estimates of phytoplankton productivity for the Caloosahatchee Estuary so that system wide comparisons to other systems could be made.

PAGE 40

40 Methods Study Area The Caloosahatchee Estuary covers an area of 62 km 2 in Lee County on the southwest coast of Florida. It is considered to be part of the larger, neighboring Charlotte Harbor National Estuary (SFWMD et al. 2009) (Figure 3 1). The Caloosa hatchee Estuary empties into and exchanges water tidally with the Gulf of Mexico through San Carlos Bay and Charlotte Harbor at Shell Point (Figure 3 2). Approximately 42 km upstream, the estuary is fed primarily by the Caloosahatchee River through the Fr anklin Lock and Dam (S 79) (Figure 3 2). The width of the estuary is irregular, ranging from 160 m near S 79 to 2,500 m in San Carlos Bay (Scarlatos 1988). The narrow section extending from S 79 to Beautiful Island has an average depth of 6 m, but the ar ea downstream of Beautiful Isl and has an average depth of 1.5 m (Scarlatos 1988). A narrow navigation channel (part of the Intracoastal Waterway) is maintained at a depth of approximately 3.5 m (Scarlatos 1988). The area experiences a combination of diur nal and mixed semi diurnal tides with a mean tidal range of 0.30 m in the middle of the estuary near downtown Fort Myers (Scarlatos 1988, NOAA 2010). Freshwater is released into the estuary through S 79 in order to maintain prescribed water levels in Lake Okeechobee, control flooding, and flush algal blooms and salt water out of the river (Flaig and Capece 1998, Doering and Chamberlain 1999). In an average year, enough freshwater flows from S 79 to fill the entire volume of the estuary over eight times (Doe ring and Chamberlain 1999). The flow through S 79 is composed of 75% rainfall runoff and 25% regulatory discharges from Lake Okeechobee (Chamberlain and Doering 1998 a ). Runoff comes from the 3,625 km 2 sized watershed

PAGE 41

41 (Knight and Steele 2005), which is ma de up of agricultural areas in the east and urban areas in the west. Additional sources of freshwater entering the estuary downstream of S 79, including the Orange River near Beautiful Island (Figure 3 2), are considered to be minor in comparison to S 79 (Scarlatos 1988, Flaig and Capece 1998, Knight and Steele 2 005). Water Sampling Four sampling sites were included in this study (Figure 3 2). Site UE was located in the Upper Estuary near Beautiful Island, which is northeast of downtown Ft. Myers. Site M E was located in the Middle Estuary near downtown Fort Myers. Site LE was located in the Lower Estuary near Iona Shores. Site BY was located in San Carlos Bay near Picnic Island. The four sites were located just outside the main navigation channel. Dep ths at the sampling sites were between 2 and 3 m. The sites were sampled once a month from February 2009 to February 2010 (excluding March 2009). For the first six monthly sampling events, water was collected from each site on four different, but cons e cutive days, within one hour before sunrise to provide water for the primary productivity experiments. Water was also collected from each site during the afternoon prior and afternoon following the pre dawn collections so that adjustments for differing w ater quality conditions during the primary productivity experiments could be made. In the later sampling months, water was collected from each site in the morning on the same day so that the productivity model could be applied to the remainder of the year All water samples were collected with a 3 m vertically integrating pole sampler to collect a mixed sample of water from the entire water column (Wetzel and Likens 1991). From each collection, sample water was retained, stored, and preserved for chemica l and phytoplankton analyses.

PAGE 42

42 Field Measurements Temperature (TEMP_W), salinity (SAL), and dissolved oxygen (DO) profiles were measured during each sampling event at 0.5 m intervals from the surface to the bottom of the water column using a H ACH (Loveland, Colorado) HQ40D meter. Light readings were taken simultaneously with paired LI 190 reference (deck) and LI 192 downwelling (2 ) L I C OR (Lincoln, Nebraska) quantum cosine corrected photosynthetically active radiation (PAR) sensors above the surface (incident irradiance) and at 0.5 meter intervals to the bottom, respectively. For the pre dawn sampling events, light attenuation wa s estimated from measurements taken during the afternoon before and after the primary productivity experiments. Light attenuation coefficients ( K T ) were calculated with the Beer Lambert relationship and taken as an average of the two coefficients obtaine d from inputting the downwelling irradiances at the 0.5 to 1.5 m depth and the 1.0 to 2.0 m depth, respectively. In cases where an average light attenuation coefficient could not be determined in this manner, K T was calculated with the same relationship u sing the incident irradiance corrected for 5% surface reflection (Ryther 1956 a ) and the downwelling irradiance at 1.0 m depth. For the pre dawn samples, light attenuation coefficients were then estimated by selecting the afternoon coefficient that most c losely represented the water conditions in the morning based on chlorophyll a (CHL A ), color dissolved organic matter (CDOM), and turbidity (TURB) values. The final K T (Table 3 1), representing light availability in the water column during the primary pro ductivity experiments, was then derived by applying correction factors for differences in CHL A (0.03; Wolfsy 1983), CDOM (0.015; McPherson and Miller 1987), and TURB (0.1152; determined from the relationship between turbidity and tripton during th e study period).

PAGE 43

43 Photic depth ( Z p ) one of the three main variables in the Cole and Cloern (1987) model, was defined as the depth of 1% incident irradiance and calculated as 4.61K T 1 (Table 3 1). Meteorological and Hydrological Data A separate LI COR (Lincoln, Nebraska) LI 190 reference (deck ) cosine corrected quantum sensor, with LI 1000 data logger, was used to determine the total PAR light flux over the incubation period of the primary productivity experiments (February through August 2009, excluding March). To test the Cole and Cloern (1987) model, t otal daytime surface irradiance ( I 0 ) was estimated by multiplying the total PAR light flux during the six hour incubations (sunrise to mid day) by 2 to represent a typical twelve hour light day (Table 3 1). To o btain productivity estimates from the BZ p I 0 model adapted to the Caloosahatchee Estuary for the entire annual study period (February 2009 through February 2010, excluding March 2009), PAR values recorded at three stations in Lee, Collier, and Hendry Counti es (Figure 3 1) were compiled from the South Instantaneous readings at fifteen minute intervals were converted into 2 1 and summed across the daylight period t o provide estimates of daily PAR flux. The maximum values of daily PAR flux across the tri county area were used to estimate monthly maximum I 0 (Table 3 1). Daily rainfall and flow values recorded at S 79 (Figure 3 2) were obtained from the South Florid (SFWMD 2010). Values were converted into metric equivalents, cm and m 3 1 respectively. These reco rds were used to identify wet and dry periods during the study since they represent the main s ources of freshwater into the estuary, which in turn were

PAGE 44

44 used to investigate seasonal changes in water quality, phytoplankton abundance and composition, and primary productivity. Chemical Analyse s Chemical analyses were conducted by the Dr. Edward J. Ph lips Laboratory at the University of Florida using methods certified by the National Environmental Laboratory Accreditation Program (NELAP). Aliquots of whole (unfiltered) and filtered sample water were retained, transported on ice, and stored frozen unti l subsequent chemical analysis. Whole water was used in the determination of chlorophyll a (CHL A ), total nitrogen (TN), total phosphorous (TP), silica (SI), and turbidity (TURB). To analyze samples for concentrations of ammonia (NH 3 ), nitrite (NO 2 ), nit rate (NO 3 ), total soluble nitrogen (TSN), total soluble phosphorous (TSP), soluble reactive phosphorous (SRP), and color dissolved organic matter (CDOM), triplicate samples of whole water were first filtered in the field using 0.7 m glass fiber filters. F ilters were stored frozen and in the dark until spectrophotometrically analyzed using a Hitachi U2810 ( Tokyo, Japan) for CHL A following hot ethanol extraction of pigments (Sartory and Grobbelaar 1984, APHA 2005). Both uncorrected and pheophytin corrected concentrations were determined. Average CHL A concentrations, based on triplicate samples, were used in the phytoplankton productivity model (Cole and Cloern 1987) as the p hytoplankton biomass parameter ( B ) (Table 3 1). Whole water samples for TN and TP analysis were digested using the persulfate oxidation method (APHA 2005) and measured colorimetrically on a Bran Luebbe AA3 auto analyzer ( Norderstedt, Germany) and a Hitachi U2810 dual beam spectrophotometer (Tokyo, Japan), respectively. Measurements of TSN and TSP were obtained in the same manner, except that filtered water was used. SI concentrations

PAGE 45

45 were determined spectrophotometrically by the molybdosilicate method with a correction for turbidity (APHA 2005). NH 3 and NO 3 were first converted int o NO 2 using an oxidation reaction with hypochlorite in an alkaline medium and a copperized cadmium reductor, respectively, and the three forms were then measured colorimetrically on the auto analyzer (Strickland and Parsons 1972, APHA 2005). Dissolved ino rganic nitrogen (DIN) was defined as the sum of NH 3 NO 2 and NO 3 SRP concentrations were determined spectrophotometrically with the ascorbic acid method (APHA 2005). CDOM values were measured against platinum cobalt color standards on the spectrophotom eter (APHA 2005). TURB was measured immediately from whole water that had not been frozen with a LaMotte (Chestertown, Maryland) 2020 turbidity meter against two (1 and 10 ntu) reference standards (APHA 2005). Phytoplankton Analyse s Whole (unfiltered) w ater samples were preserved on (APHA 2005) and analyzed microscopically for phytoplankton abundance and species composition (Susan Badylak, Dr. Edward J. Phlips Laboratory, University of Florida ). For general phytoplankton abund ance and composition, preserved samples were settled in 19 mm diameter chambers according to the Utermohl method (Utermohl 1958 ) Cells were identified and counted at 400x and 100x with a Leica (Wetzler, Germany) DMIL phase contrast inverted microscope. At 400x, a minimum of 100 cells of a single tax on and at least 30 grids were counted. If 100 cells were not counted within 30 grids, up to a maximum of 100 grids were counted until 100 cells of a single tax on was reached. At 100x, a total bottom count wa s completed for taxa greater than 30 m in size.

PAGE 46

46 Picoplankton abundances were determined using fluorescence microscopy (Fahnenstiel and Carrick 1992). Whole, unpreserved samples were filtered onto 0.2 m Nucleopore filters and mounted between a microsco pe slide and cover slip with immersion oil. Slides were stored frozen and analyzed using a Nikon (Tokyo, Japan) labophot 219146 research microscope equipped with a mercury short arc photo optic lamp (Hg 100 W) Counts for individual taxa were converted to total phytopla nkton biovolume concentrations ( BV ) (Table 3 1), using the closest geometric shape method (Smayda 1978). Biovolume was calculated for each species from specific phytoplankton dimensions measured for a minimum of 30 individuals. The desi Latin word was used to identify species that were comparable but not identical to known taxa. Primary Productivity Experiments Primary productivity experiments were conducted monthly during the first half of the study period (February to August 2009, excluding March). Primary productivity was indirectly determined from changes in DO concentrations over half day (approximately six hour) incubations using a simulated in situ light:dark bottle method (Wetzel and Likens 1991, Col e and Cloern 1987). The O 2 evolution method was selected over the 14 C uptake method used by Cole and Cloern (1987) so that estimates of both gross and net primary productivity (GPP and NPP, respectively) could be obtained. The O 2 evolution method has bee n more commonly used in shallow systems (Brawley et al. 2003), and it provided a safer, less complicated, less expensive, and more portable field approach than the 14 C uptake method (Wetzel and Likens 1991, K hler 1998).

PAGE 47

47 Incubations were conducted in a b lack, gel coated (polyester resin) fiberglass raceway (2.44 m long by 0.61 m wide by 0.1 m deep) that was placed on shore near each respective site. Clear polystyrene flasks were filled with approximately 300 mL of the morning water sample and incubated ( in triplicate) at a minimum of six and a maximum of eight light levels of varying percent transmittance under natural sunlight. Samples were not pre screened for bacteria or zooplankton due to the potential for altering the phytoplankton community in the process (Stickland 1960, Wetzel and Likens 1991). One group was the dark treatment in which the triplicate flasks were wrapped in black tape and aluminum foil prior to filling. A second group of triplicate flasks received full sunlight, representing ligh t conditions at the surface of the water column. The remaining groups of triplicate flasks received different light treatments to emulate the specific light transmittance depth profile for the water column at each site (Figure 3 3 ; parts A and B). Light levels were created using neutral density Plexiglas panels that were tinted to allow approximately 75, 50, 35, 30, 20, 15, and 5% light transmittance (Custom Glass Tinting, Gaine s ville, Florida). Twenty eight additional light levels were created by stacki ng two panels together. Flasks were surrounded by a constant flow of water from the estuary that was pumped through the raceway in order to maintain temperatures at in situ levels. The morning water samples collected for the primary productivity experim ents were approximately 1 to 2C cooler than their respective water column temperatures by the time they were used to fill the light and dark flasks. Temperatures of the samples contained in the flasks increased 1 to 5C over the course of the incubation period.

PAGE 48

48 DO concentrations in the flasks were measured at the beginning and end of the incubation period using a H ACH (Loveland, Colorado) HQ40D meter with a luminescent biological oxygen demand (LBOD) electrode. GPP for each treatment group was calculated from the oxygen fluxes in the dark and light flasks. GPP obtained from the half day incubation periods were multiplied by 2 to represent full light day rates. A molar mass ratio of 0.375 (grams per mole of carbon to grams per mole of oxygen) and the sta ndard photosynthetic quotient (moles of oxygen liberated to moles of carbon dioxide assimilated during photosynthesis) of 1.2 were applied to convert rates into units of carbon fixation (Wetzel and Likens 1991). GPP from each triplicate group were averaged to represent the total amount of carbon fixed at each corresponding light treatment level. The percent light transmittance received by each treatment group was then related back to a depth in the water column at each site using the Beer Lambert relations hip. The averaged GPP values were plotted against the calculated incubation depths to produce a curve of photosynthetic activity per unit volume of water at each site (Figure 3 3; part C). The curves were integrated over the water column (down to water d epth, Z w ) to yield da ily gross primary productivity ( GPP d ) in 2 1 for each site and experimental month (Table 3 1). NPP from each site and experimental month was also calculated for model comparison purposes by adjusting daily gross primary produc tivity for respiratory activity. Average respiratory activity in the triplicate dark flasks during the half day incubation periods was converted into full light day rates as was done for GPP except the standard respiratory quotient (ratio of moles of carb on dioxide liberated to moles of

PAGE 49

49 oxygen consumed during respiration) of 1.0 was used (Wetzel and Likens 1991). The resulting respiratory rates were integrated over the respective water column depths and subtracted from the corresponding GPP d values to yie ld daily net primary productivity ( NPP d ) in 2 1 for each site and experimental month (Table 3 1). The NPP d estimates obtained from this approach could be argued as more representative of net community productivity (NCP) since the respiratory acti vity of both bacteria and zooplankton was incorporated in the oxygen fluxes observed in the dark flasks. However, net productivity was still defined as NPP d with the assumption that respiratory activity was less than 10% of total productivity in the flask s and due to the fact that there was no ideal method for isolating phytoplankton separately from the other commonly associated non photosynthetic plankton (Ryther 1956a, Howarth and Michaels 2000). Furthermore, the term NCP could be perceived as taking in to account the respiratory fish, and other large nekton were obviously and purposefully excluded from the samples and incubation flasks. Primary Productivity Model Tes t A model of daily integrated primary productivity for the Caloosahatchee Estuary was tested using the approach described by Cole and Cloern (1987). The more commonly applied model that uses BZ p I 0 as the composite parameter (Cole and Cloern 1987) was rela ted to the original model form BI 0 K T 1 (Cole and Cloern 1984) using the relationship K T = 4.61 Z p 1 Applications of the 1984 and 1987 models were compared by dividing the slopes of the 1984 model applications by 4.61 or vice versa (Table 3 2). In this s tudy, GPP d obtained from the simulated in situ incubations ( N = 24) were regressed against the modeled values using the formula BZ p I 0 ; where B is CHL A

PAGE 50

50 concentration (an indicator of phytoplankton biomass), Z p is the photic depth, and I 0 is the total dayti me surface irradiance (Table 3 1). The model was also tested using NPP d obtained for each site and experimental month to determine if methodology (using gross vs. net rates) affected the overall relationship. The February 2009 data point from Site BY was removed from the net productivity model test ( N = 23) because it yielded an unexplainable negative NPP d value. Both the GPP d and NPP d models were tested with and without a correction for photic depth based on corresponding water column depths to determin e if this relationship affected the potential for phytoplankton productivity and, thus, the strength of the model. I n cases where the photic depth ( Z p ) e xceeded the water column depth ( Z w ), the value of Z p was cut off and made equal to Z w so that the mode l estimated phytoplankton productivity based on the actual volume of water present and the quantity of light truly available. Lastly, each of the se four model scenarios (i.e., combination of two different productivity methods and two different photic dept h approaches) were tested using CHL A concentrations that were either uncorrected or corrected for pheophytin to determine if pigment degradation products affected the overall relationship. Model parameters were determined with least squares regression a nalysis in SAS Version 9.2 (SAS Institute Inc. 2009). A p value of 0.05 was used to determine the significance of model parameters. Model strengths were evaluated using r 2 correlation coefficients of determination with values closer to 1 in (in terms of its statistical significance and strength) (in terms of its appropriateness and usefulness) model for the Caloosahatchee Estuary was selected so that the resulting adapted relationship provided the simplest, most practical

PAGE 51

51 approach to predicting phytoplankton productivity using measurements of biomass and light availability that are routinely collected (Cole and Cloern 1987). The residuals (difference in measured and predicted productivity values) BZ p I 0 model adapted to the Caloosahatchee Estuary were analyzed to identify factors contributing to the deviations from the model (Cole and Cloern 1987). Primary Productivity Model Application BZ p I 0 model adapted to the Caloosahatchee Estuary was used to estimate phytoplankton productivity for the year long study period (February 2009 to February 2010, excluding March 2009). The independent composite parameter ( BZ p I 0 ) was calculated for each month using in situ field measurements of phytoplankton biomass as CHL A ( B ) and photic depth ( Z p ) from monthly sampling events and monthly maximum values of I 0 for the area (see previous method explanations). A verage estimates of annual gross and net phy toplankton productivity GPP y and NPP y 2 1 ), respectively, were determined from modeled daily rates for each site and across all sites for sy stem comparison purposes Results Physical Chemical Conditions Daily rainfall and flow at the Franklin L ock and Dam (S 79) exhibited similar seasonal patterns during the study period (Figure 3 4). Rainfall and flow were low between February and the middle of May 2009, creating a dry period. Both rainfall and flow increased from the second half of May throu gh the beginning of October 2009, co rresponding to the normal wet period for this area (Scarlatos 1988, Chen and Gerber 1990, Stoker 1992, Chamberlain and Doering 1998 a ). From the middle of October

PAGE 52

52 2009 to the following February 2010, rainfall and flow de clined. The second dry period had somewhat higher rainfall and flow levels than the first dry period. S alinity (S AL ) ranged from near 0 to approximately 35 psu across the four sites over the study period (Figure 3 5). San Carlos Bay (BY) had the highes t and leas t variable salinities. SAL levels decreased with distance from the mouth of the estuary. SAL was greatest at all sites during the dry period and lowest at all sites during the wet period, reflecting the influence of seasonal freshwater inputs t hroughout the estuary. The largest differences between surface and bottom SAL were observed at the Upper (UE) and Middle (ME) Estuary sites in June 2009 and at the Lower Estuary (LE) site in September and October 2009. Water temperature ( TE MP_W ) showed li ttle spatial variati on and was uniform throughout the water column. However, TEMP_W gradually increased throughout the study period from a low of 18C in February 2009 to a high of 32C in August 2009 (Figure 3 6 ). Following the summer peak, TEMP_W dec reased in the fall and early winter to levels that resembled those observed in the previous spring and late winter, respectively. TEMP_W in January and February 2010 dropped to a low of 14C, well below what was seen one year prior, corresponding to recor d cold air temperatures from January to March 2010 (NCDC 2010). Total da ytime surface irradiance ( I 0 ) (i.e., PAR light flux) measured on the days of the primary productivity experiments, were higher in February, April, and May than in June, July, and Aug ust 2009 (Figures 3 7 3 8 3 9 and 3 10; part A). The decrease in PAR light flux during the summer months coincided with greater cloud cover observed during the primary productivity experiments, even though monthly maximum I 0 for the

PAGE 53

53 region ranged from 2 1 2 1 in the summer (i.e., in the absence of cloud cover) (Figure 3 11; part A). Light penetration through the water column, expressed as vertical light attenuation coefficients ( K T ) varied by site and season du ring the study period (Figure 3 12; part A). In general, light transmission through the water column was greater during the dry period than the wet period. Light attenuation increased as freshwater inflow increased, particularly at the sites in the inner estuary (UE, ME, and LE). Variability in K T was linked to changes in the concentrations of CDOM, TURB, and CHL A at each of the four sites (Figure 3 12). Light attenuation at site UE was primarily coupled with CDOM, particularly in July, August and December 2009, while CHL A and TURB played a role in February 2009 and 2010, respectively. The increase in K T at site ME in the wet summer months corresponded to an increase in TURB and CHL A in June 2009, CDOM in July 2009, CHL A in August 2009, an d TURB in October 2009. At site LE, changes in K T closely matched changes in TURB and CHL A (July 2009) with CDOM having a lesser influence in August and December 2009. The less pronounced increase in K T at site BY in August 2009 was primarily attributab le to minor increases in CDOM and CHL A Photic depth ( Z p ; depth at 1% PAR light flux) varied by date and site (Figure s 3 7 3 8 3 9 3 10 and 3 11 ; part B), while water depth ( Z w ) was consistently between 2 and 3 m at all four sites throughout the stu dy period. Z p was shallower than Z w at site UE throughout much of the study period and at sites ME and LE during the wet summer months. At site BY, Z p exceeded Z w throughout the study period with the greatest difference observed duri ng the dry months.

PAGE 54

54 C oncentrations of TSP, TP, TSN, TN, SRP, DIN, and S I were generally higher in the wet period than the dry period (Figures 3 13 3 14 and 3 15), coinciding with increased freshwater inflows from the watershed. Spatially, the peaks in nutrient concentration s appeared first and most prominently at sites UE and ME in the inner estuary. Substantial increases in concentrations were also observed at site LE, and to a lesser extent at site BY, but a month or more later. Seasonal patterns were less apparent at si te BY, which is furthest from the headwaters of the estuary and most strongly influenced by tidal water exchange. Ratios of TN to TP concentrations fluctuated around the Redfield molecular mass ratio of 7 (mostly in the range of 5 to 9) throughout the study period at all four sites (calculated ratios not shown). By comparison, ratios of DIN to SRP concentrations were consistently below the Redfield mole cular mass ratio of 7 (Figure 3 14). SRP concentrations from each site and m onth were greater than L 1 which is five times the minimum phosphorous requirement suggested for phytoplankton growth (Reynolds 2006). DIN concentrations were comparatively less than expected based on Redfield stoichiometr L 1 the nitr ogen concentration needed for sufficient phytoplankton growth in coastal waters (Reynolds 2006), were observed in February 2009 (site LE), April 2009 (site BY), May 2009 (sites UE, LE, and BY), June 2009 (site ME), and February 2010 (site BY). SI concentr ations from each site an L 1 the minimum requirement suggested for phytoplankton growth, including diatoms (Reynolds 2006). However, relatively low levels of SI were observed at site BY in June 2009, site ME in November 200 9, and site BY in January 2010 when they fell wi L 1 (Figure 3 15).

PAGE 55

55 The phytoplankton biomass potential (amount of expected CHL A ) based on observed concentrations of SRP, DIN, and SI (according to Redfield stoichiometry) a lso showed that SRP and SI are relatively more abundant than DIN in terms of the needs of phytoplankton gr owth at all four sites (Table 3 3). During most dry months (February 2009 to May 2009 and November 2009 to February 2010), the potential for addition al biomass represented by the amount of availabl e DIN was low. In some of the wet months (July to October 2009), the biomass potential was high based on DIN levels in the water column, however, the observed CHL A remained low. Phytoplankton Abundance and Composition Phytoplankton biomass ( B ) in terms of uncorrected CHL A and total phytoplankton biovolume ( BV ) concentrations were both considered as measurements of phytoplankton abundance in this study. All four sites experienced their largest B peaks in the summer (June, July, and August 2009) (Figures 3 7 3 8 3 9 3 10 and 3 11; part C). Additional peaks in B occurred at site UE in the winter (February 2009, January 2010, and February 2010) and spring (May 2009) and at site ME in the winter (January 2010). BV followed similar trends as B (Figure 3 16), although the relative peak heights differed somewhat. Overall, approximately 68% of the variability in BV was explained by the variability in B regardless if CHL A concentrations were corrected or not for pheophytin pigments (relationship not shown). The major peaks in BV were associated with dinoflagellates at sites UE and ME but with diatoms at sites LE and BY (Figures 3 16). At site UE, the dinoflagellate Akashiwo sanguinea contributed between 65 and 95% of the total phytoplankton biovolume during the five main peaks in BV (Figures 3 16 and 3 1 7). A. sanguinea contributed 95% of BV during the major peak at site ME in June 2009. Another

PAGE 56

56 dinoflagellate, Polykrikos schwartzi was responsible for the January 2010 peak at tha t site (Figures 3 16 and 3 17). At site LE, the diatom Skeletonema cf. costatum was the dominant phytoplankton species during the BV peak in July 2009, contributing up 86% of the total phytoplankton biovolume (Figures 3 16 and 3 1 7). S cf. costatum was also present at site BY, but the relatively small BV peaks in August and September 2009 were attributed to relatively large centr ic diatoms (60 to 140 m), i.e., Coscinodiscus cf. and Rhizosolenia setigera respectively. Cyanobacte ria (blue chlorophytes, euglenoids, cryptophytes, Chattonella (raphidophytes), and small flagellates, played a significant but lesser role in terms of phytoplankton biovolume in the Caloosah atchee Estuary ( Figures 3 16 and 3 17). Cyanobacteria (blue green algae) were more abundant at sites LE and BY than at sites UE and ME, while the 3 17). The at sites UE and ME during the wet summer months was attributable to small flagellates in July 2009 and to both chlorophytes and euglenoids in August 2009 (Figure 3 17). Measured Primary Productivity Measured da ily gross primary productivity ( GPP d ) fol lowed the patterns seen in phytoplankton biomass ( B ) in terms of uncorrected CHL A concentrations (Figures 3 7 3 8 3 9 and 3 10; part D ). B alone explained approximately 80% of the variability in GPP d regardless of whether or not the CHL A concentratio ns were corrected for pheophytin pigments (relationship not shown). Gross productivity estimates from the simulated in situ 2 1 (Table 3 4). The median measured productivity values increased from site BY to the inner estuary.

PAGE 57

57 Despite the substantial range of productivity estimates observed during the six month experimental period, all sites exhibited one or more dates when measured GPP d exceeded 1500 2 1 (Table 3 4). Primary Productivity Model Fit Th e linear regression relationships between measured productivity (using the simulated in situ incubation experiments) and predicted productivity (using the BZ p I 0 model parameter) were significant for all possible model scenarios (using either GPP d or NPP d v alues, calculated or corrected Z p values, and uncorrected or corrected CHL A concentrations for B ) (Table 3 5). Between 78 and 85% of the variability in daily productivity was explained by the variability in the composite parameter BZ p I 0 depending on whic h productivity method was used, whether or not the photic depth values were corrected for water depths, and whether or not the CHL A concentrations were corrected for pheophytin, as indicated by the strength of the coefficients of determination ( r 2 ). In general, the model relat having the largest r 2 values) when (1) GPP d values were used instead of NPP d values and (2) uncorrected CHL A was used instead of corrected CHL A as an indicator of biomass ( B ) Between the remainin g two model scenarios ( GPP d values for the productivity term and uncorrected CHL A concentrations for the biomass term), the r 2 values were almost equal (0.85 and 0.84) and the slopes were very similar (0.70 to 0.73) whether photic depths were corrected fo r water column depths or not, respectively. However, the use of corrected Z p values greatly increased (by a factor of 3.4) the model intercept, shifting the line further from the origin. Without the Z p correction, which was the approach presumably taken by Cole and Cloern (1987), the model relationship based on

PAGE 58

58 measured GPP d and uncorrected CHL A concentrations yielded an intercept not significantly different from zero ( p value = 0.4849) (Figure 3 18). ationship, GPP d = 75 + 0.73 BZ p I 0 differences between measured and predicted primary productivity values varied by site and season (Figure 3 19). In general, the residuals were positive from February through May 2009 and negative in July and August 2009, which indicated that the model over and underestimated primary productivity during these respective time periods. June 2009 represented a transition month in which the model produced significant over and underestimates of primary productivity at differe nt sites in the estuary. There was only one case (July 2009, site ME) when the difference between the measured and predicted productivity values was less than detection limit of the 2 (HACH 2006). Most of the resi duals were within 35 0 2 1 except for five cases (April 2009, site LE; May 2009, site BY; June 2009, site UE; July 2009, site LE; July 2009, site BY) in which the absolute values of the differences in the measured and predicted values were betwe en 523 and 724 2 1 (Figure 3 19). The measu reme nts of primary productivity obtained in these five experiments were well outside the range (higher or lower) of average estimates bounded by the 95% confidence interval of the mean (Figure 3 18). Mod eled Primary Productivity GPP d = 75 + 0.73 BZ p I 0 ) adapted to the Caloosahatchee Estuary to the year long data set of B Z p and I 0 2 1 which is a h igher and wider range of values in comparison to those obtained in the simulated in situ experiments (Table 3 4). Under modeled conditions, GPP d exceeded

PAGE 59

59 1800 2 1 at each site at least once over the course of the year, which is a higher mark of t simulated in situ experiments. The largest peaks in modeled GPP d occurred at all four sites during the summer (June, July, and August 2009) with additional peaks appearing in the win ter (February 2009, January 2010, and February 2010) at sites UE and ME (Figure 3 11 ; part D), which is the same pattern set forth in the six month experimental period. Over the course of the year, peaks in modeled GPP d corresponded to periods of increase d water temperatures, PAR influx, freshwater inflow, and/or nutrient loading. Annualized rates of modeled GPP d yielded an average annual gross phytoplankton productivity estimate ( GPPy ) 2 1 across all four sites in the Caloosahatchee Estuary for the year long study period. Since estuarine and other ecosystem comparisons are commonly based on net productivity estimates in literature, the alternative model relationship NPP d = 102 + 0.63 BZ p I 0 was also applied and yielded an average annual net phyt oplankton productivity es timate ( NP P y ) of 312 2 1 for the Caloosahatchee Estuary. Spatially, annualized mean GPP y increased from 268 2 1 (245 2 1 NPP y ) at site BY to 424 2 1 (379 2 1 NPP y ) at site UE (Table 3 6) which followed the spatial trend in bioavailable and total nutrient concentrations in the Caloosahatchee Estuary. Discussion Modeling Phytoplankton Produc t ivity The BZ p I 0 model developed by Cole and Cloern (1987) for estimating daily phytoplankton productivity yielded a strong linear relationship with simulated in situ measures of productivity in the Caloosahatchee Estuary, explaining most of the variability (84%) in daily gross productivity. This relationship is one of the

PAGE 60

60 strongest outcomes among the various efforts to apply the model to estuaries in North America and across the globe (Table 3 2). The model has been applied in a variety of estuarine types, including lagoons (South San Francisco Bay), fjords (Puget Sound), and river dominated systems (North San Francisco Bay, Narragansett Bay, Neuse River Estuary, Escambia Bay, and Tokyo Bay) (Table 3 2). While most of the applications of the model have been in temperate estuaries, the results of this study and an earlier study in Es cambia Bay, Florida (Murrell et al. 2007) indicate that the model also fits some warmer water systems. The Caloosahatchee Estuary application extends the usefulness of the BZ p I 0 model to distinct low l atitude subtropical systems, particularly those having higher annual water temperatures, higher rainfall and river flow in the summer and fall, and greater anthropogenic influence in the system. The current application of the model to the Caloosahatchee Estuary was also unique in using the O 2 evolution method to estimate both gross and net productivity, instead of the 14 C uptake method, which generally yields net productivity values, particularly in the case of extended incubation periods (Peterson 1980, Wetzel and Likens 1991). In addition to providing more information (both gross and net productivity), the O 2 evolution method was both easier and cheaper to perform than the more commonly used 14 C method, avoiding the need for hazardous chemicals and expensive equip ment (Wetzel and Likens 1991, K hler 1998). To validate the comparability between this and the other model applications, the Caloosahatchee Estuary model relationships were also examined using NPP d and the overall significance of the models did not change. The NPP and GPP based model relationship s were similar in strength, but the NPP d models had somewhat lower slopes

PAGE 61

61 than the GPP d models, while the intercepts were not affected in terms of their significance. Keller (1988) conducted a similar analysis using pooled data ( N =1010) to compare NPP an d GPP based models for Narragansett Bay, Rhode Island. The slopes of the NPP and GPP models were 0.63 and 0.70, respectively (Keller 1988), similar to the slopes observed in this study. The strength and comparability of the GPP d and NPP d Caloosahatchee E stuary models suggest that either GPP or NPP data can yield reasonable estimates of phytoplankton productivity. This study was also the first to examine some previously unaddressed complexities involved in defining the variables in the model predictor, BZ p I 0 (Cole and (Cole and Cloern 1984, 1987), data from different sources used both uncorrected and corrected CHL A concentrations as a proxy for phytoplankton biomass ( B ) In later model applications, only one study specified that corrected CHL A concentrations were used (Boyer et al. 1993), while the remaining studies gave no indication either way (Keller 1988, Murrell et al. 2007, Bouman et al. 2010). To determine i f the BZ p I 0 model was affected by the type of CHL A values used, relationships were tested for both uncorrected CHL A concentrations and values corrected for pheophytin. Although the significance of the models did not change, the coefficients of determina tion ( r 2 ) decreased slightly. Using corrected CHL A concentrations for B increased the slopes and intercepts slightly, but the increase was likely not enough to change the interpretation of the model. Since system monitoring projects often analyze and re port one of the two forms of CHL A (uncorrected or corrected for pheophytin), it is helpful

PAGE 62

62 knowing that either could be used as an indicator of phytoplankton biomass in the Cole and Cloern (1987) model without changing the overall viability of the model. Another issue examined in this study was the definition of photic depth ( Z p ) used in the model. The BZ p I 0 model assumes that Z p is completely contained withi n the total water column depth ( Z w ) because the relationship was derived and tested in relatively deep estuarine systems, which is probably why Cole and Cloern (1984, 1987) did not provide instructions on how to handle situations when this is not the case. The relative shallowness of the Caloosahatchee Estuary and t ypically low light attenuation ( K T ) during the dry months could explain why Z p was g reater than Z w at site BY in San Carlos Bay throughout the experimental period and at sites UE, ME, and LE in the inner estuary between February and May 2009. As a result, predicted productivity (using the B Z p I 0 composite parameter) was generally higher than measured productivity (using the simulated in situ incubation experiments) during these cases. Brush et al. (2002) and Boyer et al. (1993) note that the BZ p I 0 model may overestimate productivity when Z p exceeds Z w because the model assumes that the phytoplankton have access to all of the available light and, thus, that all of the photic depth was used in carbon fixation. To investigate the effect on the Caloosahatchee Estuary primary productivity relatio nships, models were run separately with Z p values cut off at Z w The resulting relationships had slightly lower, yet, comparable slopes and substantially higher intercepts that became significantly different from zero. By shifting the intercept approxima tely 2 to 3 times further from the origin, productivity estimates based on the new model predictor would be higher than those obtained using the uncorrected model predictor, assuming B and I 0 are held constant. Since the natural and more logical productiv ity model centers closer

PAGE 63

63 to the origin (i.e., zero phytoplankton biomass and/or zero light availability theoretically Z p correction was not used for the BZ p I 0 model applic ation in the Caloosahatchee Estuary, but alternative approaches should be considered in future model t ests. For example, Brush and Brawley (2009) suggested that simply substituting Z w for Z p when Z p is greater than Z w is not a valid approach because the r elationships between water column depth and irradiance and between irradiance and photosynthesis are non linear. Brawley et al. (2003) proposed a correction factor to adapt the BZ p I 0 model to shallow systems which took into account these non linear relati onships, however, it did not improve the estimates of productivity in the system in which it was tested (Waquoit Bay, Massachusetts). The authors attributed the difficulties in validating the correction factors to the limited availability of 14 C uptake da ta, variable productivity estimates, problematic estimation of light attenuation, and complex bathymetry in shallow systems (Br awley et al. 200 3 ). An approach that ultimately provided good agreement between measured and predicted productivity involved the use of ten different depth correction polynomials that were developed for varying irradiance levels (Brush and Brawley 2009). The use of these polynomials went beyond the level of analysis and interpretation desired for a simple empirical model for robus t productivity estimation in the Caloosahatchee Estuary, but it should be explored in other systems as deemed necessary. Interpretation of Model Residuals Residual outputs from model applications provide opportunities to examine factors that may be importa nt in defining the causes of variability in the relationship between independent and dependent model variables. In this application of the

PAGE 64

64 simulated in situ incubation experiments) and predicted (using the BZ p I 0 composite parameter) productivity values. The model residuals can therefore help identify factors other than CHL A and PAR availability that may affect phytoplankton productivity, aside from the inherent varianc e associated with experimental or analytical error. Cole and Cloern (1987) suggest that deviations of measured productivity from model predicted productivity are related to variability in eco physiological factors that affect photosynthetic capacity, such as nutrient availability, the quality of the underwater light field, and the presence of toxic compounds in the water column. In addition, such deviations can reflect changes in phytoplankton community structure, since genotypic differences in photosynthe tic mechanisms and growth strategies impact the responses of phytoplankton to changes in light availability. In more general terms, deviations from model predictions can be viewed as a reflection of differences in photosynthetic efficiency (Welschmeyer an d Lorenzen 1981, Geider and Osborne 1992, Geider and MacIntyre 2002). Since photosynthetic efficiency essentially represents the productivity potential of a given phytoplankton standing stock, with respect to the available light in the water column, proce sses contributing to phytoplankton biomass gains and losses, such as growth, allochthonous inputs, flushing, and grazing, have also been considered for their indirect influence on model deviations (Brush et al. 2002). assumes th at all phytoplankton biomass ( B ) expressed as CHL A concentration, is equally capable and efficient at capturing and using PAR for photosynthesis. The model also assumes that photic depth ( Z p ) and total PAR light flux ( I 0 ) accurately depict the availabili ty of photosynthetically usable radiation (PUR) in the

PAGE 65

65 water column (Kirk 1994, Falkowski 1994). However, CHL A does not necessarily provide an exact assessment of photosynthetic potential (Welschmeyer and Lorenzen 1981), and all PAR is not equally effici ent in driving photosynthesis (Kirk 1994). To summarize, phytoplankton may be more or less photosynthetically efficient due to (1) environmental changes that enhance or diminish their ability to use light for photosynthesis (e.g., changes in nutrient avai lability), (2) genotypic differences in community structure and function that alter the efficiency with which available light is used for photosynthesis (e .g., species engaging in alternative nutritional strategies), and (3) variability in underwater light quality that alter the level of PUR. An understanding of these factors is important in interpreting the output of primary productivity models (Geider and MacIntyre 2002). Residual GPP d model values followed a seasonal pattern in which measured productivi ty values were typically lower than predicted in the winter and spring (February, April, and May 2009) but higher than predicted in the summer (July and August 2009). This shift between model over and underestimates of productivity roughly aligne d with t he transition from the dry period to the wet period, with the exception of June 2009. June 2009 was an anomalous month due to strong salinity stratification in the water column. The mixing of near freshwater with saline water could account for the over e stimation of productivity by the model at site UE, relative to the in situ experiment, due to osmotic stress. The seasonal switch between the model over and underestimating primary productivity may be viewed from two main perspectives; (1) changes in en vironmental conditions, most notably increased nutrient levels, altered underwater light quality, and

PAGE 66

66 shifting salinity regimes associated with the onset of elevated rainfall and river flow in June 2009 and (2) a shift in phytoplankton composition from bio mass peaks dominated by dinoflagellates, mainly A. sanguinea at site UE and ME during the dry period and in the month of June, to peaks dominated by diatoms, such as S. cf. costatum and other centric diatoms, at sites LE and BY in July and August 2009. T hese spatial and temporal trends are, of course, inter dependent since environmental changes are primary driving factors in species succession. Model overestimation Two potential reasons for model overestimates in the dry period are the presence of relativ ely low levels of dissolved inorganic nitrogen (DIN) in the water column and the dominance of biomass peaks by A. sanguinea Nitrogen limitation, as indicated in this study by ratios of DIN:SRP below 7, can directly impact photosynthetic capacity by limit ing important cell functions, such as protein synthesis (Falkowski 1994, Bergmann et al. 2002, Geider and MacIntyre 2002). Low levels of DIN also influence photosynthetic efficiency indirectly by favoring species capable of acquiring nitrogen nutrition us ing alternative strategies, such as mixotrophy or behavioral approac hes to acquiring nitrogen, e.g., vertical migration (Smayda 1997). The dinoflagellate A sanguinea exhibits both of these capabilities (Gaines and Elbrachter 1987, Levandowsky and Kaneta 1987, Taylor 1987, Smayda 1997, Burkholder et al. 2008). Mixotrophy has been observed in many dinoflagellates, which are capable of growth based on photo autotrophic and heterotrophic modes of nutrition (Gaines and Elbrachter 1987, Smayda 1997, Burkholde r et al. 2008). A. sanguinea can incorporate dissolved organic substances, such as lysine, alanine, leucine, and phenylalanine, which are building blocks of proteins, through osmotrophy (Gaines and Elbrachter

PAGE 67

67 1987). A. sanguinea can also ingest cyanobact eria and microfaunal prey through phagotrophy (Burkholder et al. 2008). In addition, a vertical nutrient retrieval strategy is commonly used by dinoflagellates to access nutrient enriched layers in the water column (Smayda 1997). Populations of nutrient deficient A. sanguinea have been found to accumulate near the depth at which irradiance saturates photosynthesis during the daytime and then migrate below the nitracline at night, allowing for dark uptake of nitrate (Smayda 1997). Due to the relatively sh allow and generally polymictic character of the Caloosahatchee Estuary, A. sanguinea may migrate to the bottom where it could access DIN diffusing from the sediment surface (Day et al. 1989). The prominence of A. sanguinea during the dry period may also have been aided by relatively low light attenuation and long water residence times compared to the wet period. Light availability in the water column was greatest at all sites between February and May 2009, as indicated by lower values of K T and estimates of Z p being approximately equal to or greater than Z w Increased light availability during this time period coincided with reduced freshwater inflow and CDOM input, supporting phytoplankton productivity through the water column across the four sites. Wa ter residence times, defined here as the time it takes for a total mass of water to be reduced to e 1 or 37% of the initial mass, in the upper and middle regions of the Cal oosahatchee Estuary estimated at 6 to 11 days during the dry period comp ared to 1 to 4 days during the wet period due to changes in flow from S 79 ( Qiu et al. 2007 ). Since dinoflagellates typically have relatively low maximum growth rates, i.e., near or below one doubling per day (Stolte and Garcs 2008) (1.13 d 1 for A. sanguinea ; Matsu bara et al. 2007), water residence time is an important factor in their bloom

PAGE 68

68 dynamics. Decreased flushing rates would allow a slower growing species such as A. sanguinea to maintain its population and make use of the increased light availability for phot osynthesis or focus on other nutritional strategies if nutrients are limited, as they were in this case. Although not directly related to photosynthetic efficiency, m odel overestimation during the dry period could also be linked to the effect of zooplankto n grazing in the productivity experiment flasks. One of the greatest challenges associated with the use of incubation techniques is the difference between the natural ecosystem and the portion of the system that is contained in the flasks (Howarth and Mic haels 2000). Since the morning water samples used in the productivity experiments were n ot pre screened for zooplankton, to avoid damaging or altering the phytoplankton community, there was possibly a predator prey situation created in the flasks that dif fered from that in the water column. Zooplankton densities in the Caloosahatchee Estuary have been found to be highest at distances further from S 79, particularly in regions having salinities above 20 psu (Chamberlain et al. 2003). In this study, salini ties were close to or greater than 20 psu at sites ME and LE during the dry months and at site BY throughout the experimental period. Whole (unfiltered) water samples taken from these sites during these times and enclosed in the productivity flasks could have had more intense zooplankton grazing of phytoplankton due to the confined space and potentially unbalanced predator to prey ratio than in the water column. Since the model predictions were based on CHL A samples retained prior to the start of the inc ubation period, any decreases in phytoplankton biomass in the flasks due to zooplankton grazing would likely yield lower rates of productivity than the model expected. This

PAGE 69

69 likely affect the model i n the high freshwater inflows and drops in salinity (Chamberlain et al. 2003). Model underestimation In the wet period, the model most likely underestimated productivity d ue to an increase in ma cronutrient concentrations (i.e., N, P, and SI) and the shift in phytoplankton community composition towards dominance by the diatom S. cf. costatum both of which may have contributed to an increase in photosyn thetic efficiency. Du ring the wet summer months (June to August 2009), concentrations of total and inorganic nutrients increased throughout the estuary, following increased freshwater inflows from the watershed. Concentrations of DIN, which typically limit productivity in the Caloosahatchee Estuary (McPherson and Miller 1990, Doering et al. 2006, Heil et al. 2007), were introduced at levels well above the minimum requirement needed for sufficient nitrogen uptake (Reynolds 2006). SI concentrations, which were potentially limit in g to diatoms at the end of the dry period, were also added to the system at levels in excess of the suggested requirement for phytoplankton growth (Reynolds 2006). S. cf. costatum and other diatoms likely exhibited enhanced productivity following this n utrient stimulation, which is the classic response in nutrient The trade off to nutrient loading during the wet period was the decrease in light availability in the water column, which was linked to incr eased CDOM input. Phytoplankton can adapt to changing light availability at the cellular level by altering their pigment content, photosynthetic capacity, chemical composition, cell volume, and other photosynthetic responses (Falkowski and Owens 1980, Mal lin and Pearl 1992).

PAGE 70

70 These adaptations allow phytoplankton to modify their photosynthetic processes and efficiencies, given light levels that range from limiting to inhibitory (Gallagher et al. 1984, Bergmann et al. 2002). Light harvesting pigments in th e photosynthetic apparatus capture and transfer light energy to the photo chemical reaction centers where it is coupled to an electrochemical gradient (Falkowski and Owens 1980, Gallagher et al. 1984). Chlorophyll a which is most effective at absorbing b lue and red light, is the major light harvesting pigment in photosynthesis and is therefore used as both a measure of phytoplankton abundance and light absorption capability. In the ocean, red light is attenuated more rapidly than blue and green light (Ki rk 1994, Falkowski 1994), accentuating the importance of blue light with depth in terms of absorption in the water column. In inland and coastal waters, the presence of CDOM, which absorbs strongly in the blue range, further restricts light availability f or chlorophyll a (Kirk 1994). To fill the gaps between the total intensity of photosynthetically active radiation (PAR) and the portion of that which is photosynthetically usable radiation (PUR), phytoplankton may alter their pigment content to optimize the ability of the cell to harvest the available light (Falkowski and Owens 1980). Diatoms have been found to increase the concentration of chlorophyll a and fucoxanthin, an accessory carotenoid pigment, as a photoadaptive response to low light intensitie s (Falkowski and Owens 1980, Perry et al. 1981, Gallagher et al. 1984). Fucoxanthin allows increased of the available PAR in colored coastal waters (Yentsch 1980, Fal kowski 1994, Kirk 1994). To accommodate these pigment changes, diatoms, including S. cf. costatum have shown to increase the size and not the number of its photosynthetic units (PSUs)

PAGE 71

71 in order to maximize its photosynthetic efficiency in response to low light (Falkowski and Owens 1980, Perry et al. 1981, Gallagher et al. 1984, Smayda 1997, Reynolds 2006). The success of S. cf. costatum and other diatoms during the wet period is further enhanced by their resistance to increased freshwater inflows and dec reased water residence times, which likely dropped to less than 4 days based on S 79 flow rates (Qiu et al. 2007). S hortened residence times could indirectly affect the production potential, or photosynthetic efficiency, of phytoplankton by transporting b iomass downstream and out of the estuary (Day et al. 1989). The ability of S. cf. costatum and other diatoms to maintain their populations in these conditions is related to their superior growth rates. Unlike dinoflagellates, diatoms such as S. cf. costa tum are able to replenish their populations more quickly at rates that are often greater than two doublings per day (Stolte and Garcs 2008), allowing them to sustain relatively high levels of biomass and productivity during a time of increased flushing an d in an area of the estuary prone to tidal mixing. It should be noted that the observed peaks in biomass and productivity could also be attributed to populations of diatoms that are not purely autochthonous in their origin, possibly emanating from adjoini ng waters of the Gulf of Mexico or Charlotte Harbor. The dominant species of phytoplankton constituting the peaks during this time, S. cf. costatum is a common bloom forming diatom in the coastal waters of southwest Florida (Saunders et al. 1967). Other considerations The explanation for the seasonal pattern identified in the model residuals from this study has alluded to the role that nutrient availability can play in controlling phytoplankton productivity, despite the absence of a nutrient term in the r egression relationships (Brush et al. 2002). The original development and success of the BZ p I 0

PAGE 72

72 model was based on the fact that most estuaries are turbid and nutrient rich, making light availability the most important factor in controlling biomass specifi c productivity (Cole and Cloern 1987). In systems where this is not the case, one would need to adjust the potential production predicted by the model from measurements of biomass and light availability to a level of production that there are nutrients to support (Brush et al. 2002). Since no such adjustments were considered, it could be argued that one of the primary assumptions of the BZ p I 0 model was violated in this study due to the potential for nutrient limitation in the Caloosahatchee Es tuary, parti cularly during the dry period when nitrogen concentrations are proportionally less than phosphorous concentrations. Ultimately, the overall model selected for the Caloosahatchee Estuary ite being based on experimental data collected across two periods ( dry and wet) having distinct freshwater and nutrient inputs. Had the mo del been tested using just the wet period data ( N = 12), when nutrients were likely not limiting, the resulting model relationship GPP d = 280 + 0.70 BZ p I 0 ( r 2 = 0.88) would have yielded even higher overestima tes of productivity during the dry period in comparison to deviations derived from the overall model, reaffirming the influence of nutrient availability on the accur acy of the BZ p I 0 model. Comparative Rates of Producti on 2 1 ) obtained from the six month experimental period and the twelve month application of BZ p I 0 model is similar to the ranges of productivity that have been observed in other river dominated estuaries in North America, such as the Hudson River in New York and Narragansett Bay in Rhode Island (Boynton et al. 1982). These measurements and predictions vali date the appropriateness of the methodological

PAGE 73

73 techniques used in this study and the usefulness of the BZ p I 0 model in providing reasonable estimates of estuarine phytoplankton productivity in the Caloosahatchee Estuary. The peak rates of modeled GPP d exce 2 1 at all four sites demonstrate the productivity potential of the phytoplankton community in a given year. These levels of productivity compare to temperate coastal waters on the low end of the range (Ryther 1963, Boynton et al. 1982) a nd sea grass and mangrove communities on the high end of the range (Day et al. 1989, Valiela 1995). The average annual phytoplankton productivity estimate across all four sites in 2 1 GPP y 2 1 NPP y ) is greater 2 1 mean value for upwelling areas reported by Ryther (1963), reflecting the high overall productivity of the Caloosahatchee Estuary. The estimated annual rate of phytoplankton productivity in the Caloosahatchee Estuary is comparable to a number of highly productive river dominated estuaries in the southeastern United States, including those connected to the Gulf of Mexico (Table 3 6). The Caloosahatchee Estuary, and the majority of the river dominated estuaries of the Gulf of Mexico, lie between 26 and 31N latitudes, and are therefore subject to relatively high annual light flux and water temperatures (Kirk 1994, Odum et al. 1998, Pennock et al. 1999, Murrell et al. 2007), helping to support high annual productivity level s year round (Day et al. 1989). The Gulf of Mexico estuaries, including the Caloosahatchee, are also ch aracteristically microtidal and shallow, in which photic depths are frequently deeper than bottom depths, further enhancing production potential. The Ca loosahatchee Estuary, like other river dominated systems, is characterized by considerable spatial and temporal variability in phytoplankton biomass

PAGE 74

74 and productivity. The variability is in part attributable to fluctuations in salinity regimes, nutrient di stributions, light availability, and water residence times, which are driven by seasonal variability in rainfall and freshwater flow from the watershed (Boynton et al. 1982, Day et al. 1989, Pennock et al. 1999). The shallow nature of the Caloosahatchee E stuary accentuates this variability, resulting in short water residence times and sharp drops in salinity during periods of high freshwater discharge, restricting the potential for the accumulation of phytoplankton biomass (Monbet 1992, Pennock et al. 1999 ), despite the associated increase in nutrient loads from the watershed. In this study, phytoplankton biomass and productivity followed a spatial gradient and seasonal pattern that primarily corresponded to the loading and distribution of bioavailable and total nutrient concentrations. However, peaks in phytoplankton biomass and productivity were not always observed when the potential for added growth was high with respect to concentrations of DIN, SRP, and SI in the water column. Spatially, nutrient conc entrations were typically higher at the upper and middle estuary sites (UE and ME) than in the lower estuary and bay sites (LE and BY), reflecting the impact of nutrient inputs from the upstream watershed and the role of tidal mixing. Nitrogen was identif ied as the most likely limiting nutrient for phytoplankton growth, confirming the hypothesis of previous studies in the Caloosahatchee Estuary (McPherson and Miller 1990, Doering et al. 2006, Heil et al. 2007), because of the low Redfield ratios of DIN to SRP throughout the study period and the low biomass potential represented by low concentratio ns of DIN, particularly in the dry months. In comparison, proportionately higher concentrations of phosphorous were likely supported throughout the study across t he four sites by the natural geologic and

PAGE 75

75 hydrologic processes occurring in the nearby and connected Charlotte Harbor watershed, which is rich in phosphate deposits and home to an extensive mining operation (McPherson and Miller 1990). Peaks in phytoplank ton bioma ss and productivity during the dry p eriod and the beginning of the wet period, including the winter, at the upper and middle estuary sites (UE and ME) were likely supported by small pulses of nutrient enriched water from the watershed and the pres ence of flocculent muddy sediments, which could serve as an internal source of nutrients for phytoplankton growth (Day et al. 1989). With the onset of the wet period, elevated freshwater flow from the watershed and the corresponding increase in nutrient c oncentrations would have supported primary productivity throughout the estuary and the bay. However, elevated nutrient concentrations at sites UE and ME in July, August, September, and October 2009 did not correspond with high phytoplankton biomass or pro ductivity, most likely due to strong declines in salinity, light availability, and/or water residence times. With salinities at or below 5 psu, the cells of A. sanguinea the dominant species at these sites, would have likely burst and been incapable of p hotosynthesizing (Matsubara et al. 2007). Elevated levels of CDOM increased light attenuation and would have reduced the amount of usable PAR, likely affecting the productivity potential of the phytoplankton. Shortened water residence times likely hinder ed the formation and accumulation of relatively slow growing phytoplankton, such as dinoflagellates. High flushing rates and drops in salinity also likely reduced zooplankton grazing, allowing relatively fast growing phytoplankton like diatoms the opportu nity the thrive in the lower estuary and bay during this time.

PAGE 76

76 Summary This study demonstrated that the BZ p I 0 1987) can provide reasonable estimate s of phytoplankton primary productivity in the Caloosahatchee Estua ry, Florida, even given the features of the estuary that distinguish it from previously described systems, including its low latitude subtropical location, individual physical features, unique watershed characteristics, and distinctive phytoplankton commun ity structure and dynamics. The strength of the relationship supports the primary importance of biomass and light in the control of phytoplankton productivity, while the model deviations point to the secondary influence of nutrients, species composition, and light quality on the productivity potential or photosynthetic efficiency of the phytoplankton community. Water residence times, zooplankton grazing, and additional physical chemical properties were also identified for their indirect impact on the pred ictive power of the model. The results also confirmed that the model can give robust estimates of estuarine phytoplankton productivity despite differences in methodologies (using O 2 evolution versus 14 C uptake techniques, gross versus net productivity esti mates, uncorrected or corrected chlorophyll a concentrations, and calculated or corrected photic depths). Thus, model applications from this and a variety of estuaries can be used as a tool to assess productivity on larger spatial and tem poral scales to m ake system wide comparisons of productivity from data that is routinely collected in monitoring projects, saving ecologists the time and money that is required when conducting traditional field measurements. The level of uncertainty associated with the mo del as well as, the tendency of the model to deviate with seasonal shifts in water quality and phytoplankton

PAGE 77

77 community structure, brings caution to the use of the model to accurately predict primary productivity on finer scales.

PAGE 78

78 Figure 3 1. Location of the C aloosahatchee Estuary, FL. The C aloosahatchee Estuary is connected to Lake Okeechobee via the Caloosahatchee River (C 43 Canal) and to the Gulf of Mexico and Charlotte Harbor via San Carlos Bay. Figure 3 2. Location of the four sampling sites i n the Upper Estuary (UE), Middle Estuary (ME), Lower Estuary (LE), and San Carlos Bay (BY)

PAGE 79

79 Table 3 1. Definitions of variables and their units of measure. variable definition units B p h y toplankton biomass as chlorophyll a concentration m 3 B V total phytoplankton biovolume concentration 10 6 3 1 K T light attenuation coefficient m 1 I 0 total daytime surface irradiance (PAR light flux) 2 1 GPP d daily gross primary productivity m g 2 1 N P P d daily net primary productivity 2 1 Z p photic depth M Z w water depth M Table 3 estuarine productivity models. Year corresponds to date of the reference and not the date of the experiments. reference location N BZ p I 0 mo del r 2 Cole & Cloern (1984) San Francisco Bay, CA 77 P d = 3.8( BI 0 K T 1 ) + 58 = 0.82( BZ p I 0 ) + 58 0.82 Cole & Cloern (1987) San Francisco Bay, CA P uget Sound, WA Hudson River Plume, NY 211 P d = 0.73( BZ p I 0 ) + 150 0.82 Kelle r (1988) Narragan sett Bay, RI 1010 P d = 0.70( BZ p I 0 ) + 220 0.82 Boyer et al. (1993) Neuse River Estuary, NC 335 ln [( P d ) = 0.96( BZ p I 0 ) 0.08] 0.66 Kelly & Doering (1997) Massachusetts Bay, MA Boston Harbor, MA 12 P d = 0.79( BZ p I 0 ) + 285 0.66 Murrell et al. (2007) Escambia Bay, FL 22 P d = 0.59( BZ p I 0 ) + 124 0.77 Bouman et al. (2010) Tokyo Bay, Japan 72 P d = 1.87( BI 0 K T 1 ) + 72.7 = 0.41( BZ p I 0 ) + 72.7 0.52 this study Caloosahatchee Estuary, FL 24 P d = 0.7 3 ( BZ p I 0 ) + 75 0.8 4

PAGE 80

80 Figure 3 3. Schematic showing how the specific light transmittance depth profile in the water column at each site (A) was emulated with various light treatment levels in the incubation raceway (B). Primary productivity rates were then plotted for each treatment group, which correspond ed back to a water depth (C).

PAGE 81

81 Figure 3 4. Daily rainfall ( cm ) (A) and flow ( m 3 1 ) (B) from February 2009 to February 2010 at the Franklin Lock and Dam (S 79), FL, which serves as the head of the Caloosahatchee Estuary and t he predominant source of freshwater.

PAGE 82

82 Figure 3 5. Surfac e and bottom salinities (SAL in psu ) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009). Figure 3 6. Surface and bottom water temperature ( T EMP_W in C) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009).

PAGE 83

83 Figure 3 7. M odel para meters for s ite UE in the Caloosahatchee Estuary, FL. Monthly variation of (A) t otal daytime surface irradiance ( I 0 in 2 1 ) (B) photic depth ( Z p in m), and water depth ( Z w in m), (C) phytoplankton biomass ( B in 3 ) as uncorrected chloro phyll a and (D) measured daily gross primary productivity ( GPP d in 2 1 ), during the six month model test.

PAGE 84

84 Figure 3 8. M odel parameters for site ME in the Caloosahatchee Estuary, FL. Monthly variation of (A) t otal d aytime surface irradiance ( I 0 2 1 ), (B) photic depth ( Z p in m), and water depth ( Z w in m), (C) phytoplankton biomass ( B in 3 ) as uncorrected chlorophyll a and (D) measured daily gross primary productivity ( GPP d in 2 1 ), duri ng the six month model test.

PAGE 85

85 Figure 3 9. Model parameters for site LE in the Caloosahatchee Estuary, FL. Monthly variation of (A) t otal daytime surface irradiance ( I 0 2 1 ), (B) photic depth ( Z p in m), and water de pth ( Z w in m), (C) phytoplankton biomass ( B in 3 ) as uncorrected chlorophyll a and (D) measured daily gross primary productivity ( GPP d in 2 1 ), during the six month model test.

PAGE 86

86 Figure 3 10 Model paramete rs of site BY in the Caloosahatchee Estuary, FL. Monthly variation of (A) t otal daytime surface irradiance ( I 0 2 1 ), (B) photic depth ( Z p in m), and water depth ( Z w in m), (C) phytoplankton biomass ( B in 3 ) as uncorrected chlorophyll a and (D) measured daily gross primary productivity ( GPP d in 2 1 ), during the six month model test.

PAGE 87

87 Figure 3 11. Model application for the Caloosahatchee Estuary, FL based on annual variation in (A) total dayti me s urface irradiance ( I 0 in 2 1 ), (B) photic depth ( Z p in m), (C) phytoplankton biomass ( B in 3 ) as uncorrected chlorophyll a yielding (D) modeled daily gross primary productivity ( GPP d in 2 1 ), at each of the four sites from Febr uary 2009 to February 2010 (excluding March 2009).

PAGE 88

88 Figure 3 12. Estimates of light attenuation coefficients ( K T in m 1 ) and corresponding concentrations of (A) color dissolved organic matter (CDOM in pcu), (B) turbidity (TURB in ntu), (C) and uncorrected chlorophyll a (CHL A in 3 ) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009).

PAGE 89

89 Figure 3 1 ) of (A) total soluble phosphorous (TSP), (B) total phosphorous (TP), (C) total soluble nitrogen (TSN), and (D) total nitrogen (TN) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009).

PAGE 90

90 Figure 3 1 ) of (A) dissolved inorganic nitrogen (DIN) and (B) soluble reactive phosphorous (SRP) and (C) mass ratios of DIN to SRP at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009).

PAGE 91

91 Figure 3 1 ) of silica (SI) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009).

PAGE 92

92 Table 3 3. Phytoplankton biomass potential based on Redfield stoichiometric proportions of SRP, DIN, and S I with respect to observed phytoplankton biomass measured as CHL A month site biomass potential based on SRP biomass potential based on DIN b iomass potential based on SI observed biomass as CHL A Feb 09 UE 47 19 91 26 Apr 09 UE 76 12 61 8 May 09 UE 69 3 83 19 Jun 09 UE 142 16 128 39 Jul 09 UE 140 43 178 2 Aug 09 UE 132 50 172 2 Sep 09 UE 93 40 193 3 Oct 09 UE 95 29 181 5 Nov 09 UE 67 1 3 46 10 Dec 09 UE 110 63 171 5 Jan 10 UE 50 17 104 16 Feb 10 UE 48 15 107 13 Feb 09 ME 43 9 75 6 Apr 09 ME 62 9 27 4 May 09 ME 67 8 38 7 Jun 09 ME 116 4 104 39 Jul 09 ME 106 45 190 5 Aug 09 ME 82 27 187 11 Sep 09 ME 91 32 197 6 Oct 09 ME 86 15 1 47 7 Nov 09 ME 64 9 16 5 Dec 09 ME 84 39 97 7 Jan 10 ME 41 8 95 9 Feb 10 ME 37 8 52 4 Feb 09 LE 29 4 52 2 Apr 09 LE 39 9 35 2 May 09 LE 39 5 39 2 Jun 09 LE 77 6 62 4 Jul 09 LE 134 6 161 60 Aug 09 LE 132 52 195 5 Sep 09 LE 57 22 92 2

PAGE 93

93 Table 3 3. Continued. month site biomass potential based on SRP biomass potential based on DIN biomass potential based on SI observed biomass as CHL A Oct 09 LE 67 20 137 3 Nov 09 LE 42 14 29 2 Dec 09 LE 64 23 45 2 Jan 10 LE 29 7 43 2 Feb 10 LE 27 7 23 2 Feb 0 9 BY 18 5 52 2 Apr 09 BY 22 3 58 3 May 09 BY 15 2 50 2 Jun 09 BY 22 6 22 1 Jul 09 BY 31 10 77 6 Aug 09 BY 48 16 37 15 Sep 09 BY 29 11 34 4 Oct 09 BY 33 9 100 4 Nov 09 BY 28 14 32 2 Dec 09 BY 26 14 36 2 Jan 10 BY 24 6 20 2 Feb 10 BY 16 5 26 1

PAGE 94

94 Figure 3 16. Total phytoplankton biovolume concentration ( BV in 10 6 3 1 ) of dinoflagellates, diatoms, cyanobacteria, and other phytoplankton taxa at each of the four sites in the Caloosahatchee Estuary, FL from February 20 09 to February 2010 (excluding March 2009).

PAGE 95

95 Figure 3 17. Percent contribution of A. sanguinea and other dinoflagellates, S. cf costatum and other diatoms, cyanobacteria, and other phytoplankton taxa to total phytoplankton bio volume concentration ( BV in 10 6 3 1 ) at each of the four sites in the Caloosahatchee Estuary, FL from February 2009 to February 2010 (excluding March 2009).

PAGE 96

96 Table 3 4. Summary of measured and mode led daily primary productivity ( GPP d in 2 1 ) from each of the four sites in the Caloosahatchee Estuary, FL. Measured values were obtained from six monthly simulated in situ incubation experiments between February and August 2009 (excluding March). Modeled values were obtained from twelve monthly ap BZ p I 0 model between February 2009 and February 2010 (excluding March 2009). Upper Estuary ( UE ) Middle Estuary ( ME ) Lower Estuary ( LE ) San Carlos Bay ( BY ) measured GPP d minimum 322 328 153 90 median 1040 679 576 188 maximum 2120 3121 2797 1587 mean 1108 1083 847 590 modeled GPP d minimum 189 398 163 324 median 763 711 433 612 maximum 4318 4405 4061 1799 mean 1161 1096 800 735 Table 3 5. Summary of tested model relationships b ased on phytoplankton biomass ( B ) in terms of CHL A concentrations ( uncorrected or corrected for pheophytin), photic depth ( Z p ) calculate d or corrected for water depth ( Z w ), and productivity (measured gross or net daily rates). The NPP d models include d on e less data point (N = 23) than the GPP d models ( N = 24) due to the negative net productivity value observed in February 2009 at site BY, which was removed from the dataset since no explanation coul d be given. Intercepts that were significantly different from zero at a 0. 05 significance level are marked with an asterisk (*). B Z p productivity slope intercept r 2 p value unc CHL A calculated GPP d 0.73 75 0.84 <0.0001 unc CHL A calculated NPP d 0.63 102 0.82 <0.0001 unc CHL A corrected GPP d 0.70 256 0.85 <0.0001 unc CHL A corrected NPP d 0.59 262 0.81 <0.0001 cor r CHL A calculated GPP d 0.76 111 0.80 <0.0001 corr CHL A calculated NPP d 0.65 132 0.78 <0.0001 corr CHL A corrected GPP d 0.75 277 0.83 <0.0001 corr CHL A corrected NPP d 0.6 3 281 0.80 <0.0001

PAGE 97

97 Figure 3 18. Regression of d a ily gross primary productivity ( GPP d in 2 1 ), against the composite parameter BZ p I 0 for 24 incubation experiments. Uncorrected CHL A was used for B and calculat ed Z p values were not corrected for water column depths. GPP d = 75 + 0.73 BZ p I 0 (solid line); 95% confidence i nterval of mean (dashed lines) ; r 2 = 0.84; overall model and slope were significant ( p < 0.0001); intercept was not significant ( p = 0.4849). F igure 3 19. Estimated differences between predicted and measured primary productivity values for all four sites across six monthly incubation experiments.

PAGE 98

98 Table 3 6. Annual phytoplankton productivity estimates (system wide ranges or site specific means ) of various river dominated estuaries. All estimates are assumed to be net primary productivity although not all investigators specified the methodology used. Year corresponds to date of the reference and not the date of the experiments. system product i vity 2 1 sampling location reference North San Francisco Bay, CA 95 110 130 Suisun Bay San Pablo Bay Cole & Cloern (1984) Fourleague Bay, LA 120 317 Upper Bay Lower Bay Randall and Day (1987) Narragansett Bay, RI 189 308 dock Mid Bay Keller (19 88) Furnas et al. (1976) Hudson River Estuary, NY 200 370 Lower Bay Bight Malone (1977) Malone (1976) Mobile Bay, AL 242 Pennock et al. (1999) Apalachicola Bay, FL 240 255 Pennock et al. (1999) Mortazavi et al. (2000) Escambia Bay, FL 291 Murrell et a. (2007) Caloosahatchee Estuary, FL 245 266 359 379 San Carlos Bay Lower Estuary Middle Estuary Upper Estuary this study Neuse River Estuary, NC 343 456 Mallin et al. (1991) Boyer et al. (1993) Nueces River Estuary, TX 370 Pennock et al. (1999) Del aware Bay, DE 190 400 Pennock & Sharp (1986) Charlotte Harbor, FL 83 438 McPherson et al. (1990) Tokyo Bay, Japan 370 580 Bouman et al. (2010) Mid Chesapeake Bay, MD 337 782 Boynton et al. (1982)

PAGE 99

99 CHAPTER 4 NATURAL AND ANTHROPOGENIC INFLUENCES ON THE SPATIAL AND TEMPORAL PATTERNS OF PHYTOPLANKTON PRODUCTIVITY IN THE CALOOSAHATCHEE ESTUARY, FLORIDA Concepts and Applications With coastal eutrophication emerging as one of the most pressing and widespread problems of the late 20th and early 21st centu ries (Nixon 1995, Cloern 2001, Schindler 2006, Bricker et al. 2008), tracing and understanding major shifts in the trophic status of estuaries and coastal ecosystems have become the primary focus of many research and management efforts (Zingone et al. 2010 ). The trophic status of an ecosystem is best defined by its supply of organic matter as rates of primary production (Nixon 1995, Schindler 2006), but the associated terminology is more commonly used to characterize the level of anthropogenic influence on a system, primarily in the form of nutrient enrichment (Nixon 1995, Cloern 2001). Inputs of nutrients, namely nitrogen and phosphorous, have accelerated in coastal waters due to human activiti es in the surrounding watershed, such as land clearing, fertil izer production and application, animal consumption, sewage emission, and fossil fuel combusti on (Nixon 1995, Vitousek et al. 1997b, Cloern 2001). Nutrient levels in coastal ecosystems are also affected by natural processes like geochemical weathering and internal recycling, which are largely regulated by climate and weather (Schindler 2006). A major consequence of c oastal eutrophication is the over production and accumulation of plant biomass, which can come in the form of toxic, harmful, and/ or nuisance algal blooms that have been associated with hypoxic conditions, fish kills, community shifts, human health risks, and a suite of other responses (Nixon 1995, Cloern 2001, Bricker et al. 2008).

PAGE 100

100 Because the trophic status of an ecosystem can change as a res ult of both natural and anthropogenic influences, it should be treated more as a dynamic process rather than a static description of the conditions of an ecosystem (Nixon 1995). The ability to fairly assess and manage changes in the trophic status of an e c osystem thus requires long time series that encompass seasonal, interannual, and decadal variations in primary producti on (Zingone et al. 2010). Long time series of primary productivity data from estuaries and coastal ecosystems are rare due to the histo rical focus of aquatic research in freshwater ecosystems (Boynton et al. 1982, Cloern 2001), as well as the logistical limitations of the traditional methods used to measure photosynthetic rates (Ryther 1956b, Ryther and Yentsch 1957, Tilzer 1989, Sand Jen sen 1997). Attempts to determine and compare the trophic status of aquatic ecosystems have primarily relied on a suite of biological and physical chemical parameters and/or symptoms that serve as proxies for primary production rates and indica tors of wate r quality including concentrations of chlorophyll a levels of dissolved oxygen, concentrations of limiting nutrients, measures of water transparency, and appearances of key plant and animal species (Ryding and Rast 1989, Schindler 2006, Bricker et al. 200 8, Boyer et al., 2009). Chlorophyll a has been the more widely used and accepted index of productivity and trophic status in estuaries and coastal ecosystems because it is easily and routinely measured in oceanographic investigations, and it provides a ge nerally good assessment of the structure and function of both lower and higher order trophic levels and the associated biogeochemical processes (Cloern 2001, Boyer et al. 2009, Cloern and Jassby 2010, Zingone et al. 2010). However, very few time series cl early document trends of increasing chlorophyll a in concert with trends of increasing

PAGE 101

101 nutrients associat ed with coastal eutrophication and, in some cases, the opposite relationship is observed (Boynton et al. 1982, Cloern 2001). Inconsistencies in the re sponses of estuaries and coastal ecosystems to natur al and anthropogenic stressors are due to a number of morphological, physical, chemical, and biological factors that are unique to each ecosystem and have the potential to regulate photosynthetic biomass and production (Brylinsky and Mann 1973, Boynton et al. 1982, Day et al. 1989, Cloern 2001). As an indicator of phytoplankton biomass, chlorophyll a represents the net result of both growth and loss processes, so using it alone to characterize trophic sta tus may not accurately portray the productivity potential of an ecosystem. To overcome the challenges encountered when using traditional methods for measuring primary productivity and assessing trophic status on the basis of more descriptive indices, inv estigators have turned to ecosystem modeling to identify factors that govern production and obtain accurate production estimates (Odum 1968, Vollenwider 1969, Brylinsky and Mann 1973, Boyer et al. 1993, Scardi 1996). The BZ p I 0 eloped by Cole and Cloern (1987) is an example of a simple, empirical relationship founded on long standing paradigms that emerged in the mid twentieth century that address the primary factors controlling phytoplankton productivity in aquatic ecosystems (R yther 1956b, Ryther and Yentsch 1957). This approach has provided robust estimates of phytoplankton productivity in estuaries and coastal ecosystems around the world (Cole and Cloern 1987, Keller 1988, Mallin et al. 1991, Boyer et al. 1993, Kelly and Doer ing 1997, Murrell et al. 2007, Bouman et al. 2010). Most of the model applications were in temperate systems, except for Escambia Bay, Florida, which was classified as being subtropical despite its annual temperature,

PAGE 102

102 rainfall, and river flow patterns tha t are more characteristic of temperate environments (Murrell et al. 2007). The recent adaptation of the BZ p I 0 model to the Caloosahatchee Estuary, Florida extended its applicability to a truly subtropical system and one with additional distinct features (Chapter 3). The Caloosahatchee Estuary is a shallow, microtidal subtropical system that has been extensively modified from its natural state since the late 1800s. It is one of many ecosystems making up the Florida Everglades that have b een originally used, subsequently altered, and currently maintained in order to accommodate agriculture and urban development in an otherwise uninhabitable area prone to prolonged flooding. The Caloosahatchee Estuary was once fed by a small, sinuous river that originated from (Teeter 1980, Flaig and Capece 1998, Knight and Steele 2005). The estuary is now artificially connected to Lake Okeechobee, the headwaters of the Flo rida Everglades and the heart of a major agriculture, cattle, and dairy industry, via a trans state canal that is deeper, wider, and straighter than the original Caloosahatchee River. Three lock and dam structures control the flow of water between Lake Ok eechobee, the Caloosahatchee Estuary, and ultimately the Gulf of Mexico. The trophic status of the Caloosahatchee Estuary has become a growing concern over the last few decades due to the contributi on of hypertrophic water s, released from Lake Ok eechobee and flushed downstream, to coastal eutrophication and the prevalence of harmful algal blooms (Perry 2008). The purpose of this study was to use the BZ p I 0 C loern 1987) previously adapted to the Caloosahatchee Estuary, Fl orida (Chapter 3) to

PAGE 103

103 obtain long term estimates of da ily gross primary productivity ( GPP d ) for the system. Model parameters were extracted from an existing discontinuous twenty five year water quality data set, thereby extending the usefulness of the asso ciated monitoring status. Multivariate statistical methods were employed to build a complete subset of data in order to identify potential environmental drivers of phytopla nkton productivity in the Caloosahatchee Estuary. Variability in chlorophyll a associated with seasonal cycles, annual disturbances (i.e., human actions, climatic shifts), and residual events (i.e., phytoplankton blooms), was also explored using the metho ds of Cloern and Jassby (2010), since phytoplankton biomass was previously identified as the main explanatory variable of GPP d (Chapter 3). Long term changes in GPP d status were examined for their connection to both natural and a nthropogenic influences in the surrounding watershed on the structure and function of the entire ecosystem. Methods Study Area The Caloosahatchee Estuary is an extensively modified, shallow, microtidal, subtropical system located on the southwest coast of Florida in Lee County (Figure 4 1). The upper reach of the Caloosahatchee Estuary is connected to Lake Okeechobee via the Caloosahatchee River (C 43 Canal), and the lower reach of the estuary is connected to the Gulf of Mexico and Charlotte Harbor via San Carlos Bay. The physical shape and hydrology of the system has been modified from its natural state since the late 1800s. In addition to a series of canals and pumping stations, three lock and dam structures now control the flow of water between Lake Ok eechobee and the Caloosahatchee Estuary along the C 43 Canal. The Franklin Lock and Dam (S 79)

PAGE 104

104 artificially serves as the head of the estuary and main source of freshwater approximately 42 km upstream of Shell Point (Figure 4 2). The width of the estuary is irregular, ranging from 160 m near S 79 to 2,500 m in San Carlos Bay (Scarlatos 1988). The narrow section extending from S 79 to Beautiful I sland has an average depth of 6 m, but the area downstream of Beautiful Island has an average depth of 1.5 m (S carlatos 1988). A narrow navigation channel (part of the Intracoastal Waterway) is maintained at a depth of approximately 3.5 m (Scarlatos 1988). The area experiences a combination of diurnal and mixed semi diurnal tides with a mean tidal range of 0.30 m in the middle of the estuary near downtown Fort Myers (Scarlatos 1988, NOAA 2010). Freshwater flows through S 79 at a rate great enough to fill the entire volume of the estuary over eight times in an average year (Doering and Chamberlain 1999). Flow rate s are set by the South Florida Water Management District (SFWMD) and the United States Army Corps of Engineers (USACE) to maintain prescribed water levels in Lake Okeechobee, control flooding, and flush algal blooms, salt water, and other contaminants out of the Caloosahatchee River (Stoker 1992, Flaig and Capece 1998, Doering and Chamberlain 1999, Rand and Bachman 2008). The estuary also receives an average of 137 cm of rainfall per year (NCDC 2011), of which the majority (75%) arrives during the months o f May throu gh October, defining a typical wet season (Stoker 1992). Surface water runoff accompanying this rainfall passes through agricultural areas in the east and urban areas in the west. The combination of rainfall, runoff, and releases from the wate rshed and canals causes large fluctuations in the quantity, quality, timing, and distribution of freshwater inflow to the estuary, which has affected

PAGE 105

105 the str ucture, function, and overall health of the ecosystem (Chamberlain and Doering 1998a, Barnes 2005, Knight and Steele 2005, SFWMD et. al 2009). Water Quality Data Sources and Manipulations A twenty five year data set (January 1 986 through December 2010) was compiled using water quality data collected in the vicinity of the Caloosahatchee Estuary durin g five monitoring programs/research projects either conducted or supported by the South Florida Water Management District, West Palm Beach, Florida (SFWMD 2011) No water quality data were available for the periods of June 1989 through September 1994 and July 1997 through March 1999. Additional months in the twenty five year period were also missing water quality data, although they occurred at random times and do not represent significant gaps in the data set. Water quality data collected at sites outsi de the b ounds of the defined study area ( i. e., the Caloosahatchee Estuary between S 79 and Shell Point) and San Carlos Bay, were not included in the combined data set, unless otherwise noted. The five monitoring programs/research projects that contributed water quality data include: 1. The Caloosahatchee Estuary (CAL) program sampled eighteen stations (1 through 18) in the Caloosahatchee Estuary (S 79 to Shell Point), San Carlos Bay, Matlacha Pass, and Pine Island Sound. The two stations in Pine Island Sound (15 and 16) were excluded from the data set due to their distance from S 79 and physical separation from the other stations in the study area by Pine Island. The station described as being in Matlacha Pass (14) was still included in the data set due to it s close vicinity and connectivity to San Carlos Bay. Stations were sampled monthly from December 1985 to May 1989, although the December 1985 measurements were not included in the data set. Sampling continued at a subset of stations from October 1994 to August 1996. Additional samples were collected at a select few sites in April 1997 and April 1999. 2. The Coastal Charlotte Harbor Monitoring (CCHM) project was created for the Charlotte Harbor National Estuary Program. The Caloosahatchee Estuary (S 79 to S hell Point) was divided into fifty grid squares (3.4 km 2 each) according to the

PAGE 106

106 Florida Fish and Wildlife Conservation Commission Florida Marine Research Institute Fisheries Independent Monitoring Program Grid System (Corbett 2004), and sampling occurred at a random location within the grid squares. Stations were registered according to the GPS location of the bottom right hand corner of the grid in which the sample was collected. Samples were collected monthly within five randomly selected grids from S eptember 2006 to September 2009. 3. The Center for Environmental Studies (CES) program was initially developed as an on going water quality monitoring program, but it was later redesigned and renamed the Caloosahatchee Estuary Water Quality Monitoring Progra m (CESWQ). A total of eleven stations (1 through 11) in the Caloosahatchee River (upstream of S 79), Caloosahatchee Estuary (S 79 to Shell Point), and San upstream of S 79 (1) wa s excluded from the data set for being outside the boundaries of the defined study area. Stations were sampled monthly between April 1999 and March 2002. Monthly sampling resumed in May 2002 and continues today at a smaller subset of stations. Only meas urements through December 2010 were included in the data set. Additional sampling occurred at selected times and at specific sites to capture event driven effects of freshwater releases from Lake Okeechobee to the Caloosahatchee Estuary. 4. The Environmental Research and Design (ERD) program sampled fifteen (1 through 15) sites in the Caloosahatchee Estuary (S 79 to Shell Point), San Carlos Bay, and Gulf of Mexico. The two stations in the Gulf of Mexico on the south side of the Sanibel Causeway (1 and 2) wer e excluded from the data set for being beyond the boundaries of the defined study area. Stations were sampled four times (approximat ely ten days apart) during the wet and dry seasons over a three year period from 2000 to 2002. 5. The Harbor Branch (HB) pro ject included seven (1 through 7) stations in the Caloosahatchee Estuary (S 79 to Shell Point), San Carlos Bay, and Pine Island Sound. The station in Pine Island Sound (7) was excluded from the data set due to its distance from S 79 and physical separatio n from the other stations by Pine Island. Stations were visited between May 1996 and March 1997. The frequency of sampling was not consistent; samples were collected on a weekly or bi weekly schedule for some months, while other months had only one colle ction or none at all. The resulting twenty five year data set, hereaft a compilation of all discontinuous sampling efforts from the five monitoring programs/research projects. All monthly, bi weekly, weekly, and event dr iven sample collections were included in the data set so that the most comprehensive span of water quality, phytoplankton biomass, and primary productivity data could be obtained. To

PAGE 107

107 address the issue of inconsistent sampling frequencies, monthly means we re taken across the measurements obtained from multiple collections at the same site and used to represent conditions there for a given month. When sites were only sampled once in a given month, then the measurements from the single collection were used t o represent average conditions there for the entire month. Measurements from individual collections were also considered separately (and noted when done so) to identity the timing of bloom events. The data set was further manipulated to prevent errors wi th mathematical computations and statistical analyses. Any reported negative values were replaced with zero. Values recorded in conjuncti NH 3 < 5) were taken as the value alone (e.g., NH 3 = 5). T he compiled data set e ncompassed the sampling efforts at a total of eighty different sites, having unique GPS (latitude and longitude) coordinates, not counting multiple locations that were used in more than one project. Because sampling efforts were not continuous at each of the sites over the extent of the data set, sites were classified using two systems (geographic and physical chemical) so that spatial, in addition to temporal, variation in water quality, phytoplankton biomass, and primary productivity in the estuary could be addressed. The validity of using a site classification scheme to pool data was tested with a hierarchical agglomerative clustering analysis and corresponding similarities profile analysis (further details provided in the Statistical Analyses for Ident ification of Environmental Drivers section). S ites were grouped into four geographic regions according to their distance from S 79 (Doering and Chamberlain 1999, Doering et al. 2006), The four regions were classified as the Upper Estuary (UE), Middle Estu ary (ME), Lower Estuary (LE), and

PAGE 108

108 San Carlos Bay (BY) (Figure 4 2), and sites lying between 0 to 14, 14 to 28, 28 to 40, and 40 to 48 km from S 79 were assigned to each regio n accordingly (Table 4 1). The division of the Caloosahatchee Estuary into segmen ts of varying distance from S 79 has been the traditionally used approach in system monitoring and modeling efforts th ere because it accounts for spatial differences in geology, bathymetry, and hydrology (Scarlatos 1988, Bierman 1993 Doering and Chamberla in 1999, Doering et al. 2006 ). Sites were also assigned to one of five salinity zones using the Venice System (Anonymous 195 9 ) based on water column conditions at the time of sample collection. Sites having measured salinities between 0 to 0.5, 0.5 to 5, 5 to 18, 18 to 30, and 30 to 40 psu were categorized in salinity zones freshwater, oligohaline, mesohaline, polyhaline, and euhaline, respectively (Table 4 2). This approach essentially divide d the Caloosahatchee Estuary into ecoclines governed by gradual and progressive gradients of salinity (Quinlan and Phlips 2007). Here, salinity was assumed to be the dominant environmental factor defining a water mass and regulating the phytoplankton community structure and function in the Caloosahatchee Estuary sinc e it corresponds directly to changes in freshwater inflow and tidal exchange. The five s alinity zones roughly encompassed the previously described median salinity values for each of the four geographic regions (Table 4 2). Although the four geographic reg ions (UE, ME, LE, and BY) represented a fixed spatial classification system based on sampling location, sites were expected to shift between the five salinity zones (freshwater, oligohaline, mesohaline, polyhaline, and euhaline) due to variation in freshwa ter inflow and tidal influence in the estuary. For example, the location of the freshwater saltwater interface (the point at which the salinity is 0.5 psu)

PAGE 109

109 moves upstream and downstream daily with the tide and seasonally with the volume of freshwater infl ow (Stoker 1992). Water Sampling and Analyses For each of these monitoring programs/research projects, water samples were collected at a depth of 0.5 m using a van Dorn or similar horizontal sampling device (Doering and Chamberlain 1999, Corbett 2004, Doe ring et al. 2006). Any measurements on water collected at depths other than 0.5 m were excluded from the data set to maintain consistency in methodology across the twenty five year time period. Depending on the scope of the respective monitoring programs /research projects, a number of physical and chemical measurements were taken directly in the field or from laboratory analyses. At each site, vertical profiles of temperature (TEMP_W), salinity (SAL), and dissolved oxygen (DO) were recorded using a Hydr olab or YSI. Only the readings taken at a depth of 0.5 m were included in the data set to correspond to the collected water sample since bottom measurements were not consistently provided. Estimates of photic depth ( Z p ; de pth of 1% incident irradiance) w ere derived from secchi disk depths ( Z s ) using the relationship Z p = 2* Z s (Wetzel 1983). For the ERD project, Z s measurements were not available, so Z p estimates were calculated as 4.61/ K T Light attenuation coefficients ( K T ) were calculated using the Be er Lambert relationship ln( I z / I 0 ) = K T Z based on simultaneous photosynthetic active radiation (PAR) readings taken at the surface ( I 0 ) and a depth of 1.0 m ( I z ) with deck (reference) and downwelling cosin e corrected quantum sensors (Li Cor or equivalent) Surface (incident) irradiance readings ( I 0 ) were corrected for 5% reflection prior to calculating K T

PAGE 110

110 Samples for the determination of color (CDOM) were passed through a 0.45 m filter using a syringe in the field. Unfiltered water samples were retai ned for total phosphorous (TP), total Kjeldahl nitrogen (TKN), silica (SI), turbidity (TURB), total suspended solids (TSS), and chlorophyll a (CHL A ). All samples were stored on ice until their return to the laboratory. CHL A samples were filtered in the laboratory within 24 hours of collection. Water quality parameters were analyzed by standard methods (APHA 2005) using a spectrophotometer or auto analyzer. Concentrations of CHL A not corrected for pheophytin degradation pigments (if the distinction wa s noted by the respective monitoring programs/research projects) were used as a p roxy for phytoplankton biomass ( B ), was defined to indicate the top 10% of CHL A concentrations observed during the twe nty five year period (Phlips et al. 2010). All laboratory analyses were conducted in accordance to the specifications of the South Florida Water Management District Quality Assurance Plan and in compliance with the National Environmental Laboratory Accred itation Program (Doering and Chamberlain 1999, Corbett 2004, Doering et al. 2006). Meteorological and Hydrological Data Sources Measurements of PAR light flux recorded at three stations in Lee, Collier, and Hendry Counties (Figure 4 1) were compiled from t he South Florida Water Management fifteen minute intervals we 2 1 and summed across the daylight period to provide estimates of to tal daytime surface irradiance ( I 0 ) The maximum values of I 0 across the tri county area were plotted and used to extrapolate monthly maxim a I 0 Monthly maximum I 0 values from February 2009 to January 2010,

PAGE 111

111 which were verified with field measurements and observatio ns (Chapter 3), were applied across the long term data set to predict phytoplankton productivity using the BZ p I 0 approach. Daily measurements of flow recorded at S 7 7, S 78, and S 79 (Figures 4 1 and 4 2) were obtained from the South Florida Water Manageme environmental database (SFWMD 2011). Flow values were converted into metric equivalents (m 3 1 ) and then averaged across each collection month (January 1986 through December 2010). Monthly means, opposed to daily rates, took into co nsideration the t ravel time of freshwater from S 79 through the estuary and the effect of preceding flow conditions (Stoker 1992). Mean monthly flow has been previously determined as an acceptable approach to analyzing freshwater inflow effects on water q uality because it adequately represents the approximate expected residence time for a variety of flow regimes observed in the system (Chamberlain and Doering 1998a). The mean monthly flows were used to identify int ra and in ter annual discharge pattern s an d unusual water release or withdrawal events in the long term data set. Daily inflows from the three water control structures were also used to identify the timing and sources of freshwater into the estuary. Monthly rainfall totals and average air tempera tures recorded in Fort Myers and Punta Gorda (Figure 4 1) were obtained from the National Climatic Data Center (NCDC 2011). Punta Gorda values were only used when Fort Myers readings were unavailable, mainly between October 1995 and February 1998. Values were converted into metric equivalents, cm and C, respectively. Departures from normal were taken as the average value of the meteorological element over a recent thirty year time period

PAGE 112

112 (NCDC 2011). Monthly rainfall totals and departures from normal w ere used to identify intra and interannual prec ipitation patterns and potential flood and drought conditions in the long term data set. Monthly average air temperatures and departures from normal were us ed to identify int r a and int er annual temperature p atterns and unusual warm and cold periods in the long term data se t. Daily sums of rainfall at S 79 were also obtained (SFWMD 2011), conve rted into metric equivalents as befo re and used to identify the timing and sources of freshwater into the estuary. Bi monthly values of the Multivariate ENSO (El Nio/Southern Oscillation) Index (MEI) were obtained from the Earth System Research Laboratory (ESRL 2011). MEI values were used to identify periods of ENSO based on six main observed variables over the tropical Pacific, including (1) sea level pressure, (2) zonal and (3) meridional components of the surface wind, (4) sea surface temperature, (5) surface air temperature, and (6) to tal cloudiness fraction of the sky (ESRL 2011). Monthly indices were taken as the b i monthly value (i.e., December/January, January/February, February/March, etc.) o f the listed second month (i.e., January, February, March, etc.) since there is a n underst ood time lag required for the global atmosphere to respond to tropical sea surface temperature anomalies (ESRL 2011). Positive MEI values represented the warm ENSO phase (El Nio), while negative values represented the cold ENSO phase (La Nia) (ESRL 2011 ). The severity of ENSO events were compared using MEI ranks, a number between 1 and 62, which were based on the MEI values of similar bimonthly seasons since the beginning of record in December 1949/January 1950 (ESRL 2011). Monthly ranks were taken fro m the bi monthly values

PAGE 113

113 using the same approach as the indices. Months having ranks 1 through 6 were classified as strong La Nia event s, while months having ranks 56 through 62 were classified as strong El Nio events, representing the top ten percent of each (ESRL 2011). Weak La Nia periods corresponded to ranks 7 through 19, and weak El Nio periods corresponded to ranks 44 through 55 (ESRL 2011). Months ha ving ranks in the middle (20 through 43) were classified as near normal periods (ESRL 2011). Pr imary Productivity Estimates Estimates of d aily gross primary productivity ( GPP d ) of the phytoplankton community in the water column were obtained using the model relationship GPP d = 75 + 0.73 BZ p I 0 which was previously derived for the Caloosahatchee Estua ry (Chapter 3). Model predicted GPP d was calculated across the twenty five year data set for each given site and month as long as there were corresponding values for B Z p and I 0 available. E stimates of GPP d were averaged across years and converted into estimates of ann ual gross primary productivity ( GPP y ) evaluated and system wide comparisons could be made. The trophic scheme proposed by Nixon (1995) for estuaries and coastal ecosystems was used to classify annual 2 1 ), mesotrophic (between 100 and 300 2 1 ), eutrophic (between 301 and 500 2 1 ), or hypertrophic (greater than 500 2 1 ). Seasonal contributions to GPP y were obtained b y averaging GPP d across wet and dry months and multiplying by the number of days in each season in an average year (181 days in the dry season between November 1 and April 30; 184 days in the wet season between May 1 and October 31).

PAGE 114

114 Decomposition of Phyto plankton Biomass Variability The methods of Cloern and Jassby (2010) were followed to examine the long term time series of monthly mean CHL A concentrations and isolate the underlying mechanisms controlling phytoplankton biomas s and, thus, phytoplankton pr oductivity by inference, at different timescales in the four regions of the Caloosahatchee Estuary. Phytoplankton biomass, in terms of CHL A was previously identified as the primary model parameter explaining most of the variability in GPP d (Chapter 3). The analysis required that each region have at least 8 years of data with at least 10 months of data for each year (Cloern and Jassby 2010). This latter criterion was relaxed to 9 months of data for each year as long as the three missing months were not f rom the same season (winter, spring, summer, or fall) so that the minimum number of years could be met. It was further required that each month be represented by at least 6 years of data from each region (Cloern and Jassby 2010). These criteria were sele cted to obtain the longest time series of CHL A for each region of the estuary while reducing bias associated with missing data. The multiplicative model expressed as c ij = Cy i m j i j was used to partition each CHL A time series into a long term mean and t hree other Jassby 2010). Here, c ij was defined as the CHL A concentration in year i and month j ; C was the long term CHL A mean of the time series; y i was the annual effec t in year i defined as the annual mean Y i divided by C ; m j was the seasonal (monthly) effect in month j taken as the mean CHL A concentration for each calendar month j over all years of the CHL A concentrations from each month j and year i divided by the Y i ; and ij was the residual term (Cloern and Jassby 2010) Values of y i > 1, m j >1, and ij >1 were used to identify years with above average mean CHL A from the long term mean,

PAGE 115

115 months with above average mean CHL A from the annual mean, and individual obs ervations greater than the expected value for a given month and year, respectively (Cloern and Jassby 2010). The standard deviations of y i m j and ij were used as coefficients of variation, allowing comparisons of phytoplankton biomass variability to be made within and across different ecosystems (Cloern and Jassby 2010). Statistical Analyses for Identification of Environmental Drivers Multivariate statistical analyses were performed using PERMANOVA+ for PRIMER software (Plymouth, United Kingdom) and w ere focused on two objectives: addressing discontinuities in the data, and ultimately, identifying key environmental drivers of phytoplankton productivity in the Caloosahatchee Estuary. Because the parameters of interest were not sampled each month across the 80 unique sites over the twenty five year time period, steps were taken to select closely matched data that would fill in gaps to create a useable matrix of measurements across a consistent spatial extent and temporal duration. The results of this ap proach were suitable for further investigation to identify correlations among potential environmental drivers and estimates of GPP d An initial hierarchical agglomerative clustering (Anderson et al. 2008) grouped sites according to similarities calculat ed across 9 physical chemical variables (i.e., SAL, TEMP_W, DO, TP, TKN, SI, CDOM, TURB, and Z p ). The target data set for this analysis was the largest possible block of data for all 9 parameters gathered in consecutive months at 80 sites spanning the Cal oosahatchee Estuary. Where necessary, missing values were replaced by shifting multiple samples collected early or late in one month to the previous or following month, interpolating one to three month gaps linearly, inserting a value from a neighboring m onth or site, or using means for

PAGE 116

116 multiple samples in a given month. Values in the resulting square matrix were range standardized, which converted measurements from multiple scales to a more comparable, dimensionless scale ranging from 0 to 1 (Anderson et al. 2008). Bray Curtis similarity coefficients (Bray and Curtis 1957) were calculated between pairs of sites, with a similarity coefficient, S of 1 representing 100% or total similarity, and a value of 0 representing total dissimilarity (Anderson et al. 2008). Hierarchical agglomerative clustering based on a group average algorithm was performed, and statistically significant clusters were identified by a similarity profile permutation test (Anderson et al. 2008). Within the selected site clusters, dis continuous strings of data from the rele vant sites could be combined (i.e ., by calculating mean values) to improve the spatial temporal continuity of the data set documenting potential environmental drivers and GPP d The continuity of the resulting data se t was further improved by clustering sampling events according to Bray Curtis similarities calculated on range standardized values. For each site group, the data comprised of 9 physical chemical parameters (i.e., SAL, TEMP_W, DO, TP, TKN, CDOM, TURB, TSS, and Z p ) and 3 meteorological and hydrological forcing factors that operated across all sites (i.e., MEI, total monthly rainfall, and mean monthly flow). Measures of SI were excluded from this analysis because they were too sparse. The inclusion of TSS i n this step was done to decouple its possible influence on water quality from that of TURB. A matrix of site groups and year and month combinations between January 1986 and December 2010 was created from the means of values across sites within groups and linear interpolation across 1 to 2 month gaps. As before, hierarchical agglomerative clustering and similarities profile

PAGE 117

117 analysis were used to identify groups of sampling events having similar spatial patterns in water quality and meteorological and hydro logical conditions (Anderson et al. 2008). Selected groups of sampling events augmented site groups as a basis for replacing missing values with values from a sampling period deemed similar or means of values from multiple, similar sampling periods. In c onjunction with pooling across sites, combining data from sampling events generated a more robust data set for investigation of correlations with GPP d Ultimately, two new data sets were generated for combinations of site and time groups. One data set con tained 14 water quality, meteorological, and hydrological parameters (i.e., SAL, TEMP_W, DO, TP, TKN, SI, CDOM, TURB, TSS, Z p I 0 MEI, total monthly rainfall, and mean monthly flow) and the other one contained the potential biological response variable, G PP d The variables Z p and I 0 were excluded from correlation analyses involving GPP d because they were parameters used in conjunction with CHL A to estimate gross primary productivity. These data sets maximized the spatial and temporal coverage of the ori ginal twenty five year data set for the Caloosahatchee Estuary. Values within the data sets were range standardized across combinations of site and parameter for all relevant time periods. Resemblance matrices were calculated for response variables using Bray Curtis similarity coefficients (Bray Curtis 1957), and a stepwise permutation procedure (BVSTEP; Anderson et al. classified as the minimum number of drivers that yiel ded the highest Spearman correla tion coefficient ( ) adjusted for tied ranks (Kendall 1970) when the resemblance

PAGE 118

118 drivers to explain the biological response was terminated when > 0.95 or when the change in between successive comparisons was less than 0.001 Values of for 99 random permutations provided a distribution for evaluating statistical significance, with a statistically significant result being a value greater th an those generated from the randomized data. Results Spatial and Temporal Similarity Groupings of Data Only 12 out of 80 sites were sampled relatively consistently by the five monitoring programs/research projects that were active during the twenty five y ear time period. After missing values were addressed, data from these sites provided continuous time series for 9 parameters across 40 months that were used to calculate resemblances and evaluate clusters of sites. All of the 9 physical chemical paramete r s (i.e., SAL, TEMP_W, DO, TP, TKN, SI, CDOM, TURB, and Z p ) used to identify spatial groups in the data contributed to similarity among sites within groups, without a dominant contribution from any parameter. The results of the similarities profile test i ndicated six clusters of paired sites lying along a spatial gradient characterized by distances from S 79. Two pairs of sites at distances of 6 and 14 km, respectively, clustered at a similarity of 76%. Additionally, two pairs of sites at 21 and 26 km, r espectively, clustered at a similarity of 80%. Given the degree of similarity among data from these sites and the fact that they fell within the existing geographic boundaries of the Upper Estuary (UE) and Middle Estuary (ME) regions, respectively, the pa irs of sites were combined in further analyses. The remaining pairs of sites that clustered at 83 and 67% similarity aligned with the geographic boundaries of the Lower Estuary (LE) and San

PAGE 119

119 Carlos Bay (BY), respectively. These results confirmed a previou sly documented gradient in water quality in the Caloosahatchee Estuary based on distances from S 79 (Doering and Chamberlain 1999, Doering et al. 2006); therefore, further analyses treated sites within the four regions (UE, ME, LE, and BY) as replicates. B etween January 1986 and December 2010, samples were collected in at least one region of the Caloosahatchee Estuary in 197 of the possible 300 months covered by the twenty five year time period. A smaller number of months (76) yielded nearly complete data for 12 parameters across each of the four regions, and these data were used to calculate resemblances for use in hierarchical agglomerative clustering. None of the 9 physical chemical (i.e., SAL, TEMP_W, DO, TP, TKN, CDOM, TURB, TSS, and Z p ) and 3 meteoro logical or hydrological parameters ( i.e., MEI, total monthly rainfall, and mean monthly flow) represented a dominant cause of similarity among monthly sampling events. A similarities profile analysis indicated 33 clusters comprising one to seven sampling events, with clusters separating at similarities of 76.5% or greater. Because the 76 months of data used in this analysis included a wide range of meteorological and hydrological conditions, the corresponding patterns in water quality across the four regi ons of the Caloosahatchee Estuary were considered representative of the potential natural and anthropogenic drivers of primary productivity during the entire twenty five year time period. Thus, samples from the same site group and time period were treated as replicates in order to build a complete subset of data needed to test for correlations between the environmental and biological response variables. Meteorological and Hydrological Observations Between January 1986 and December 2010, there were appr oximately 15 alternating El Nio/Southern Oscillation (ENSO) phases based on switches between

PAGE 120

120 positive (El Nio) and negative (La Nia) Multivariate ENSO Index (MEI) values (Figure 4 3 ; part A). There was a greater occurrence of El Nio conditions, but no t all of these events were considered to be significant. In comparison to past ENSO events occurring in the same month since January 1950 (ESRL 2011), conditions were considered to be near normal during 111 of the 300 months (37%) of the twent y five year period (Figure 4 3; part B). An additional 26% and 19% of months experienced weak El Nio and La Nia conditions, respectively. Strong El Nio and La Nia events occurred during 14% and 4% of the 300 months, respectively. The summer and fall of 1997 was ranked the strongest El Nio event, while the summer and fall of 2010 was ranked the strongest La Nia event during the twenty five year period. Average monthly air temperatures typically ranged from 18 to 28C on an annual basis between January 1986 and December 2010, reflecting the subtropical climate of the Calo osahatchee Estuary (Figure 4 4; part A). Summer air temperature averages were maintained around 28C with smaller departures from normal, while winter air temperature averages showed greater variability (Figure 4 4). Unusually low winter air temperature averages below 16C were observed in December 1989, January 2001, January 2003, December 2003, January 2010, February 2010, and December 2010, corresponding to 2 to 5C departures below norma l, respectively. During the January 2003 and December 2010 cold spells, water temperatures (TEMP_W) in the Caloosahatchee Estuary fell to a twenty five year low of 14C (Figure 4 5). These cooler than normal winter temperatures were typically observed du ring El Nio or near normal conditions (Figures 4 3, 4 4, and 4 5). Unusually high winter air temperatures with monthly averages reported at 3 to 5C above normal were observed

PAGE 121

121 in November and December 1986, January 1989, January and February 1990, Januar y 1991, January 1993, February 1994, February 1997, and February 2001 (Figure 4 4). During the unusually warm late fall/early winter of 1986, water temperatures between 25 and 29C were observed in the Caloosahatchee Estuary when they typically do not exc eed 24C during that time of year (Figure 4 5). The January 1989, February 1997, and February 2001 warmer than normal winter temperatures occurred under La Nia conditions (Figures 4 3 and 4 4). Monthly rainfall totals ranged from 0 to 48 cm and averaged 12 cm over the twenty five year study period (Figure 4 6 ; part A). Rainfall was typically higher from May through October (the defined wet season) than November through April (the defined dry season). Wet season months showed the greatest variability in rainfall, ranging from flood to drought conditions with respect to normal (Figure 4 6). The months of September 1988, September 1992, June 1994, August 1996, August 1997, June 2001, July 2003, June 2009 and July 2009 showed well below normal levels of rai nfall (Figure 4 6 ; part B). These drier than normal months typically occurred during La Nia, near normal, or weak El Nio periods (Figures 4 3 and 4 6). Flooding conditions were more common during the positive (El Nio) phases, including the summer of 1 986 to the fall of 1987, the summer of 1990 to the summer of 1992, the late spring and summer of 1995, the fall and winter of 1997, the late summer of 2001, the late summer of 2003 to the early fall of 2005, the summer of 2006, an d the spring of 2010 (Figu res 4 3 and 4 6). Unusually high rainfall totals in the wet season months often, though not always, corresponded to the occurrence of nearby episodic storm events, including an unnamed storm in late October 1987, Tropical Storm Ana in July

PAGE 122

122 1991, Hurricane Gabrielle in September 2001, Hurricane Erika in August 2003, and Hurricanes Charley and Frances in August 2004 (NOAA 2012). Other episodic storm events, such as the Category 5 Hurricane Andrew in August 1992, did not bring above average levels of rainfal l to the Caloosahatchee Estuary, although it may have to other parts of south Florida. Mean monthly flow from S 79 into the Caloosahatchee Estuary was typically greater in the summer and fall months of each year between January 1986 and December 2010, prov iding drainage for the upper Caloosahatchee River watershed and Lake Okeechobee when rainfall totals were at their highest (Figure 4 7). The major peaks in mean monthly flow from S 79 occurred immediately prior to or following an episodic storm event, inc luding Hurricane Erin and Tropical Storm Jerry (July and August 1995), Hurricane Erika (August 2003), Hurricanes Charley, Frances, Ivan, and Jeanne (August and September 2004), and Hurricanes Katrina and Wilma (August and October 2005) (NOAA 2012). Althou gh it occurred outside of the usual wet/storm season, the January to March 1998 peak in mean monthly flow into the Caloosahatchee Estuary from S 79 was traced upstream to large releases of freshwater from Lake Okeechobee at S 77 and the upper Caloosahatche e River at S 78, during a strong El Nio period with higher than norma l rainfall levels (Figures 4 3, 4 6 4 7 and 4 8). Other periods of high rainfall did not always correspond to high mean monthly flows even though there were extreme single day freshwa ter releases contributing to the monthly mean, which was the case in June 1992 (Figures 4 6 4 7 and 4 8). In general, the daily flow rates at S 77 and S 78 were almost identical, while the volume released at S 79 was often three to four times greater th an that upstream around the

PAGE 123

123 same time periods (Figure 4 8). In comparison to the distribution of mean monthly inflows recommended for the Caloosahatchee Estuary by the South Florida Water Management District (SFWMD), more than 50% of the mean monthly infl ows observed between January 1986 and December 2010 were at levels recommended less than one percent of the time (Table 4 3). The closest agreement among the distributions of observed and recommended mean monthly inflows during the twenty five year period was for moderate rates between 22.7 to 42.5 m 3 1 Water Quality Variations Salinities (SAL) ranged from 0 to 38 psu in the Caloosahatchee Estuary over all observations from the twenty five year period. In general, salinities increased with distance fr om S 79 towards San Carlos Bay. Each region (UE, ME, LE, BY) encountered the entire range of salinities from freshwater to euhaline, except for region BY, which did not fall into th e freshwater zone (Figures 4 9 and 4 10). Oligohaline salinities were obs erved at the mouth of the estuary in region BY in June 2003, September 2003, July 2005, and November 2005, corresponding to mean monthly inflows at S 79 greater than 167 m 3 1 Euhaline salinities were observed in region UE in July 2007 when the average inflow was approximately 7 m 3 1 or less for that month and the six months prior. These extremely low and high regional salinities occurred in months having positive (El Nio) and negative (La Nia) ENSO phase conditions with above and below normal rainf all eve nts, respectively (Figures 4 3, 4 6 and 4 9). At other times, these non recommended high and low flow levels corresponded to salinities within the expected range for each region (Table 4 3; Figures 4 10 and 4 11). Total phosphorous (TP) concentrat 1 over the twenty 1 across the five

PAGE 124

124 salinity zones or four regions of the Caloosahatchee Estuary (Figure 4 12). Concentrations were generally lower and less variable in th e higher salinity zones (polyhaline and euhaline), particularly when these zones were contained within regi on BY (Figure 4 12; Tables 4 4 and 4 5). The highest monthly means were generally observed during the wet seasons of 2000, 2001, and 2007 correspond ing to periods of above average rainfall and/or inflow from S 79 in otherwise dr y La Nia phases (Figures 4 3, 4 6 4 7 4 8 and 4 12). The isolated TP peak in oligohaline waters of region UE in November 1996 occurred during a relatively dry month with r espect to rainfall and inflow levels, but it lagged a couple of weeks behind a period of heightened rainfall and freshwater releases at S 78 and S 79 (Figures 4 3, 4 6 4 7 4 8 and 4 12). 1 overall, 1 during the twenty five year period in the C aloosahatchee Estuary (Figure 4 13). Concentrations below this range generally occurred in the LE and BY regions when salin ities exceeded 18 psu (Figure 4 13; Tables 4 4 and 4 5). Concentrations above this range were more common in the UE and ME regions under freshwater and oligohaline conditions. Similar to the pattern observed with TP, the highest TKN monthly means (greater than 2 1 ) were associated with the wet seasons of 2000 and 2001. TKN concentrations also peaked in the UE and ME regions during the springs and early summers of 1987 and 1988, following several weeks of unusually high freshwater d ischarge that originated from S 77 during an otherwise dry period with below average rainfall (Figures 4 3 4 6 4 8 and 4 13).

PAGE 125

125 1 over the twenty five year period, 1 across the different regions and salinity zones (Figure 4 14). The largest variability was associated with the mesohaline salinity zone, which spanned all four regions of the estuary at some point between January 1986 and December 2010. In general, the highest SI concentrations (gre ater 1 ) were consistently found in the UE and ME regions of the estuary, respectively, when salinities were less than 5 psu (Figure 4 14; Tables 4 4 and 4 5). The major SI peaks were observed in the late summer/early fall of 1986 and the summe r of 2000 in all three inner regions of the estuary (UE, ME, and LE) when salinities ranged from 0 to 18 psu. The 1986 SI peaks followed a period of above average rainfall in June and several large pulses of freshwater released from S 79 in August, which both corresponded to the onset of a strong E l Nio event (Figures 4 3, 4 6, 4 7, 4 8, and 4 14). Photic depth ( Z p ) fluctuated both spatially and temporally with respect to salinity in the Caloosahatchee Estuary during the twenty five year time period (Figu re 4 15). Light availability in the water column was typically greater, as indicated by larger Z p values, at higher salinities (polyhaline and euhaline zones) and at distances farther from S 79 (regions LE an d BY) (Figure 4 15; Tables 4 4 and 4 5). Howev er, light conditions in these regions were also more variable between seasons and across years with values ranging from near 0 to approximately 6 m given a range of salinities created at the interface of freshwater outflow and tidal inflow. The monthly me an Z p values of 0.02 m 1 observed in the polyhaline and euhaline zones of regions BY and LE, respectively, were suspicious in that they corresponded to secchi disk depths ( Z s ) of

PAGE 126

126 0.01 m, which is the approximate thickness of a standard disk Photic depths were consistently shallow with respect to average water column depths and less variable across seasons and years in regions UE and ME, particularly under freshwater conditions. Mean concentrations of TSS over the twenty five year period showed the same p attern as Z p with the highest, long term averages associated with euhaline waters and regions LE and BY. Mean concentrations of CDOM followed an opposite pattern with the highest, twenty five year averages found in freshw ater and region UE (Tables 4 4 and 4 5). The long term averages of TURB decreased with distance from S 79 until region BY, while the highest mean concentrations were observed in both oligohaline and euhaline waters (Table 4 4 and 4 5). Phytoplankton Biomass and Productivity Patterns Ph ytoplankton biomass ( B ) in terms of chlorophyll a (CHL A ) concentration 3 ov 3 in 90% of individual collections from the twenty five year period. Blooms with mean CHL A concentrations greater than 3 o ccurred in all four regions of the Caloosahatchee Estuary and under all salinity conditions below 30 psu, excluding the euhaline zone (Figure 4 16). Freshwater blooms were primarily confined to region UE, although they were often flushed downstream into r egion ME where they were exposed to salinities up to 18 psu, which was the case in May and June 2000. Blooms with the highest CHL A concentrations were most often associated with oligohaline salinities, particularly in region ME in the early 2000s, and th is pattern was reflected in the spatial gradient of the mean CHL A concentrations over the twenty five year period with respect to salinity and distance from S 79 (Figure 4 16; Tables 4 4 and 4 5). The high summer and early fall peaks in CHL A from the ea rly 2000s occurred after rainfall and/or freshwater inflow increased in

PAGE 127

127 conjunction with the sporadic changes from dry La Nia to wet El Nio conditions (Figures 4 3, 4 6, 4 7 4 8 and 4 16). Under polyhaline and euhaline salinity conditions, there appea red to be limited potential for phytoplankton blooms in the Caloosahatchee Estuary during the twenty five year period (Figure 4 16). However, an unusual peak in B was detected in the polyhaline zone of region UE and nowhere else in the estuary at that mag nitude in November 2007, which was a very dry La Nia month with essentially no rainfall and no freshwater inflow from S 79 (Figures 4 3, 4 6 4 7, 4 8 and 4 16). Similarly, the maximum monthly mean CHL A concentration observe d in salinities greater tha n 30 3 in region LE in June 2008, corresponding to the end of an extended dry La Nia period with nearly no rainfall or freshwater inflow from S 79 (Figures 4 3, 4 6 4 7 4 8 and 4 16). Daily estimates of gr oss phytoplankton productivit y ( GPP d ) averaged 694 2 1 overall but exceeded 2000 2 1 in 5% of the samples collected during the twenty five year period (Figure 4 17). Monthly mean GPP d estimates greater than 2000 2 1 were observed in all four regions and ass ociated with all salinities, except in the euhaline zone. The maximum monthly mean GPP d estimate that occurred in salinities greater than 30 psu was 1200 2 1 from region LE in August 2001. The UE region exhibited most of the highest mean estimat es of GPP d and these outlying peaks were predominately found i n the mesohaline zone (Figure 4 17). The excessive GPP d peak in April 1989 under mesohaline conditions of region UE did not correspond to any unusually high inputs of freshwater from rainfall or S 79, but the one detected in June 2002 followed a dramatic increase in rainfall and inflow relative to the dry La Ni a period preceding it (Figures 4 3

PAGE 128

128 4 6, 4 7 4 8 and 4 17). The oligohaline zone had the highest average GPP d estimates overall, and the peaks associated with those salinities were observed in all three regions of the inner estuary (UE, ME, and LE) at various times between January 1986 and Decembe r 2010 (Figure 4 17; Tables 4 4 and 4 5). The spikes in GPP d in 1986, 1988, 1995, 1996, 20 00, 2001, 2009, and 2010 all occurred within a month or two of a notable switch in the ENSO cycle (Figu re 4 3 and 4 17). The BY region only exhibited mean GPP d 2 1 under polyhaline conditions, and these peaks also occu rred concurrently with El Nio and La Nia cycle changes. A close up view of the actual (as opposed to mo nthly mean) observations of CHL A and corresponding estimates of GPP d in two contra sting periods were compared to identify phytoplankton community resp onses to shifting salinity zones and associated water quality conditions given both natural and anthropogenic variations in freshwater inputs (Figures 4 18, 4 19 4 20 and 4 21). The late spring and early summer of 2000 coincided with an abrupt shift in the ENSO cycle from extended dry La Nia to temporary wet El Nio conditions (Figures 4 3, 4 6 4 7 and 4 8). Starting in April 2000, below average rainfall and low inflows from S 79 restricted phytoplankton biomass and primary productivity in the predom inantly mesohaline to euhaline salinities across the four regions of the Caloos ahatchee Estuary (Figures 4 18 and 4 19). With small inputs of rainfall and gradual increases in river inflows, salinities began to decline to freshwater and oligohaline levels in region UE and to mesohaline levels in region LE. Phytoplankton biomass and productivity showed an initial increase in oligohaline waters with the first blooms appearing at distances closest to S 79. By the end of May/beginning of June, both CHL A and GPP d had reached their seasonal peaks, which

PAGE 129

129 were most pronounced in regions UE and ME in fresh and oligohaline waters, respectively. The return of La Nia conditions caused rainfall, river flow, salinity, CHL A and GPP d to return to dry season levels d espite it being the month of June. The summer and fall of 1995, in contrast, was the final wet season in a five and half year long El Nio phase with well above average rainfall and excessive inflows from S 79 (Figures 4 3, 4 6 4 7 and 4 8). At the beg inning of June 1995, daily pulses of freshwater in the form of rainfall and river flow were promoting high levels of phytoplankton biomass and productivity in the oligohaline and mesohaline waters of the UE and ME regio ns, respectively (Figures 4 20 and 4 21). By early August, excessive inflows in conjunction with additional rainfall reduced salinities throughout the estuary from freshwater in region UE to polyhaline conditions in region BY. Blooms were flushed downstream and out to San Carlos Bay, where they reached peak GPP d levels CHL A concentrations for that region (BY). Continued freshwater inputs suppressed salinities below 18 psu, even in region BY, which restricted phytoplankton productivity and biomass accumulation through the end of the wet sea son. Spatial and Temporal Changes in Trophic Status The average, overall trophic status of the Caloosahatchee Est uary was mesotrophic (253 2 1 ) across space and time when all model estimates of GPP d from the twenty five year period were considered together and converted into average estimates of annual gross primary productivity ( GPP y ). When the estuary was divided into regions, the overall GPP y averages for regions UE and ME were higher (295 and 282 2 1 respectively), while the overall GP P y averages for regions LE and BY were lower (220 and 183 2 1 ), than that for the entire system, creating a general decreasing spatial gradient of mesotrophic productivity levels with distance from S 79

PAGE 130

13 0 (Figure 4 22; Table 4 4). When overall GPP y averages were determined for each salinity zone, the average trophic status of the freshwater, polyhaline, and euhaline zones of the Caloosahatchee Estuary was mesotrophic (194, 218, and 155 2 1 respectively), while the average trophic status of t he oligohaline and mesohaline zones of the system was eutrophic (350 and 322 2 1 respectively) across the twenty five year period (Figure 4 23; Table 4 5). From both spatial perspectives, GPP y ranged from oligotrophic to hypertrophic levels acros s the four regions or the five salinity zones within the Caloosahatchee Estuary between 1986 and 2010 (Figure s 4 22 and 4 23). These average, minimum, and maximum estimates of GPP y were taken as measures of the spatial and temporal changes in the trophic status of the Caloosahatchee Estuary despite the bias associated with missing data. Out of the twenty five year period, GPP y could not be determined in any region or zone of the Caloosahatchee Estuary for at least six of those years due to the discontinua tion of all sampling efforts (1990, 1991, 1992, 1993, and 1998) or the lack of sufficient measurements (2003) of CHL A and/or Z p required by th (Figures 4 22 and 4 23). An additional number of annual estimates were missing for some or all of the regions (UE, ME, LE, and BY) or salinity zones (freshwater, oligohaline, mesohaline, polyhaline, and euhaline). Furthermore, the values of GPP y for each year in each location were dependent on the number and seasonality of the monthly mean e stimates of GPP d used to calculate the annual means. For example, the minimum annual rates across the regions or zones were based on single, dry season month estimates of GPP d (i.e., November 1994 from region UE; April 1999 from the euhaline zone), while the maximum annual rates across the regions or zones were

PAGE 131

131 heavily weighted by wet season month estimates of GPP d (i.e., June, July, and August 1996 from region LE; June and July 2002 from the mesohaline zone). Considering the 13 years (12 for region BY) out of the twenty five year period that had seasonally diverse (both wet and dry season) estimates of primary productivity contributing to GPP y in the four regions of the Caloosahatchee Estuary, the system still fluctuated between states of oligotrophy (r egion UE in 2005) and hypertrophy (region LE in 1996) (Figure 4 24). Regions UE and BY dropped to minimum GPP y levels of oligotrophy (in 2005 and 2009, respectively), while regions ME and LE maintained mesotrophic levels of GPP y at a minimum over time. M aximum GPP y levels from each region corresponded to periods of hypertrophy in regions UE and LE (in 2000 and 1996, respectively), eutrophy in region ME (in 1996), and mesotrophy in region BY (in 2001). The years 1988, 1995, 1996, and 2000 produced eutroph ic and/or hypertrophic levels of GPP y in at least two of the three regions of the inner estuary (UE, ME, or LE). Eutrophic levels of GPP y were also detected in 2008 and 2009 in region UE and in 1986 in region ME. A smaller number of years had both wet an d dry season estimates of primary productivity contributing to GPP y in each salinity zone (Figure 4 25) due to insufficient sampling and/or shifting water quality conditions given interannual variations in rainfall and flow (i.e., certain salinity zones ma y have not existed in the estuary in a given year). Regardless, the Caloosahatchee Estuary showed the same range of GPP y (oligotrophy to hypertrophy) across the range of salinities (freshwater to euhaline). Minimum oligotrophic levels of GPP y produced in the freshwater and euhaline zones were comparable to those observed in regions UE and BY, while minimum mesotrophic levels of GPP y maintained in the oligohaline, mesohaline, and polyhaline zones were

PAGE 132

132 similar to those observed in regions ME and LE. The yea rs having eutrophic and/or hypertrophic levels of GPP y in the inner regions of the Caloosahatchee Estuary were associated with salinities between 0.5 and 30 psu. The oligohaline zone produced the highest maximum GPP y across both seasons in the year 2000. Overall, the breakdown of GPP y into seasonal averages for each year and region or zone revealed that wet season productivity was generally higher and more variable than dry season productivity across the Caloosahatchee Estuary during the twenty five year time period (Figure s 4 24 and 4 25). T he productivity potential of the Caloosahatchee Estuary was also examined given CHL A concentrations at the threshold for a violation of the Florida Department of Environmental Protection (FDEP) Impaired Water Rule ( Bailey et al. 2009, SFWMD et 3 the overall trophic status of the Caloosahat chee Estuary was expected to be meso trophic given average values of Z p (given the corresponding levels of phytoplank ton biomass) and I 0 observed during the twenty five year period (Figure 4 26). When spatial differences in photic depth were taken into consideration the productivity potential of region UE and the mesohaline zone increased to eu trophic levels overall. When predicted GPP y values from each region and zone were broken down into the seasonal contributions to productiv ity across the wet and dry month s given spatial and seasonal differences in Z p and I 0 the trophic status of the Caloosahatchee Estuary was ma intained at mesotrophic levels or lower. Across the spatial and seasonal differences, wet season productivity was consistently higher than dry season productivity.

PAGE 133

133 Phytoplankton Biomass Variability at Different Scales Patterns of phytoplankton biomass v ariability were examined according to the approach of Cloern and Jassby (2010) using a subset of regional monthly mean CHL A concentrations selected from the original Caloosahatchee Estuary data set according to the three data criteria. Only 8 out of 25 y ears of data for each region had at least 9 monthly CHL A values (taken as means when there were multiple collections in the same month) with each season ( i.e., winter, spring, summer, and fall) represented and each of these months sampled in at least 6 of those years. Regional monthly mean CHL A concentrations from the years 1986, 1987, 1988, 1995, 2004, 2005, 2006, and 2010 showed notable similarities and distinctions with respect to the annual, seasonal, and residual components of phytoplankton biomass variability in the Caloosahatchee Estuary (Figures 4 27 4 28 4 29 and 4 30). Each of these years had at least one prominent peak in the monthly mean CHL A time series that stood out against the long term averages taken over the eight years of data in a t least one region of the estuary. Several of these peaks were shared between regions, occurring simultaneously or within a span of one or two months of being first observed. Over the eight years of data included in this analysis, region ME had the high est long term mean CHL A 3 (Fi gures 4 27 4 28 4 29 and 4 30; part A). Region UE had the next highest long term CHL A 3 3 3 respectively. These long term CHL A means taken across the eight year subset of data were comparable in magnitude and spatial distribution to the average CHL A concentrations calculated over all individual samples collected in each region of the Caloosahatchee Estuary during the twenty five year period (Table 4 4). Thus, the patterns in phytoplankton biomass

PAGE 134

134 variability observed in the eight year subset of data were considered representative of the general responses of the phytoplankton community to the variety of mete orolog ical and hydrological conditions observed in the original twenty five year Caloosahatchee Estuary data set. The annual ( y ) component of phytoplankton biomass variability showed a relatively narrow range across four regions of the Ca loosahatchee Estuary ( Figures 4 27, 4 28 4 29 and 4 30 ; part C). Annual variability in CHL A in the Caloosahatchee Estuary was small ( y did not deviate much from 1) except for several years that stood out in the eight year data subset. The highest ( y >>1) mean CHL A concentr ations in the eight year subset were observed in 2004 in region ME. Above average CHL A concentrations (relative to the long term mean over the eight years) were also observed at a lesser magnitude in 2004 in regions LE and UE, respectively, while the sam e year brought below average CHL A concentrations to region BY. Concentrations of CHL A were relatively high in 1995 in all regions of the Caloosahatchee Estuary, particularly in regions BY and UE, respectively. The lowest ( y <1) mean CHL A concentrations in the eight year subset were observed in region UE and ME in 2005, although this year brought above average CHL A concentrations to reg ion BY. Below average mean CHL A was observed at similar levels in all four regions of the Caloosahatchee Estuary in t he year 2006. From a seasonal perspective, the monthly ( m ) component of phytoplankton biomass variability was fairly uniform across the four regions of the Caloosahatchee Estuary with the wet season months having mean CHL A concentrations above th e annua l average (Figures 4 27, 4 28 4 29 and 4 30 ; part D). Throughout the estuary,

PAGE 135

135 there were one to three consecutive wet season months th at maintained above average CHL A concentrations ( m >1), and the timing of these peaks moved downstream with the flow of freshwater. The highest CHL A averages occurred first in May, June, and July in region UE and were followed by above average CHL A concentrations in June, July, and August in region ME. Region LE had its seasonal increase in phytoplankton biomass in Jul y and August, and region BY had the latest rise in CHL A concentrations in August and September. The regions differed seasonally with respect to the timing of below average CHL A concentrations. Regions LE and BY had the lowest mean CHL A concentrations in the winter months (December, January, February), but regions UE and ME had their lowest mean CHL A concentrations in the fall (September, October, November). The residual ( ) component of phytoplankton biomass variability identified that 40 to 45% of th e monthly mean CHL A concentrations in the eight year subset were higher than expected for the given month and year ( >1) within each of the four regions of the Caloo sahatchee Estuary (Figures 4 27, 4 28 4 29 and 4 30 ; part B). Region ME had two months with unusually high CHL A monthly means ( >3), but only one of these months (June 1995) corresponded to a prominent peak in the CHL A time series. The unusually high CHL A concentration detected in region ME in January 2005 was overshadowed by the magnitu de of other monthly mean spikes in the phytoplankton biomass time series of that and other regions. The residual component confirmed that several of the other noticeable monthly mean peaks (August 1988 in regions ME and BY, December 1988 in region UE, Jun e 1995 in regions UE, June/July/August 2004 in regions ME and LE, and April 2006 in region ME) were also high ( >2) for the given

PAGE 136

136 month and year. A number of other deviations from the annual and seasonal patterns of phytoplankton biomass variability did not correspond to recognizable bloom events (i.e. monthly mean CHL A 3 ), a nd most of these cases occurred in the cooler and drier months of the year (November through April) when concentrations were typically lower in comparison. The standard deviations of y (SD y ) and m (SD m ) were taken as measures of annual and seasonal variabi lity, respectively (Cloern and Jassby 2010). Values of SD y showed a broader range (0.32 to 0.51) than the values of SD m which were more uniform (0.36 to 0.43) across the four regions of th e Caloosahatchee Estuary (Table 4 6). Seasonal variability had a larger influence on phytoplankton biomass patterns than annual variability overall and in all regions of the Caloosahatchee Estuary except for region ME. The residual component was more variable than the annual component y and seasonal component m with a standard deviation (SD ) range of 0.45 to 0.71, and the consistently larger values of SD indicate there we re sources of CHL A variability beyond the average seasonal patterns and annual mean fluctuations (Cloern and Jassby 2010). In comparison to the p atterns of phytoplankton biomass variability analyzed by Cloern and Jassby (2010) across 84 sites sampled within 51 estuarine coastal ecosystems, the four regions of the Caloosahatchee Estuary had moderate (regions UE, LE, and BY) to high (region ME) level s of annual variability, moderate levels of seasonal variability (regions UE, ME, LE, and BY), and low (region BY), moderate (regions UE and LE), and high (region ME) levels of residual variability (Table 4 6). The values of SD y and SD m position the four regions of the Caloosahatchee Estuary in the middle of the pool of 84 other sites sampled within 51

PAGE 137

137 other estuarine coastal ecosystems, in terms of the degree of natural and/or human disturbance and the influence of annual climate cycles, respectively, on CHL A variability (Figure 4 31). The circle sizes representing the median SD value for the four regions of the Caloosahatchee Estuary wer e also moderate in comparison to levels of nutrient enrichment in other systems around the world (Figure 4 31). Environmental Drivers of Phytoplankton Productivity Data sets created for combinat ions of 4 regions and 14 parameters across 33 time periods were used in analyses designed to identify environmental drivers correlated with patterns in GPP d Correlations with patterns in phytoplankton productivity were explained using a pool of 48 enviro nmental driver combinations representing the possible influences of 12 of the 14 different environmental variables (i.e., SAL, TEMP_W, DO, TP, TKN, SI, CDOM, TURB, TSS, MEI, total monthly rainfall, and mean monthly flow) in the 4 different regions of the C aloosahatchee Estuary. The parameters Z p and I 0 were not considered because they were used along with CHL A to estimate GPP d A combination of 11 variables across the four regions yielded the highest correlation ( = 0.525) with the spatial temporal patte rn in GPP d (Table 4 7). In region UE, the variables MEI, total monthly rainfall, CDOM, TSS, TEMP_W, and TURB were correlated with estimates of GPP d Concentrations of S I and TEMP_W were correlated with estimates of GPP d in region ME. Values of MEI and S AL correlated with primary productivity in region LE. In region BY, MEI values were most strongly correlated with primary productivity.

PAGE 138

138 Discussion Relationships between Phytoplankton Productivity and Environmental Variables Correlations between GPP d and s elected environmental parameters, encompassing a variety of meteorological, hydrological, and physical chemical conditions that have been traditionally considered in similar analyses (Boynton et al. 1982), were examined to gain insights into possible drivi ng factors of phytoplankton productivity in the Caloosahatchee Estuary. Out of 48 possible driver combinations, representing the potential influence of 12 environmental parameters on GPP d in each of the 4 regions, 11 provided the best overall correlation to the spatial temporal patterns in phytoplankton productivity in the Caloosahatchee Estuary. Some regional differences were observed in the nature of these relationships. Climate and weather had a broad influence in the Caloosahatchee Estuary on monthly mean GPP d estimates. ENSO patterns (with respect to MEI values) were correlated to primary productivity throughout the Caloosahatchee Estuary, except in region ME, while rainfall played an important role in region UE only. Water temperature, in combinat ion with other variables, was correlated to monthly mean GPP d estimates in regions UE and ME. Salinity influenced the pattern in primary productivity in region LE only. Monthly mean estimates of GPP d were partially correlated to light availability (with respect to CDOM, TSS, and TURB) in region UE only. Lastly, SI was the only nutrient correlated to the patterns in primary productivity, and this relationship was significant in region ME only. Climate and weather Climate and weather are naturally interrel ated due to the impact of alternating phases of the ENSO cycle on the displacement of atmospheric heat overlaying warm ocean waters, which change global atmospheric circulation and, thus, regional weather

PAGE 139

139 conditions ( TAO 2011 ). Because ENSO cycles are cha racterized by changes in wind, temperature, and rainfall, it is difficult to dissect the correlations found between MEI values and patterns in phytoplankton productivity in the Caloosahatchee Estuary. However, the impacts of alternating ENSO cycles and the ir respective strengths on the climate and weather in regions outside the equatorial Pacific are evident in shifting seasonal temperatures and precipitation patterns (IRI 2007). Rainfall and temperature were unique variables in this correlation analysis a s they were considered to be inherent components of the MEI metric of ENSO and independent environmental drivers of phytoplankton productivity; they were therefore examined separately. Wind was not considered as a separate variable in the correlation anal ysis since the Caloosahatchee Estuary is shallow and presumably well mixed under normal conditions with the help of tidal exchange (McPherson et al. 1990). In addition, variations in wind fields operate at very short time scales, making comparisons with o ther variables in the data set tenuous. Temperature The effect of temperature on phytoplankton productivity is largely due to the role of enzymes in photosynthesis. Photosynthetic activity increases with temperature until an optimum level is reached. Opt im um temperature levels vary among species in different environments, but a range of 20 to 25C is widely accepted for phytoplankton (Day et al. 1989). As a subtropical system, seasonal variations in water temperatures in the Caloosahatchee Estuary are le ss dramatic than in temperate estuaries, providing greater year round photosynthetic potential. The highest productivity rates and biomass concentrations were observed in the warmest months of the year as expected (Boynton et al. 1982), however, blooms wi th CHL A 3 were

PAGE 140

140 also detected in cooler months, confirming year round potential for high rates of productivity. The impacts of El Nio and La Nia on temperatures in the Caloosahatchee Estuary were more apparent in the winter, as is the case in other systems at latitudes above the tropics ( TAO 2011 ). Cooler than normal winters were more common in El Nio periods, and warmer than normal winters were more common in La Nia periods, following the expected outcomes of alte rnating ENSO phases in Florida (Jones et al. 1999) and in temperate areas (IRI 2007 TAO 2011 ). Variations in air temperatures are quickly matched in the water column of a system as shallow as the Caloosahatchee Estuary, potentially augmenting or hinderin g primary productivity. Strong terrestrial watershed inflows and small tidal exchanges in regions UE and ME make these areas even more susceptible to shifts in water temperatures. Rainfall Rainfall plays an important and widely recognized role in the r egulation of phytoplankton biomass and primary productivity in estuaries around the world (Mallin et al. 1993, Zingone et al. 2010). The delivery of freshwater to estuaries via rainfall naturally replenishes nutrient supplies, increases turbidity, and aid s in the transport of phytoplankton biomass (Zingone et al. 2010). However, the degree to which rainfall impacts estuarine water quality and corresponding responses of the phytoplankton community is dependent on anthropogenic manipulations of freshwater i nflow (Alber 2002). In estuaries with relatively few upstream modifications, rainfall is directly coupled with flow, with inputs following a natural, seasonal, and relatively consistent pattern with respect to the timing, quantity, quality, and distributi on of the delivery. The situation in the Caloosahatchee Estuary, and the one commonly associated with many estuaries,

PAGE 141

141 involves dams, diversions, and withdrawals, which result in unnatural freshwater inflow delays, shortages, and/or purges, depending on th e needs of the surrounding human population (Alber 2002). The Caloosahatchee Estuary is a classic example of an extensively modified system with obstructed and managed inflows that either hinder or exacerbate the natural patterns of rainfall. Patterns i n rainfall and flow were relatively consistent and well coupled on a seasonal basis, with both peaking during the wet season months of May through October. However, unusually high rainfall events associated with strengthening El Nio phases and episodic s torms did not always translate to high mean monthly inflows. In these cases, rainfall was likely retained in the Caloosahatchee Watershed and used for agriculture and public consumption. Conversely, mean monthly inflows were sometimes high when rainfall was at or below average. High volume releases done outside of the normal seasonal and annual cycles are typically used to flush algal blooms, salt water, and other contaminants out of the Caloosahatchee River, which was likely the case in the spring of 19 98. As an independent variable in the examination of environmental drivers of phytoplankton productivity, total rainfall was only correlated to GPP d in region UE of the Caloosahatchee Estuary. This region encompasses a 14 km stretch of the narrowest and d eepest portion of the estuary closest to S 79 (Scarlatos 1988). The upper estuary also has the least amount of industrial, commercial, and residential development in comparison to the downstream regions, and it is home to a national wildlife refuge (SFWMD et al. 2009). The delivery of freshwater via runoff from the northern and southern shores of this region is likely synchronized with minimal time delays of actual

PAGE 142

142 correspondi ng changes in water quality. This correlation was less evident in the downstream regions of the Caloosahatchee Estuary possibly due to the extensive urban channelization network in Cape Coral and Fort Myers (Stoker 1992), which causes rainfall runoff that would have been historically retained by an undeveloped watershed to reach the estuary in higher volumes and in less time (Barnes 2005). The fact that mean monthly inflows were not correlated to patterns in phytoplankton productivity, while measures of EN SO and rainfall were, is most likely related to the timing discrepancies between weather events, regulated flow schedules, and sample collections. In this study, mean monthly inflows were used to represent the hydrological conditions in the Caloosahatchee Estuary for an entire month, while the collection of samples occurred erratically between the beginning, middle, and end of each month over the twenty five year time period. Depending on the variability of flow and the corresponding changes in water resi dence times and water column characteristics (i.e., salinity, nutrient levels, light attenuation), GPP d estimates may have occurred under flow conditions not accurately depicted by a monthly mean. For example, samples collected within the first few days o f a month may have been better correlated to average flow rates from one month prior. Similarly, samples collected at the end of a month under flow conditions that were vastly different than the rest of the month would have been tested for correlations ag ainst higher/lower mean rates than the true conditions at the time of collection. Salinity Fluctuating salinity is a major variable in the Caloosahatchee Estuary in terms of the structure and function of phytoplankton, as well as other key biota (SFWMD e t al.

PAGE 143

143 2009). Salinities that exceed or fall short of the physiological requirements of organisms can cause stress and/or mortality. For example, oysters and submerged aquatic vegetation (tapegrass and seagrasses) have been identified as key species and e xtensively studied for their adaptability to altered salinity in the Caloosahatchee Estuary (Chamberlain and Doering 1998a). High salinities increase the susceptibility of oysters to predation and parasites and decrease the distribution of tapegrass ( Vall isneria americana ). On the other hand, prolonged low salinities can kill oysters, shoal grass ( Halodule wrightii ), and turtle grass ( Thalassia testudinum ). These key species are valuable ecosystem components (VECs) that provide important habitats, nurser y grounds, and food sources for many fish, invertebrates, birds, and other organisms in the Caloosahatchee Estuary, making it a recreationally and commercially valued system worth studying and monitoring. Phytoplankton are also sensitive to changes in sali nity, although many estuarine species are euryhaline and capable of adapting to natural variations in salinity (Brand 1984). Two of the dominant cosmopolitan species observed in the Caloosahatchee Estuary (Chapter 3), Akashiwo sanguinea and Skeletonema cf costatum are able to grow in salinity ranges of 10 to 40 psu and 15 to 45 psu, respectively, although their growth rates peak at salinities of 20 and 25 psu, respectively (Brand 1984, Matsubara et al. 2007). With these salinity preferences, it is not s urprising that these two species are common features of the Caloosahatchee Estuary (Chapter 3, Saunders et al. 1967, McPherson et al. 1990, Montgomery et al. 1991). Phytoplankton assemblages in the Caloosahatchee Estuary often follow a salinity gradient, from more fresh to more marine dominated communities, which is the case in

PAGE 144

144 other river dominated estuaries in Florida (Quinlan and Phlips 2007). The spatial and temporal extent of phytoplankton assemblages are altered by seasonal patterns and episodic eve nts that change freshwater inflow and the associated physical chemical properties of the water column (Quinlan and Phlips 2007). Spatial and temporal variation in salinity in the Caloosahatchee Estuary is primarily related to the quantity, quality, timing and distribution of freshwater via rainfall and river flow. Tidal exchange is believed to play a secondary role in the determination of salinity in the estuary (Bierman 1993, Chamberlain and Doering 1998b), particularly in the Bay and the lower regions of the system. During the twenty five year period, the Caloosahatchee Estuary shifted between fresh and euhaline conditions, with the size and position of the different salinity zones changing in response to rainfall, freshwater inflows, and tidal exchang es. Lower salinities and the associated higher nutrient concentrations and shallower photic depths were generally observed closer to S 79 but spread throughout the estuary during flood periods, while higher salinities and the associated lower nutrient con centrations and deeper photic depths were observed closer to the Gulf of Mexico and throughout the estuary during drought periods. Patterns in primary productivity were partially explained by salinity in region LE only. The relationship between GPP d an d salinity in this region was negative, so that higher phytoplankton productivity was associated with lower salinities. The highest peaks in GPP d in region LE were more commonly observed under mesohaline conditions. A wide range of moderate to excessive S 79 flows (14.2 to 127.4 m 3 1 ) produce mesohaline salinities in region LE (Bierman 1993) due to the larger, and more variable influence of tidal mixing in the lower extent of the Caloosahatchee Estuary.

PAGE 145

145 Phytoplankton productivity in this region respond s t o shifts in nutrient and light availability and the corresponding changes in water residence times given the balance between freshwater inflows and tidal exchanges. Lower (or higher) flow rates in the given range would increase (or decrease) the extent of tidal mixing in the lower estuary, decreasing (or increasing) nutrient availability while increasing (decreasing) light availability through the dilution of the nutrient rich, highly colored freshwater inputs. Residence times under this range of flows would fluctuate between 1 and 3 days (Qiu et al. 2007) in region LE, accommodating the growth of both faster (diatoms) and slower (dinoflagellates) species under varying conditions. Light Both the quantity and quality of light penetrating the water column is important to phytoplankton productivity. The quantity of light reaching the surface waters of a system varies naturally on a daily and seasonal basis. Because of the subtropical location of the Caloosahatchee Estuary, this system receives relatively h igh levels of incident solar radiation year round. The transmission of light through the water column depends on the levels of light absorbing constituents. The water itself, color dissolved organic matter (CDOM), algae, and tripton (i.e., non algae susp ended solids) create a disparity between the total intensity of photosynthetically active radiation (PAR) and the portion that is photosynthetically usuable radiation (PUR) depending on the absorption spectrum of the phytoplankton (Kirk 1994), driving phyt oplankton adaptations at the cellular level and/or species succession at the community level. Light availability in the water column varied spatially and temporally with respect to Z p values, representing the depth of 1% surface irradiance where gross prod uction and algal respiration rates are equal (Day et al. 1989). Photic depths were deeper in

PAGE 146

146 the water column in higher salinity zones of the Caloosahatchee Estuary, particularly in regions LE and BY. Tripton was the major component of light attenuation in these regions, as reflected in the relatively high values of TSS and TURB relative to CHL A and CDOM. In the freshwater and upper portions of the Caloosahatchee estuary, photic depths were shallower and less variable given the large fluctuation in rain fa ll and freshwater inflows via S 79 during the twenty five year period. Smaller values of Z p at distances closer to S 79 and in lower salinities corresponded to high average concentrations of CDOM, confirming the influence of highly colored humic waters on light attenuation under these conditions in the Caloosahatchee Estuary (Doering and Chamberlain 1999, Doering et al. 2006). Parameters affecting light availability provided important correlations to GPP d throughout the Caloosahatchee Estuary, except i n region BY. Photic depths in San Carlos Bay were typically deep enough to preclude significant periods of light limitation of photosynthesis due to the dilution of humics via tidal mixing. As a result, phytoplankton productivity was likely limited by fa ctors other than light in region BY, such as nutrient availability or water residence times. In contrast, patterns in GPP d were negatively correlated with patterns in CDOM in region UE, highlighting the influence of humics on the attenuation of PAR, as ob served in other highly colored aquatic environments (Kirk 1994, Falkowski 1994). Nutrients Nitrogen is the most important nutrient limiting factor for phytoplankton production in the Caloosahatchee Estuary (McPherson and Miller 1990, Doering et al. 2006, H eil et al. 2007), as in many other estuaries and coastal ecosystems worldwide (Boynton et al 1982, Cloern 2001). Being a subtropical system, the Caloosahatchee

PAGE 147

147 Estuary is acutely vulnerable to nitrogen pollution via agricultural runoff since lower latitude systems are exceeding global consumption trends in fertilizer use (Beman et al. 2005). The potential for nitrogen limitation is accentuated by excess phosphorous loads from natural weathering and anthropogenic mining of local phosphate rich limestone roc k (Odum et al. 1998). Silica has been considered for its secondary role in shaping the structure of the phytoplankton community given its use in the growth and maintenance of diatom cells. Previous researchers in the Caloosahatchee Estuary have found t hat the distribution of nutrients in this system is linked to riverine inputs, freshwater runoff, tidal exchange, and nutrient recycling (McPherson and Miller 1990). The Caloosahatchee River via S 79 plays the dominant role in nutrient distribution by co ntributing substantial nutrient loads and flushing nutrients downstream and out of the system (McPherson and Miller 1990). Rainfall works in conjunction with river flow to control the supply of new nutrients to the Caloosahatchee Estuary given the urban, agricultural, and environmental demands on water in the system and surrounding watershed. Six wastewater treatment facilities also discharge directly in the Caloosahatchee Estuary, but their combined average daily nutrient loads are less than those of S 79 by an order of magnitude in both the wet and dry seasons (Doering et al. 2006, Bailey et al. 2009). Higher concentrations of TKN, TP, and SI were associated with lower salinities and at distances closer to S 79, reflecting the influence of freshwater inflow on nutrient loading and spatial distribution in the Caloosahatchee Estuary between January 1985 and December 2010. Similar results have been previously reported for shorter timescales (McPherson and Miller 1990, Doering and Chamberlain 1998, Doerin g and

PAGE 148

148 Chamberlain 1999, Doering et al. 2006). Long term trends in the concentrations of these macronutrients were not statistically tested in this study due to the discontinuity of the pooled data set, but no visual patterns were evident in the twenty fiv e year time series of monthly means. Previous efforts to identify changes in nutrient levels in the Caloosahatchee Estuary over time was accomplished by comparing three hydrologically different periods of discontinuous data, including a relatively dry peri od from 1986 to 1989, a relatively wet period from 1994 to 1996, and an intermediate period from 1999 to 2003 (Doering et al. 2006). Across the four regions of the Caloosahatchee Estuary, salinity was lowest during the wet period, while concentrations of dissolved inorganic nitrogen (DIN) and dissolved inorganic phosphorous (DIP) were lowest and highest, respectively, during the dry period (Doering et al. 2006). These changes in water quality were not reflected in concentrations of CHL A (Doering et al. 2 006), possibly indicating the role of other factors in the regulation of phytoplankton biomass and, thus, primary productivity in the Caloosahatchee Estuary. In this study, silica was the only nutrient correlated with the patterns in phytoplankton produc tivity in the Caloosahatchee Estuary. Higher average estimates of GPP d were generally observed in conjunction with higher average concentrations of SI, but the relationship in region ME was less clear. A slight negative correlation between monthly mean c oncentrations of SI and monthly mean estimates of GPP d in this part of the estuary may point to cases of model over simplification caused by differences in photosynthetic efficiency among the species present in the Caloosahatchee Estuary (Chapter 3). It i s possible to hypothesize that diatoms, which typically dominate in regions LE and BY (Chapter 3), are less productive in region ME in comparison to other

PAGE 149

149 phytoplankton groups, regardless of the availability of silica, if some other factor, such as salinit y stress or light quality affects their growth potential. Patterns of Phytoplankton Biomass Variability As one of the major variables used in the adapted BZ p I 0 model to estimate GPP d in this study, it was important to consider trends in CHL A levels in t he Caloosahatchee Estuary. Phytoplankton biomass, in terms of CHL A concentrations, explained 80% of the variability in GPP d during the model development period (Chapter 3), so variations in this parameter alone may partially explain variations in GPP d ov er the twenty five year period. The partitioning of the variability associated with the eight year subset of regional monthly mean CHL A concentrations revealed that phytoplankton biomass in the Caloosahatchee Estuary fluctuated both temporally and spatia lly in response to seasonal cycles, annual disturbances, and residual events. The greatest contribution to phytoplankton biomass variability in the Caloosahatchee Estuary was the residual component, with the seasonal and annual components explaining appro ximately one third to one half of the CHL A pattern observed during the eight years selected from the original twenty five year data set, respectively. In comparison to the median values d within 51 estuarine coastal ecosystems worldwide, the Caloosahatchee Estuary as a whole had relatively moderate influences of annual, seasonal, and residual variability on the patterns in phytoplankton biomass. Annual variability In general, the annua l component explained the smallest portion of the phytoplankton biomass variability in the Caloosahatchee Estuary during the eight year subset. However, in region ME, annual variability had a greater influence than the

PAGE 150

150 seasonal component. The findings of Cloern and Jassby (2010) suggest that sites having large annual variability have been disturbed by climatic shifts or human actions. Large scale changes in climatic conditions (i.e., temperature, winds, pressure, etc.) caused by oscillating phases of glob al climate patterns have altered pelagic and benthic grazing pressures on phytoplankton biomass and, consequently, the long term patterns in CHL A in places such as Gulf of Naples (Italy), Narragansett Bay (Rhode Island), Massachusetts Bay (Massachusetts) and South San Francisco Bay (California) (Cloern and Jassby 2010). Anthropogenic disturbances (e.g., habitat modifications, invasive species, etc.) that also altered trophic interactions (e.g., bivalve grazing) or nutrient availability have been linked to deviations from long term CHL A trends in Florida Bay (Florida), North San Francisco Bay (California), Ringkobing Fjord (Denmark), and Tampa Bay (Florida) (Cloern and Jassby 2010). In the Caloosahatchee Estuary, the year 2004 produced unusually high C HL A concentrations in region ME and, to a lesser extent, in regions LE and UE. Elevated CHL A concentrations above the annual mean were observed in the inner regions of the Caloosahatchee Estuary just prior to the passing of Hurricanes Charley, Frances, Ivan, and Jeanne. In conjunction with the start of the wet season and a strengthening El Nio, the estuary went from a very dry state in the month of May with almost no rainfall (total of 1.2 cm at 7.5 cm below normal) and river flow (monthly mean of 7.5 m 3 1 ) to a very wet state in the month of June with well above average rainfall (total of 38.3 cm at 13.5 cm above normal) and steady, moderate level releases of freshwater (monthly mean of 19.8 m 3 1 ). The corresponding delivery of nutrients to the es tuary likely caused the initial increase in CHL A concentrations observed in regions UE and ME in

PAGE 151

151 June. By mid July and into the beginning of August, average flow rates were maintained between 20 and 80 m 3 1 in preparation for the arrival of the first s torm. The resulting higher volume inflows would have reduced water residence times in the upper estuary to one or two days (Qiu et al. 2007) and flushed the nutrients and the phytoplankton downstream where light conditions would have been more favorable a fter dilution by tidal exchange, helping to explain the large CHL A peaks in the latter months in regions ME and LE. The annual decline in CHL A that was detected further downstream in region BY that same year could be explained by a subsequent crash of t he phytoplankton bloom since the water mass would have been depleted of its nutrients and unable to sustain production once it reached San Carlos Bay. Additionally, the high concentrations of phytoplankton biomass may have been grazed down by zooplankton, which are typically more densely concentrated in the higher salinity zones of the lower and outer regions of the Caloosahatchee Estuary (Chamberlain et al. 2003, Tolley et al. 2010). However, the zooplankton (and the phytoplankton themselves) would have been flushed out to the Gulf of Mexico under flows greater than 80 m 3 1 (Chamberlain et al. 2003, Tolley et al. 2010), which started in mid August and continued through the end of the storm season. A similar series of events occurred in 1995 when above a verage CHL A concentrations were observed in all four regions of the Caloosahatchee Estuary, although the largest annual deviation from the long term mean occurred in region BY. High monthly CHL A means were observed throughout the estuary as early as Apr il 1995, which corresponded to an early start to the wet season with above average rainfall (total of 13.6 cm at 10.9 cm above normal) during an El Nio period.

PAGE 152

152 Phytoplankton blooms were likely triggered by the supply of nutrients delivered through rainfa ll runoff, which was again replenished in June and sustained through August with the arrival of Hurricane Erin and Tropical Storm Jerry. The above average rainfall levels were coupled with subsequent increases in freshwater releases from Lake Okeechobee ( via S 77) and the upper watershed (via S 78) into the Caloosahatch ee Estuary (via S 79), which would have carried the phytoplankton biomass to the lower regions of the system and Bay. Because the high CHL A concentrations were first observed in the freshw ater and oligohaline zones, it is likely that the bloom originally formed upstream of S 79 and was subsequently flushed out to the Gulf of Mexico through the estuary. This latter hypothesis is supported by reports of the presence of blooms of the toxic bl ue green Microcystis during that summer in Lake Okeechobee and the Caloosahatchee River, which ultimately impacted the estuary (Burns 2008). At the same time, a major red tide Karenia brevis bloom was reported off the west coast of Florida between Tampa B ay and Charlotte Harbor (FDEP 2005, Walsh et al. 2006), which may have infiltrated region BY on incoming tides, explaining the presence of high CHL A concentrations in that region in 1995. Seasonal variability A general underlying seasonal pattern was dete cted in the CHL A time series of the four regions of the Caloosahatchee Estuary, but its contribution to the overall variability in phytoplankton biomass was relatively small and influence on the magnitude of mean CHL A concentrations changed over time. S uch is the case in many estuaries and coastal ecosystems worldwide, where regular seasonal patterns are commonly absent, less pronounced, or increasingly found to shift over time periods longer than a decade (Cloern and Jassby 2010). Irregularity in the t iming and amplitude of

PAGE 153

153 phytoplankton biomass seasonal variability has been attributed to the shallowness and connectivity of estuarine coastal ecosystems to land and sea, allowing unique and complex processes (e.g., upwelling, tidal mixing, wind driven re suspension, river flow, hydraulic manipulations, nutrient inputs, benthic grazing, species introductions, etc.) to interject the annual climatologically driven cycles of biomass growth and senescence that function in the open ocean and on land (Cloern and Jassby 2008). In the Caloosahatchee Estuary, regional monthly mean CHL A concentrations were typically higher in the wet summer and early fall months, and the timing of these peaks followed the downstream flow of freshwater from the upper estuary (region UE) to San Carlos Bay (region BY). The recurrence of this pattern was attributed to the seasonal increase in rainfall during the months of May through October and the subsequent need for more frequent and higher volume freshwater discharges through S 79 as a way to maintain prescribed water levels in Lake Okeechobee and control flooding in the upper watershed. Rainfall runoff and regulatory releases provide a seasonal supply of new nutrients to the Caloosahatchee Estuary and the resident phytoplankton co mmunities, stimulating primary productivity there (Doering et al. 2006). Here, the wet season also coincides with annual peaks in temperature, solar radiation, and photoperiod, providing a combination of ideal phytoplankton bloom conditions. Phytoplankto n biomass and productivity tend to be highest during the warmer months of the year ( i.e., May through October) in a broad spectrum of estuarine coastal ecosystem types ( e.g., river dominated, embayments, lagoons, and fjords) (Boynton et al. 1982), but thes e annual climatology cycles have been more commonly associated with regular and high amplitude seasonal patterns in temperate estuaries worldwide,

PAGE 154

154 including South San Francisco Bay (California), North Inlet (South Carolina), and Oosterschelde (Netherlands) (Cloern and Jassby 2010). At 26N latitude, the influence of these climatologically driven forces on phytoplankton seasonal patterns in the Caloosahatchee Estuary is weakened (i.e. levels are relatively high throughout the year), and, therefore, cannot be solely responsible for the observed trends. Regional differences in the timing of the seasonal declines in mean CHL A concentrations further suggest the role of both natural and anthropogenic processes in the regulation of phytoplankton biomass in the Caloosahatchee Estuary. Low mean CHL A concentrations in the lower estuary (region LE) and San Carlos Bay (region BY) during the winter were likely tied to natural declines in temperature, solar radiation, and nutrient supplies due to decreased rainfall and river inflow. However, the seasonally low mean CHL A concentrations in the upper and middle estuary (regions UE and ME) 2006) caused by excessive anthropogenic fl ushing of rainfall reserves at the end of the wet season with the biggest events occurring in the months of August, September, and October. Alternatively, the inflow of nutrient rich, colored water may have first aided the accumulation of phytoplankton bi omass at the beginning of the season but later hindered the maintenance of blooms due to depleted nutrient supplies and/or decreased light availability from elevated CDOM concentrations and shading by the phytoplankton themselves. Phytoplankton biomass an d productivity declines in the late wet season have also been observed in neighb oring Charlotte Harbor and have been attributed to reduced light availability from increased river inflow color (McPherson et al. 1990). Zooplankton were likely not responsibl e for the decline of phytoplankton biomass in the

PAGE 155

155 late wet season months in regions UE and ME since their densities typically reach reported lows during this time and in these areas in response to increased freshwater inflow and reduced salinity (Chamberla in et al. 2003, Tolley et al. 2010). Residual variability The major contribution of CHL A variability from the residual component indicated that there was a high occurrence of deviations fr om average seasonal patterns of, and annual shifts in, productivit y in the Caloosahatchee Estuary (Cloern and Jassby 2010). This outcome can be considered the norm based on the findings of Cloern and Jassby (2010) in their analysis of 84 different CHL A time series from within 51 estuarine coastal ecosystems worldwide. Variability of this type can be great when the seasonal pattern changes strongly from year to year (Cloern and Jassby 2010). This explanation did not seem to apply in the Caloosahatchee Estuary since CHL A concentrations were consistently highest in the wet season months (May through October), creating a relatively uniform seasonal pattern across the twenty five year time period. Therefore, blooms, appearing as spikes in the CHL A time series (Cloern and Jassby 2010). The magnitude of these bloom events were greater in the inner regions of the Caloosahatchee Estuary than in San Carlos Bay (region BY). This p attern was likely explained by the fact that region BY had very low CHL A concentrations throughout the study period due to its relatively low concentrations of nutrients (i.e., TP, TKN, and SI). The few phytoplankton blooms that did occur in region BY we re likely allochthonous in nature since they were previously or simultaneously observed in upstream regions of the Caloosahatchee Estuary under natural and/or anthropogenic flushing events. In

PAGE 156

156 general, the formation and accumulation of phytoplankton bioma ss blooms throughout the Caloosahatchee Estuary were most commonly associated with periods of above average rainfall and/or freshwater inflow that stemmed from changes in ENSO conditions and/or episodic storm events. The major effect of weather and/or man agement action was the creation of high nutrient habitats, which facilitated the development of phytoplankton biomass blooms in this and other estuaries worldwide (Cloern and Jassby 2010). In all four regions of the Caloosahatchee Estuary, the residual c omponent of phytoplankton biomass variability identified unusually high mean CHL A concentrations for a given month and year that did not correspond to spikes in the time series. The occurrence of residual events outside of these prescribed bloom conditio ns was most likely the result of sampling errors that can be large when single measurements are used as estimators of monthly mean CHL A concentrations (Cloern and Jassby 2010), which was the approach used in this analysis. Infrequent and inconsistent sam pling efforts miss phytoplankton community responses to changes in water quality that occur on timescales shorter than a month, causing discrepancies in the scales at which variability can be explained (Cloern and Jassby 2010). Variations in Trophic Statu s The trophic status of the Caloosahatchee Estuary varied spatially and temporally between 1986 and 2010 given a wide range of meteoro logical, hydrological, physical, chemical, and biological factors likely influencing phytoplankton productivity. Taking i nto account all model estimates of phytoplankton productivity observed during the twenty five year period, average long term conditions in the Caloosahatchee Estuary were me s o 2 1 ). This overall trophic status classification is lower

PAGE 157

157 t 2 1 ) derived during the model development study (February 2009 to Feb r uary 2010, excluding March 2009) (Chapter 3), reflecting the added influence of interannual and decadal changes in water quality, beyond the previ ously detected seasonal/intraannual influence, on patterns of productivity in the Caloosahatchee Estuary. The previous annual estimate of phytoplankton productivity in the Caloosahatchee Estuary was based on field tested measurements taken across a phase shift between a relatively weak La Nia period to a relatively strong El Nio period corresponding to seasonally low and high inputs of freshwater into the system, respectively. In comparison, the long term overall estimate of GPP y derived from this study encompassed the effect of fifteen alternating ENSO phases, as well as the passing of more than twenty named storms, suggesting the importance of natural and anthropogenic inputs of freshwater on the regulation of phytoplankton productivity in the Caloosah atchee Estuary. When temporal and spatial differences in phytoplankton biomass and light availability (the primary variables used to estimate GPP y ) were considered, the trophic status of the Caloosahatchee Estuary varied between oligotrophic and hypertrop hic years between 1986 and 2010. Years producing eutrophic and/or hypertrophic levels of GPP y were associated with average annual CHL A concentrations greater than 11 3 the threshold set by the Florida Department of Environmental Protection (FDEP) f or defining impaired estuaries and coastal waters (Bailey et al. 2009, SFWMD et al. 2009). These high annual levels of productivity primarily occurred in years experiencing climatic shifts with respect to alternating ENSO phases, which is a

PAGE 158

158 recognized sou rce of annual variability in phytoplankton biomass and, thus, primary productivity (Cloern and Jassby 2010). Under the current total maximum daily loads (TMDLs) set by FDEP, the annual ave rage CHL A 3 has been linked to elevated nutrie nt levels (primarily TN) and reduced DO concentrations that are above and below, respectively, accumulations of phytoplankton biomass, and the high levels of phytoplankton pr of a healthy, well 2009). If CHL A levels are maintained below this water quality standard, the trophic status of the Caloosahatchee Estuary should not exceed mesotrophic levels given the spatial and temporal variations in light availability expected in the system. This target conform s to the Comprehensive Everglades Restoration Program (CERP) goal of avoiding transformations of oligotrophic systems into eutrophic ones (Boyer et al. 2009). These variations in trophic status observed during the twenty five year period may not fairly represent the range of productivity potential in the Caloosahatch ee Estuary due to the bias associated with missing data across both space and time. A better representation of the actual temporal and spatial variations in GPP y and the corresponding trophic status of the Caloosahatchee Estuary would have been depicted h ad the annual rates for each region or zone incorporated monthly estimates of productivity across each of the calendar seasons ( i.e., wint er = December through February, spring = March through May, summer = June through August, fall = September through Nov ember), or at the very least, from each of the more inclusive wet

PAGE 159

159 and dry seasons ( i.e., wet = May through October dry = November through April). The limitation of the available data is a (potentially large) source of error in the analysis of seasonal, i nterannual, and decadal trends in phytoplankton product ivity (Winder and Cloern 2010). Summary This study extended the usefulness of a lo ng term water quality data set encompassing the monitoring and research efforts of five programs/projects occurring dis continuously between January 1986 and December 2010 in the Caloosahatchee Field measurements of phytoplankton biomass, in terms of chlorophyll a concentrations, and li ght availability, with respect to incident PAR flux and photic depths, were compiled and used to estimate phytoplankton productivity using the BZ p I 0 approach (Cole and Cloern 1987) over the twenty five year period. The GPP d model previo usly adapted to the Caloosahatchee Estuary depicted variations in the seasonal, interannual, and decadal responses of the phytoplankton community to natural and anthropogenic influences in the surrounding watershed. Correlations between GPP d and selected environmental parameters were used to identify possible drivers of phytoplankton productivity in the Caloosahatchee Estuary. Climate and weather, with respect to ENSO cycles, rainfall, and water temperatures, largely influenced the patterns in phytoplank ton productivity throughout the system. Salinity, nutrients, and light availability also partially explained the patterns in GPP d although their influence varied spatially. Flow through the main water control structure was not correlated to GPP d likely due to the time lag between natural freshwater inputs,

PAGE 160

160 anthropogenic manipulations, and sampling events, although changes in productivity in response to river releases were observed on shorter time scales. Variations in CHL A with respect to seasonal patterns, annual disturbances, and residual events, were considered as likely explanations for variations in GPP d Spatial and temporal deviations in phytoplankton biomass and, thus, primary productivity, were primarily linked to temporally irregular, exc essive phytoplankton blooms that followed natural and anthropogenic inputs of new nutrients. Seasonal peaks in CHL A coupled with peaks in GPP d occurred during the warmer, wetter months, given elevated temperatures, light levels, and freshwater inputs. Years having the greatest concentrations of CHL A matched by high estimates of GPP d had been disturbed by climatic shifts in the ENSO cycle. These patterns in phytoplankton biomass and primary productivity corresponded to spatial and temporal shifts i n the trophic status of the Caloosahatchee Estuary from oligotrophic to hypertrophic levels. Allochthonous blooms, originating from the upper watershed and coastal waters likely contributed to the high levels of productivity through flushing events and ti dal exchanges. Mesotrophic levels of productivity are theoretically achievable when CHL A standard for water quality, protecting the ecological and economic value of the system. However, a more complete time series of water quality data and in depth analysis of these and additional factors controlling phytoplankton productivity in the Caloosahatchee Estuary is needed to confirm the appropriateness of current management and restoration effort s.

PAGE 161

161 Figure 4 1. Location and connection of the C aloo sahatchee E stuary, FL with respect to Lake Okeec hobee via the Caloos ahatchee River (C 43 Canal) and the Gulf of Mexico and Charlotte Harbor via San Carlos Bay. Figure 4 2. Four regions of the C aloosahatchee Estuary sampled in long term water quality monitoring programs/research projects.

PAGE 162

162 Table 4 1. Sampling sites used in water quality monitoring programs/research projects in the four regions of the Caloosahatchee Estuary, defined by their distan ce in kilometers from S 79, the Franklin Lock and Dam. For the CCHM project, site numbers represent reference points for the location of sampling grids. Samples for this project were taken randomly within a particular grid. For all other projects, site numbers represent specific and consistent locations (based on registered GPS points) where samples were collected. p roject Upper Estuary (UE) Middle Estuary (ME) Lowe r Estuary (LE) San Carlos Bay (BY) CAL 1, 2, 3, 4 5, 6, 7, 8 9, 10, 17 11, 12, 13, 14, 18 CES 2, 3, 4 5, 6 7, 8 9, 10, 11 CCHM 450, 451, 453, 454, 455, 456, 457, 461, 462, 463, 467, 468, 469, 470, 471, 472 413, 414, 417, 418, 419, 422, 423, 427, 428, 429, 430, 434, 436, 437, 441, 442, 443, 444, 447, 448, 449, 452 386, 387, 388, 389, 391, 3 92, 393, 394, 395, 396, 404, 411, 412, 464 --ERD 12, 13, 14, 15 9, 10, 11 6, 7, 8 3, 4, 5 HB 1 2, 3 4, 5 6 km from S 79 0 14 14 28 28 40 40 48 Table 4 2. Median salinity values observed in the four regions of the Caloosahatchee Estuary from December 1985 to May 1989, November 1994 to August 1996, and April 1999 to June 2003 (Doering et al. 2006) and the corresponding salinity zone classification. Upper Estuary (UE) Middle Estuary (ME) Lower Estuary (LE) San Carlos Bay (BY) salinity (psu ) --4.1 10.1 22.6 29.4 salinity z one 0 0.5 (A) f reshwater 0.5 5 (B) o ligohaline 5 18 (C) m esohaline 18 30 (D) p olyhaline 30 40 (E) e uhaline

PAGE 163

163 Figure 4 3. (A) Multivariate ENSO (El Nio/Southern Oscillation) Index (MEI) between January 1986 and December 2010. Positive MEI values represent the warm ENSO phase (El Nio), while negative MEI values represent the cold ENSO phase (La Nia). (B) Corresponding MEI ranks based on MEI values across the same bimonthly seasons since the beginning of r ecord in December 1949/January 1950.

PAGE 164

164 Figure 4 4. Average monthly air temperatures (C) (A) and departures from normal (C) (B) recorded at the meteorological station in Fort Myers, FL (from January 1986 to December 2010 (after U.S. National Climatic Da ta Center, Florida Climatological Data Annual Summaries).

PAGE 165

165 Figure 4 5. Actual water temperatures (TEMP_W in C), observed in the Caloosahatchee Estuary, FL between January 1986 and December 2010.

PAGE 166

166 Figure 4 6. Monthly rainfall totals (cm) (A) and depa rtures from normal (cm) (B) recorded at the meteorological station in Fort Myers, FL from January 1986 to December 2010 (after U.S. National Climatic Data Center, Florida Climatological Data Annual Summaries).

PAGE 167

167 Figure 4 7. Mean monthly flow (m 3 1 ) fro m S 79 into the Caloosahatchee Estuary, FL between January 1986 and December 2010.

PAGE 168

168 Figure 4 8. Daily flows (m 3 1 ) along the Caloosahatchee River between Lake Okeechobee and the Caloosahatchee Estuary at water control structures S 77 (A), S 78 (B), an d S 79 (C).

PAGE 169

169 Table 4 3. Recommended (SFWMD et al. 2009) and observed (January 1986 to December 2010) frequency distribution of mean monthly inflows to the Caloosahatchee Estuary, FL from S 79. flow l evel f low r ange (m 3 s 1 ) from S 79 recommended percent di stribution of f lows from S 79 observed p ercent d istribution of f lows from S 79 1 0 to 12.7 0% 30% 2 12.7 to 14.2 42.8% 4% 3 14.2 to 22.7 31.7% 12% 4 22.7 to 42.5 19.2% 14% 5 42.5 to 79.3 5.6% 16% 6 79.3 to 127.4 0.7% 13% 7 >127.4 0% 11%

PAGE 170

170 Figure 4 9. S alinity (SAL in psu) ranges divided into five zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) observed in the four regions (UE, ME, LE, BY) of the Caloosahatchee Estuary, FL between January 1986 and December 2010.

PAGE 171

171 Figure 4 10. S alinity (SAL in psu) observed in each of the four regions of the Caloosahatchee Estuary, FL between January 1986 and December 2010 with respect to mean monthly inflow (m 3 1 ) of freshwater at S 79.

PAGE 172

172 Figure 4 11. Predicted steady state salinity (SAL in psu) distribution along a spatial gradient in the Caloosahatchee Estuary, FL for a range of freshwater inflows (m 3 1 ) from S 79 (Bierman 1993).

PAGE 173

173 Table 4 4. Mean (with standard deviations) physical chemical, biomass, and producti vity values associated with the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL between January 1986 and December 2010. Standard deviations are provided in parentheses. Upper Estuary (UE) Middle Estuary (ME) Lower Estuary (LE) San C arlos Bay (BY) SAL (psu) 4.14 (5.78) 10.0 (8.49) 20.1 (9.91) 26.7 (7.05) DO 1 ) 6.36 (2.00) 7.45 (1.74) 7.06 (1.47) 7.52 (1.63) TEMP_W (C) 26.0 (3.97) 25.6 (4.12) 25.8 (4.23) 25.9 (4.25) CDOM (pcu) 88.0 (58.2) 63.2 (52.1) 37.8 (40.4) 17.8 (17.8) TURB (ntu) 4.35 (4.22) 4.20 (3.28) 4.18 (3.15) 5.12 (3.26) TS S 1 ) 8.55 ( 11.6) 11.8 (11.0) 18.4 (19.5) 18.2 (13.0) Z p (m 1 ) 1.77 (1.04) 2.05 (0.91) 2.57 (1.04) 2.55 (1.40) TP 1 ) 0.14 (0.08) 0.12 (0.06) 0.09 (0.07) 0.05 (0.02) TKN 1 ) 1.16 (0.45) 1.03 (0.47) 0.75 (0.49) 0.71 (0.44) SI 1 ) 6.34 (3.24) 4.64 (3.23) 3.18 (2.41) 1.53 (1.47) CHL A 3 ) 11.4 (13.4) 12.3 (22.8) 7.27 (11.1) 4.91 (5.05) GPP d (mg 2 1 ) 809 (1057) 774 (727) 602 (576) 500 (392 )

PAGE 174

174 Table 4 5. Mean (with standard deviations) physical chemical, biomass, and productivity values associated with the five salinity zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) of the Caloosahatchee Estuary, FL between January 1986 and December 2010. Standard deviations are provided in parentheses. freshwater o ligohali ne mesohaline polyhaline euhaline SAL (psu) 0.22 (0.10) 2.42 (1.30) 11.4 (3.77) 24.1 (3.48) 32.6 (1.79) DO 1 ) 5.57 (1.88) 7.24 (1.83) 7.49 (1.6 6) 7.40 (1.36) 7.53 (1.57) TEMP_W (C) 27.0 (3.65) 26.2 (4.07) 25.3 (4.07) 25.1 (4.28) 25.1 (4.30) CDOM (pcu) 135 (61) 88.2 (45.3) 50.4 (23.4) 25.0 (22.8) 11.0 (10.0) TURB (ntu) 4.10 (3.37) 5.00 (5.04) 4.59 (3.75) 3.95 (2.79) 4.74 (3.40) TSS (m 1 ) 5.25 (5.45) 8.09 (10.5) 12.7 (11.3) 17.1 (15.5) 24.0 (21.2) Z p (m 1 ) 1.38 (0.54) 1.74 (0.70) 2.13 (1.10) 2.64 (1.07) 3.07 (1.43) TP 1 ) 0.14 (0.08) 0.15 (0.10) 0.12 (0.06) 0.08 (0.05) 0.05 (0.04) TKN 1 ) 1.20 (0.41) 1.24 (0.42) 1. 04 (0.46) 0.74 (0.42) 0.55 (0.42) SI 1 ) 8.20 (2.42) 7.45 (2.87) 4.50 (2.60) 2.11 (1.62) 1.45 (1.15) CHL A 3 ) 8.94 (13.70) 15.0 (16.6) 11.4 (11.0) 5.90 (6.08) 3.29 (2.26) GPP d m 2 1 ) 531 (772) 958 (897) 882 (999) 598 (468) 424 (25 2)

PAGE 175

175 Figure 4 12. Monthly mean total phosphorous (TP in 1 ) across five salinity ( SAL ) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of the four regions (UE, ME, LE, BY) of the Caloosahatchee Estuary, FL between January 1986 and December 2010.

PAGE 176

176 Figure 4 13. Monthly mean total Kjeldahl nitrogen (TKN in 1 ) across five salinity ( SAL ) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of the four regions (UE, ME, LE, BY) of the Caloosahatchee Es tuary, FL between January 1986 and December 2010

PAGE 177

177 Figure 4 14. Monthly mean silica (SI in 1 ) across five salinity ( SAL ) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of the four regions (UE, ME, LE, BY) of the Caloosaha tchee Estuary, FL between January 1986 and December 2010

PAGE 178

178 Figure 4 15. Monthly mean photic depth ( Z p in m 1 ) across five salinity ( SAL ) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of the four regions (UE, ME, LE, BY) of th e Caloosahatchee Estuary, FL between January 1986 and December 2010

PAGE 179

179 Figure 4 16. Monthly mean chlorophyll a (CHL A in m 3 ) across five salinity ( SAL ) zones (freshwater, oligohaline, mesohaline, polyhaline, euhaline) in each of the four regions (UE, ME, LE, BY) of the Caloosahatchee Estuary, FL between January 1986 and December 2010

PAGE 180

180 Figure 4 17. Monthly mean daily gross primary productivity ( GPP d in 2 1 ) across five salinity ( SAL ) zones (freshwater, oligohaline, mesohaline, polyhaline, eu haline) in each of the four regions (UE, ME, LE, BY) of the Caloosahatchee Estuary, FL between January 1986 and December 2010

PAGE 181

181 Figure 4 18. Close up view of phytoplankton bio mass responses (in terms of CHL A 3 ) in variable sali nities (F = freshwater, O = oligohaline, M = mesohaline, P = polyhaline, E = euhaline) from April to June 2000 give n daily S 79 inflows and rainfall inputs in the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL.

PAGE 182

182 Figure 4 19. Close u p view of phytoplankton productivity responses (as GPP d estimates 2 1 ) in variable salinities (F = freshwater, O = oligohaline, M = mesohaline, P = polyhaline, E = euhaline) from A pril to June 2000 given daily S 79 inflows and rainfall inputs i n the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL.

PAGE 183

183 Figure 4 20. Close up view of phytoplankton bio mass responses (in terms of CHL A 3 ) in variable salinities (F = freshwater, O = oligohaline, M = mesohal ine, P = polyhaline, E = euhaline) from June to November 1995 given d aily S 79 inflows and rainfall inputs in the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL.

PAGE 184

184 Figure 4 21. Close up view of phytoplankton productivity responses (a s GPP d 2 1 ) in variable salinities (F = freshwater, O = oligohaline, M=mesohaline, P=polyhaline, E=euhaline) from June to November 1995 given daily S 79 inflows and rainfall inputs in the four regions (UE, ME, LE, and BY) of the Caloo sahatchee Estuary, FL.

PAGE 185

185 Figure 4 22. Modeled average ann ual gross primary productivity ( GPP y in 2 1 ) in the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL between 1986 and 2010. The numbers above the bars indicate the numbe r of months used to calculate the annual average. Estimates of GPP y could not be obtained using the BZ p I 0 model for years with insufficient measure ments of phytoplankton biomass ( B ) as CHL A con centration and/or photic depth ( Z p ) The solid and dashed li nes represent the overall, long term averages for the specific regions and the entire estuary, respectively.

PAGE 186

186 Figure 4 23. Modeled average ann ual gross primary productivity ( GPP y in 2 1 ) by salinity zone (freshwater, oligohaline, mesohaline, polyhaline, and euhaline) in the Caloosahatchee Estuary, FL between 1986 and 2010. The numbers above the bars indicate the number of months used to calculate the annual average. Estimates of GPP y could not be obtained using the BZ p I 0 model for years with insufficient measure ments of phytoplankton biomass ( B ) as CHL A con centration and/or photic depth ( Z p ) The solid and dashed lines represent the overall, long term averages for the specific zones and the entire estuary, respectively.

PAGE 187

187 Figure 4 24. Modeled average ann ual gross primary productivity ( GPP y in 2 1 ) in the four regions (UE, ME, LE, and BY) of the Caloosahatchee Estuary, FL between 1986 and 2010 with respect to season (DRY = November through April, WET = May through October). The diamonds represent the overall annual GPP y averages for the individual regions.

PAGE 188

188 Figure 4 25. Modeled average ann ual gross primary productivity ( GPP y in 2 1 ) by salinity zones (freshwater, oligohaline, mesohaline, polyhaline, and euhaline) of t he Caloosahatchee Estuary, FL between 1986 and 2010 with respect to season (DRY = November through April, WET = May through October). The diamonds represent the overall annual GPP y averages for the individual zones.

PAGE 189

189 Figure 4 26. A verage expected annual gross primary productivity ( GPP y in 2 1 ) for the Caloosahatchee Estuary, FL overall and for each region (UE, ME, LE, and BY) and salinity zone (freshwater, oligohaline, mesohalin e, polyhaline, and euhaline) Estimated c ontributions to the annual averages from each season (dry or wet) were based on 181 or 184 days per season, respectively Model estimates were calculated using the Florida Department of Environmental Regulation (FDEP) annual threshold for CHL A (SFWMD e t al. 2009) and mean photic depths and light flux levels observed in the given region, zone, and/or season. Colo rs identify the corresponding trop hic status (blue = oligotroph ic; yellow = mesotrophic; green (1995) classification scheme

PAGE 190

190 Figure 4 2 7 Region UE m onthly mean chlorophyll a concentrations ( CHL A in 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C), and seasonal, ( m ) (D) components of phytoplankton biomass variability in the Caloosahatchee Estuary, FL across the years 1986, 1987, 1988, 1995, 2004, 2005, 2006, and 2010.

PAGE 191

191 Figure 4 2 8 Region ME m onthly mean chlorophyll a conce ntrations ( CHL A in 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C), and seasonal, ( m ) (D) components of phytoplankton biomass variability in the Caloosahatchee Estuary, FL across the years 1986, 1987, 1988, 1995, 2004, 200 5, 2006, and 2010.

PAGE 192

192 Figure 4 2 9 Region LE m onthly mean chlorophyll a concentrations ( CHL A in 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C), and seasonal, ( m ) (D) components of phytoplankton biomass variability in the Caloosahatchee Estuary, FL across the years 1986, 1987, 1988, 1995, 2004, 2005, 2006, and 2010.

PAGE 193

193 Figure 4 30 Region BY m onthly mean chlorophyll a concentrations ( CHL A in 3 ) and the long term mean ( C ) (A) with the residual ( ) (B), annual ( y ) (C) and seasonal, ( m ) (D) components of phytoplankton biomass variability in the Caloosahatchee Estuary, FL across the years 1986, 1987, 1988, 1995, 2004, 2005, 2006, and 2010.

PAGE 194

194 Table 4 6. Standard deviations of the annual ( SD y ), seasonal ( SD m ), and residual ( SD ) components of phytoplankton biomass variability in the Caloosahatchee Estuary, FL in comparison to the median values of SD y SD m and SD obtained from the analysis of chlorophyll a variability in 84 sites sampled within 51 estuarine coastal ecosystems (Cloern and Jassby 2010). Upper Estuary (UE) Middle Estuary (ME) Lower Estuary (LE) San Carlos Bay (BY) m edian (Caloosahatchee Estuary) m edian (Cloern and Jassby 2010) SD y 0.32 0.51 0.36 0.33 0.35 0.30 SD m 0.36 0.43 0.43 0.39 0.41 0.39 SD 0.64 0.71 0 .61 0.45 0.63 0.59

PAGE 195

195 Figure 4 3 1 Bubble plot comparing patterns of chlorophyll a ( CHL A ) variability across 84 sites sampled within 51 estuarine coastal ecosystems. X axis measures variability of the annual component as standard deviation of y ( SD y ), hypothesized to be an index of disturbance from human actions and shifts in the climate system. Y axis measures variability of the seasonal component as standard deviation of m ( SD m ), hypothesized to be an index of the importance of the annual climate cycle. Circle size measures variability of the residual component as standard deviation of ( SD ), hypothesized to be an index of nutrient enrichment. Method and site key taken from Cloern and Jassby (2010). The four regions of the Caloosahatc hee Estuary, FL have been added as circles 85 (UE), 86 (ME), 87 (LE), and 88 (BY). Median values of SD y SD m and SD for the Caloosahatchee Estuary, FL are represented by the star symbol.

PAGE 196

196 Table 4 7. Best combinations of environmental drivers tha t explained patterns in daily gross primary productivity ( GPP d ) in the four regions of the Caloosahatchee Estuary, FL. Upper Estuary (UE) Middle Estuary (ME) Lower Estuary (LE) San Carlos Bay (BY) GPP d = 0.525 MEI RAINFALL TEMP_W TURB CDOM TSS TEMP_W S I MEI SAL MEI

PAGE 197

197 CHAPTER 5 CONCLUSION The structure, function, and overall health of the Caloosahatch ee Est uary, located on t he southwest coast of Florida in the USA have been jeopardized by human activities going on in its watershed since the expansion of urban a nd agricultural development began there in t he late 1800s. Modifications to the physical shape and hydrology of the system, including the impoundment of Lake Okeechobee, the creation of a trans state navigable canal, and the installation of thr ee water control structures along the Caloosahatchee River have subsequently altered the quantity, quality, timing, a nd distribution of freshwater flow into and through the estuary Natural underlying seasonal and annual variations in salinity and nutrie nt load s in particular, have been amplified or truncated by anthropogenic control over the delivery of freshwater into the Caloosahatc hee Estuary affecting the ecological b alance among and economic value of downstream organisms and the eco system overall One associated impact of these changes is the i ncrease in the frequency and distribution o f algal blooms which h as brought attention to water quality and processes affecting phytoplankton production and biomass accumulation ther e. While phytoplankton p roduction and biomass support natural food webs and the global carbon cycle, ha rmful, toxic, and/or nuisance algal blooms have been found to directly and indirectly harm aquatic plants, invertebrates, fish, birds, mammals, and humans, raising concerns and stimulating action amongst the scientific commun ity, managing authorities, and general public. Because of the potentially widespread impact of phytoplankton dynamics on the health of an entire ecosystem, m easurements of phytoplankton productivity in the Ca loosahatchee Estuary, and other coastal ecosys tems, are thus desired so that the

PAGE 198

198 structure and function of these ecosystems can be defined, better understood and compared However, t he methods traditionally employed to gain such knowledge have been limit ed in their spatial and temporal extent due to the amount of money, labor, and time required to assess productivity at the s cale of e ntire ecosystem s over an extended period of time. A pproaches that instead estimate production using easily obtained and re adily available paramete rs have become popula r tools for building long time series of productivity data across large areas which, in turn, can b e used to establish ecosystem us. The BZ p I 0 d Cl oern 1987) is one such approach that predicts daily rates of phytoplankton production from measurements of chlorophyll a photic depth, and PAR light flux, which are routinely collected in oceanographic in vestigations and system wide monitoring efforts. The model was founded on l ong standing principles that underline the major r ole of plant biomass and light availabilit y i n the control of primary production The original development and subsequ ent use s of the model to provide reasonable estimates of phytoplankton productivity ( i.e., with fitted relationships having r 2 values greater than 0.50) have involved a variety of estuary types having a w ide range of conditions P revious model tests demonstrated the applicability of the BZ p I 0 re lationship to estuaries having temperate to warm temperate/ subtropical climates, shallow to deep water depths, and low to hi gh tidal ranges. Th is study was unique in that the Caloosahatchee Estuary has a combination of featu res that distinguish it from other estuaries and coastal ecosystems in which the BZ p I 0 had been previously adapted.

PAGE 199

199 Being a low latitude subtropical estuary the Caloosahatchee has warmer water temperatur es, stronger light levels, and higher inputs of rainfall and river flow in the summer and fa altered hydrology create a pr esumably ideal recipe for phytoplankton blooms. Together, t h e climate/weather, bathymetry, and hydrology of the system were expected to hinder relationship yielded a stronger fit than that of other model applications. The adaptabili ty of the model to the Caloosahatche e Estuary extended the utility of the BZ p I 0 parameter for predicting phytoplankton productivity to systems having features commonly associated with tropical systems, including those that hav e been extensively modified an d are currently maintained to meet the immediate needs and priorities of nearby residents and businesses. The methods employed during the model adaptation in the Caloosahatchee Estuary also provided clarity regarding the effect of methodological differen ces and Estimates of both gross and net primary productivity provided statistically comparable models in terms of their strengths ( coefficients of determination), intercepts, and slopes. T he use of either uncorrected or corrected (for pheophytin degradation pigments) chlorophyll a concentrations as a proxy for phytoplankton biomass was also accommodated by the s flexibility. On the other hand, a djusting the p hotic depths to account f or shallow water column depths resulted in a significantly different model relationship having an intercept that did not support the natural and logical expected outcome in productivity when there is ze ro biomass or light available. By allowing the use of O 2 evolution or 14 C

PAGE 200

200 fixation techniques to measure productivity, uncorrected or corrected chlorophyll a to represent biomass, and photic depths not requiring any post hoc correction, the BZ p I 0 or proved to be a simple and flexible to ol that provide d reasonable estimates of phytoplankton productivity that were comparable a c ross different studies despite these frequently encountered differences in methodology. The relationship ultimately selected as the (in terms of its statist ical significance and strength) (in terms of its appropriateness and usefulness) model for the Caloosahatchee Estuary c onfirmed the primary importance of biomass and light in the control of ph ytoplankton productivity. Model deviations, o r differences in measured and predicted productivity, indicated the secondary influence of nutrients, light quality, and species composition on the productivity potential or photosynthetic efficiency of the phytoplankton community. Seasonal variations in freshwater inflow (i.e., rainfall and river flow) created distinct dry and wet periods with lower and higher nutrient loa ds and wider and narrower underwater light fie lds, which when coupled with longer and shorter water residence times, supported dominati ng communities of dinoflagellates and diatoms, respectively. The dynamics of phytoplankton community responses to these and other natural and anthropogenic influences in the Caloosahatchee Estuary were further explored by applying the adapted model over a discontinuous twenty five year period using data collected during routine monitoring programs and other research projects. The estimates of phytoplankton productivity obtained from this model application identified spat ial, as well as, seasonal, intera nnu al, and deca dal sources of variation in the rates of production and, thus, the trophic status of the Caloosahatchee Estuary.

PAGE 201

201 P hytoplankton p roductivity was consistently highest in the upper and middle regions of the estuary closest to S 79, the Franklin L ock and Dam primarily due to the influx of new nutrients that accompan y freshwater releases from the Caloosahatchee River and access to nutrient reserves in the flocculent muddy sediment s there Blooms of phytoplankton having chlorophyll a concentrations 3 representing the top 10% of observations from the twenty five year period, were also more frequently observed in o ligohaline and mesohaline sa linitie s. On a seasonal basis, productivity and biomass were typically highest in the su mmer and fall wet season, given the high water temperatures, PAR flux levels, and nutrient inputs. However, years that had wetter than normal wet seasons due to above average rainfall and/or excessive flushing events reduced salinities, water residence ti mes, and the quantity and quality of light needed for photosynthesis throughout the estuary, restricting phytoplankton produc tivity and biomass accumulation. Cyclical El Nio phases and episodic storm events contributed to the interannual and decadal vari ability in from oligotrophic to hypertrophic levels due to their direct effect on rainfall and temperatures and their indirect effect on any corresponding dependent factors. Ultimately, this study confirmed the importance of lo ng term data sets in tracing, understanding, and predicting shifts in the structure and function of phytoplankton communities and, thus, entire ecosystems in response to natural and anthropogenic changes oc curring at multiple time scales. The morphology a nd hy drology of the Caloosahatchee Estuary were linked to variations in tidal mixing, water residence times, and light availability, which act as filters that enhance or mask the responses of the phytoplankton community to nutrient enrichment and other str essors (e.g., climate

PAGE 202

202 change, hydrologic manipulations, etc.). These interactions have profound impacts on the economic and social utility of the Caloosahatchee Estuary and other coastal ecosyst ems and, ultimately, the vitality of humanity due to the depe ndency on these ecosystems (Cloern 2001). These and other concepts addressed in this study are central to the development of a broader conceptual model of coastal eutrophication that is needed to guide the development of strategies and tools to be used in the management and restoration of damaged estuaries and coastal ecosystems worldwide (Cloern 2001).

PAGE 203

203 LIST OF REFERENCES Adler, R. W., Landman, J. C., Cameron, D. M. (1993). The Clean Water Act 20 years later. Island Press, Washington, D.C. Alber, M. (2002 ). A conceptual model of estuarine freshwater inflow management. Estuaries 25(6B): 1246 1261 American Public Health Association (APHA). (2005). Standard methods for the examination of water and wastewater, 21 st ed. American Public Health Association, Wash ington, D.C. Anderson, J., Gorley, R. N., Clarke, K. R. (2008). PERMANOVA+ for PRIMER: Guide to software and statistical methods. PRIMER E Ltd, Plymouth, United Kingdom Anonymous. (1959). The final resolution of the symposium on the classification of brack ish waters. Archo Oceanogr. Limnol. 11(supplement): 243 248 Bailey, N., Magley, W., Mandrup TMDL report: Nutrient TMDL for the Caloosahatchee Estuary (WBIDS 3240A, 3240B, and 3240C). Florida Department of Environmental Protection, Tallahassee, Florida Barbour, M. G., Burk, J. H., Pitts, W. D. (1987). Terrestrial plant ecology, 2 nd ed. The Benjamin/Cummings Publishing Company, Menlo Park, California Barnes, T. (2005). Caloosahatchee Estuary conceptual ecolo gical model. Wetlands 25(4): 884 897 Beman, J. M. (205). Agricultural runoff fuels large phytoplankton blooms in vulnerable areas of the ocean. Nature 434: 211 214 Bergmann, T., Richardson, T. L., Paerl, H. W., Pinckney, J. L., Schofield, O. (2002). Synerg y of light and nutrients on the photosynthetic efficiency of phytoplankton populations from the Neuse River Estuary, North Carolina. J. Plankton Res. 24(9): 923 933 Bierman, V. J. Jr. (1993). Performance report for Caloosahatchee Estuary salinity modeling. Limno Tech, Inc., Ann Arbor, Michigan Borum, J. (1996). Shallow waters and land/sea boundaries. In: Jrgensen, B. B., Richardson, K. (ed.) Eutrophication in coastal marine ecosystems. American Geophysical Union, Washington, D.C., p. 179 203 Bouman, H. A., Nakane, T., Oka, K., Nakata, K., Kurita, K., Sathyendranath, S., Platt, T. (2010). Environmental controls on phytoplankton production in coastal ecosystems: A case study from Tokyo Bay. Estuar. Coast. Shelf Sci. 87: 63 72

PAGE 204

204 Boyer, J. N., Christian, R. R., S tanley, D. W. (1993). Patterns of phytoplankton primary productivity in the Neuse River estuary, North Carolina, U.S.A. Mar. Ecol. Prog. Ser. 97: 287 297 Boyer, J. N., Kelble, C. R. Ortner, P. B., Rudnick, D. T. (2009). Phytoplankton bloom status: Chloroph yll a biomass as an indicator of water quality condition in southern estuaries of Florida, U.S.A. Ecol. Indicat. 9(6, Supplement): 56 57 Boynton, W. R., Kemp, W. M., Keefe, C. W. (1982). A comparative analysis of nutrients and other factors influencing est uarine phytoplankton production. In: Kennedy, V. S. (ed.) Estuarine comparisons. Academic Press, New York, p. 69 90 Brand, L. E. (1984). The salinity tolerance of forty six marine phytoplankton isolates. Estuar. Coast. Shelf Sci. 18: 543 556 Brand, L. E., Compton, A. (2007). Long term increase in Karenia brevis abundance along the Southwest Florida Coast. Harmful Algae 6: 232 252 Brawley, J. W., Brush, M. J., Kremer, J. N., Nixon, S. W. (2003). Potential applications of an empirical phytoplankton production model to shallow water ecosystems. Ecol. Model. 160: 55 61 Bray, J. R., Curtis, J. T. (1957). An ordination of the upland forest communities of Southern Wisconsin. Ecol. Monogr. 27: 325 349 Bricker, S. B., Longstaff, B., Dennison, W., Jones, A., Boicourt, K., Wicks, C., Woerner, change. Harmful Algae 8:21 32 phytoplankton primary production to shal low marine ecosystems. J. Mar. Syst. 75: 227 235 Brush, M. J., Brawley, J. W., Nixon, S. W., Kremer, J. N. (2002). Modeling phytoplankton production: problems with the Eppley curve and an empirical alternative. Mar. Ecol. Prog. Ser. 238: 31 45 Brylinsky, M ., Mann, K. H. (1973). An analysis of factors governing productivity in lakes and reservoirs. Limnol. Oceanogr. 18(1): 1 14 Burkholder, J. M., Glibert, P. M., Skelton, H. M. (2008). Mixotrophy, a major mode of nutrition for harmful algal species in eutroph ic waters. Harmful Algae 8: 77 93 Burns, J. (2008). Toxic cyanobacteria in Florida Waters. In: Hudnell, H. K. (ed.) Cyanobacterial harmful algal blooms: State of the science and research needs. Springer, New York.

PAGE 205

205 Carriker, R. R., Borisova, T. (2009). Pu blic policy and water in Florida. EDIS document FE799, Food and Resource Economics Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida Chamberlain, R.H., Doering, P.H. (1998a). Freshwater inflow to the Caloosahatchee Estuary and the resource based method for evaluation. In: Treat, S.F. (ed.) Proceedings of the Charlotte Harbor public conference and technical symposium, March 15 16, 1997, Punta Gorda, Florida. Charlotte H arbor National Estuary Program Technical Report No. 98 02, South Florida Water Management District, West Palm Beach, Florida, p. 81 90 Chamberlain, R.H., Doering, P.H. (1998b). Preliminary estimate of optimum freshwater inflow to the Caloosahatchee Estuary : A resource based approach. In: Treat, S.F. (ed.) Proceedings of the Charlotte Harbor public conference and technical symposium, March 15 16, 1997, Punta Gorda, Florida. Charlotte Harbor National Estuary Program Technical Report No. 98 02, South Florida W ater Management District, West Palm Beach, Florida, p. 121 130 Chamberlain, R. H., Doering, P. H., Haunert, K. M., Crean, D. (2003). Impacts of freshwater inflows on the distribution of zooplankton and ichthyoplankton in the Caloosahatchee Estuary, Florida In: Technical documentation to support development of minimum flows and levels for the Caloosahatchee River and Estuary. Florida Bay and Lower West Coast Division, Southern District Restoration Department, South Florida Water Management District, West Pa lm Beach, Florida, App. C. Chen, E., Gerber, J. F. (1990). Climate. In: Myers, R. L., Ewel, J. J. (ed.) Ecosystems of Florida. University of Central Florida Press, Orlando, Florida, p. 11 34 Cloern, J. E. (2001). Our evolving conceptual model of the coasta l eutrophication problem. Mar. Ecol. Prog. Ser. 210: 223 253 Cloern, J. E., Jassby, A. D. (2008). Complex seasonal patterns of primary producers at the land sea interface. Ecol. Lett. 11: 1294 1303 Cloern, J. E., Jassby, A. D. (2010). Patterns and scales o f phytoplankton variability in estuarine coastal ecosystems. Estuaries Coasts 33:230 241 Cole, B. E., Cloern, J. E. (1984). Significance of biomass and light availability to phytoplankton productivity in San Francisco Bay. Mar. Ecol. Prog. Ser. 17: 15 24 C ole, B. E., Cloern, J. E. (1987). An empirical model for estimating phytoplankton productivity in estuaries. Mar. Ecol. Prog. Ser. 36: 299 305 Corbett, C.A. (2004). Coastal Charlotte Harbor monitoring network: Description and standard operating procedures Technical Report 02 03, Charlotte Harbor National Estuary Program, North Fort Myers, Florida

PAGE 206

206 Crossett, K. M., Culliton, T. J., Wiley, P. C., and Goodspeed, T. R. (2004). Population trends along the coastal Unites States: 1980 2008. Coastal Trends Report Series, U.S. Department of Commerce, National Oceanic and Atmospheric Administration, National Ocean Service, Washington, D.C. Day, J. W. Jr., Hall, C. A. S., Kemp, W. M., Yanez Arancibia A. (1989). Estuarine ecology. John Wiley and Sons, New York Degrove, B. (1981). Caloosahatchee River waste load allocation documentation: Lee County. Water Quality Technical Series, vol. 2, no. 52, Florida Department Environmental Regulation, Tallahassee, Florida Doering, P. H., Chamberlain, R. H. (1998). Water quality in the Caloosahatchee Estuary, San Carlos Bay and Pine Island Sound, Florida. In: Treat, S. F. (ed.) Proceedings of the Charlotte Harbor Public conference and technical Symposium, March 15 16, 1997, Punta Gorda, Florida. Charlotte Harbor National Estuary Prog ram Technical Report No. 98 02, South Florida Water Management District, West Palm Beach, Florida, p. 229 240 Doering, P. H., Chamberlain, R. H. (1999). Water quality and source of freshwater discharge to the Caloosahatchee Estuary, Florida. J. Am. Water R esour. As. 35(4): 793 806 Doering, P. H., Chamberlain, R. H., Haunert, K. M. (2006). Chlorophyll a and its use as an indicator of eutrophication in the Caloosahatchee Estuary, Florida. Fla. Sci. 69: 51 72 Earth System Research Laboratory (ESRL). (2011). M ultivariate ENSO index (MEI). Retrieved November 21, 2011, from http://www.esrl.noaa.gov Fahnenstiel, G. L., Carrick, H. J. (1992). Phototrophic picoplankton in Lakes Huron and Michigan: Abundance, distribution, composition, and contribution to biomass and production. Can. J. Fish. Aquat. Sci. 49: 379 388 Falkowski, P. G. (1994). The role of phytoplankton photosynthesis in global biogeochemical cycles. Photosyn. Res. 39: 235 258 Falkowski, P. G., Owens, T. C. (1980). Light shade adaptation: Two strategies i n marine phytoplankton. Plant Physiol. 66: 592 595 Flaig, E. G., Capece, J. (1998). Water use and runoff in the Caloosahatchee Watershed. In: Treat, S. F. (ed.) Proceedings of the Charlotte Harbor Public conference and technical Symposium, March 15 16, 199 7, Punta Gorda, Florida. Charlotte Harbor National Estuary Program Technical Report No. 98 02, South Florida Water Management District, West Palm Beach, Florida, p. 73 80

PAGE 207

207 Florida Department of Environmental Protection (FDEP) (2005). Water quality assessmen t report: Caloosahatchee. Florida Department of Environmental Protection, Tallahassee, Florida Foster, C. E., Wessel, R. A. (2007). Legacy of the Caloosahatchee. Caloosahatchee River Citizens Association, Inc. Retrieved February 6, 2013, from http://crca. caloosahatchee.org/about/?show=legacy Furnas, M. J., Hitchcock, G. L., Smayda, T. J. (1976). Nutrient phytoplankton relationships in Narragansett Bay during the 1974 summer bloom. Estuar. Processes 1: 119 133 Gaines, G., Elbrachter, M. (1987). Heterotroph ic nutrition. In: Taylor, F. J. R. (ed.) The biology of dinoflagellates. Blackwell Scientific, Oxford, p. 224 268 Gallagher, J. C., Wood, A. M., Alberte, R. S. (1984). Ecotypic differentiation in the marine diatom Skeletonema costatum : Influence of light i ntensity on the photosynthetic apparatus. Mar. Biol. 82(2): 121 134. Geider, R. J., MacIntyre, H. L. (2002). Physiology and biochemistry of photosynthesis and algal carbon acquisition. In: Williams, P. J., Thomas, D. N., Reynolds, C. S. (ed.) Phytoplankton productivity: Carbon assimilation in marine and freshwater ecosystems. Blackwell Science, Oxford, p. 44 77 Geider, R. J., Osborne, B. A. (1992). Algal photosynthesis: The measurement of algal gas exchange. Chapman and Hall, New York Guest, G. D. (2001). T his time for sure: A political and legal history of water control projects in Lake Okeechobee and the Everglades. St. Thom. L. Rev. 13(3):645 667 Heil, C. A., Revilla, M., Glibert, P. M., Murasko, S. (2007). Nutrient quality drives differential phytoplankt on community composition on the southwest Florida shelf. Limnol. Oceanogr. 52(3): 1067 1078 Hinrichsen, D. (1999). The coastal popula tion explosion. In: Cicin Sain, B., Knecht, R. W ., Foster, N. (ed.) Trends and future challenges for U.S. nat ional ocean an d coastal Policy: Proceedingsof a workshop, January 22, 1999, Washington, D.C. National Oceanic and Atmospheric Administration, National Ocean Service, Washington, D.C., p. 27 29 Howarth, R. W., Michaels, A. F. (2000). The measurement of primary production in aquatic ecosystems. In: Sala, O. E., Jackson, R. B., Mooney, H. A., Howarth, R. W. (ed.) Methods in ecosystem science. Springer Verlag, New York, p. 72 85 International Research Institute for Climate and Society (IRI). (2007). The Influence of ENSO on climate. Retrieved December 19, 2012, from http://iri.columbia.edu/climate/ENSO/globalimpact/overview/overview.html

PAGE 208

208 fire in Florida. Fla. Geogr. 30: 55 69 Keller, A E. (1988). Estimating phytoplankton productivity from light availability and biomass in the MERL mesocosms and Narragansett Bay. Mar. Ecol. Prog. Ser. 45: 159 168 Kelly, J. R., Doering, P. H. (1997). Monitoring and modeling primary production in coastal waters: Studies in Massachusetts Bay 1992 1994. Mar. Ecol. Prog. Ser. 148: 155 168 Kendall, M. G. (1970). Rank correlation methods, 4 th ed. Griffin, London Kirk, J. T. O. (1994). Light and photosynthesis in aquatic ecosystems, 2 nd ed. University Press, Cam bridge Knight, R., Steele, J. (2005). Caloosahatchee River/Estuary nutrient issue white paper. Wetland Solutions, Inc., Gainesville, Florida Khler, J. (1998). Measurement of photosynthetic oxygen production. Rostock. Meeresbiol. Beitr. 6: 17 21 Kranzer, B (2003). Everglades restoration: Interactions of population and environment. Popul. Environ. 24(6): 455 484 Lapointe, B. E., Bedford, B. J. (2007). Drift rhodophyte bloom emerge in Lee County, Florida, USA: Evidence of escalating coastal eutrophication. H armful Algae 6:421 437 Levandowsky, M., Kaneta, P. J. (1987). Behaviour in dinoflagellates. In: Taylor, F. J. R. (ed.) The biology of dinoflagellates. Blackwell Scientific, Oxford, p. 360 397 Macfayden, A. (1948). The meaning of productivity in biological systems. J. Anim. Ecol. 17(1): 75 80 Mallin, M. A., Paerl, H. W. (1992). Effects of variable irradiance on phytoplankton productivity in shallow estuaries. Limnol. Oceanogr. 37(1): 54 62 Mallin, M. A., Paerl, H. W., Rudek, J. (1991). Seasonal phytoplankton composition, productivity, and biomass in the Neuse River Estuary, North Carolina. Estuar. Coast. Shelf Sci. 32: 609 623 Mallin, M. A., Paerl, H. W., Rudeck, J., Bates, P. W. (1993). Regulation of estuarine primary production by watershed rainfall and riv er flow. Mar. Ecol. Prog. Ser. 93: 199 203.

PAGE 209

209 Malone, T. C. (1976). Phytoplankton productivity in the apex of the New York Bight: Environmental regulation of productivity/chlorophyll a In: Gross, M.G. (ed.) Am. Soc. Limnol. Oceanog. Spec. Symp. 2: 260 272 M alone, T. C. (1977). Environmental regulation of phytoplankton productivity in the lower Hudson Estuary. Estuarine Coastal Mar. Sci. 5: 157 171 Matsubara, T., Nagasoe, S., Yamasaki, Y., Shikata, T., Shimasaki, Y., Oshima, Y., Honjo, T. (2007). Effects of t emperature, salinity, and irradiance on the growth of the dinoflagellate Akashiwo sanguinea J. Exp. Mar. Bio. Eco. 342: 226 230 McPherson, B. F., Miller, R. L. (1987). The vertical attenuation of light in Charlotte Harbor, a shallow, subtropical estuary, south western Florida. Estuar. Coast. Shelf Sci. 25: 721 737 McPherson, B. F., Miller, R. L. (1990). Nutrient distribution and variability in the Charlotte Harbor Estuarine Sytem, Florida. Water Resour. Bull. 26(1): 67 80 McPherson, B. F., Montgomery, R. T ., Emmons, E. E. (1990). Phytoplankton productivity and biomass in the Charlotte Harbor Estuarine System, Florida. Water Resour. Bull. 26(5): 787 800 Monbet, Y. (1992). Control of phytoplankton biomass in estuaries: A comparative analysis of microtidal and macrotidal estuaries. Estuaries 15(4): 563 571 Montgomery, R. T., McPherson, B. F., Emmons, E. E. (1991). Effects of nitrogen and phosphorous additions on phytoplankton productivity and chlorophyll a in a subtropical estuary, Charlotte Harbor, Florida. Wa ter Resources Investigations Report 91 4077, United States Geological Survey, Tallahassee, Florida Mortazavi, B., Iverson, R. L., Landing, W. M., Lewis, F. G., Huang, W. (2000). Control of phytoplankton production and biomass in a river dominated estuary: Apalachicola Bay, Florida, USA. Mar. Ecol. Progr. Ser. 198: 19 31 Murrell, M. C., Hagy, J. D. III, Lores, E. M., Greene, R. M. (2007). Phytoplankon production and nutrient distributions in a subtropical estuary: Import ance of freshwater flow. Estuaries Coa sts 30(3): 390 402 National Climatic Data Center (NCDC). (2010). 2009/2010 Cold season: Special report. Retrieved June 20, 2011, from http://www.ncdc.noaa.gov/special reports/2009 2010 cold season.html National Climatic Data Center (NCDC). (2011). Climatol ogical data: Annual summary, Florida, 1985 2010. Retrieved December 1, 2011, from http://www7.ncdc.noaa.gov/IPS National Oceanic Atmospheric Administration (NOAA) (2010). Tides and currents. Retrieved December 1, 2010, from http://tidesandcurrents.noaa.go v

PAGE 210

210 National Oceanic and Atmospheric Administration (NOAA). (2012). Historical hurricane tracks. Retrieved February 6, 2012, from http://www.csc.noaa.gov/hurricanes Nixon, S. W. (1995). Coastal marine eutrophication: A definition, social causes, and future c oncerns. Ophelia 41: 199 219 Odum, E. P. (1968). Energy flow in ecosystems: A historical review. Am. Zool. 8:11 18 Odum, H. T., Odum, E. C., Brown, M. T. (1998). Environment and society in Florida. Lewis Publishers, Boca Raton, Florida Paerl, H. W., Ross ignol, K. L., Hall, N., Peierls, B. L., Wetz, M. S. (2010). Phytoplankton community indicators of short and long term ecological change in the anthropogenically and climatically impacted Neuse River Estuary, North Carolina, USA. Estuaries Coasts 33: 485 4 97 Pennock, J. R., Boyer, J. N., Herrera Silveira, J. A., Iverson, R. L., Whitledge, T. E., Mortazavi, B., Comin, F. A. (1999). Nutrient behavior and phytoplankton production in Gulf of Mexico estuaries. In: Bianchi, T. S., Pennock, J. R., Twilley, R. R. ( ed.) Biogeochemistry of Gulf of Mexico estuaries. John Wiley & Sons, New York, p. 109 162 Pennock, J. R., Sharp, J. H. (1986). Phytoplankton production in the Delaware Estuary: temporal and spatial variability. Mar. Ecol. Prog. Ser. 41: 187 203 Perry, M. J ., Talbot, M. C. Alberte, R. S. (1981). Photoadaptation in marine phytoplankton: Response of the photosynthetic unit. Mar. Biol. 62: 91 101 Perry, W. B. (2008). Everglades restoration and water quality challenges in south Florida. Ecotoxicology 17: 569 578 Peterson, B. J. (1980). Aquatic primary productivity and the 14 C CO 2 method: A history of the productivity problem. Ann. Rev. Ecol. Syst. 11: 359 85 Phlips, E. J., Badylak, S., Christman, M. C., Lasi, M. A. (2010). Climatic trends and temporal patterns of phytoplankton composition, abundance, and succession in the Indian River Lagoon, Florida, USA. Estuaries Coasts 33:498 512 PRIMER E Ltd. (2012). PRIMER 6 Software. Plymouth, United Kingdom Qiu, C., Sheng, Y., Zhang, Y. (2007). Development of a hydrodynam ic and salinity model in the Caloosahatchee and Estero Bay, Florida. In: Spaulding, M. L. (ed.) Estuarine and coastal modeling, Proceedings of the tenth international conference, November 5 7, 2007, Newport, Rhode Island. American Society of Civil Engineer s, Reston, Virginia, p. 106 123

PAGE 211

211 Quinlan, E. L., Phips, E. J. (2007). Phytoplankton assemblages across the marine to low salinity transition zone in a blackwater dominated estuary. J. Plankton Res. 29(5): 401 416 Rand, G. M., Bachman, P. M. (2008). South Fl orida ecosystems. Ecotoxicology 17: 565 568 Randall, J. M., Day, J. W. (1987). Effects of river discharge and vertical circulation on aquatic primary production in a turbid Louisiana (USA) estuary. Net. J. Sea. Res. 21: 231 242 Restoration Coordination and Verification (RECOVER). (2011). Scientific and technical knowledge gained in Everglades restoration (1999 2009). Restoration Coordination and Verification, U.S. Army Corps of Engineers, Jacksonville, Florida, and South Florida Water Management District, W est Palm Beach, Florida Reynolds, C. (2006). The ecology of phytoplankton. University Press, Cambridge Ryding, S. O., Rast, W. (1989). The control of eutrophication of lakes and reservoirs. UNESCO, Paris Ryther, J. H. (1956a). Photosynthesis in the ocean as a function of light intensity. Limnol. Oceanogr. 1(1): 61 70 Ryther, J. H. (1956b). The measurement of primary production. Limnol. Oceanogr. 1(2):72 83 Ryther, J. H. (1963). Geographical variations in productivity. In: Hill, M.N. (ed.) The sea, vol. 2, Wiley Interscience, New York, p. 347 380 Ryther, J. H., Yentsc h, C. S. (1957). The estimation of phytoplankton production in the ocean from chlorophyll and light data. Limnol. Oceanogr. 2(3): 281 286 Sand Jensen, K. (1997). Broad scale comparison of pho tosynthesis in terrestrial and aquatic plant communities. Oikos 80(1): 203 208 Sartory, D. P., Grobbelaar J. U. (1984). Extraction of chlorophyll a from freshwater phytoplankton for spectrophotometric analysis. Hydrobiologia 114: 177 187 SAS Institute Inc (2009). SAS 9.2 Software. Cary, North Carolina Saunders, R. P., Birnhak, B. I., Davis, J. T., Wahlquist, C. L. (1967). Seasonal distribution of diatoms in Florida inshore waters from Tampa Bay to Caxambas Pass, 1963 1964. In: Red tide studies, Pinellas to Collier Counties, 1963 1966, a symposium. Florida Bd. Conserv. Mar. Lab. Pr of. Paper Ser. No. 9. p. 48 78

PAGE 212

212 Scardi, M. (1996). Artificial neural networks as empirical models for estimating phytoplankton production. Mar. Ecol. Progr. Ser. 139: 289 299 Scarlatos, P. D. (1988). Caloosahatchee estuary hydrodynamics. Technical Publicat ion 88 7, South Florida Water Management District, West Palm Beach, Florida. Schindler, D. W. (2006). Recent advances in the understanding and management of eutrophication. Limnol. Oceanogr. 51(1, part 2): 356 363 Smayda, T. J. (1978). From phytoplankter s to biomass. In: Sournia, A. (ed.) Phytoplankton manual. UNESCO, Paris, p. 273 279 Smayda, T. J. (1997). Harmful algal blooms: Their ecophysiology and general relevance to phytoplankton blooms in the sea. Limnol. Oceanogr. 42(5, part 2): 1137 1153 Sorokin Y. I. (1999). Radioisotopic methods in hydrobiology. Springer Verlag, Berlin, Germany South Florida Water Management District (SFWMD). (2008). Rules of the South Florida Water Management Disrict: Minimum flows and levels. Chapter 40E 8, F.A.C., South Fl orida Water ManagementDistrict, West Palm Beach, Florida South Florida Water Management District (SFWMD) (2010). DBYDRO environmental database Re trieved September 1, 2010, from http://www.sfwmd.gov/dbhydroplsql/show_dbkey_info.main_menu South Florida Wate r Management District (SFWMD ). (2011). DBYDRO environmental database. Retrieved September 1, 2011 from http://www.sfwmd.gov/dbhydroplsql/show_dbkey_info.main_menu South Florida Water Management District (SFWMD), Florida Department of Environmental Protecti on (FDEP), Florida Department of Agriculture and Consumer Services (FDACS). (2008). Lake Okeechobee Watershed construction project: Phase II technical plan. West Palm Beach, Florida South Florida Water Management District (SFWMD), Florida Department of Env ironmental Protection (FDEP), Florida Department of Agriculture and Consumer Services (FDACS). (2009). Caloosahatchee River Watershed protection plan. West Palm Beach, Florida Stoker, Y. E. (1992). Salinity distribution and variation with freshwater inflow and tide, and potential changes in salinity due to altered freshwater inflow in the Charlotte Harbor Estuarine System, Florida. Water Resources Investigations Report 92 4062, United States Geological Survey, Tallahassee, Florida Stolte, W., Garcs, E. (20 08). Ecological aspects of harmful algal in situ population growth rates. In: Graneli, E. Turner, J. T. (ed.) Ecological studies 189: Ecology of harmful algae. Springer, Berlin, p. 139 152

PAGE 213

213 Strickland, J. D. H. (1960). Measuring the production of marine phy toplankton. Fisheries Research Board of Canada, Ottawa. Strickland, J. D. H., Parsons, T. R. (1972). A Practical Handbook of Seawater Analysis, 2 nd ed. Fisheries Research Board of Canada, Ottawa, Canada, p. 9 14 Taylor, F. J. R. (1987). Ecology of dinoflag ellates: General and marine ecosystems. In: Taylor, F. J. R. (ed.) The biology of dinoflagellates. Blackwell Scientific, Oxford, p. 399 502 Teeter, J. W. (1980). Ostracoda of the Lake Flirt Formation (Pleistocene) of southern Florida. Micropaleontology 26( 4): 337 355 Tilzer, M. M. (1989). Introduction. In: Tilzer, M. M. (ed.) Production and pathways of freshly produced organic matter within the pelagic zone. Hydrobiologia 173: 85 87 Tolley, S. G., Fugate, D., Parson, M. L., Burghart, S. E., Peebles, E. B. ( 2010). The responses of turbidity, CDOM, benthic microalgae, phytoplankton and zooplankton to variation in seasonal freshwater inflow to the Caloosahatchee Estuary. South Florida Water Management District, West Palm Beach, Florida Tropical Atmosphere Ocean Project (TAO). (2011). El Nio theme page. Retrived November 21, 2011, from http://www.pmel.noaa.gov/tao/elnino Utermohl, H. (1958). Zur vervollkommnung der quanti tative phytoplankton methodik. Mit. Int. Ver. Theo. Ager. Limnol. 9: 1 38 Valiela, I. (1995 ). Marine ecological process, 2 nd ed. Springer Verlag, New York Vitousek, P. M., Mooney, H. A., Lubchenco, J., Melillo, J. M. (1997a). Human 499 Vitousek, P. M., Schlesinger, W. H., Tilman, D. G. (1997b). Human alteration of the global nitrogen cycle: Sources and consequences. Ecol. Appl. 7(3): 737 750 Volety, A. K. (2008). Effects of salinity, heavy metals, and pesticides on health and physiology of oysters in the Caloosahatchee Estuary, Florida. Ecotoxic ology 17: 579 590 Vollenwider, R. A. (1969). Objective. In: Vollenwider, R. A. (ed.) A manual on methods for measuring primary production in aquatic environments. International Biological Programme, London, p. 1 3 Walsh, J. J., Jolliff, J. K., Darrow, B. P ., Lenes, J. M., Milroy, S. P., Remsen, A., Dieterle, D. A., Carder, K. L., Chen, F. R., Vargo, G. A., Weisberg, R. H., Fanning, K. A., Muller Karger, F. E., Shinn, E., Steidinger, K. A., Heil, C. A., Tomas, C. R., Prospero, J. S., Lee, T. N., Kirkpatrick, G. J., Whitldege, T. E.,

PAGE 214

214 Stockwell, D. A., Villareal, T. A., Jochens, A. E., Bontempi, P. S. (2006). Red tides in the Gulf of Mexico: Where, when, and why? J. Geophys. Res. 111: 1 46 Welschmeyer, N. A., Lorenzen, C. J. (1981). Chlorophyll specific photosy nthesis and quantum efficiency at subsaturating light intensities. J. Phycol. 17: 283 293 Wessel, A., Capece J (2007). What are the oxbows. Caloosahatchee River Citizens Association. Retrieved February 6, 2013, from http://crca.caloosahatchee.org/oxbows /?show=about Wetze l, R. G. (1983). Limnology, 2 nd ed. Saunders College Publishing, Philadelphia Wetzel, R. G., Likens, G. E. (1991). Limnological Analyses, 2 nd ed. Springer Verlag, New York Winder, M., Cloern, J. E. (2010). The annual cycles of phytoplankt on biomass. Phil. Trans. R. Soc. B 365: 3215 3226 Wolfsy, S. C. (1983). A simple model to predict extinction coefficients and phytoplankton biomass in eutrophic waters. Limnol. Oceanogr. 28: 114 155 Yentsch, C. S. (1980). Light attenuation and phytoplankto n photosynthesis. In: Morris, I. (ed.) The physiological ecology of phytoplankton. Univ. of California Press, Berkeley, p. 95 127 Zingone, A., Phlips, E. J., Harrison, P. J. (2010). Multiscale variability of twenty two coastal phytoplankton time series: A global scale comparison. Estuaries Coasts 33: 224 229

PAGE 215

215 BIOGRAPHICAL SKETCH Loren Mathews grew up in Peachtree City, Georgia, a suburb of Atlanta. She attended the University of Miami where she received a Bachelor of Science degree in m arine s cience and b iology in 2005 and a Master of Arts degree in m arine a ffairs and p olicy in 2007. During her time in Miami, Loren worked full time at Miami Seaquarium, the oldest, privately owned marine life park in the United States. She worked first as an Education Pr ograms Lead and was later promoted to the Park Operations Supervisor, where she oversaw day to day park activities and helped plan the functioning of a new swim with the there and focused on the role of aquariums in public education and conservation efforts. In 2007, Loren moved to Gainesville, Florida with her husband, so he could attend law school at the University of Florida. Loren obtained a part time position as a laboratory assista nt for Edward Phlips in the Fisheries and Aquatic Sciences Program. She soon became the full time Quality Assurance/Quality Control Officer, for which she er sampling and c hemical analyse s. Interested in staying current in her field, Loren became a graduate student in the Fisheries and Aquatic Sciences Program in 2008 under the guidance of Edward Phlips. A grant from the South Florida Water Management District of West Palm Beach, Florida provided her the opportunity to lead a phytoplankton productivity study in the Caloosahatchee Estuary, Florida. The results of that project were the basis of her Doctor of Philosophy degree research and this dissertation. Loren moved to S avannah, Georgia in 2012 with her husband, who is a practicing attorney, and their one year old son.