USF Libraries
USF Digital Collections

A three-dimensional biophysical model of light, nutrient, and grazing controls on phytoplankton competition affecting re...

MISSING IMAGE

Material Information

Title:
A three-dimensional biophysical model of light, nutrient, and grazing controls on phytoplankton competition affecting red tide maintenance on the west Florida shelf
Physical Description:
Book
Language:
English
Creator:
Milroy, Scott P
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Mathematical model
Karenia brevis
Gulf of Mexico
Upwelling
Zooplankton
Dissertations, Academic -- Marine Science -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: A coupled, three-dimensional, time-dependent numerical model of water circulation, spectral light, plankton dynamics, nutrient/CDOM loadings, and zooplankton grazing provided an assessment of the factors affecting the growth and maintenance of red tides on the west Florida shelf (WFS). The coupled biophysical model consisted of state variable quantities for temperature, salinity, horizontal/vertical velocity components, turbulent diffusion, spectral light, colored dissolved organic matter (CDOM), dissolved organic and inorganic carbon, particulate silica, four dissolved inorganic nutrient pools (nitrate, ammonium, phosphate, and silicate), and four phytoplankton groups (diatoms, microflagellates, non-toxic dinoflagellates, and the red tide organism Karenia brevis).The model also included a complex grazing scheme that utilized thirteen different zooplankton groups to explore the effects of selective herbivory, feeding periodicity, diel vertical migration, fecal pellet egestion, and ammonium/phosphate excretion within a diverse zooplankton community. Over the shelf and slope of the eastern Gulf of Mexico, from the Mississippi River delta to the Florida Keys, four cases of the model were run during August -- November to explore the dynamics of red tide maintenance with respect to: (1) no refuge from grazing for K. brevis; (2) grazer avoidance of K. brevis during CDOM shading; (3) grazer avoidance of K. brevis in Case II waters; and (4) increased grazing stress on K. brevis competitors. NEGOM and ECOHAB data sets during July -- November 1999 were used to establish the initial/boundary conditions and provided validation data for the coupled model as well.Model results indicate that the red tide of 5.9 x 10^6 cells L-1 witnessed offshore Sarasota, Florida on 07 October 1999 was initiated by an inoculum of K. brevis observed in near-bottom waters above the 30 m isobath offshore Sarasota on 31 August 1999. Flowfields measured at moored ADCPs, observations from AVHRR satellite imagery, and west Florida shelf circulation models indicate that conditions of coastal upwelling existed during the period of bloom development, such that the K. brevis inoculum was delivered to the coast in the bottom Ekman layer. As a shade-adapted species capable of vertical migration, K. brevis cells aggregated near the bottom in order to escape photo-inhibitive light intensities in the overlying water column during the day and harvested the recycled nitrogen excreted by zooplankton grazers.This concomitant relaxation of light inhibition and nitrogen-limitation ultimately led to the growth and maintenance of the red tide, constrained in near-bottom waters during much of the day and preferentially advected inshore as a result of coastal upwelling. As K. brevis was advected inshore, self-shading, CDOM, and suspended inorganic particulates all contributed to the prevention of photo-inhibitive light intensities that, in combination with the excretion of recycled ammonium, ultimately led to the maintenance of a significant red tide at the coast.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Scott P. Milroy.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 157 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001919507
oclc - 184842615
usfldc doi - E14-SFE0002068
usfldc handle - e14.2068
System ID:
SFS0026386:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001919507
003 fts
005 20071218123237.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 071218s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002068
035
(OCoLC)184842615
040
FHM
c FHM
043
n-us-fl
049
FHMM
090
GC11.2 (ONLINE)
1 100
Milroy, Scott P.
2 245
A three-dimensional biophysical model of light, nutrient, and grazing controls on phytoplankton competition affecting red tide maintenance on the west Florida shelf
h [electronic resource] /
by Scott P. Milroy.
260
[Tampa, Fla.] :
b University of South Florida,
2007.
520
ABSTRACT: A coupled, three-dimensional, time-dependent numerical model of water circulation, spectral light, plankton dynamics, nutrient/CDOM loadings, and zooplankton grazing provided an assessment of the factors affecting the growth and maintenance of red tides on the west Florida shelf (WFS). The coupled biophysical model consisted of state variable quantities for temperature, salinity, horizontal/vertical velocity components, turbulent diffusion, spectral light, colored dissolved organic matter (CDOM), dissolved organic and inorganic carbon, particulate silica, four dissolved inorganic nutrient pools (nitrate, ammonium, phosphate, and silicate), and four phytoplankton groups (diatoms, microflagellates, non-toxic dinoflagellates, and the red tide organism Karenia brevis).The model also included a complex grazing scheme that utilized thirteen different zooplankton groups to explore the effects of selective herbivory, feeding periodicity, diel vertical migration, fecal pellet egestion, and ammonium/phosphate excretion within a diverse zooplankton community. Over the shelf and slope of the eastern Gulf of Mexico, from the Mississippi River delta to the Florida Keys, four cases of the model were run during August -- November to explore the dynamics of red tide maintenance with respect to: (1) no refuge from grazing for K. brevis; (2) grazer avoidance of K. brevis during CDOM shading; (3) grazer avoidance of K. brevis in Case II waters; and (4) increased grazing stress on K. brevis competitors. NEGOM and ECOHAB data sets during July -- November 1999 were used to establish the initial/boundary conditions and provided validation data for the coupled model as well.Model results indicate that the red tide of 5.9 x 10^6 cells L-1 witnessed offshore Sarasota, Florida on 07 October 1999 was initiated by an inoculum of K. brevis observed in near-bottom waters above the 30 m isobath offshore Sarasota on 31 August 1999. Flowfields measured at moored ADCPs, observations from AVHRR satellite imagery, and west Florida shelf circulation models indicate that conditions of coastal upwelling existed during the period of bloom development, such that the K. brevis inoculum was delivered to the coast in the bottom Ekman layer. As a shade-adapted species capable of vertical migration, K. brevis cells aggregated near the bottom in order to escape photo-inhibitive light intensities in the overlying water column during the day and harvested the recycled nitrogen excreted by zooplankton grazers.This concomitant relaxation of light inhibition and nitrogen-limitation ultimately led to the growth and maintenance of the red tide, constrained in near-bottom waters during much of the day and preferentially advected inshore as a result of coastal upwelling. As K. brevis was advected inshore, self-shading, CDOM, and suspended inorganic particulates all contributed to the prevention of photo-inhibitive light intensities that, in combination with the excretion of recycled ammonium, ultimately led to the maintenance of a significant red tide at the coast.
502
Dissertation (Ph.D.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 157 pages.
Includes vita.
590
Advisor: John J. Walsh, Ph.D.
653
Mathematical model.
Karenia brevis.
Gulf of Mexico.
Upwelling.
Zooplankton.
0 690
Dissertations, Academic
z USF
x Marine Science
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2068



PAGE 1

A Three-Dimensional Biophysical Model of Light, Nut rient, and Grazing Controls on Phytoplankton Competition Affecting Red Tide Mainte nance on the West Florida Shelf by Scott P. Milroy A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Major Professor: John J. Walsh, Ph.D. Kendall L. Carder, Ph.D. Gary J. Kirkpatrick, Ph.D. Gabriel A. Vargo, Ph.D. Robert H. Weisberg, Ph.D. Date of Approval: April 23, 2007 Keywords: Mathematical Model, Karenia brevis Gulf of Mexico, Upwelling, Zooplankton Copyright 2007, Scott P. Milroy

PAGE 2

Dedication To my beautiful wife Helen, who has given to me al l the things in my life I cherish most: the inspiration and support to pursu e my career as a scholar, the excitement I feel for each new day that we are together, and m ost of all, our wonderful son Ricky.

PAGE 3

Acknowledgments Were it not for the love and support from my family parents Michael and Tish and sister Libby, I would not have been blessed wit h the fortunes which have allowed me to succeed throughout the years. Were it not for m y loving wife Helen and our precious son Ricky, I would not have possessed the imaginati on, the drive, nor the stamina, to complete this odyssey. Their contributions to this document, and to my future, are incalculable. Of course, I also owe a great deal of thanks to my major professor and advisor, John J. Walsh. His support, his insights, his guid ance, and his dedication to excellence have cut within me a more perfect prism through whi ch I view the natural world. So too must I pay homage to the scholarly contributions of my graduate committee members, Kendall L. Carder, Gary J. Kirkpatrick, Gabriel A. Vargo, and Robert H. Weisberg. Finally, I would be remiss if I did not properly cr edit, and thereby reveal, Dwight Dieterle as the wizard behind the curtain in our emerald lab oratory. This work would not have been possible without supp ort from the National Oceanic and Atmospheric Administration (NA76RG0463 and NA96OP0084), the Environmental Protection Agency (R827085-01-0), the National Aeronautics and Space Administration (NAG5-6449), the Office of Naval Res earch (1435-0001-30804, N0001499-1-0212, and N00014-98-1-0158), and the State of Florida.

PAGE 4

i Table of Contents List of Tables ii List of Figures iv Abstract xi Chapter 1: Introduction 1 1.1. Observations 6 1.2. Objectives 11 Chapter 2: Methods 13 2.1. Physical Model 13 2.2. Ecological Model 19 2.2.1. State Equations and Functional Groups 21 2.2.2. Light Availability and Utilization 26 2.2.2.1. Spectral Irradiance 32 2.2.2.2. Spectral Absorption 37 2.2.2.3. Spectral Scattering 40 2.2.3. Nutrient Availability and Utilization 41 2.2.4. Zooplankton Grazing 47 2.3. Initialization and Validation of the Biophysic al Model 60 2.4. Numerical Experiments: Simulation Conditions 65 Chapter 3: Results 69 3.1. Physical Forcings 69 3.1.1. Physical Model Results 69 3.2. Ecological Forcings 71 3.2.1. Ecological Observations 71 3.2.2. Ecological Model Results 82 3.2.2.1. Case I – “No Refuge for K. brevis ” 82 3.2.2.2. Case II – “Grazer Avoidance of K. brevis ” 93 3.2.2.3. Case III – “Increased Shading for K. brevis ” 102 3.2.2.4. Case IV – “Increased Shading/Increased Grazing” 112 Chapter 4: Discussion 120 4.1. Discussion Summary 137 Chapter 5: Conclusions 139 References 141 Appendix: Model Variables, Constants, and Coefficie nts 155 About the Author End Page

PAGE 5

ii List of Tables Table 1. Data sources and dates sampled for all rel evant observations used for model initialization and validation. 9 Table 2. Competition parameters of the phytoplankto n functional groups during an annual cycle of thermally-modula ted (10 C – 30 C) growth in west Florida shelf waters (from Shanley and Vargo 1993, Penta 2000, and Walsh et al. 2001). 23 Table 3. Simplified pigment ratio matrix used durin g CHEMTAX analysis to estimate the proportional abundance o f diatoms (Pd), microflagellates (Pf), non-toxic dinoflagellates (Pn), and K. brevis (Pb) in mixed phytoplankton samples collected on the west Florida shelf, 07 July 1999. 33 Table 4. Sources of average daily streamflow data f rom United States Geological Survey (USGS) freshwater discharge gau ging stations, 31 August – 08 November 1999. 42 Table 5. Mean inorganic nutrient and suspended sedi ment concentrations of modeled estuarine systems (EPA 2006, SFWMD 2006, USGS 2006, Walsh et al. 2007). 43 Table 6. Incidence () or absence ( ) of selective herbivory (SH) and/or vertical migration (VM) behaviors among th e dominant zooplankton grazers sampled on the west Florida shelf, August – November 1999. 49 Table 7. Mean zooplankton dry weight biomass (Leste r 2005) and associated ingestion rates for the non-selective, non-migrating zooplankton grazers (Group III) sampled on the we st Florida shelf, August – November 1999. 50 Table 8. Mean zooplankton dry weight biomass (Leste r 2005) and associated ingestion rates for the selective, non -migrating zooplankton grazers (Group II) sampled on the wes t Florida shelf, August – November 1999. 51 Table 9. Mean zooplankton dry weight biomass (Leste r 2005) and associated ingestion rates for the selective, mig rating zooplankton grazers (Group I) sampled on the west Florida shelf, August – November 1999. 52

PAGE 6

iii List of Tables, continued Table 10. Model comparisons between algal standing stock, biomass accumulation due to growth, biomass removal due t o grazing stress, ammonium accumulation due to grazer excre tion, and the maximum near-bottom scalar irradiance from si mulation Cases I – IV averaged over the entire water column above the Sarasota 30 m and 10 m isobaths at 12:00 loca l hour on 07 October 1999. 90

PAGE 7

iv List of Figures Figure 1. NEGOM (blue) and ECOHAB (red) survey stat ions throughout the eastern Gulf of Mexico, Summer & Fall of 1999. 7 Figure 2. Time series of K. brevis cell counts (cells L-1) observed on the west Florida shelf, including: A) a near-bottom “i noculum” of 33,000 cells L-1 on 31 August 1999; B) a moderate patch (151,000 cells L-1) near Captiva Island, 11–14 September 1999; C) the significant red tide (1.9 x 106 cells L-1 at the 10 m Sarasota isobath, 5.9 x 106 cells L-1 near Big Sarasota Pass) offshore Sarasota, 05–07 October 1999; and D) the termination of red tide within coastal waters of the west Florida shelf by 06–08 November 1999. Assuming 300 pg C cell-1 and 30 pg C pg-1 chla for K. brevis (Walsh et al. 2001), 100,000 cells L-1 = 1.0 mg chla L-1. 10 Figure 3. The orthogonal curvilinear grid mesh of t he threedimensional coupled biophysical model of the east ern Gulf of Mexico, including the west Florida shelf. 15 Figure 4. Graphical representation of the ecologica l model (green), coupled to the WFS-POM circulation model (orange) which contains the minimal set of state variables used for hindand forecasts of red tides on the west Florida shelf using a variety of numerical schemes to estimate the relative sig nificance of grazing stress (red), spectral light (yellow), an d nutrient fluxes (purple). 20 Figure 5. Spectrally-dependent A) chlorophyll-speci fic and B) carbonspecific absorption spectra for diatoms (Pd), microflagellates (Pf), non-toxic dinoflagellates (Pn), and K. brevis (Pb), derived from HPLC and CHEMTAX analysis of in vivo phytoplankton samples collected on the west Florida shelf, 07 J uly 1999. 34 Figure 6. Geographical location of the estuarine sy stems used to estimate discharge to the west Florida shelf as a function of the average daily streamflows (m3 sec-1), 31 August – 08 November 1999. 44 Figure 7. Near-surface (A) and near-bottom (B) chlo rophylla stocks (mg L-1) observed within the eastern Gulf of Mexico, 17 – 28 August 1999 (NEGOM) and 07 – 10 September 1999 (ECOHAB). 62

PAGE 8

v List of Figures, continued Figure 8. Estimates of initial metazoan grazer abun dance (# ind. m-3) within the eastern Gulf of Mexico (based on ECOHA B zooplankton data, 07 – 10 September 1999). 64 Figure 9. Inorganic nutrient concentrations (mmol kg-1) observed within near-surface and near-bottom waters of the eastern Gulf of Mexico, 17 – 28 August (NEGOM) and 07 – 1 0 September 1999 (ECOHAB). 66 Figure 10. WFS-POM estimates of the average daily c ross-shelf transport (cm s-1) along the A) Tampa, B) Sarasota, and C) Ft. Myers transects throughout September 1999 (positive values of the dashed (-) isotachs indicate on shore transport, negative values of the solid (—) isotachs denote offshore transport, from He 2001). 70 Figure 11. AVHRR satellite images (JHU/APL 2006) of sea surface Temperature (SST) over the west Florida shelf on: A) 11 September 1999; B) 14 September 1999; C) 17 September 1999; D) 20 September 1999. 72 Figure 12. Spatial patterns of observed: A) nitrate (mmol 3NO kg-1) at the surface; B) nitrate (mmol 3NO kg-1) at the bottom; C) phosphate (mmol 3 4PO kg-1) at the surface; and D) phosphate (mmol 3 4PO kg-1) at the bottom during 07 – 10 September 1999. 73 Figure 13. Spatial patterns of observed : A) surfac e concentrations of the diazotroph Trichodesmium erythraeum (# colonies L-1); B) surface concentrations of dissolved organic ni trogen (mmol DON kg-1); C) surface salinity (psu); and D) depthaveraged concentrations of metazoan zooplankton ( # ind. m-3) on the west Florida shelf during 07 – 10 Septembe r 1999. 75 Figure 14. Spatial patterns of observed: A) chlorop hyll-a stocks (mg L-1) at the surface; B) chlorophyll-a stocks (mg L-1) at the bottom; C) silicate (mmol 4SiO kg-1) at the surface; and D) silicate (mmol 4SiO kg-1) at the bottom during 07 – 10 September 1999. Dashed (-) contour lines denote K. brevis abundance (cells L-1) measured at ECOHAB stations indicated in red. 76

PAGE 9

vi List of Figures, continued Figure 15. Spatial patterns of observed: A) nitrate (mmol -3NO kg-1) at the surface; B) nitrate (mmol -3NO kg-1) at the bottom; C) phosphate (mmol -3 4PO kg-1) at the surface; and D) phosphate (mmol -3 4PO kg-1) at the bottom during 05 – 07 October 1999. 78 Figure 16. Spatial patterns of observed: A) chloro phyll-a stocks (mg L-1) at the surface; B) chlorophyll-a stocks (mg L-1) at the bottom; C) silicate (mmol 4SiO kg-1) at the surface; and D) silicate (mmol 4SiO kg-1) at the bottom during 05 – 07 October 1999. 80 Figure 17. Spatial patterns of observed: A) nitrat e (mmol -3NO kg-1) at the surface; B) nitrate (mmol -3NO kg-1) at the bottom; C) phosphate (mmol -3 4PO kg-1) at the surface; and D) phosphate (mmol -3 4PO kg-1) at the bottom during 06 – 08 November 1999. 81 Figure 18. Spatial patterns of observed: A) chloro phyll-a stocks (mg L-1) at the surface; B) chlorophyll-a stocks (mg L-1) at the bottom; C) silicate (mmol 4SiO kg-1) at the surface; and D) silicate (mmol 4SiO kg-1) at the bottom during 06 – 08 November 1999. 83 Figure 19. Case I simulation results of the: A) near-surface and B) nearbottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999. 85 Figure 20. Results from the Case I simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelfbreak on 07 October 1999. 86 Figure 21. Results from the Case I simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida on 07 October 1999. 87

PAGE 10

vii List of Figures, continued Figure 22. Results from the Case I simulation of the: A) particulate organic nitrogen (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) near-bottom DIN:DIP ratios; and D) maximum nea rbottom scalar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis 89 Figure 23. Case I simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at midnigh t (A) and noon (B) along the Sarasota transect, 07 October 1999. 91 Figure 24. Depth profiles of total K. brevis accumulation due to new production and the growth limitation conditions f or K. brevis along the Tampa and Sarasota transect lines, from simulation Cases I – II at 12:00 local hour on 07 October 1999. Panel H contour lines indicate K. brevis standing stock (mg chl-a L-1). 92 Figure 25. Depth profiles of total diatom accumulat ion (mg chl-a L-1) due to new production, total ammonium accumulatio n (mmol N kg-1) due to zooplankton excretion, total K. brevis accumulation of particulate organic nitrogen (mmol N kg-1) due to new production, and standing stock of diss olved inorganic nitrogen (mmol N kg-1) from simulation Cases I – II at 12:00 local hour on 07 October 1999. 94 Figure 26. Case II simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999. 95 Figure 27. Results from the Case II simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelfbreak on 07 October 1999. 97

PAGE 11

viii List of Figures, continued Figure 28. Results from the Case II simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida on 07 October 1999. 98 Figure 29. Case II simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at midnigh t (A) and noon (B) along the Sarasota transect, 07 October 1999. 99 Figure 30. Results from the Case II simulation of the: A) particulate organic nitrogen (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) near-bottom DIN:DIP ratios; and D) maximum nea rbottom scalar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis. 101 Figure 31. Surface and near-bottom salinity fields, as computed by the coupled physical model (A-B) or observed during C TD casts (C-D) at ECOHAB stations during the period of red tide development (06 – 10 September 1999). 103 Figure 32. Case III simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999. 104 Figure 33. Results from the Case III simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelf-break on 07 October 1999. 105

PAGE 12

ix List of Figures, continued Figure 34. Results from the Case III simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida on 07 October 1999. 1 06 Figure 35. Case III simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at midnigh t (A) and noon (B) along the Sarasota transect, 07 October 1999. 108 Figure 36. Depth profiles of total K. brevis accumulation due to new production and the growth limitation conditions f or K. brevis along the Tampa and Sarasota transect lines, from simulation Cases III – IV at 12:00 local hour on 07 October 1999. Contour lines in panels C, D, G, and H indicate K. brevis standing stock (mg chl-a L-1). 109 Figure 37. Depth profiles of total diatom accumulat ion (mg chl-a L-1) due to new production, total ammonium accumulation (mmol N kg-1) due to zooplankton excretion, total K. brevis accumulation of particulate organic nitrogen (mmol N kg-1) due to new production, and standing stock of diss olved inorganic nitrogen (mmol N kg-1) from simulation Cases III – IV at 12:00 local hour on 07 October 1999. 110 Figure 38. Results from the Case III simulation of the: A) particulate organic nitrogen (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) near-bottom DIN:DIP ratios; and D) maximum nea rbottom scalar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis. 111 Figure 39. Case IV simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999. 113

PAGE 13

x List of Figures, continued Figure 40. Results from the Case IV simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelf-break on 07 Octobe r 1999. 115 Figure 41. Results from the Case IV simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida on 07 October 1999. 116 Figure 42. Case IV simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at midnigh t (A) and noon (B) along the Sarasota transect, 07 October 1999. 118 Figure 43. Results from the Case IV simulation of the: A) particulate organic nitrogen (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) near-bottom DIN:DIP ratios; and D) maximum nearbottom scalar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis. 119 Figure 44. K. brevis counts (cells L-1) from Florida Marine Research Institute (FMRI) sampling stations prior to 1998 (cumulative from 1954-1997). Numbers in red indicate the tot al number of samples collected from each station. 121 Figure 45. Weekly distributions of K. brevis above the 30 m, 20 m, and 10 m isobaths of the Sarasota transect during Aug ust – December 1999. The sampling interval was ~7 days (www.floridamarine.org), but only the occasions w here K. brevis were present (>1000 cells L-1) are shown. 123

PAGE 14

xi A Three-Dimensional Biophysical Model of Light, Nut rient, and Grazing Controls on Phytoplankton Competition Affecting Red Tide Mainte nance on the West Florida Shelf Scott P. Milroy ABSTRACT A coupled, three-dimensional, time-dependent numeri cal model of water circulation, spectral light, plankton dynamics, nut rient/CDOM loadings, and zooplankton grazing provided an assessment of the factors affec ting the growth and maintenance of red tides on the west Florida shelf (WFS). The cou pled biophysical model consisted of state variable quantities for temperature, salinity horizontal/vertical velocity components, turbulent diffusion, spectral light, colored dissol ved organic matter (CDOM), dissolved organic and inorganic carbon, particulate silica, f our dissolved inorganic nutrient pools (nitrate, ammonium, phosphate, and silicate), and f our phytoplankton groups (diatoms, microflagellates, non-toxic dinoflagellates, and th e red tide organism Karenia brevis). The model also included a complex grazing scheme th at utilized thirteen different zooplankton groups to explore the effects of select ive herbivory, feeding periodicity, diel vertical migration, fecal pellet egestion, and ammo nium/phosphate excretion within a diverse zooplankton community. Over the shelf and slope of the eastern Gulf of Me xico, from the Mississippi River delta to the Florida Keys, four cases of the model were run during August – November to explore the dynamics of red tide mainte nance with respect to: (1) no refuge from grazing for K. brevis; (2) grazer avoidance of K. brevis during CDOM shading; (3) grazer avoidance of K. brevis in Case II waters; and (4) increased grazing stres s on

PAGE 15

xii K. brevis competitors. NEGOM and ECOHAB data sets during Ju ly – November 1999 were used to establish the initial/boundary conditi ons and provided validation data for the coupled model as well. Model results indicate that the red tide of 5.9 x 106 cells L-1 witnessed offshore Sarasota, Florida on 07 October 1999 was initiated by an inoculum of K. brevis observed in near-bottom waters above the 30 m isobath offsho re Sarasota on 31 August 1999. Flowfields measured at moored ADCPs, observations f rom AVHRR satellite imagery, and west Florida shelf circulation models indicate that conditions of coastal upwelling existed during the period of bloom development, suc h that the K. brevis inoculum was delivered to the coast in the bottom Ekman layer. As a shade-adapted species capable of vertical mig ration, K. brevis cells aggregated near the bottom in order to escape photo -inhibitive light intensities in the overlying water column during the day and harvested the recycled nitrogen excreted by zooplankton grazers. This concomitant relaxation o f light inhibition and nitrogenlimitation ultimately led to the growth and mainten ance of the red tide, constrained in near-bottom waters during much of the day and prefe rentially advected inshore as a result of coastal upwelling. As K. brevis was advected inshore, self-shading, CDOM, and suspended inorganic particulates all contributed to the prevention of photo-inhibitive light intensities that, in combination with the excretion of recycled ammonium, ultimately led to the maintenance of a significant red tide at the coast.

PAGE 16

1 Chapter 1: Introduction Although the concept of species succession was or iginally developed from research in terrestrial ecology, the foundation for a wide variety of plankton species succession models has been well laid by several aut hors in past years (Conover 1956, Margalef 1968, Smayda 1980, Karentz & Smayda 1984, Evans 1988, Sommer 1989). While the differences among these models are as sub tle as they are numerous, the fundamental forcings at work within these successio n paradigms remain the same: whether random or periodic in occurrence, any ecolo gical event which alters the physical, chemical, and/or biological environment will also l ead to differential selective pressure on the plankton species within the affected communi ty, thereby modulating the population dynamics of the plankton species therein Thus, a thorough effort to measure and evaluate pertinent environmental conditions sho uld enable predictions of recurrent successional patterns and/or the potential for epis odic blooms of particular planktonic species. Among these models of phytoplankton succession, the development of accurate successional models of harmful algal blooms is of p aramount importance, particularly in regions that are most susceptible to their detrimen tal effects. This is especially true along the west Florida shelf (WFS) and coast, where episo dic bloom events of the unarmored dinoflagellate Karenia brevis (=Gymnodinium breve, Ptychodiscus brevis) has caused mass mortalities of over 100 species of marine life (Steidinger & Ingle 1972, Steidinger 1983, Sakamoto et al. 1987), the results of which r outinely impact the valuable commercial fisheries and tourism industries of west Florida.

PAGE 17

2 The red tide organism K. brevis has been linked to these mortalities either direct ly due to the production and release of a variety of n eurotoxins (Shimizu et al. 1995) or indirectly due to the depletion of dissolved oxygen during periods of decreased production and increased respiration (Simon & Dauer 1972). The bioaccumulation of sub-lethal brevitoxins in fish and mollusk tissues has led to significant dolphin and seabird mortalities (Steidinger & Haddad 1981) and has even been linked to neurotoxic shellfish poisoning (NSP) in humans (Shimizu et al. 1986). Brevitoxins released from lysed cells may also become airborne, and in the pr esence of breaking waves and/or onshore winds, may cause severe respiratory irritat ion to humans along the coast (Pierce et al. 1990). Thus, when considering the multitude of ecological, commercial, and public health hazards associated with these blooms, the be nefits of a reliable predictive model of K. brevis population dynamics are obvious. Until recently, the causative elements that underli e the initiation and dispersal of red tides had remained enigmatic. Several paradigm s had been developed throughout the years by a multitude of researchers, all of which s ought to reduce the complexities of K. brevis bloom dynamics to a simple index of one or very fe w requisite environmental predictors. Unfortunately, these efforts succeeded only in providing a woeful array of contradictory information in the literature. For example, anthropogenic eutrophication of coasta l waters has often been cited as a main cause of red tides (Satsmadjis & Friligos 1983, Cannon 1990, Pagou & Ignatiades 1990); however, since toxic dinoflagella tes such as K. brevis exhibit several competitive disadvantages compared to other phytopl ankton in nutrient replete environments (Penta 2000, Walsh et al. 2001), eutro phication would favor the initiation

PAGE 18

3 of a K. brevis bloom only in the absence of other phytoplankton c ompetitors. However, many areas suffering from eutrophication have not w itnessed any obvious changes in red tide frequency (Wyatt 1990). In fact, Ryther (1955 ) has even suggested that high nutrient concentrations are not at all required for the deve lopment of a K. brevis bloom. Specific to the west Florida shelf, research indicates that oligotrophic (rather than eutrophic) conditions correlate well with periods of red tide initiation (Steidinger & Haddad 1981, Vargo et al. 2003). Of course, eutrophication of coastal waters via nat ural processes (i.e. terrestrial rainfall and subsequent run-off) has also been expl ored as a potential cause of red tides, particularly along the WFS. Toward this end, rainf all measurements were employed in the development of a “Red Tide Index” by Chew (1956 ) with the hope that riverine efflux data could be used to accurately predict K. brevis blooms as a function of estuarine nutrient loading via terrestrial run-off and subseq uent transport offshore. Although the method of nutrient delivery in Chew’s model seemed somewhat encouraging in its simplicity, the differential utilization of those n utrients within the phytoplankton community (i.e. the role of inter-specific competit ion) was not addressed sufficiently to yield a reliable predictor of red tide initiation. Certainly, the role of nutrient availability alone cannot adequately presage the development of a red tide. Other methods used to predict red tide development have focused on the complementary variations in temperature and salinit y in an effort to develop a reliable link to the surrounding physical environment. Of c ourse, virtually all phytoplankton groups exhibit accelerated growth rates as temperat ures increase, so it is no surprise that several authors (Chew 1956, Fraga et al. 1990, Pago u & Ignatiades 1990) had observed

PAGE 19

4 positive correlations between red tides and warm te mperatures. Preliminary investigations into the salinity requirements for r ed tide initiation yielded a 21-37 ppt favorable growth range for K. brevis (Rounsefell & Dragovich 1966, Dragovich et al. 1968), although subsequent work had narrowed the op timal range to 32-35 ppt (Wilson 1967), which is consistent with the assertion that K. brevis is indigenous to oligotrophic offshore waters (Steidinger & Haddad 1981, Tester & Steidinger 1997). A salinity relationship has even been used in the analysis of nearshore processes, particularly in the context of K. brevis coincidence with salinity fronts (Vargo et al. 200 3). Indeed, the significance of coastal density front s and various upwelling paradigms would seem to offer the most promising av enues for red tide prediction. Several authors have associated red tide initiation with periods of decreased vertical mixing, such that stability of the water column cor relates well with red tide initiation and persistence (Margalef et al. 1979, Cannon 1990, Wya tt 1990). The increased growth capacity of red tide organisms has also been linked to upwelling, even in stratified waters where the subsequent accumulation of biomass along density fronts has led to localized blooms (Margalef et al. 1979, Taft & Martin 1986, F raga et al. 1990, Franks 1992, Vargo et al. 2003). In fact, the physical processes which give rise to the accumulation of biomass at these density fronts may be as important as the bio logical processes of bloom initiation and growth, especially on the WFS. If coastal red tides are indeed initiated offshore (Steidinger & Haddad 1981, Tester & Steidinger 1997 ), upwelling events would facilitate the cross-shelf transport of those material propert ies necessary to sustain a coastal red tide. For the case of wind-driven upwelling, along shore wind stress may cause a drop in

PAGE 20

5 the coastal sea level as surface waters are transpo rted offshore, resulting in an alongshore coastal jet and an associated onshore transport wit hin the bottom Ekman layer (Weisberg et al. 2000). If the background hydrographic condi tions are favorable for cold bottom water to be advected shoreward and mixed to the sur face, these upwelling events may be indicated by changes in the sea surface temperature (SST). On the WFS, coastal jets wrought from wind-driven u pwelling are typically centered between the 20 m and 30 m isobaths, where onshore across-isobath flows are maximal (Weisberg et al. 2000). Since stratificati on provides a buoyant inhibition to upwelling, water over the inner shelf must enter a shallower depth before turbulent mixing is sufficient for it to reach the surface (W eisberg et al. 2004). Given the unique geometry of the WFS isobaths, stratification ultima tely leads to a greater upwelling response on the inner shelf compared to the downwel ling response during oscillatory winds (Weisberg et al. 2001). Although the Lagrang ian trajectories that feed the inner shelf between Tampa Bay and Charlotte Harbor typica lly originate in the Big Bend (Weisberg & He 2003), it is precisely the bottom Ek man layer which provides the conduit for across-shelf transport. In fact, the n earshore delivery of a red tide initiated on the WFS would require the cross-shelf transport of the inoculum in the bottom Ekman layer. Thus, it is necessary to consider both the biological and physical processes which affect red tide delivery and maintenance near the c oast. Despite these promising leads, the multitude of bio logical, chemical, and physical interactions that define the phytoplankton communit y of the WFS waters necessarily require a complex analysis of their contributions in toto. As our understanding of phytoplankton ecology increases, so too does the pr edictive power of the more recent

PAGE 21

6 (and decidedly more complex) ecological simulations Chief among these is the biophysical model offered by Walsh et al. (2007) ou tlining several prerequisites for the successful competition of K. brevis in west Florida shelf waters, whereby the initiati on and persistence of red tides follow the strict chro nology: 1) A mid-shelf, phosphorus-rich nutrient supply (at low DIN/DIP ratios) 2) Aeolian delivery of iron-rich dust to alleviate diazotroph Fe-limitation 3) Relaxation of N-limitation due to diazotroph rel ease of DON 4) Co-aggregation of sun-adapted diazotrophs and sh ade-adapted K. brevis 5) Vertical migration of K. brevis into near-bottom Ekman layers 6) Coastal upwelling of K. brevis into coastal, CDOM-rich surface waters 7) Release of ichthyotoxic exudates from K. brevis, causing fish-kills 8) In situ growth of self-shaded K. brevis, fed by decaying diazotrophs/fish While the biophysical model of Walsh et al. (2007) has been able to explain red tide initiation along the west Florida shelf, the factor s affecting red tide maintenance require further investigation. As this and other hindcast models serve to further that end, it is hoped that an operational predictive model of red t ide dynamics may be developed. 1.1. Observations Cross-shelf sections taken to the 1000 m isobath wi thin the eastern Gulf of Mexico during August – November 1999 were availabl e as part of the concurrent NOAA/EPA Ecology and Oceanography of Harmful Algal Blooms (ECOHAB) and the MMS Northeastern Gulf of Mexico (NEGOM) research in itiatives (Fig. 1). Hydrography from these cruises provided the initialization and validation data for the hindcast analysis

PAGE 22

7 F i g u r e 1 N E G O M ( b l u e ) a n d E C O H A B ( r e d ) s u r v e y s t a t i o n s t h r o u g h o u t t h e e a s t e r n G u l f o f M e x i c o S u m m e r & F a l l o f 1 9 9 9 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 NEGOM Hydrography Hydrography/K. brevis ZooplanktonK. brevisADCPs

PAGE 23

8 of red tide maintenance during August – November 19 99 (Table 1) and included measures of: 1) biophysical and bio-optical oceanographic parame ters, such as temperature, salinity, density, light attenuation, and transmiss ion (ECOHAB, NEGOM); 2) in situ nutrient concentrations, including ammonium, nitra te, nitrite, phosphate, silicate, and DON (ECOHAB, NEGOM); 3) in vivo chlorophyll-a and HPLC spectral absorption data (ECOHAB, NEGOM); 4) offshore and coastal K. brevis cell counts (ECOHAB); and 5) zooplankton diversity, biomass, and abundance (E COHAB). On 31 August 1999, a deep patch of K. brevis (33,000 cells L-1) was sampled in the near-bottom waters at the 30 m isobath offshore Sarasota (Fig. 2), just 37 days prior to a large bloom event (5.9 x106 cells L-1) observed at the coast during 05 – 07 October 1999. If the deep-water patch of K. brevis was the inoculum for the coastal bloom event, it would have required the maintenance of K. brevis in the near-bottom Ekman layers and subsequent advection into nearshore waters during a period of coastal upwelling. This paradigm would also require a light and nutrient cl imate favorable to the net growth of the K. brevis biomass on the inner shelf, particularly in light of potential losses due to zooplankton grazing. By 08 November 1999, K. brevis stocks had returned to background concentrations offshore Sarasota, indica ting that the conditions that originally favored red tide maintenance in nearshor e waters had changed. Since the forcings that contributed to the transport, mainten ance, and ultimate fate of the red tide

PAGE 24

9 Table 1. Data sources and dates sampled for all re levant observations used for model initialization and validation. Initialization Data Sources Data Dates NEGOM NO3, NH4, PO4, SiO2, Chla, temperature, salinity 17-28 AUG 1999 ECOHAB Phytoplankton pigments; spectral absorption 07 JUL 1999 “ K. brevis cell counts 24 AUG – 10 SEP 1999 “ NO3, PO4, SiO2, Chla, temperature, salinity, zooplankton species richness, abundance, and biomas s 07-10 SEP 1999 V alidation Data Sources Data Dates ECOHAB Monthly mean currents, temperature, salinity SEP – NOV 1999 “ NO3, PO4, SiO2, Chla temperature, salinity, K. brevis zooplankton species richness, abundance, and biomas s 05-07 OCT, 06-07 NOV 1999 “ K. brevis cell counts 01 SEP – 08 NOV 1999 JHU/APL (2006) AVHRR sea surface temperature (SST) 11-17 SEP 1999

PAGE 25

10 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) 31 AUG 19993 3 0 0 0 ma x -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 5 9 0 0 0 0 0 m a x C) 05-07 OCT 1999 B) 11-14 SEP 1999 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 1 5 1 0 0 0 m a x D) 06-08 NOV 1999 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Figure 2. Time series of K. brevis cell counts (cells L-1) observed on the west Florida shelf, including: A) a near-bottom “inoculum” of 3 3,000 cells L-1 on 31 August 1999; B) a moderate patch (151,000 cells L-1) near Captiva Island, 11–14 September 1999; C) the significant red tide (1.9 x 106 cells L-1 at the 10m Sarasota isobath, 5.9 x 106 cells L-1 near Big Sarasota Pass) offshore Sarasota, 05–07 Oc tober 1999; and D) the absence of red tide within coastal waters of the west Florida shelf by 06–08 November 1999. Assuming 300 pg C cell-1 and 30 pg C pg-1 chl-a for K. brevis (Walsh et al. 2001), 100,000 cells L-1 = 1.0 mg chl-a L-1.

PAGE 26

11 were temporally constrained to 31 August – 08 Novem ber 1999, a coupled biophysical model was used for the hindcast analysis of this pa rticular red tide and the maintenance factors which contributed to the event. 1.2. Objectives Indeed, the factors affecting the maintenance of red tides are less well-known than those of onset. While previous one-dimensional (Pe nta 2000) and three-dimensional (Walsh & Steidinger 2001, Walsh et al. 2001, Walsh et al. 2002, Walsh et al. 2003) models have offered insight into the K. brevis dynamics on the WFS, it is difficult to embark upon exhaustive ecological simulations witho ut the benefit of a robust dataset to initialize and validate such models. Fortunately, hydrographic data from several cruises on the west Florida shelf were available for the de velopment of this model (Table 1). Originally, a three-dimensional biophysical model w as used in the analysis of a 1979 red tide of K. brevis to explore the mechanisms of bloom initiation, tra nsport, and fate (Walsh et al. 2002). Significant modification s to the Walsh et al. (2002) model were made in an effort to explore the events affecting t he maintenance and inshore transport of the October 1999 red tide offshore Sarasota, Florid a to address the following questions: Are the measured and modeled flow fields of a magn itude and direction such that a red tide initiated from an offshore ino culum could be advected onshore in keeping with observed events? Is in situ growth of K. brevis sufficient to reach the observed bloom proportions without having to invoke paradigms of f rontal aggregation?

PAGE 27

12 To what extent does the differential utilization o f spectral irradiance contribute to phytoplankton competition and red tid e maintenance? To what extent do disparate supplies of inorganic nutrients along the west Florida shelf affect phytoplankton competition and red tide maintenance? To what extent is zooplankton grazing a significan t contributor to red tide growth maintenance and/or bloom termination? Results from four numerical experiments were analy zed with the goal of elucidating which state variables were most predict ive of the observed K. brevis dynamics and therefore critical to the development of an operational forecast model of red tides on the west Florida shelf. These experim ents were performed under the following conditions: Case I ) A diverse (13-member) grazer community capable o f differential ingestion as a function of prey selection, verti cal migration, and feeding periodicity; no refuge from grazing for K. brevis; alleviation of light stress derived from fixed, shelf-wide estimates of dissolved organic carbon (DOC). Case II ) Case I conditions but with the removal of all grazing str ess from K. brevis; alleviation of light stress derived from variable shelfwide estimates of DOC as a function of ambient sa linity. Case III ) Case II conditions but with the further alleviation of lig ht stress due to the effects of coastal turbidity inshore of the 30 m isobath. Case IV ) Case III conditions but with a simplified (2-member) grazer community with generalized metazoan and protozoa n grazers.

PAGE 28

13 Chapter 2: Methods 2.1. Physical Model The physical circulation model used to resolve time -dependent flows on the WFS is an adaptation of the Blumberg and Mellor (1987) Princeton Ocean Model (POM), used extensively to model WFS circulation (Li & Weisberg 1999a, Li & Weisberg 1999b, Weisberg et al. 2000, Weisberg et al. 2001, He & We isberg 2002, He & Weisberg 2003, Weisberg & He 2003, Weisberg et al. 2004). The WFS -POM consists of a domain which extends from the Florida Keys to the Mississippi Ri ver delta (Fig. 3), partitioned into an orthogonal curvilinear grid with a horizontal resol ution varying between ~2 km near the coast and ~6 km near the open boundary (He & Weisber g 2003). A topographyfollowing sigma coordinate system, where s = (z-h) (H+h)-1, is employed in the vertical dimension with 21 discrete layers distributed non-u niformly to better resolve the nearsurface and near-bottom frictional boundary layers (He & Weisberg 2003). The circulation model utilized in this study has pr eviously been found to be quite suitable for modeling large-scale, quasi-horizontal flows under a variety of conditions. While Yang et al. (1999) studied the WFS response t o climatological monthly mean wind forcing, the monthly mean wind stress alone was not sufficient to account for all observed currents (He & Weisberg 2001). Further, m odeling investigations of the forcing due to synoptic scale winds under constant density (Li & Weisberg 1999a, b) confirmed a simple Ekman-geostrophic route to spin-up and ident ified regional upwelling centers promoted by coastline and isobath geometries (He & Weisberg 2001). Thus, Weisberg et al. (2001) were able to show an asymmetrical respon se to up/downwelling-favorable winds on the inner WFS under stratified conditions. Ultimately, these investigations

PAGE 29

14 were used to improve the fidelity of the circulatio n model employed by He & Weisberg (2001, 2003), which investigated the relative effec ts of buoyancy and wind forcing over a broader region of the WFS, thereby elucidating the exchange of water between the deep ocean and different regions of the shelf (He & Weis berg 2001). The equations that govern general shelf circulation and surface elevation within the WFS-POM as originally defined by Blumberg and M ellor (1987) and used in Yang and Weisberg (1999): 0 ) ( ) ( = + + + t y DV x DU h s w (1) x MF U D K x gD fVD U y UVD x UUD t UD + = + + + + s s h s w) ( ) ( ) ( ) ( (2) y MF V D K y gD fUD V y VVD x UVD t VD + = + + + + + s s h s w) ( ) ( ) ( ) ( (3) where U and V are the horizontal velocity components (assuming c onstant density), w is the vertical velocity component normal to the sigma surface, D = H + h, KM is a variable vertical kinematic viscosity modeled according to t he level 2.5 turbulence closure scheme of Mellor and Yamada (1982), f is the Coriolis parameter, and g is the acceleration due to gravity. The horizontal friction terms, Fx and Fy, were defined as: () ()yx xx xD y D x Ft t + = (4) () ()yy xy yD y D x Ft t + = (5)

PAGE 30

15 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 LongitudeLatitude Figure 3. The orthogonal curvilinear grid mesh of the three-dimensional coupled biophysical model of the eastern G u l f o f M e x i c o i n c l u d i n g t h e w e s t F l o r i d a s h e l f

PAGE 31

16 such that: x U AM xx = 2t (6) + = = x V y U AM yx xyt t (7) y V AM yy = 2t (8) where AM is the horizontal viscosity modeled according to t he Smagorinsky (1963) formulation. The He and Weisberg (2003) formulation of the WFS-P OM is initialized at rest with horizontally uniform stratification, swiftly g enerating baroclinicity in balance with the wind, air pressure, and buoyancy forcing. Tida l forcing is excluded since, with tidal currents of only a few cm s-1, the related tidal mixing is weak in comparison wi th other sources for mixing (He & Weisberg 2003). Atmospher ic forcing is accomplished by interpolating the wind, air pressure, and net surfa ce heat flux fields from the National Center for Environment Prediction (NCEP) to generat e realistic baroclinicity in keeping with the spatial scale of the WFS-POM (He & Weisber g 2003) following the methodology first described by He and Weisberg (200 2). This includes a relaxation of the model sea surface temperature (SST) to an obser ved SST to account for errors in the surface heat flux. To accomplish this, the OI-SST product of He et al. (2002) is employed. The model uses an Arakawa C grid conserving both linear and quadratic quantities such as mass and energy (Blumberg & Mell or 1987). A mode-splitting technique was employed to solve the two-dimensional external mode equations in short

PAGE 32

17 (12 sec) time steps, while the three-dimensional in ternal mode equations were solved in longer (360 sec) time steps (He & Weisberg 2003), b oth of which fall within the CourantFriedrichs-Levy (CFL) stability constraints (Roach 1976). A centered leapfrog technique was utilized with spa tial gradients approximated by centered differencing, while contributions from the advective terms were obtained using a scheme developed by Easter (1993) to solve multi-di mensional transport equations. This particular scheme uses both the positive definite u pstream advection algorithm of Bott’s (1989) and the time-splitting mode suggested by Roa ch (1976) to reduce phase and amplitude errors, as well as numerical diffusion (W alsh et al. 2002). Boundary conditions were established such that w = 0 at the sea surface and seafloor; at the sea surface, U and V are related to the wind stress components tx and ty: [ ] 0 as 1 = n s t t r s soy x MV U D K (9) where ro is the water density. At the seafloor, the horizo ntal velocity components are zero and vertical shear is specified by: [ ] ()1 as ,2/1 2 2 + = n -s s sV U V U C V U D KZ M (10) where: () {} n + =-0025 .0, / 1 ln2 1 2osz H k Ckb ZMAX (11) k = 0.4 is the von Karman constant, zo is the roughness parameter, and kb – 1 is the vertical level next to the seafloor (Blumberg & Mel lor 1987). Thus, while the sea surface stress is specified, the bottom stress is determine d internally within the physical model.

PAGE 33

18 From a state of rest, the surface motion is first d riven by the specified surface wind stress, which generates an interior flow through the pressu re gradient and the vertical mixing of momentum according to the Mellor and Yamada (1982) turbulence closure scheme. The motion near the sea floor then generates the bottom stress by Eq. 11 with the velocity equal to zero at the sea floor (Blumberg & Mellor 1 987). Horizontal boundary conditions over land are imple mented by a land mask which ensures that the normal velocity along the coastlin e is set to zero. Impacts from an idealized Loop Current (LC) intrusion are mimicked by imposing a Gaussian-shaped sea surface height perturbation along the open boundary west of the Florida Keys, where geostrophic in/outflows may occur on either side of the boundary, ensuring a freelyevolving current within the model domain (He & Weis berg 2003). Other than the imposed pressure perturbation, the open boundary is treated with an Orlansky (1976) radiation condition. Additionally, the interpolate d monthly mean freshwater mass fluxes from the Mississippi, Mobile, Apalachicola, Suwanne e, Hillsborough, Peace, and Shark rivers are input to the top sigma level at the grid cells closest to the rivers’ locations (He & Weisberg 2003). Because of the large width of the WFS combined wit h its unique coastline/isobath geometries (Weisberg et al. 2000, Weisberg et al. 2 001, He & Weisberg 2003, Weisberg & He 2003, Weisberg et al. 2004), sporadic intrusio ns of slope water near the shelf break (Li & Weisberg 1999a, Weisberg et al. 2000, He & We isberg 2003, Weisberg & He 2003, Weisberg et al. 2004), and differential sensi tivity to local wind forcing during periods of stratification (He & Weisberg 2003, Weis berg & He 2003), the WFS is host to a wide variety of transient physical phenomena whic h influence the overall shelf

PAGE 34

19 circulation patterns in a very complex way. Among the most ecologically-important shelf circulation patterns are those which result i n up/downwelling. These events are significant because they provide the physical means for the across-shelf transport of discrete water masses (Li & Weisberg 1999a, Weisber g et al. 2000, Weisberg et al. 2001, Weisberg & He 2003). Since the WFS circulation is intimately tied to the ecological processes which affect phytoplankton competition an d red tide maintenance, physical model estimates of the daily horizontal ( u v ) and vertical ( w ) flows, as well as the temperature ( T ) and salinity ( S ) fields, were interpolated throughout the spatial domain in 360 sec time steps and were ultimately used to forc e the coupled ecological model. 2.2. Ecological Model A total of 22 state variables were used in the eco logical simulation (Fig. 4), including the carbon-specific biomass among four co mpeting phytoplankton groups (diatoms, microflagellates, non-toxic dinoflagellat es, and Karenia brevis ) which were subjected to grazing losses. Four dissolved inorga nic nutrient pools ( SiO4 3NO + 4NH 3 4PO ), two dissolved carbon pools (DIC, DOC), particula te silica (PSi), and the zooplankton excretion of + 4NH and3 4PO were also among the state variables employed in the ecological model, in addition to the total s calar irradiance at depth (Eo). Within the framework of the ecological simulation, calculations were performed as a synthesis of three distinct numerical schemes. The first was a detailed light scheme used to simulate the overall carbon-specific growth of the phytoplankton groups as functions of the periodicity, intensity, and spectr al composition of the available light

PAGE 35

20 Figure 4. Graphical representation of the ecologic al model (green), coupled to the WFSPOM circulation model (orange), which contains the minimal set of state variables used for hindand forecasts of red tides on the west Fl orida shelf using a variety of numerical schemes to estimate the relative significance of gr azing stress (red), spectral light (yellow), and nutrient fluxes (purple).

PAGE 36

21 field. Phytoplankton growth is of course mitigated by the uptake kinetics of dissolved inorganic nutrients, the availability of which was calculated in the nutrient scheme as a complex function of in situ and estuarine supplies, advective transport, bacte rial remineralization of particulate organic material, a nd zooplankton excretion. Paradigms of grazer migration and prey selection w ere explored first with a complex grazing scheme, employing different grazer species within three generalized groups (selective migrators, selective non-migrator s, and non-selective non-migrators) to most accurately represent the true species richness and abundance within the zooplankton community measured at paired ECOHAB stations (Table 1). A second grazing scheme, which calculated grazing losses from a community of just two grazers (protozoan and metazoan), was later used to test the sensitivity o f the computed grazing losses to the measures of zooplankton diversity. 2.2.1. Ecological Model: State Equations and Funct ional Groups The state equation which governs phytoplankton gro wth (Penta 2000, Walsh et al. 2001, Walsh et al. 2002) is of the form: i i s i i i i i i i i i iP w z P dl P d P de P d dP Tr t P_ * + = g m (12) where the time-dependent change in the carbon-speci fic biomass of the ith phytoplankton group ( t Pi /) is defined as the advective and diffusive transp ort ( idP Tr) of the phytoplankton biomass and the realized net growth (i iP d*m) thereof, less any losses due

PAGE 37

22 to phytoplankton excretion (i iP de ), grazing (i iP dg), cell lysis (i iP dl ), and sinking ( i i sP w z_/ ), if applicable. Since the carbon-specific losses due to phytoplan kton excretion (i iP de ) and cell lysis (i iP dl ) were assumed to be included implicitly in the cal culation of the realized net growth (i iP d*m) for each phytoplankton group, Eq. 12 simply becom es: i i s i i i i i iP w z P d P d dP Tr t P_ * + = g m (13) where the physical advective transport is defined a s: () () () n + + = B h h vdB h udB h h h dB Tradvw s V x2 1 1 2 2 11 (14) and the diffusive transport is: =s sB d K dB Trh diff (15) where B represents any of the state variables subje ct to physical transport. The functional groups of phytoplankton in the sim ulations were based upon their historical prevalence in WFS waters, the availabili ty and extent of in situ data, and their disparate sensitivities to spectral light, dissolve d inorganic nutrients, and grazing losses within the context of competition. These competiti on parameters follow previous ecological simulations of WFS waters and the phytop lankton groups found therein (Penta 2000, Walsh & Steidinger 2001, Walsh et al. 2001, W alsh et al. 2002) and are summarized in Table 2. Additional ecological model variables, constants, and coefficients are explained in the Appendix.

PAGE 38

23 Table 2. Competition parameters of the phytoplankt on functional groups during an annual cycle of thermally-modulated (10 C – 30 C) growth in west Florida shelf waters (from Shanley & Vargo 1993, Penta 2000, and Walsh et al. 2001). CompetitionDiatomsMicroflagellatesDinoflagellates K. brevis Parameter(Pd)(Pf)(Pn)(Pb)Diameter (mm): 2522525 ws (m day-1): 0.030 1 m hr-11 m hr-1mmax (day-1 @ 10C): 0.550.50.30.2mmax (day-1 @ 20C): 1.10.90.60.4mmax (day-1 @ 30C): 2.21.81.20.8 Ecomp (mE m-2 s-1): 6666 Esat (mE m-2 s-1): 19027515065 Einhib (mE m-2 s-1): 1940281015301530 knitrate (mmol m-3): 1.050.21.80.5 kammonium (mmol m-3): 1.50.20.90.5 kphosphate (mmol m-3): 0.50.30.20.2 ksilicate (mmol m-3): 1.15N/AN/AN/A C/cell (pg):8756600300C/chl a (pg pg-1): 5010010030 C/N/P (molar):100 / 15 / 1100 / 15 / 1100 / 15 / 11 00 / 15 / 1

PAGE 39

24 The diatom group (Division Chromophyta, Class Bacil lariophyceae) is included due to their competitive advantage in the autotroph ic biomass of west Florida shelf waters (Saunders & Glenn 1969) as a result of their high growth rates under high light saturation intensities (Penta 2000, Walsh et al. 20 01). Diatoms are unique in that they require silicate ( SiO4) in order to construct their cellular frustules; t hus, a lack of dissolved silica in the water column can significan tly impact diatom growth. Therefore, diatom success is generally dependent upon silica a vailability during periods of nutrient enrichment (Jeffrey & Vesk 1997), especially at hig h irradiances. The microflagellates represent those small autotrop hic and mixotrophic eukaryotes from the Divisions Chlorophyta (Classes Euglenophyceae, Prasinophyceae, and Chlorophyceae) and Chromophyta (Classes Cryptop hyceae, Raphidophyceae, Chrysophyceae, Dictyochophyceae, and the non-coccol ithophorid Prymnesiophyceae) which possess one or more flagella (Penta 2000). D ue to their small size, they are assumed to have negligible swimming speeds and are neutrally buoyant in the water column. Like the diatoms, the competitive advantag e enjoyed by the microflagellates is conferred largely due to their high growth rates du ring periods of increased irradiance. However, microflagellates typically cannot compete against the diatoms unless silica limitation is in effect (Walsh et al. 2003). While their small size also affords them protection from the larger metazoan grazers (to whi ch all the other functional groups are differentially susceptible), microflagellates are s ubjected to heavy grazing pressure by protozoan grazers. Historically, the non-toxic dinoflagellates (Divisi on Dinophyta) are less prevalent within the WFS phytoplankton community. However, t hey are included due to their

PAGE 40

25 persistence in the background population of WFS aut otrophs (Steidinger & Joyce Jr. 1973) and to explore paradigms of differential succ ess between the toxic and non-toxic dinoflagellates. Although the dinoflagellate group s have relatively moderate growth rates and nutrient uptake efficiencies, they represent th ose groups which actively migrate to or below the nutricline (Kamykowski et al. 1998) for t he absorption of nutrients at night and return to the surface for photosynthesis during the day. Although the toxic dinoflagellate K. brevis would seem to be less competitive than the non-toxic dinoflagellates (as evidenced by its slightly lower growth rate), it is highly shade-adapted, may reach nitrogen-saturation amid lower concentrations of DIN, and is of course highly toxic to many species, incl uding potential grazers (Stoecker & Sanders 1985, Huntley et al. 1986). Since K. brevis cells exhibit a light-induced negative geotaxis (H eil 1986, Kamykowski et al. 1998), this typically results in a daylight ascent in the water column. Because K. brevis also exhibits a “stop response” (Baden & Mende 197 8) under light intensities of ~150 mE m-2 s-1 (Shanley 1985), a previous model of K. brevis vertical migration (Walsh et al. 2002) had imposed an upward movement at light levels <65 mE m-2 s-1, except during night-time hours when a positive ge otaxis was used to simulate nocturnal convection and wind-induced mixing. In t his model, the vertical migration of K. brevis was controlled solely by the ambient light field, such that light levels below the saturation intensity ( Esat) of 65 mE m-2 s-1 induced upward movement while those the beyond Esat reversed the direction of vertical migration.

PAGE 41

26 2.2.2. Ecological Model: Light Availability and Ut ilization Since the growth of the phytoplankton biomass wit hin each functional group is limited by the least-available growth factor (Walsh & Dieterle 1994), the effects of light and nutrient limitation are not assumed to be multi plicative. Thus, the realized net growth rate (* im) is defined as the lesser of the light-limited (i ll_m) and nutrient-limited (i nl_m) growth rates for each phytoplankton group: r r =i nl i ll iMIN_ *m m m (16) The determination of i ll _m and i nl _m begins the effects of ambient water temperature on balanced carbon-specific growth, rel ative to the maximum theoretical growth rate (i max_m) for each group (Table 2). Using the i max_m values for each functional group at varying temperatures (10 C, 20 C, and 30 C), the temperaturedependent maximum growth rates (i T maxm) for the diatoms (Pd), microflagellates (Pf), non-toxic Dinoflagellates (Pn), and K. brevis (Pb) were estimated via quadratic regression, such that: 3 5 2 15 max10 2.9 0027 .0 0733 .0 10 6.4 T x T T xPd T-+ + =m (17) 3 4 2 15 max10 0.1 0035 .0 0750 .0 10 3.3 T x T T xPf T-+ + =m (18) 3 5 2 16 max10 0.5 0015 .0 0400 .0 10 3.9 T x T T xPn T-+ + =m (19) 3 5 2 15 max10 3.3 0010 .0 0267 .0 10 3.3 T x T T xPb T-+ + =m (20) where T represents the ambient water temperature in degrees Celsius.

PAGE 42

27 Although i T maxm is given in units of day-1, growth does not take place throughout the entire day; in fact, the model simulation of ph ytoplankton growth must be adjusted so that the “daily” growth is calculated only during t he light cycle of the daily photo-period, which is a fraction of the 24-hour day. Of course, the length of the light cycle (lf) within a particular day’s photo-period is a function of so lar declination (relative to the time of year) and the latitude at which the calculation is being performed. Thus, the julian day (jd) was used to express each day as a fraction of the earth’s orbital period (jd): = 365 2j jdp d (21) thereby enabling the calculation (Gregg & Carder 19 90, Walsh & Dieterle 1994) of solar declination (solarD, in radians) as: ) 3( 00148 .0 ) 3( 002697 .0 ) 2( 000907 .0 ) 2( 006758 .0 ) ( 070257 .0 ) ( 399912 .0 006918 .0j j j j j j solarSIN COS SIN COS SIN COS Dd d d d d d + + + = (22) The number of daylight hours within the light cycle (lf) at the reference latitude (refLAT ) was then determined according to the relationship : n = 1 180 24solar ref lD TAN LAT TAN ACOSp p f (23) enabling the temperature-dependent maximum growth r ate (i T_ maxm) for each phytoplankton group to be modified to yield a maxim um hourly growth rate (i T_ maxm ¢ ): = ¢l i T i Tf m m_ max max (2 4)

PAGE 43

28 which was applied to growth over each daylight hour The light-limited growth rate was calculated at eac h grid and time interval when the quantity of scalar irradiance at depth (Eo) exceeded compensation irradiance (Ecomp); otherwise, i ll _m was set to zero. Jassby & Platt (1976) define i ll _m by the hyperbolic tangent function: ( ) i T i T comp o i i llE E TANH_ max max _m m a m¢ n n ¢ = (25) Shanley and Vargo (1993) indicate that the compensa tion intensity (Ecomp) for K. brevis is ~6.0 mE m-2 s-1, such that K. brevis is the only shade-adapted phytoplankton in the simulation. Thus, it was assumed that 6.0 mE m-2 s-1 was below the compensation intensity for the remaining phytoplankton groups. Photosynthetic efficiency (ia) is calculated as a function of maximum quantum yield (iF) relative to the carbon-specific absorption coeffi cient (* ia) over all photosynthetically active radiation (PAR) wavelengt hs for each phytoplankton group. According to Kirk (1994), this relationship is simp ly: i i ia F =a (26 ) Quantum yield is generally defined to be the number of CO2 molecules fixed in the phytoplankton biomass per quantum of light absorbed While Kirk (1994) suggests a maximum theoretical value of ~0.1 mmol C (mmol quanta)-1 for quantum yield, this approximation ignores the extent to which ia can vary in nature. Thus, the more conservative value of 0.0833 mmol C (mmol quanta)-1, as suggested by Bissett et al. (1999b), was chosen for application in Eq. 26 for a ll phytoplankton groups.

PAGE 44

29 Although Bissett et al. (1999a) also suggest that the calculation of the carbonspecific absorption coefficients (* ia) must necessarily include any contributions caused by an intra-cellular “package effect”, their method assumes that the carbon-to-chlorophyll ratios are variable through time. Since the carbon -to-chlorophyll ratios for all phytoplankton groups used in this ecological simula tion were fixed (Table 2), the effects of photo-acclimation were not included explicitly i n the model. Since the model does not specify a particular species for each functional gr oup (save that of K. brevis ), it was assumed that absorption spectra determined from HPL C pigment analysis of a natural phytoplankton population (Cannizzaro, personal comm unication) would be more representative of each generalized functional group rather than a monoalgal laboratory culture maintained under artificial conditions. Ad ditionally, it was assumed that contributions to the absorption spectra due to the package effect were inherent to the phytoplankton samples gathered in vivo From these spectra, the spectrally-dependent carbon-specific absorption coefficients (* ia) for each phytoplankton group were determined for use in Eq. 26. Of course, natural phytoplankton assemblages are only rarely monoalgal, thus making it difficult to discriminate the specific co ntributions to an absorption spectrum caused by one particular phytoplankton species or g roup. However, the photopigments unique to a particular phytoplankton species may be used to discriminate it from the suite of photopigments within a diverse phytoplankton com munity. This practice of diagnostic photopigment analysis has been widely used to deter mine the presence and relative abundance of algal groups in mixed species assembla ges (Wright et al. 1991, Jeffrey et al. 1999, Wright & van den Enden 2000, Ornolfsdotti r et al. 2003), and is particularly

PAGE 45

30 effective when using CHEMTAX (Mackey et al. 1996) t o analyze high-performance liquid chromatography (HPLC) measurements of in vivo phytoplankton pigments. According to Mackey et al. (1996), phytoplankton community composition can be resolved from photopigment profiles using the chemi cal taxonomy program, CHEMTAX. The CHEMTAX algorithm partitions the phyt oplankton biomass within a sample by comparing the pigment composition of the sample with the diagnostic pigment composition of predetermined phytoplankton groups ( Ornolfsdottir et al. 2003). Of course, a more specific determination of phytoplank ton community structure is possible when a particular species or group possesses a uniq ue diagnostic photopigment. Therefore, HPLC samples gathered from west Florida shelf waters (Cannizzaro et al. 2003) prior to the red tide event (07 July 1999) we re used to derive the chlorophylland carbon-specific absorption spectra of the functiona l groups employed within the ecological model. These spectra were used to deter mine the carbon-specific absorption coefficients (* ia, Eq. 26) for each phytoplankton group over all PAR wavelengths as an initial condition. Prior to performing the CHEMTAX analysis, all of the HPLC samples (which consisted of 27 different photo-active pigments) re quired a simplification so as to eliminate all but those pigments that were represen tative of a specific phytoplankton group. Thus, an analysis of the unique pigment sui te within each of the phytoplankton functional groups was necessary in order to determi ne the proper diagnostic pigment for use in the CHEMTAX discrimination procedure Regardless of species, members of the Bacillariophy ta possess as their major photopigments a complex suite of chlorophylls ( a c1, c2, and c3) and carotenoids (b,b-

PAGE 46

31 carotene, fucoxanthin, diatoxanthin, and diadinoxan thin) specific to diatoms (Jeffrey & Vesk 1997). Among the carotenoids of the phytoplan kton groups used in the ecological model, only fucoxanthin was unique to the Bacillari ophyta. Since the diatoms also lack significant contributions from chlorophyllb these two photopigments were used within CHEMTAX to discriminate diatom biomass within HPLC samples (Pederson, personal communication). Although members of the microflagellate group inc lude the Chlorophyta and Chromophyta, each group possesses chlorophyllb as a significant photosynthetic pigment; hence, chlorophyllb was diagnostic of microflagellate biomass (Pederso n, personal communication). Within the non-toxic dino flagellates (Division Dinophyta), the carotenoid peridinin is not only a primary light-ha rvesting pigment (Jeffrey et al. 1975, Millie et al. 1993), it is also exclusive to member s of the Dinophyta (Jeffrey & Vesk 1997); therefore, peridinin was used to define the non-toxic dinoflagellate biomass (Pederson, personal communication). Despite the fact that K. brevis is also a member of the Dinophyta, they do not possess peridinin but instead have fucoxanthin and 19-acylofucoxanthins as primary light-harvesting pigments (Jeffrey et al. 1975). W hile the fucoxanthin signal could potentially interfere with the discrimination of th e Bacillariophyta within CHEMTAX, K. brevis cells also contain the carotenoid gyroxanthin-dies ter as an accessory photopigment (Johnsen & Sakshaug 1993, Hansen et al 2000). Since the only known gyroxanthin-containing dinoflagellates in the Gulf of Mexico are Karenia (= Gymnodinium ) mikimotoi and K. brevis it was employed as a diagnostic carotenoid for K. brevis (Pederson, personal communication).

PAGE 47

32 Using the diagnostic pigments indicated for each of the four phytoplankton groups within the ecological simulation, the initia l pigment ratio matrix used for WFS phytoplankton assemblages (Pederson, personal commu nication) was employed (Table 3) for the CHEMTAX analysis of HPLC samples to yield c hlorophyll-specific absorption spectra for each of the four phytoplankton groups u sed in the ecological simulation. These spectra (Fig. 5) were then converted to spect rally-dependent carbon-specific absorption coefficients (* ia) and ultimately used in Eq. 26 to calculate photos ynthetic efficiency (ia) for each of the phytoplankton groups. After defining the temperature-dependent maximum th eoretical growth rate (i T maxm ¢ ), photosynthetic efficiency (ia), and compensation irradiance (Ecomp), the lightlimited growth rate (i ll _m) was calculated for each functional group using Eq 25. However, when the total scalar irradiance at depth (Eo) was greater than saturation intensity (Esat), a different formulation (Bissett et al. 1999) of i ll _m was used to include the effects of photo-inhibition, such that: ( ) ()sat o sat inhibE E E E i T i T comp o i i lle E E TANH-¢ n n ¢ =7 max max _m m a m (27) 2.2.2.1. Light Availability and Utilization: Spect ral Irradiance Calculation of the total scalar irradiance at dep th (Eo) used a modification of the Gregg and Carder (1990) RADTRAN model of spe ctral solar irradiance for cloudless maritime atmospheres. Sensitivity analys es (data not shown) indicated that RADTRAN simulations of the total spectral (350 – 70 0 nm) downwelling irradiance at

PAGE 48

33 peridinin fucoxanthin gyroxanthin chlb chla Pd 0.00000 0.75540 0.00000 0.00000 1.00000 Pf 0.00000 0.00000 0.00000 0.26274 1.00000 Pn 1.06452 0.00000 0.00000 0.00000 1.00000 Pb 0.00000 0.73040 0.04000 0.00000 1.00000 Table 3. Simplified pigment ratio matrix used duri ng CHEMTAX analysis to estimate the proportional abundance of diatoms (Pd), microflagellates (Pf), non-toxic dinoflagellates (Pn), and K. brevis (Pb) in mixed phytoplankton samples collected on the west Florida shelf, 07 July 1999.

PAGE 49

34 0.00000 0.00002 0.00004 0.00006 0.00008 0.00010 0.00012 350400450500550600650700 Pd Pf Pn Pb 0.0000000 0.0000005 0.0000010 0.0000015 0.0000020 0.0000025 0.0000030 0.0000035 350400450500550600650700 Pd Pf Pn Pb Figure 5. Spectrally-dependent A) chlorophyll-spec ific and B) carbon-specific absorption spectra for diatoms (Pd), microflagellates (Pf), non-toxic dinoflagellates (Pn), and K. brevis (Pb), derived from HPLC and CHEMTAX analysis of phytop lankton samples collected on the west Florida shelf, 07 Jul y 1999. Wavelength ( l ) A) B) Wavelength ( l ) m2 m mol C-1 m2 m g chla-1

PAGE 50

35 hyper-spectral (1 nm) resolution only differed by a n average of 0.16% from identical runs at 15 nm resolution. Hence, irradiance within the UV wavelengths (350 – 400 nm) was calculated at 50 nm resolution, while irradiance wi thin the PAR wavelengths (400 – 700 nm) was calculated at 15 nm resolution, in the inte rest of computational efficiency. Since the data sets of the atmospheric forcings use d in the RADTRAN simulation were discontinuous in both time and space across th e model domain, irradiances were calculated using the RADTRAN default settings for a tmospheric pressure (29.92” Hg), relative humidity (80%), air mass type (1 = maritim e), precipitable water vapor (1.5 cm), mean wind speed (4.0 m s-1), current wind speed (6.0 m s-1), and visibility (15 km). Contributions due to ozone were calculated automati cally by the RADTRAN simulation, which estimated the solar zenith angle (Q) and both the diffuse (Edif 0-) and direct (Edir 0-) components of spectral downwelling irradiance just beneath the sea surface at each time interval for each surface coordinate in the grid sp ace. Estimating surface irradiance across the entire m odel domain proved to be very time intensive, when combined with the other calcul ations of the ecological model. Several RADTRAN test cases indicated that surface i rradiance values calculated throughout the entire spatial domain differed by on ly 3% of the median, which was consistently located at (28.4394 N, -82.8225 W) w ithin the model domain. Thus, all RADTRAN simulations were performed using these coor dinates, with the results applied uniformly to the entire domain at each time interva l. While this resulted in a loss of spatial variability in the surface irradiance estim ates, the sub-surface light field swiftly became non-uniform in spectral character and intens ity due to the high variability of the inherent and apparent optical properties at each gr id and time interval.

PAGE 51

36 To calculate the sub-surface light field, the exp onential attenuation of spectral light was calculated using Beer’s Law, where the sp ectral downwelling irradiance at depth z (Ed_z) was determined from the spectral downwelling irra diance at some lesser depth (Ed_ z -1), less the exponential attenuation (Kd) of the light field over the distance D z, so that: z K z d z dde E ED -=1 _ (28) The diffuse attenuation coefficient for downwelling plane irradiance (Kd) was approximated (Gordon 1989, Mobley 1994, Bissett et al. 1999b) as a function of the total spectral absorption (atotal) and total spectral backscattering (btotal) relative to the average downwelling cosine (i.e. solar angle) within the water column (d _m), where: ()_/d total total db a Km+ = (29) The mean cosine beneath the water surface (0 dm) was determined by: ( ) ( ) ( ) ( ) [ ] -+ n + =0 0 0 0 0 0 0 _/dif dir dif dif dir dir dE E E Em m m (30) where the RADTRAN model provided estimates of the d iffuse (0 difE ) and direct (0 dirE) components of spectral downwelling irradiance jus t below the water surface. While the average cosine of the diffuse downwelling irradiance (0 difm) was approximated by the value 0.86 (Morel 1991, Mobley 1994), the average cosine of the direct downwelling irradiance (0 dirm) was calculated as a function of the solar zenith

PAGE 52

37 angle ( Q ), which was also provided by the RADTRAN model. T hus, 0 dirm was calculated according to Snell’s law as: n Q =-SIN n SIN COSw dir11 0 _m (31) where the index of refraction for seawater ( nw) was approximated by the value 1.34 (Mobley 1994). As 0 dm only defines the average cosine just beneath the w ater surface, d _m will vary at all depths relative to the different optical properties witnessed at depth. Since the calculation of d _m at all depths throughout the model domain was also time intensive, it was assumed that -0 _ d dm m throughout the water column. 2.2.2.2. Light Availability and Utilization: Spect ral Absorption Total spectral absorption ( atotal) is the sum of the spectral absorption due to seawater ( awater), phytoplankton ( aphyto), and colored dissolved organic matter ( aCDOM) over all wavelengths throughout the domain, summari zed by: CDOM phyto water totala a a a + + = (32) The spectral absorption of seawater ( awater) was taken from Pope and Fry (1997) and interpolated to 15 nm resolution. Absorption due t o phytoplankton ( aphyto) was calculated by using the carbon-specific absorption spectra (Fi g. 5B) for each phytoplankton group interpolated to 15 nm resolution and multiplied by the carbon-specific biomass, such that: ==4 1 i i i phytoP a a (33)

PAGE 53

38 The calculation of spectral absorption due to color ed dissolved organic matter ( aCDOM), including particulate detritus, was first calcul ated as a function of the in situ dissolved organic carbon (DOC) with the method sugg ested by Bissett et al. (1999a). Since the spectral signal of the colored dissolved organic matter (CDOM) is assumed to be controlled by the concentration and type (Bisset t et al. 1999a) of colored dissolved organic carbon (CDOC), the absorption of light by C DOM was simply: ==2 1 i i CDOC CDOMa a ( 34) where ( ) [ ] l-=nm S i CDOC nm i CDOCie a a410 410 (35) described the spectral absorption of both the labil e and relict fractions of CDOC, where Slabile = 0.014 nm-1 and Srelict = 0.025 nm-1 (Bissett et al. 1999a). Due to the exponential decay of CDOC absorption with increasing wavelength (Carder et al. 1989, Kirk 1994, Bissett et al. 1999a), absorption at 410 nm can be used to determine CDOC absorption across all wavelengths relative to the spectral slo pe ( Si) using Eq. 35. Therefore, it was necessary to determine i CDOC nma_ 410 by the weight-specific absorption of CDOCi at 410 nm (* 410 i CDOC nma) relative to the in situ CDOCi concentration, such that: i i CDOC nm i CDOC nmCDOC a a* 410 410= (36) where C g m alabile CDOC nm 1 2 6 41010 4587 .1- =mand C g m arelict CDOC nm 1 2 7 41010 724 .4- =m (Bissett et al. 1999a). In order to estimate CDOCi, it was assumed that 80% of the total DOC was reli ct ( FRACrelict = 0.8), while the remaining 20% was labile (Kirchm an et al. 1991, Carlson &

PAGE 54

39 Ducklow 1995, Darrow et al. 2003). Further, the co lored component of DOC (i.e. the fraction of DOC which affects light absorption) was assumed to be 50% of the total (relict + labile) DOC (Darrow et al. 2003), such th at: ( ) DOC FRAC CDOClabile labile5.0 = (37) ( ) DOC FRAC CDOCrelict relict5.0 = (38) Based on earlier measures of DOC in west Florida sh elf waters (Carder et al. 1989), Darrow et al. (2003) suggest that total DOC concentrations above the 30 m isobath can reach 912.6 mg L-1 in surface waters (182.6 mg L-1 DOClabile and 730 mg L-1 DOCrelict) with only 8.0 mg L-1 DOCrelict (with no labile DOC) at depths below 5 m. Thus, t he method assumes FRAClabile = 0.0 and FRACrelict = 1.0 at all depth intervals below 5 meters. Reductions of DOC from photochemical degradation we re parameterized as fractional reductions of in situ DOC assuming an average photolysis rate of 0.02 mmol DOC (Mopper et al. 1991) relative to a mean surface concentration of 83 mM (Guo et al. 1994) per daylight hour. Additional sources of DOC were calculated as an estimate of DOC loading in nearshore waters due to estuarine di scharge; ultimately, these allochthnous sources of DOC were distributed throug hout the model domain via advective and diffusive transport schemes (Eqs. 14 – 15). DOC minima were imposed throughout the model domain, where DOCmin = 0.9126 mg C m-3 in near-surface waters while at all depths below 5 m, DOCmin = 0.008 mg C m-3 (or ~1% of the near-surface DOCmin), as suggested by Darrow et al. (2003).

PAGE 55

40 To test the necessity of an explicit calculation of DOC to provide estimates of CDOM absorption, CDOMa was also determined by an empirical algorithm. Us ing ~120 nearshore measurements of CDOM absorption relative to salinity (Del Castillo et al. 2000), the salinity fields ( S ) from the physical model were used to estimate aCDOM according to the following regressions (Walsh et al 2003), assuming a spectral slope of 0.020 nm-1: { } ‰ 5. 36 ‰ 0. 28 095 .0 470 .3) 443 (< < = S S anm CDOM (39) { } ‰ 0. 28 ‰ 0. 24 003 .0 892 .0) 443 (< < = S S anm CDOM (40) { } ‰ 0. 24 ‰ 0.0 060 .0 250 .2) 443 (< < = S S anm CDOM (41) 2.2.2.3. Light Availability and Utilization: Spect ral Scattering The calculation of Kd (Eq. 29) also required a parameterization of the t otal spectral scattering (btotal), which is the sum of the total spectral scatterin g due to seawater (water waterb b ~ ), biogenic particulates (phytob), and lithogenic particulates (lithb) over all PAR wavelengths throughout the domain (Mobley 1994), su mmarized as: lith phyto water water totalb b b b b + + = ~ (42) The spectral scattering of seawater (bwater) was taken from Smith and Baker (1981) and interpolated to 15 nm resolution, while waterb ~ was defined by Mobley (1994) as: 2 ~water waterb b = (43) Biogenic particulate scattering (bphyto) was estimated as a spectrally-dependent function of the chlorophyll-specific biomass of eac h phytoplankton group (Morel 1988):

PAGE 56

41 [ ] bp i phytof a b62.0chl 30.0=, (44) [] n + =lnm 550 chl log 4 1 2 1 02.0 002 .0i bpa f (45) Lithogenic particulate scattering (blith) was assumed to be zero throughout the model domain except at coastal boundaries where estuarine discharge into west Florida shelf waters was significant. Average daily streamflows from USGS freshwater disc harge gauging stations (Table 4) were used for the Anclote, Apalachicola, Caloosahatchee, Suwannee, and Withlacoochee Rivers as well as Charlotte Harbor, T ampa Bay, and Sarasota Bay, along with suspended sediment load data (Table 5) to dete rmine the location and extent of estuarine efflux into WFS waters (Fig. 6). These s pecific regions within the model domain were assumed to be Case II waters; thus, blith was parameterized with the spectrally-dependent Loisel-Morel near-surface powe r law (Mobley 2000) using the in situ concentration (g m-3) of suspended sediment [SS], such that: [] =lnm 550 407 .0795 .0SS blith (46) 2.2.3. Ecological Model: Nutrient Availability and Utilization Growth of the carbon-specific phytoplankton bioma ss was also mitigated by the effects of nutrient limitation. Within the microfl agellate (Pf), non-toxic dinoflagellate (Pn), and K. brevis (Pb) groups, a reduction of the temperature-dependent maximum growth rate (i T maxm) was a function of the pools of phosphate (3 4PO) and dissolved inorganic nitrogen (DIN), either as nitrate (3NO) or ammonium (+ 4NH). Within the

PAGE 57

42 River/Estuary Source Anclote River USGS 02310000 Apalachicola River USGS 02359170 Caloosahatchee River USGS 02292900 Charlotte Harbor Myakka River USGS 02298830 Myakka River, Venice USGS 02298928 Deer Prairie Slough USGS 02299120 Big Slough Canal USGS 02299410 Shell Creek USGS 02298202 Peace River USGS 02296750 Joshua Creek USGS 02297100 Horse Creek USGS 02297310 Prairie Creek USGS 02298123 Sarasota Bay Whitaker Bayou USGS 02299864 Walker Creek USGS 02299861 Phillippi Creek USGS 02299780 Catfish Creek USGS 02299741 South Creek USGS 02299737 Cow Pen Slough USGS 02299700 Suwannee River USGS 02323500 Tampa Bay Hillsborough River USGS 02304500 Delaney Creek USGS 02301750 Sweetwater Creek USGS 02306647 Henry Street Canal USGS 02306654 Brushy Creek USGS 02306950 Bullfrog Creek USGS 02301700 Cypress Creek USGS 02303800 Alafia River USGS 02301500 Little Manatee River USGS 02300500 Sixmile Creek USGS 02301800 Braden River USGS 02300042 Manatee River USGS 02299950 Withlacoochee River USGS 02313230 Table 4. Sources of average daily streamflow data from United States Geological Survey (USGS) freshwater discharge gauging stations, 31 Au gust – 08 November 1999.

PAGE 58

43 River/Estuary System NO3 mM N NH4 mM N PO4 mM P SiO2 mM Si DOC g m-3 SS mg L-1 Anclote River (28.1521 N, -82.8436 W) 10.71 5.00 3.23 151.45 9.25 9.07 Apalachicola River (29.6812 N, -84.9438 W) 23.21 1.79 0.31 108.44 3.58 43.50 Caloosahatchee River (26.4378 N, -81.9683 W) 21.43 5.00 4.26 166.43 5.47 10.00 Charlotte Harbor (26.7369 N, -82.1808 W) 63.12 4.77 20.18 123.58 25.28 22.70 Sarasota Bay (27.2442 N, -82.5806 W) 38.90 8.47 10.48 119.55 7.36 5.10 Suwannee River (29.2255 N, -83.1255 W) 35.46 2.85 4.32 109.16 23.02 9.07 Tampa Bay (27.5855 N, -82.7077 W) 58.93 67.22 22.26 111.09 9.25 5.00 Withlacoochee River (29.0309 N, -82.7806 W) 6.66 2.98 2.13 76.25 22.83 3.08 Table 5. Mean inorganic nutrient and suspended sed iment concentrations of modeled estuarine systems (EPA 2006, SFWMD 2006, USGS 2006, Walsh et al. 2007).

PAGE 59

44 0 50 100 150 200 250 300 242245248251254257260263266269272275278281284287290 293296299302305308Julian Day Daily Streamflow Anclote Apalachicola Caloosahatchee Charlotte Harbor Sarasota Bay Suwannee Tampa Bay Withlacoochee -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Figure 6. Geographical location of the estuarine s ystems used to estimate discharge to the west Florida shelf as a function of the average daily streamflows (m3 sec-1), 31 August – 08 November 1999. NOV 07OCT 05AUG 30SEP 07SEP 14

PAGE 60

45 diatom (Pd) group, silica dependence (4SiO) was also included. Thus, nutrient-limited carbon growth rates were calculated at each grid an d time interval, where: r r r r r r r r r r r r r r + + r r r r + + =+ + 4 4 max 3 4 3 4 max 4 4 max 3 3 max _SiO k SiO PO k PO NH k NH NO k NO MAX MINPd silicate Pd T i phosphate i T i ammonium i T i nitrate i T i nlm m m m m (47) using the half-saturation constant (i Bk_) specific to each phytoplankton group (Table 2). In situ concentrations of dissolved inorganic nutrients we re calculated as a function of advective and diffusive transport ( dB Tr), nutrient uptake due to realized phytoplankton growth (BdG), dissolution/remineralization of zooplankton feca l pellets (BdD), zooplankton excretion (Bdc), nitrification (dNIT), estuarine discharge (Bd Y), sediment exchange (BdSX), and atmospheric exchange (BdAX), if applicable: -+ Y + + = -3 3 33 3NO NO NOdSX d dNIT dG dNO Tr t dNO (48) + + + + ++ Y + + D + = + +4 4 4 4 44 4NH NH NH NH NHdSX d dNIT d d dG dNH Tr t dNHc (49) -+ Y + + D + = -3 4 3 4 3 4 3 4 3 43 4 3 4PO PO PO PO POdSX d d d dG dPO Tr t dPOc (50) 4 4 44 4SiO SiO SiOdSX d dG dSiO Tr t dSiO+ D + = (51) DIC DIC DIC DIC DICdAX dSX d d dG dDIC Tr t dDIC+ + + D + = c (52)

PAGE 61

46 The advective/diffusive transport ( dB Tr) of each inorganic nutrient was calculated using Eqs. 14 – 15, while nutrient uptake due to ph ytoplankton growth was determined from: i i R BP C B d dG*m = (53 ) where ( ) RC B was the intracellular Redfield ratio of inorganic nutrient “B”. Bacterial conversion of + 4NH to 3NO within the water column was calculated using the modified Michaelis-Menton function: + ++ =4 4NH k NH r d dNITnit nit (54) Estuarine discharge (Bd Y) of inorganic nutrients was determined as a functi on of the daily riverine/estuarine streamflow (R, Fig. 6) from USGS freshwater discharge gauging stations (Table 4) for the Anclote, Apalach icola, Caloosahatchee, Suwannee, and Withlacoochee Rivers as well as Charlotte Harbor, T ampa Bay, and Sarasota Bay relative to the average concentration of nutrient “B” (RB) within the riverine/estuarine effluent (Table 5) and the volume of water at the model grid point (ijV) into which the effluent was added, such that: ( ) ij R R BV B d d = Y ( 55) Nutrient flux at the water-sediment interface (BdSX) was defined as: ( ) Sd pv BB B k d dSX= (56) using an estimate of the mean bottom piston velocit y (pvk) for the exchange of dissolved

PAGE 62

47 inorganic nutrients with the overlying water colum n relative to the gradient between the nutrient concentration within the water column ( ) B and within the sediments ( ) SdB. Estimates ofSdBwere determined as a function of algal biomass and fecal pellet degradation, bioturbation, physical dissolution, an d bacterial remineralization of organic nutrients using the method prescribed by Walsh et a l. (2003). Nutrient flux via atmospheric exchange (BdAX) was only applicable to DIC in the surface layer and was calculated as a function of the CO2 gas exchange coefficient (2_CO xk) and CO2 solubility (2_CO spk) relative to the gradient between partial pressure s of CO2 in the atmosphere and in the surface layer through out the domain. Thus, the atmospheric exchange of DIC was estimated by: ( ) 2 2 _2 2PCO PCO A k k d dAXATM ij CO sp CO x DIC= (57) using the method suggested by Peng et al. (1987), i gnoring the contributions of3 4PO, 4SiO, and O H2 to the total alkalinity of seawater. Additions to the inorganic nutrient pool due to zoo plankton excretion (Bdc) and the physical dissolution and bacterial remineraliza tion of zooplankton fecal pellets (BdD) were also calculated explicitly. However, since th ese contributions were more intimately tied to the grazing scheme, they are discussed at l ength in the next section. 2.2.4. Ecological Model: Zooplankton Grazing. Estimates of carbon-specific grazing losses (i iP dg) were estimated using adult copepod species, numerical abundance, and biomass d ata (Lester 2005) obtained from

PAGE 63

48 oblique bongo tows at paired ECOHAB stations above the 10 m, 25 m, and 50 m isobaths (Fig. 1) during monthly cruises from August – Novem ber 1999 (Table 1). With rare exceptions, ~94% of the total grazer abundance belo nged to thirteen (13) distinct genera and/or species. These 13 representative grazers we re further categorized according to prey selection and/or diel vertical migration behav iors (Table 6). Thus, members within the mixed grazer population we re categorized either as selective migrators (Group I), selective non-migrat ors (Group II), or non-selective nonmigrators (Group III). Carbon-specific ingestion r ates for the non-selective nonmigrators (Group III) were assumed to be insensitiv e to the type of phytoplankton prey eaten (Table 7), regardless of prey palatability/to xicity. However, the carbon-specific ingestion rates among the selective non-migrators ( Group II) were variable relative to the type of phytoplankton prey (Table 8). Among the se lective migrating zooplankton (Group I), ingestion rates were variable according to the type of phytoplankton prey as well (Table 9). Ultimately, the volumetric abundance of zooplankton (ZV, ind m-3) was used to calculate the total carbon grazed from each of the four phytoplankton groups, such that: ==13 1 ), (Z i Z c Z diel Z i Z i Z i ik V P E IR P dg (58) where factors such as prey electivity (i ZE) and feeding periodicity (diel ZP) were used to impose fractional reductions in the overall ingesti on rates (i ZIR) among the Group I zooplankton as a consequence of their vertical migr ation behaviors. Since verticallymigrating zooplankton such as A. tonsa exhibit a 2-3 fold increase in gut contents and

PAGE 64

49 Species SH Source VM Source Acartia tonsa (Turner & Tester 1989) (Fulton 1984) (Calanoid Copepods) (Besiktepe & Dam 2002) (Cervetto et al. 1995) Heterotrophic Ciliates (Kamiyama 1994) (Dolan & Marrase 1995) (Protozoans) (Kamiyama & Arima 2001) (Dolan et al. 1999) Clausocalanus furcatus (Mazzocchi & Paffenhofer 1998) (Fragopoulu et al. 2001) (Calanoid Copepods) Corycaeus spp. (Sutton et al. 2001) (Fulton 1984) (Cyclopoid Copepods) Euterpina acutifrons (Nassogne 1969) (Fulton 1984) (Harpacticoid Copepods) (Bagoien et al. 1996) Copepodites (Roff et al. 1995) (Durbin et al. 2000) (Immature Copepods) Oikopleura dioica (Alldredge 1981) (Checkley et al. 1992) (Appendicularians) (Fernandez et al. 2004) Oithona spp. (Calbet et al. 2000) (Fulton 1984) (Cyclopoid Copepods) Oncaea spp. (Turner & Tester 1989) (Ohtsuka et al. 1996) (Cyclopoid Copepods) (Sutton et al. 2001) Paracalanus spp. (Turner & Tester 1989) (Checkley et al. 1992) (Calanoid Copepods) (Santos 1992) (Tang et al. 1994) Parvocalanus crassirostris (Calbet et al. 2000) (Checkley et al. 1992) (Calanoid Copepods) (Liu & Wang 2002) (Tang et al. 1994) Penilia avirostris (Wong et al. 1992) (Checkley et al. 1992) (Cladocerans) Temora turbinata (Turner 1984) (Bird 1983) (Calanoid Copepods) (Turner et al. 1998) Table 6. Incidence () or absence ( ) of selective herbivory (SH) and/or vertical migration (VM) behaviors among the dominant zooplan kton grazers sampled on the west Florida shelf, August – November 1999.

PAGE 65

50 Species Ingestion Rates (ZIRi) Source (ZIRi) Corycaeus spp. 4.9 =i ZIR (Sutton et al. 2001) (1.00 x 10-4 mg d.w. ind-1) ng Chl-a eaten ind-1 day-1 Oikopleura spp. 728 16 =i ZIR (Alldredge 1981) (4.30 x 10-3 mg d.w. ind-1) mg C eaten ind-1 day-1 Oithona spp. 4.9 =i ZIR (Sutton et al. 2001) (1.40 x 10-3 mg d.w. ind-1) ng Chl-a eaten ind-1 day-1 Parvocalanus crassirostris [ ] i i ZB IR 6004 .6 31. 4126 + = (Calbet et al. 2000) (8.00 x 10-4 mg d.w. ind-1) cells eaten ind-1 day-1, where Bi is cells mL-1 Penilia avirostris 7. 34 =i ZIR (Wong et al. 1992) (1.70 x 10-3 mg d.w. ind-1) ng Chl-a eaten ind-1 day-1 Temora turbinata [ ] 20083 .0 8512 45 2. 1240i i i ZB B IR + = (Turner et al. 1998) (2.00 x 10-4 mg d.w. ind-1) cells eaten ind-1 day-1, where B is cells mL-1 Table 7. Mean zooplankton dry weight biomass (Lest er 2005) and associated ingestion rates for the non-selective, non-migrating zooplank ton grazers (Group III) sampled on the west Florida shelf, August – November 1999.

PAGE 66

51 Species Ingestion Rates (ZIRi) Source (ZIRi) Clausocalanus furcatus [ ] Pd Pd ZB IR 0438 .0 8. 10 + = (Mazzocchi & Paffenhofer 1998) (5.00 x 10-2 mg d.w. ind-1) cells eaten ind-1 day-1, where B is cells L-1 [ ] Pf Pf ZB IR 0438 .0 8. 10 + = ” cells eaten ind-1 day-1, where B is cells L-1 [ ] Pn Pn ZB IR 0438 .0 8. 10 + = ” cells eaten ind-1 day-1, where B is cells L-1 0 =Pb ZIR ” cells eaten ind-1 day-1 Copepodites [ ] Pd Pd ZB IR 0166 .0 8842 .0 + = (Roff et al. 1995) (4.00 x 10-3 mg d.w. ind-1) cells eaten ind-1 day-1, where B is cells L-1 [ ] Pf Pf ZB IR 0166 .0 8842 .0 + = ” cells eaten ind-1 day-1, where B is cells L-1 [ ] Pn Pn ZB IR 0166 .0 8842 .0 + = ” cells eaten ind-1 day-1, where B is cells L-1 0 =Pb ZIR ” cells eaten ind-1 day-1 Paracalanus spp. [ ] 20166 .0 4374 32 823 253Pd Pd Pd ZB B IR + = (Santos 1992) (1.64 x 10-2 mg d.w. ind-1) cells eaten ind-1 day-1, where B is cells mL-1 [ ] 20166 .0 4374 32 823 253Pf Pf Pf ZB B IR + = ” cells eaten ind-1 day-1, where B is cells mL-1 [ ] 20166 .0 4374 32 823 253Pn Pn Pn ZB B IR + = ” cells eaten ind-1 day-1, where B is cells mL-1 0 =Pb ZIR ” cells eaten ind-1 day-1 Protozoans 0 =Pd ZIR (Kamiyama & Arima 2001) (3.00 x 10-6 mg d.w. ind-1) cells eaten ind-1 day-1 (Dolan et al. 1999) [ ] Pf Pf ZB IR 624 .0 56. 28 + = ” cells eaten ind-1 day-1, where B is cells mL-1 0 =Pd ZIR ” cells eaten ind-1 day-1 0 =Pb ZIR ” cells eaten ind-1 day-1 Table 8. Mean zooplankton dry weight biomass (Lest er 2005) and associated ingestion rates for the selective, non-migrating zooplankton grazers (Group II) sampled on the west Florida shelf, August – November 1999.

PAGE 67

52 Species Ingestion Rates (ZIRi) Source (ZIRi) Acartia tonsa ( ) [ ] PdB Pd Ze IR009 .01 05.7-= (Besiktepe & Dam 2002) (1.64 x 10-2 mg ind-1) mg C eaten ind-1 day-1, where B is mg C L-1 ( ) [ ] PfB Pf Ze IR001 .01 5. 32-= ” mg C eaten ind-1 day-1, where B is mg C L-1 ( ) [ ] PnB Pn Ze IR006 .01 9. 10-= ” mg C eaten ind-1 day-1, where B is mg C L-1 [ ] Pb Pb ZB IR 8.0 2. 129 + = (Turner & Tester 1989) (103) cells eaten ind-1 hr-1, where B is (103) cells mL-1 Euterpina acutifrons [ ] 20005 .0 8932 .1 1. 74Pd Pd Pd ZB B IR + = (Nassogne 1969) (5.70 x 10-3 mg ind-1) cells eaten ind-1 day-1, where B is cells mL-1 [ ] 20028 .0 6642 23 05. 529Pf Pf Pf ZB B IR + = ” cells eaten ind-1 day-1, where B is cells mL-1 [ ] 20005 .0 8932 .1 1. 74Pn Pn Pn ZB B IR + = ” cells eaten ind-1 day-1, where B is cells mL-1 0 =Pb ZIR ” cells eaten ind-1 day-1, where B is cells mL-1 Oncaea spp. 4.9 =Pd ZIR (Sutton et al. 2001) (2.00 x 10-4 mg ind-1) ng Chl-a eaten ind-1 day-1 4.9 =Pf ZIR ” ng Chl-a eaten ind-1 day-1 4.9 =Pn ZIR ” ng Chl-a eaten ind-1 day-1 [ ] Pb PbB IR Z 47.1 0. 2248 + = (Turner & Tester 1989) (103) cells eaten ind-1 hr-1, where B is (103) cells mL-1 Table 9. Mean zooplankton dry weight biomass (Lest er 2005) and associated ingestion rates for the selective, migrating zooplankton graz ers (Group I) sampled on the west Florida shelf, August – November 1999.

PAGE 68

53 feeding rates over the diel period (Cervetto et al. 1995), periodicity of the grazing response was introduced to the Group I zooplankton in the form of a three-fold increase in the periodicity coefficient (diel ZP) to 0.75 at night, which resulted in 75% of the da ily grazing pressure being applied during the nighttime hours (Dagg, 1995) ; of course, 25.0=diel ZP during the day. For those zooplankton groups that did not exhibit diel vertical migration (i.e. Groups II and III), diel ZP was set to the value 1.0 as a default condition. Within the selective grazers, it was assumed that t he differences between the prey-specific ingestion rates (Tables 8 – 9) were a dequate to describe prey electivity; thus, 0.1=Z iE was set as the default condition for those grazers that selectively ingest non-toxic prey items. While Acartia tonsa and Oncaea spp. will select K. brevis cells for ingestion (Table 9), Acartia tonsa will ingest K. brevis only when no other prey items are available (Turner & Tester 1989). Therefore, the p rey electivity coefficient for Group I zooplankton was shifted to 01.0=Z iE when any of the non-toxic phytoplankton prey were above their refuge concentrations; otherwise, Z iE remained at 1.0. Volumetric abundance of each zooplankton species wa s obtained from monthly zooplankton counts (Table 1) at paired ECOHAB stati ons (Fig. 1) and interpolated throughout the model domain as a proportional chang e to the mean zooplankton abundance for each species over the simulation peri od (31 August – 08 November 1999). Non-migrating zooplankton (Groups II and III) were assumed to be distributed evenly throughout the water column. However, the distribu tions of vertically-migrating species (Group I) were calculated explicitly as a sinusoida l function relative to the time of day in

PAGE 69

54 adherence to the classic “predator avoidance” parad igm (Haney 1988, Loose & Dawidowicz 1994). Thus, the Group I zooplankton ag gregated near the bottom during the simulated daylight hours. This pattern was rev ersed during nighttime hours, when the Group I zooplankton engaged in an upward migration at dusk and remained in the nearsurface waters until dawn. Hence, the vertical distribution of migrating zoopl ankton was calculated from Vinogradov (1970), where the volumetric abundance w ithin each discrete sigma layer () (sZV) was: ) () (Z Z Z Zd SIN A dVj ws s= (59) where the areal abundance of zooplankton species Z (ZA, ind m-2) defined the amplitude of the sine curve. Using a wave period (Zw) of 0.5 and the thickness of each sigma layer that the zooplankton occupy (sd), time of day was used to define the phase shift (Zj) of the function. Thus, the areal abundance of zooplan kton does not change relative to timeof-day, but volumetric abundance within each sigma layer varies according to the idealized nocturnal migration scheme, ensuring that : ==21 1 ) (s s sd V AZ Z (60) While some ecological models treat grazing losses as a proportional reduction of the realized growth rate for each of the phytoplank ton functional groups susceptible to grazing stress (Walsh et al. 2001, Darrow et al. 20 03, Walsh et al. 2003), this strategy is typically employed as a closure scheme when zooplan kton abundance and ingestion data are unavailable (Bissett et al. 1999b). Although t he availability of zooplankton data for

PAGE 70

55 the 31 August – 08 November 1999 simulation (Table 1) enabled the development of a grazing scheme to investigate complex predator-prey relationships within a diverse grazer community (Eq. 58), a more simplified grazing schem e was also necessary to test the sensitivity of the model to the variable distributi on and diversity of the grazers. Thus, the initial 13-member grazer community (Table 6) was simplified into a single “mixed” grazer population, where prey-specif ic carbon ingestion rates (i ZIR) were calculated from each of the 13-member ingestion fun ctions (Tables 7 – 9) assuming a prey concentration of 1.0 mmol C kg-1. These results were converted to consistent units using kc(Z,i) and averaged to yield mean carbon ingestion rates f or each of the non-toxic phytoplankton prey species ( iIR) within the simulation. Total grazer abundance wa s also simplified into metazoan (Zm) and protozoan (Zp) fractions. Thus, Eq. 58 was simplified to: ==3 1 ), (i i Z c i i ik IR P dg (6 1) where = - =12 1 410 1.8m m d PZ P IRd (62) () + == 12 1 4 610 5.8 10 6.2m m f p f PZ P Z P IRf (63) = - =12 1 410 7.8m m n PZ P IRn (64) This simplified grazing scheme assumed that the mix ed grazer population would avoid ingesting K. brevis and that all members of the grazer community exhib ited linear

PAGE 71

56 ingestion rates that were adequately represented by = 4 1i iIR. Since the published ingestion rates for Acartia tonsa, Euterpina acutifrons, Temora turbinata, and members of the genus Paracalanus are non-linear (Tables 7 – 8 ), the linear functio ns of iIR (Eqs. 62 – 64) are most likely an under-estimation of gra zing losses compared to the more complex formulation of Eq. 58; nonetheless, these c omparisons were made to test the sensitivity of the ecological model to the two form ulations of grazing stress. Ultimately, the zooplankton ingestion of the phytop lankton biomass is used as a loss term (i iP dg) in the phytoplankton state equation (Eq. 13). Ho wever, it is not sufficient to simply remove phytoplankton biomass f rom the model domain, as this would ignore the significance of zooplankton excret ion (Bdc) and egestion (BdD) as they pertain to the nutrient dynamics discussed ear lier (Eqs. 48 – 52). When phytoplankton biomass is consumed, the parti culate organic C/N/P of the ingested biomass may either be assimilated or egest ed by the zooplankton grazers. Particulate wastes eliminated as fecal pellets are subject to dissolution and bacterial remineralization (BdD), where they may re-enter the dissolved inorganic nutrient pools. That portion of the phytoplankton biomass which is assimilated may be metabolized into dissolved inorganic nutrients via respiration and/o r excretion (Bdc), or it may ultimately be incorporated into the zooplankton biomass as gro wth (Parsons et al. 1984). Of the total ingested carbon (i iP dg), it was assumed that 85% was assimilated by zooplankton; the remainder was egested as fecal mat erial (Walsh et al. 2001). Hence, the concentration of fecal material within a particular grid cell was determined from the

PAGE 72

57 amount of phytoplankton carbon from each functional group that was grazed and egested, less the fraction of fecal material lost due to dis solution (i PFECdD), plus or minus any fecal material sinking into or out of each grid cel l. Thus, i i i P iP P FEC FEC i i PFEC w z d P dFEC_15.0 D =g (65) Although fecal pellet size and density can vary wid ely, a review of fecal pellet sinking rates (Smayda 1969) indicated that even the smallest fecal pellets sink at 36 m day-1. Since the horizontal scale of the grid spacing r anged from 2 – 6 km, advective and diffusive transport in the horizontal dimension wer e ignored; thus, the transport of fecal material was limited to the vertical dimension as a consequence of sinking. The total egested carbon (EC) from zooplankton grazing among all four phytoplankton groups was calculated simply as: ==4 1i PiFEC dEC (66) where the total egested N/P/Si were ultimately dete rmined from the Redfield proportions of the nutrient in question relative to the total e gested carbon, so that: d d R R RP C Si dESi EN N P dEP EC C N dENg15.0 , = = = (67) Ultimately, the amount of a particular inorganic nu trient released from the fecal pellet mass as a result of dissolution and/or bacterial re mineralization (BdD) was determined from the actual abundance of the egested nutrient r elative to the dissolution and/or remineralization rate of that nutrient (RtFECB) to ensure that nutrient dissolution could not exceed the nutrient abundance within the fecal material. Thus, dissolved inorganic

PAGE 73

58 nutrient pools derived from fecal pellet dissolutio n/remineralization were calculated as: ( ) ( ) ( ) P PO N NH C DICRtFEC EP d RtFEC EN d RtFEC EC d= D = D = D+3 4 4, (68) ( ) Si SiORtFEC ESi d= D4 (69) Since 85% of the total ingested carbon is ultimatel y assimilated by the zooplankton grazers, the determination of assimilated phytoplan kton carbon is: ==4 185.0i i iP dACg ( 70) where the total amount of assimilated N/P is calcul ated similarly to Eq. 67, so: AN N P dAP AC C N dANR R = = (71) Of course, the total amount of carbon assimilated e ffectively limits the metabolic losses (respiration) and gains (reproduction/growth ) of carbon within the zooplankton population. Although zooplankton metabolism is hea vily influenced by the ambient water temperature, a variety of regression equation s for the log-specific zooplankton respiration rates ( R ¢ ) within boreal, temperate, and tropical marine eco systems (Ikeda 1970) indicate that the dry weight of the animal is a sufficient determiner of the respiration rate, at least for comparisons within t he same habitat. Therefore, it was assumed that the WFS zooplankton community within t he model followed Ikeda’s (1970) tropical species regression, where the log-specific respiration rate ( R ¢ ) was estimated as: 874 .0 464 .0 + = ¢ W R ( 72) using only the dry weight (W) of the zooplankton grazer. While this regression yielded a respiration rate in terms of mL O2 consumed per unit time, it was ultimately converte d to

PAGE 74

59 units of mg C produced per unit time (Parsons et al 1984), where: R RQ R¢ = 4. 22 12 ( 73) A value of 0.97 was chosen for the respiratory quot ient (RQ), as it is indicative of ammonium-excreting herbivores (Omori & Ikeda 1984) like those within the simulated grazer population. Zooplankton excretion of nitrogen (as+ 4NH) and phosphorus (as3 4PO) was estimated using regressions of inorganic nutrient e xcretion within a mixed zooplankton community. According to Pagano et al. (1993), the rate of ammonium excretion (mg at-N ind-1 day-1, as+4NHER) is related to the dry weight (Wmg) of the zooplankton grazer, where: g NHW ERm1278 .04=+ (74) Similarly, Uye et al. (1990) found that the rate of phosphate excretion (ng at-P ind-1 hr-1, as -3 4POER) is also related to the dry weight of the grazer a ccording to the log function: ( ) 377 .0 log 486 .0 log13 4+ =--g POW ERm (75) Thus, mean zooplankton biomass values (Tables 7 – 9 ) were utilized in Eqs. 72 – 75 to yield respired CO2 (DICdc), excreted ammonium (+4NHdc), and excreted phosphate (-3 4POdc), such that: R V k dZ DICDICcc= (76) + + +=4 4 4NH Z NHER V k dNHcc (77) -=3 4 3 4 3 4PO Z POER V k dPOcc (78)

PAGE 75

60 Since the values for AC / AN / AP (Eqs. 70 – 71) represent the upper limit of +3 4 4/ /PO NH DICc c c respectively, the quantity of C/N/P excreted was n ever allowed to exceed the quantity of C/N/P assimilated. However, when assimilated C/N/P exceeded the amount of C/N/P excreted, it was assumed that t he balance was incorporated into the zooplankton biomass as egg production and/or tissue growth. Since the mortality and fecundity of the grazer population was not a compon ent within the ecological model, those nutrients sequestered within the grazer bioma ss represented a “permanent” loss of nutrients from the model domain. 2.3. Initialization and Validation of the Biophysic al Model All hydrographic data for model initialization and validation (Table 1) were determined from CTD casts at discrete stations, exc ept for surface K. brevis cell counts at select coastal stations and satellite-relayed telem etry of wind velocity, current velocity, sea level, and temperature data from moored ADCP ar rays (Fig. 1). Inorganic nutrients (SiO4 -3NO, +4NH, -3 4PO) were determined at micromolar levels using standa rd methods (Atlas et al. 1971, Gordon et al. 1994). E xtracted chlorophyll stocks were measured with both the Holm-Hansen/Welschmeyer fluo rometric protocols (Heil et al. 2002) and HPLC assays (Wright et al. 1991), which y ielded similar estimates of the chlorophyll-a concentrations. Ultimately, these data were used to establish the initial and boundary conditions for the variety of state variab les computed within the hindcast simulation of the 31 August – 08 November 1999 red tide (Fig. 2).

PAGE 76

61 Of course, the mathematical simulation was dependen t upon initial conditions, boundary constraints, and how the physical and ecol ogical schema (q.v. Sections 2.1. – 2.2.) calculated changes in the relevant parameters as the biophysical model progressed forward in time. Thus, the spatially heterogeneous distributions of total chlorophyll-a, metazoan grazers, and dissolved inorganic nutrients such as SiO4 -3NO, +4NH, -3 4PO in the eastern Gulf of Mexico in August 1999 were inte rpolated in three dimensions over the entire model domain and used to initialize four sim ulation cases (q.v. Section 2.4.). Initial stocks of chlorophyll-a within shelf waters of the eastern Gulf of Mexico ranged from 0.25 – 3.0 mg chl-a L-1, with maximum concentrations near the coast (Fig. 7). Without HPLC pigment data to analyze the 17 – 28 August 1999 (NEGOM) or 07 – 10 September 1999 (ECOHAB) chl-a (Table 1), it was impossible to determine the proportional abundance of each phytoplankton group from the available chlorophyll-a data. Therefore, at those stations where K. brevis abundance was zero, it was assumed that 90% of the chl-a standing stock was attributable to the diatom biom ass, with the nontoxic dinoflagellates and the microflagellates comp rising equal proportions of the remainder. At those stations with measured K. brevis abundances, the estimated fraction of K. brevis chl-a was first calculated assuming 300 pg C cell-1 and 30 pg C pg-1 chl-a (Table 2). Then, 90% of the balance was attributed to the diatom biomass, while the remaining 10% was shared evenly among the non-toxic dinoflagellates and the microflagellates. Since cell counts for each of the other phytoplankt on groups were unavailable, the chlorophyll-a concentrations of the diatom, microflagellate, and non-toxic dinoflagellate

PAGE 77

62 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Near-surface chlorophyll-a -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 B) Near-bottom chlorophyll-a Figure 7. Near-surface (A) and near-bottom (B) chl orophyll-a stocks (mg L-1) observed within the eastern Gulf of Mexico, 17 – 28 August 1 999 (NEGOM) and 07 – 10 September 1999 (ECOHAB).

PAGE 78

63 groups were converted to carbon-specific phytoplank ton biomass (mmol C kg-1) using the unique C:cell and C:chl-a ratios for each functional group of the model (Tab le 2) and the standard density of seawater (1.027 kg L-1, Knauss 1997). Additionally, phytoplankton refugia were established throughout the model domai n, where the phytoplankton biomass was not permitted to drop below the “background” co ncentration, which was assumed to be ~250 cells L-1 for each group, or 1.77 x 10-2 mmol C kg-1 for the diatoms, 1.21 x 10-4 mmol C kg-1 for the microflagellates, 1.21 x 10-2 mmol C kg-1 for the non-toxic dinoflagellates, and 6.07 x 10-3 mmol C kg-1 for K. brevis. Since the diversity and relative abundance of zoopl ankton species on the west Florida shelf are segregated into estuarine, shelf, and offshore assemblages (Sutton 2001, Lester 2005), spatial variability among the WFS met azoan grazers (Table 6) was first interpolated as a function of water depth throughou t the domain. Thus, initial distributions of the selective migrating (Group I), selective non-migrating (Group II), and non-selective non-migrating (Group III) zooplankton groups on 31 August 1999 were estimated (Fig. 8) using data from the 07 – 10 Sept ember 1999 zooplankton survey (Lester 2005). Since zooplankton fecundity and mortality were not calculated in the biological model, the monthly metazoan grazer abundance data ( Table 1) measured at paired ECOHAB zooplankton stations (Fig. 1) were interpola ted daily to introduce temporal differences in the WFS grazer community from 31 Aug ust – 08 November 1999. Protozoan grazer abundance (540 cells L-1) was assumed to be uniform throughout the domain, regardless of depth (Dolan & Marrase 1995).

PAGE 79

64 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 C) Non-selective Non-migrating Grazers (Group III) -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 B) Selective Non-migrating Grazers (Group II) -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Selective Migrating Grazers (Group I)Figure 8. Estimates of initial metazoan grazer abu ndance (# ind. m-3) within the eastern Gulf of Mexico (based on ECOHAB zooplankton data, 0 7 – 10 September 1999).

PAGE 80

65 Initial pools of the inorganic nutrients SiO4 -3NO, +4NH, and -3 4PO (Fig. 9) were determined from both the NEGOM (17 – 28 August 1999 ) and ECOHAB (07 – 10 September 1999) cruises (Table 1). Nutrient distri butions thereafter were calculated from the biophysical model without periodic data infusio n. The observations of inorganic nutrients from subsequent ECOHAB surveys on 05 – 07 October and 06 – 08 November 1999 were instead used for model validation. The results from the coupled physical model defined the three-dimensional model domains of advective/diffusive transport, temperatu re, and salinity. Validation of the physical model was performed by the USF Ocean Circu lation Group (OCG) in a manner consistent with their previous models of WFS circul ation (Li & Weisberg 1999a, Li & Weisberg 1999b, Weisberg et al. 2000, Weisberg et a l. 2001, He & Weisberg 2002, He & Weisberg 2003, Weisberg & He 2003, Weisberg et al. 2004). Monthly CTD data from ECOHAB stations (Table 1) were used to validate thr ee-dimensional estimates of the salinity fields simulated by the physical model. S patial patterns of sea surface temperature were examined using Advanced Very High Resolution Radiometer (AVHRR) satellite data (JHU/APL 2006) to further ev aluate the performance of the physical model with respect to simulated upwelling events. 2.4. Numerical Experiments: Simulation Conditions Four numerical experiments were performed with the coupled biophysical model to test the response of K. brevis growth to different light, nutrient, and grazing c onditions. Unless otherwise noted, all simulations were perfor med from 31 August – 08 November 1999 using the fixed parameters described in Table 2 and in the Appendix.

PAGE 81

66 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Near-surface NO -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 B) Near-bottom NO3 3 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 C) Near-surface NH -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 D) Near-bottom NH4 4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 E) Near-surface PO -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 F) Near-bottom PO4 4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 H) Near-bottom SiO4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 G) Near-surface SiO4 Figure 9. Inorganic nutrient concentrations (mmol kg-1) observed within near-surface and near-bottom waters of the eastern Gulf of Mexico, 1 7 – 28 August (NEGOM) and 07 – 10 September 1999 (ECOHAB).

PAGE 82

67 The Case I (“No Refuge for K. brevis ”) simulation was performed using the methods and initial/boundary conditions outlined in Sections 2.1. – 2.3., where K. brevis was given no refuge from grazing pressure within a complex zooplankton community (Eq. 58) throughout the simulation period. Shelfwi de estimates of in situ DOC (Darrow et al. 2003) were used for the calculation of CDOCi and the inherent optical properties derived therefrom (Eqs. 34 – 38). The Case II (“Grazer Avoidance of K. brevis ”) simulation involved the removal of all grazing pressure on the red tide organism. Cells of K. brevis were afforded protected status within the phytoplankton community thereby enabling a sensitivity analysis of the impact of selective grazing. Since the shelfwide DOC estimates (Darrow et al. 2003) used in the Case I simulation were inhomogeneous in the vertical dime nsion only, an alternative formulation of aCDOM (Eq. 39 – 41) was chosen to test the phytoplankton response to spatially variable CDOM f ields. The Case III (“Increased Shading for K. brevis ”) simulation was a modification of Case II conditions that assumed the inherent optical prope rties of west Florida shelf waters were much more heavily influenced by the con centration of suspended inorganic particulates [ ] SIP than originally parameterized. Using the inverse function: [ ] ( ) 1107 .0 984 .20 1=-dz SIP (79) where the depth of the water column (0 1-dz) was used to estimate [ ] SIP in g m-3 at all grid points inshore of the 30m isobath. Although the av erage suspended sediment concentration can exceed 5.0 g m-3 in fluvial zones within the southeastern Gulf of

PAGE 83

68 Mexico (Carranza et al. 1993), Eq. 79 assumed [ ] SIP= 0.05 g m-3 at the 20 m isobath and [ ] SIP= 0.5 g m-3 at the 5 m isobath on the west Florida shelf. Since these suspended inorganic particulates contri buted to both the absorption (sipa) and backscattering (sipb) of spectral light (Mobley 2000), each was compute d (Gallie & Murtha 1992, Bukata et al. 1995) using th e regressions: ( ) [ ] SIP x asip2 610 0.3 0036 .0 1444 .1l l-+ = (80) ( ) [ ] SIP bsipl388 207 1818 .0+ = (81) Ultimately, the equations which defined total absor ption (Eq. 32) and total backscattering (Eq. 42) were amended to include those contribution s from sipa and sipb as a test case to investigate the K. brevis response to a “shallow-water shading” paradigm whi ch included the effects of increased CDOM and suspended sediments near the coast. The Case IV (“Increased Shading/Increased Grazing”) simulation was a modification of the Case III conditions but assumed that the grazing losses (an d the attendant estimates of zooplankton excretion/egesti on) could more easily be determined from a simplified grazing scheme (Eq. 61) employing mean carbon ingestion rates within a community of just two generalized grazers: metaz oa and protozoa (Eqs. 62 – 64). This simulation was used to determine whether the phytop lankton competitors on the west Florida shelf were insensitive to the diversity of the grazers utilized within the more complex (Eq. 58) grazing scheme in Cases I III.

PAGE 84

69 Chapter 3: Results 3.1. Physical Forcings 3.1.1. Physical Model Results The estimated flowfields from the physical model as sumed a Cartesian coordinate system with vector components oriented to true east (u) and north (v), such that an analysis of the cross-shelf transport requires a co astal boundary normal to the original east (u) coordinate axis. Since the Florida coast and the isobaths of the west Florida shelf are generally oriented at 333 (Yang et al. 1999), a recalculation of the vector components of the modeled currents (relative to an axis reoriented –27 from true north, to 333) yielded estimates of the average daily cro ss-shelf transport along the ECOHAB transects off Tampa, Sarasota, and Fort Myers, Flor ida (Fig. 10). Results from this analysis indicated that the simul ated cross-shelf flows were offshore in the surface waters along the three tran sects (as the solid, negative isotachs of Fig. 10A-C) throughout September 1999. Within each cross-shelf section, the model exhibited onshore transports within the bottom Ekma n layer as the positive, dotted isotachs. Onshore flows increased from north to so uth, with the greatest onshore flow of >1 cm sec-1 off Ft. Myers (Fig. 10), where the zero isotach in each section above the 20 – 25 m isobaths denotes near-bottom convergence front s that would facilitate the physical accumulation of red tides and other particulates. The cross-shelf transport of near-bottom water was observed during this period of bloom maintenance, where satellite images of sea su rface temperature (SST) from the AVHRR sensor (JHU/APL 2006) were indicative of stro ng coastal upwelling as the cold bottom waters of the west Florida shelf were upwell ed at the coastal boundary near

PAGE 85

70 01020304050607080 -50 -40 -30 -20 -10 0 C) Ft. Myers TransectDepth (m)Distance from shore (km) 01020304050607080 -50 -40 -30 -20 -10 0 B) Sarasota TransectDepth (m) 01020304050607080 -50 -40 -30 -20 -10 0 A) Tampa Transect D e p t h ( m ) Figure 10. WFS-POM estimates of the average daily cross-shelf transport (cm s-1) along the A) Tampa, B) Sarasota, and C) Ft. Myers transec ts throughout September 1999 (positive values of the dashed (-) isotachs ind icate onshore transport, negative values of the solid (—) isotachs denote offshore transport from He 2001).

PAGE 86

71 Ft. Myers, Sarasota, Tampa Bay, and the Big Bend ar ea during 17 – 20 September 1999 compared to 11 – 14 September 1999 (Fig. 11A-D). T his upwelling was in response to shelfwide wind events, which initiated coastal upwe lling as the surface waters were transported offshore, thereby creating a divergent flow near the coast. Ultimately, an upward vertical flow was required to replace the wa ter lost by the surface divergence, causing upwelling at the coast. 3.2. Ecological Forcings 3.2.1. Ecological Observations Surface and near-bottom measures of in situ nitrate (-3NO) and phosphate (-3 4PO) during 07 – 10 September 1999 indicated that west F lorida shelf waters were nitrogenlimited with respect to dissolved inorganic nitroge n (DIN) and phosphorus (DIP) stocks, with concentrations of -3NO < 0.05 mmol kg-1 throughout the water column over much of the inner shelf (Fig. 12A-B). Surface and near-bot tom stocks of -3 4PO were similar to -3NO (also < 0.05 mmol kg-1), except for the increased supplies of -3 4PO near the coast as a result of estuarine discharge (Fig. 12C-D). With DIN:DIP ratios of ~0.2 – 2.0 throughout much of the west Florida shelf, any phyt oplankton species utilizing the dissolved inorganic nutrient concentrations of the mid-shelf would be nitrogen-limited, except for those capable of nitrogen fixation. The surface concentrations of the diazotroph Trichodesmium erythraeum were near background concentrations (0.1 – 0.5 colonies L-1) at most stations. But at two near

PAGE 87

72 Figure 11. AVHRR satellite images (JHU/APL 2006) o f sea surface temperature (SST) over the west Florida shelf on: A) 11 September 199 9; B) 14 September 1999; C) 17 September 1999; D) 20 September 1999.

PAGE 88

73 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Surface NO ( mol kg )3m-1 A) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Bottom NO ( mol kg )3m-1 B) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Surface PO ( mol kg )4m-1 C) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Bottom PO ( mol kg )4m-1 D) Figure 12. Spatial patterns of observed: A) nitrat e (mmol -3NO kg-1) at the surface; B) nitrate (mmol -3NO kg-1) at the bottom; C) phosphate (mmol -3 4PO kg-1) at the surface; and D) phosphate (mmol -3 4PO kg-1) at the bottom during 07 – 10 September 1999.

PAGE 89

74 the coast (Fig. 13A), their excretion of dissolved organic nitrogen (DON) may have provided a source of new nitrogen (Lenes et al. 200 1), thereby alleviating conditions of nitrogen limitation for the other phytoplankton. H owever, the surface concentrations of DON during this period (Fig. 13B) were instead attr ibuted to the low salinity plumes (Fig. 13C) from Tampa Bay and Charlotte Harbor (Len es et al. 2001). Since estuarine supplies of DON are more refractory than DOP (Walsh et al. 2007), the remineralization rates for DOPDIP would exceed those of DONDIN, thereby favoring the preferential accumulation of D IP in estuarine waters in all but the most N-enriched systems. Since both the Tampa Bay and Charlotte Harbor estuaries exhibit Redfield ratios of 3.0 – 5.6 for end-member concentrations of -3NO:-3 4PO (Table 5), increased output from these estuaries would fav or the enrichment of nearshore waters with “new” DIP (relative to DIN), thereby maintaini ng conditions of nitrogen limitation for most phytoplankton on the inner-shelf. Metazoan zooplankton abundance was greatest at near shore stations (Fig. 13D), similar to the distribution of the algal biomass (F ig. 14A-B), where stocks ranged from 1.0 – 3.0 mg chl-a L-1. Although very low surface concentrations of K. brevis were noted near the 45 m and 25 m isobaths of the Sarasota and Ft. Meyers transects respectively (Fig. 14A), K. brevis cells were found near the bottom of the Sarasota 4 5 m isobath as well (Fig. 14B). Since the mean concentration of S iO2 within the Tampa Bay and Charlotte Harbor estuaries is 111.09 mM and 123.58 mM respectively, the silicate pools on the inner WFS (Fig. 14C-D) were indicative of es tuarine supplies of “new” silicate

PAGE 90

75 A)B) C)D) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999(colonies L )-1T. erythraeum -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Surface Salinity (psu) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Surface DON ( mol kg )m-1 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Zooplankton (ind. m )-3 Figure 13. Spatial patterns of observed : A) surfa ce concentrations of the diazotroph Trichodesmium erythraeum (# colonies L-1); B) surface concentrations of dissolved organic nitrogen (mmol DON kg-1); C) surface salinity (psu); and D) depth-averaged concentrations of metazoan zooplankton (# ind. m-3) on the west Florida shelf during 07 – 10 September 1999.

PAGE 91

76 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Bottom chl( g L )m-1a A)B) C)D) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Surface chl( g L )m-1a -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Surface SiO ( mol kg )4m-1 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 07-10 September 1999Bottom SiO ( mol kg )4m-1 Figure 14. Spatial patterns of observed: A) chloro phyll-a stocks (mg L-1) at the surface; B) chlorophyll-a stocks (mg L-1) at the bottom; C) silicate (mmol 4SiO kg-1) at the surface; and D) silicate (mmol 4SiO kg-1) at the bottom during 07 – 10 September 1999. Dashed ( ) contour lines denote K. brevis abundance (cells L-1) measured at ECOHAB stations indicated in red.

PAGE 92

77 associated with a low salinity plume (Fig. 13C); th us, diatom growth on the inner WFS could not have been limited by silicate during 07 – 10 September 1999. By 05 – 07 October 1999, the surface stocks of DIN were below detection limits near the coast between Tampa Bay and Charlotte Harb or (Fig. 15A). Near-bottom nitrate stocks were <0.1 mmol kg-1 over much of the west Florida shelf, except near t he mouth of Charlotte Harbor and a patch of nitrate-enriched (0 .7 mmol kg-1) water located 55 km offshore Tampa Bay (Fig. 15B). Although surface concentrations of T. erythraeum and DON were not measured during 05 – 07 October 1999, the WFS pools of nitra te and phosphate indicated that inorganic nitrogen limitation was still in effect ( Fig. 15) for those phytoplankton incapable of nitrogen-fixation. Under conditions o f inorganic nitrogen limitation, the fraction of dissolved inorganic nitrogen within any nutrient source (whether from estuarine discharge, zooplankton excretion, or the bacterial remineralization of organic material) would be preferentially utilized by the p hytoplankton biomass, allowing the DIP fraction to accumulate in the water column. Indeed, the nitrate and phosphate surface maxima we re displaced, with the greatest surface concentration of -3 4PO (0.7 mmol kg-1) located near the coast between Tampa Bay and Charlotte Harbor (Fig. 15C), a region where the surface stocks of -3NO were below detection limits (Fig. 15A) and K. brevis abundance exceeded 2.0 x 106 cells L-1 (Fig. 2C). Beyond the 60 m isobath, the near-bott om -3NO and -3 4PO had been nearly exhausted by 05 – 07 October 1999 (Fig. 15B, D) com pared to those the previous month (Fig. 12B, D), indicating significant consumption o f both DIN and DIP beneath the

PAGE 93

78 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) 05-07 Octber 1999Surface NO ( mol kg )3m-1 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 05-07 Octber 1999Bottom NO ( mol kg )3m-1 B) -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 05-07 Octber 1999Surface PO ( mol kg )4m-1 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 05-07 Octber 1999Bottom PO ( mol kg )4m-1 C) D) Figure 15. Spatial patterns of observed: A) nitrat e (mmol -3NO kg-1) at the surface; B) nitrate (mmol -3NO kg-1) at the bottom; C) phosphate (mmol -3 4PO kg-1) at the surface; and D) phosphate (mmol -3 4PO kg-1) at the bottom during 05 – 07 October 1999.

PAGE 94

79 nutricline over the intervening 25 – 30 days. Despite the paucity of DIN in coastal waters during September – October 1999, nearshore chlorophyll-a concentrations rose in excess of 10 mg L-1 from Sarasota to the mouth of Charlotte Harbor by 05 – 07 October 1999, with a maximum surface concentration of 27.6 mg chl-a L-1 found near Big Sarasota Pass (Fig. 16A-B). Assumi ng a C:chl-a ratio of 50.0 for diatoms (Table 2) and a C:Si mol ar ratio of 6.67 (Walsh et al. 2003) for balanced growth, a diatom bloom of this s ize would have required nearly 18 mmol SiO4 kg-1. Since dissolved silicate concentrations in Octob er 1999 (Fig. 16C-D) were relatively unchanged from the previous month ( Fig. 14C-D), the increase in chlorophyll-a stocks could not have been due to a diatom bloom. Instead, cell counts from the coastal and Sarasota transects confirmed that the chlorophyll-a stocks were attributed to a K. brevis bloom (Fig. 2C). In fact, surface samples of K. brevis at the Sarasota 10 m isobath indicated that a red tide of this size (1.9 x 106 cells L-1) would yield ~19 mg chl-a L-1, assuming 30 mg C mg-1 chl-a and 300 pg C cell-1 for K. brevis (Table 2). Since the chlorophyll-a concentration measured at the Sarasota 10 m isobath was 19.12 mg L-1, it was presumed that most of the 05 – 07 October 1999 chlorophyll-a stocks (Fig. 16A-B) were attributed to the red tid e. By 06 – 08 November 1999, surface concentrations of DIN were still <0.1 mmol -3NO kg-1 over much of the shelf (Fig. 17A). While a near-b ottom patch of relatively nitrate-enriched water ( >0.5mmol -3NO kg-1) was observed 55 km offshore of Tampa Bay during 05 – 07 October 1999 (Fig. 15B), it was advected southeast in the bottom Ekman layer (Fig. 10) and later observed 65 km offs hore of Sarasota on 06 – 07

PAGE 95

80 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) B) 05-07 October 1999Surface chl( g L )m-1a05-07 October 1999Bottom chl( g L )m-1a -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 C)D) 05-07 October 1999Surface SiO ( mol kg )4m-105-07 October 1999Bottom SiO ( mol kg )4m-12 7 6 M A X Figure 16. Spatial patterns of observed: A) chlor ophyll-a stocks (mg L-1) at the surface; B) chlorophyll-a stocks (mg L-1) at the bottom; C) silicate (mmol 4SiO kg-1) at the surface; and D) silicate (mmol 4SiO kg-1) at the bottom during 05 – 07 October 1999.

PAGE 96

81 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 06-07 November 1999Surface NO ( mol kg )3m-1 A) 06-07 November 1999Bottom NO ( mol kg )3m-1 B) C)D) 06-07 November 1999Surface PO ( mol kg )4m-106-07 November 1999Bottom PO ( mol kg )4m-1 Figure 17. Spatial patterns of observed: A) nitra te (mmol -3NO kg-1) at the surface; B) nitrate (mmol -3NO kg-1) at the bottom; C) phosphate (mmol -3 4PO kg-1) at the surface; and D) phosphate (mmol -3 4PO kg-1) at the bottom during 06 – 08 November 1999.

PAGE 97

82 November 1999 (Fig. 17B). It still exhibited maxim al concentrations of ~0.7 mmol -3NO kg-1, indicating that either nitrate uptake was minimal or that removal and regeneration in the near-bottom waters were near steady-state. Spatial patterns of the surface and near-bottom phosphate pools on the inner WFS on 06 – 08 November 1999 (Fig. 17C-D) were similar to those of the previous month (Fig. 1 5C-D), although DIP consumption on the inner WFS had reduced surface maxima to 0.2 – 0 .3 mmol -3 4PO kg-1 near the coast (Fig. 17D). Cell counts indicated that by 06 – 08 November 19 99 K. brevis stocks were at background concentrations of <1,000 cells L-1 (Fig. 2D), compared to a maximum surface concentration of 5.9 x 106 cells L-1 a month earlier (Fig. 2C). The distribution of chlorophyll-a near the coast during 06 – 08 November 1999 (Fig. 18A-B) was presumed to be due to diatom production instead, as evidence d by a draw-down of the nearshore silicate stocks (Fig. 18C-D) compared to previous m onths (Figs. 14C-D, 16C-D). 3.2.2. Ecological Model Results 3.2.2.1. Ecological Model Results: Case I – “No R efuge for K. brevis ” The Case I simulation tested the hypothesis that a red tide i nitiated offshore could be maintained by the nearshore pools of ammonium an d nitrate, despite maximum grazing pressure (i.e. given no refuge from the metazoan grazers). Altho ugh the small red tide (33,000 cells K. brevis L-1, or 0.33 mg chl-a L-1) measured in the near-bottom waters at the Sarasota 30 m isobath on 31 August 19 99 (Fig. 2A) was presumably initiated by the release of labile DON by T. erythraeum cells occupying the overlying

PAGE 98

83 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A)B) C)D) 06-07 November 1999Surface chl( g L )m-1a -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 06-07 November 1999Bottom chl( g L )m-1a06-07 November 1999Surface SiO ( mol kg )4m-106-07 November 1999Bottom SiO ( mol kg )4m-1 Figure 18. Spatial patterns of observed: A) chlor ophyll-a stocks (mg L-1) at the surface; B) chlorophyll-a stocks (mg L-1) at the bottom; C) silicate (mmol 4SiO kg-1) at the surface; and D) silicate (mmol 4SiO kg-1) at the bottom during 06 – 08 November 1999.

PAGE 99

84 water column, the simulation results indicated that K. brevis was incapable of any net production throughout the model domain under maxima l grazing losses (Fig. 19). With the exception of a small, non-toxic dinoflagellate bloom off Pensacola, the diatoms were responsible for the bulk of all net production, mos t of which was limited to an area near the Florida panhandle (Fig. 19) shoreward of the sh elf-break on 07 October 1999, where the estuarine DIN:DIP discharge was >16 (i.e. no DIN-limitation). Shelfwide concentrations of dissolved inorganic nut rients in the Case I simulation indicated that west Florida shelf waters above the 100 m isobath were replete with phosphate, relative to nitrate and ammonium stocks, except at the shelf-break (Fig. 20); thus, DIN-limitation was in effect for most of the phytoplankton over much of the photic zone. Despite heavy grazing losses (and the concom itant excretion of dissolved inorganic nutrients), ammonium concentrations were very low (0.01 – 0.03 mmol kg-1) compared to nitrate in shelf waters (Fig. 20A-B). Comparisons between Case I simulation results and observed near-bottom nutrient stocks indicated that nitrate availability was over-estimated by the model while phosphate and silicate stocks were under-estimated (Fig. 21) compared to those measured on 07 October 1999 (Figs. 15-16). The Case I simulation also failed to replicate the 19 mg chl-a L-1 of the red tide at the surface of the Sarasota 10 m isobath (Fig. 2C). Assuming background concentrations of 0.00607 mmol C kg-1 and a C:chl-a ratio of 30 for K. brevis cells (Table 2), the computed K. brevis biomass was at the refuge concentration (~0.0025 mg chl-a L-1) throughout the domain by 07 October 1999 (Fig. 19).

PAGE 100

85 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 B) Near-bottom chl-am( g/L) A) Near-surface chl-am( g/L)Case I Diatoms Microflagellates Dinoflagellates K. brevis Case I Diatoms Microflagellates Dinoflagellates K. brevis nn Figure 19. Case I simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999.

PAGE 101

86 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case I n -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case I n -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case I n -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case I nA) Bottom NH4B) Bottom NO3C) Bottom PO4D) Bottom SiO4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32E) Bottom DIN:DIP Case I n Figure 20. Results from the Case I simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelf-break on 07 October 1999.

PAGE 102

87 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case I n B) Bottom NO3 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case I n C) Bottom PO4 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) Bottom NH4D) Bottom SiO4 Case I nCase I n 0 0 1 Figure 21. Results from the Case I simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida o n 07 October 1999.

PAGE 103

88 Assuming a C:N molar ratio of 6.67 for K. brevis (Table 2), the mean nitrogen content within the K. brevis biomass averaged over the entire water column was <1 nM at-N (Fig. 22A) despite the relaxation of DIN-limit ation in nearshore waters (Fig. 22C) due to the excretion of 0.2 – 0.7 mmol kg-1 of recycled +4NH (Fig. 22B). Since the initial red tide was only 33,000 cells L-1 on 31 August 1999 (Fig. 2A), the metazoan zooplankton within the simulation had swiftly graze d the initial K. brevis cells and prevented any significant net production in west Fl orida shelf waters, regardless of nutrient availability. In fact, the Case I simulation of mean grazing loss to the algal bioma ss, averaged throughout the water column at the Sarasota 30 m is obath, was 0.738 mg chl-a eaten L-1 by 07 October 1999 (Table 10). Although grazing lo sses to K. brevis were only 4% (~0.03 mg chl-a eaten L-1) of the total grazing losses, the computed grazing stress on K. brevis was sufficient to reduce the initial red tide (0.3 3 mg chl-a L-1) on 31 August 1999 (Fig. 2A) to just 30% of the mean algal standing st ock (~0.0025 mg chl-a L-1) at the Sarasota 30 m isobath by 07 October 1999 (Table 10) Since K. brevis concentrations were calculated at the refuge threshold (~0.0025 mg chl-a L-1) throughout the water column along the Sarasota line (Fig. 23), the total accumulation of new K. brevis biomass as a consequence of production was <0.006 mg chl-a L-1 (Fig. 24A-B). In addition to the effects of grazing stress, K. brevis growth was inhibited by the relatively high light intensities of the west Flori da shelf (Fig. 24C-D). Since the Case I simulation assumed that DOC concentrations were 912 .6 mg L-1 at the surface and 8.0 mg L-1 for all depths below 5 m, the calculation of spect ral absorption due to the colored

PAGE 104

89 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case I n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case I n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 C) Bottom DIN:DIP Case I nCase I n D) Bottom E (PAR)oA) K. brevisPON B) Cum. Excretion NH4 Figure 22. Results from the Case I simulation of the: A) particulate organic nitroge n (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) nearbottom DIN:DIP ratios; and D) maximum near-bottom s calar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis.

PAGE 105

90 Table 10. Model comparisons between algal standing stock, biomass accumulation due to growth, biomass removal due to grazing stress, a mmonium accumulation due to grazer excretion, and the maximum near-bottom scalar irrad iance, from simulation Cases I-IV averaged over the entire water column at the Saraso ta 30 m and 10 m isobaths at 12:00 local hour on 07 October 1999. Computed Value Case I Case II Case III Case IV Sarasota 30 m isobath Mean algal standing stock (mg chla L-1): 0.0090.8830.8830.892 % Diatom biomass52%1%1%1% % Microflagellate biomass0%0%0%0% % Non-toxic Dinoflagellate biomass18%0%0%0% % K. brevis biomass 30%99%99%99%Mean biomass accumulation due to growth (mg chla L-1): 0.0921.0861.0861.082 % Diatom biomass92%17%17%17% % Microflagellate biomass1%0%0%1% % Non-toxic Dinoflagellate biomass4%0%0%1% % K. brevis biomass 3%83%83%81%Mean biomass removal due to grazing (mg chla eaten L-1): 0.7380.8190.8190.826 % from Diatom biomass87%92%92%90% % from Microflagellate biomass4%4%4%5% % from Non-toxic Dinoflagellate biomass5%4%4%5% % from K. brevis biomass 4%0%0%0%Mean NH 4 accumulation due to excretion (mmol NH4 kg-1):0.3450.4010.4010.399 Maximum near-bottom scalar irradiance (mE m-2 s-1): 48.62112.53112.53112.725Sarasota 10 m isobath Mean algal standing stock (mg chla L-1): 0.0090.0350.1720.403 % Diatom biomass52%13%3%20% % Microflagellate biomass0%0%0%0% % Non-toxic Dinoflagellate biomass18%4%0%0% % K. brevis biomass 30%83%97%80%Mean biomass accumulation due to growth (mg chla L-1): 0.3451.1001.6661.231 % Diatoms98%97%90%74% % Microflagellates1%<1%<1%2% % Non-toxic Dinoflagellates1%<1%<1%1% % K. brevis0%2%9%23%Mean biomass removal due to grazing (mg chla eaten L-1): 1.1931.9372.4491.787 % from Diatom biomass92%95%96%93% % from Microflagellate biomass4%3%2%4% % from Non-toxic Dinoflagellate biomass4%2%2%3% % from K. brevis biomass 0%0%0%0%Mean NH 4 accumulation due to excretion (mmol NH4 kg-1):0.5890.9741.2370.816 Maximum near-bottom scalar irradiance (mE m-2 s-1): 183.121179.03438.53534.737

PAGE 106

91 A)K. brevis(Tampa Transect)Case II 07 October 1999 01020304050607080 -50 -40 -30 -20 -10 0 Case I 07 October 1999 Sarasota TransectA)K. brevis(24:00 local hour) 01020304050607080 -50 -40 -30 -20 -10 0 B)K. brevis(12:00 local hour)Case I 07 October 1999 Sarasota Transect 0.0025 g L chlthroughoutm-1a0.0025 g L chlthroughoutm-1a Figure 23. Case I simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at mi dnight (A) and noon (B) along the Sarasota transect, 07 Oc tober 1999.

PAGE 107

92 01020304050607080 -50 -40 -30 -20 -10 0 010203040506070 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 ( g chlL )biomassm-1E) New K. brevisa( g chlL )biomassm-1F) New K. brevisaG) Limiting FactorK. brevisfor H) Limiting FactorK. brevisfor 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 C) Limiting FactorK. brevisfor D) Limiting FactorK. brevisfor( g chlL )biomassm-1A) New K. brevisa < 0.005 throughout < 0.006 throughout Case I n rnnCase I n rn( g chlL )biomassm-1B) New K. brevisaCase I n rn 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation Case I n rnnCase II n rnnCase II n rnCase II n rnCase II n rnn Figure 24. Depth profiles of total K. brevis accumulation due to new production and the growth limitation conditions for K. brevis along the Tampa and Sarasota transect lines, from simulation Cases I – II at 12:00 local hour on 07 October 1999. Panel H c ontour lines indicate K. brevis standing stock (mg chl-a L-1).

PAGE 108

93 fraction of DOC (Eqs. 37 – 38) led to relatively lo w CDOM absorption coefficients throughout the spatial domain, particularly near th e shore. As a result, maximum nearbottom scalar irradiance at the Sarasota 30 m isoba th was 48.6 mE m-2 s-1 (Table 10) and within the range of saturation intensity (45 – 65 mE m-2 s-1, Shanley & Vargo 1993) for K. brevis (Table 2). However, K. brevis was unable to maintain maximal growth inshore of the 27 m isobath (Fig. 22D) without suffering from photo-inhibition for at least part of the day. Since the maximum near-bottom scalar irradiances in shore of the 20 m isobath ranged from 100 – 250 mE m-2 s-1 (Fig. 22D), only those phytoplankton groups with h igh saturation (Esat) and high photo-inhibition (Einhib) irradiances (e.g. the diatoms and microflagellates) were not light-inhibited. This, coupled with the high grazing stress on K. brevis, led to the diatom group contributing to 92 – 98% of all new phytoplankton production on the inner shelf (Table 10) in the Case I simulation. As a result, the bulk of the ammonium excreted by the zooplankton grazers wa s associated with regions of diatom production (Fig. 25A-B), where this recycled nitrogen was added to the dissolved inorganic nitrogen stocks on the inner shelf (Fig. 25D), either as ammonium or as remineralized nitrate. 3.2.2.2. Ecological Model Results: Case II – “Gra zer Avoidance of K. brevis ” With the removal of all grazing stress on K. brevis in the Case II simulation, the slow-growing red tide organism was able to support surface concentrations of up to ~0.8 mg chl-a L-1 on the west Florida shelf by 07 October 1999 (Fig. 26A). Sub-surface

PAGE 109

94 Figure 25. Depth profiles of total diatom accumula tion (mg chl-a L-1) due to new production, total ammonium accumulation (mmol N kg-1) due to zooplankton excretion, total K. brevis accumulation of particulate organic nitrogen (mmol N kg-1) due to new production, and standing stock of dissolved inorgan ic nitrogen (mmol N kg-1) from simulation Cases I – II at 12:00 local hour on 07 October 1999. 01020304050607080 -50 -40 -30 -20 -10 0 A) New Diatom Biomass 01020304050607080 -50 -40 -30 -20 -10 0 B) Excreted NH4Case I n rnnCase I n rnn 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 < 0.0025 throughout Case I n rnnC) New PONK. brevisD) Total DINCase I n rnn 01020304050607080 -50 -40 -30 -20 -10 0 Case II n rnnE) New Diatom Biomass 01020304050607080 -50 -40 -30 -20 -10 0 F) Excreted NH4Case II n rnn 01020304050607080 -50 -40 -30 -20 -10 0 Case II n rnnG) New PONK. brevis 01020304050607080 -50 -40 -30 -20 -10 0 Case II n rnnH) Total DIN

PAGE 110

95 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 B) Near-bottom chl-am( g/L) A) Near-surface chl-am( g/L)nnCase II Diatoms Microflagellates Dinoflagellates K. brevis Case II Diatoms Microflagellates Dinoflagellates K. brevis Figure 26. Case II simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999.

PAGE 111

96 populations of K. brevis were also present near the bottom, where concentra tions reached 0.4 mg chl-a L-1 on the inner shelf (Fig. 26B). Despite increased production by K. brevis, the pool of dissolved inorganic nitrogen on the outer shelf was relatively unchange d (Fig. 27A-B) from that computed by the Case I simulation (Fig. 20A-B). While increased productio n had reduced nearshore inorganic nutrient stocks by as much as 0.384 mmol -3NO kg-1 and 0.028 mmol -3 4PO kg-1 offshore Sarasota (Fig. 28B-C), increased grazing s tress on the diatoms (Table 10) led to the addition of ~0.7 mM recycled ammonium (Fig. 28A) as a consequence of zooplankton excretion. The Case II simulation indicated that the growth and maintenan ce of the 31 August 1999 red tide was significantly influenced b y reduced grazing stress, such that the maximum computed surface concentrations of K. brevis had reached 17.2 mg chl-a L-1 along the Sarasota transect by 07 October 1999 (Fig 29A). During the daylight hours, the simulated K. brevis biomass could retreat to the bottom Ekman layer as a consequence of diel vertical migration (Fig. 29B), where the near-bottom accumulation of new K. brevis biomass (Fig. 24F) was affected more by nitrogen-l imitation than by photo-inhibition (Fig. 24H). Maintenance within th e sub-surface layers along the Sarasota transect ultimately led to the onshore tra nsport of K. brevis (Fig. 10B), where surface concentrations of 1 mg chl-a L-1 of red tide had been advected to within 28 km of the shore by 07 October 1999 (Fig. 29A). Since the Case II simulation used salinity fields from the physical model to estimate absorption due to CDOM within shelf waters overall estimates of aCDOM (and

PAGE 112

97 Figure 27. Results from the Case II simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelf-break on 07 October 1999. -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Bottom NH4 Case II n -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case II nB) Bottom NO3 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case II nC) Bottom PO4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case II nD) Bottom SiO4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32E) Bottom DIN:DIP Case II n

PAGE 113

98 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) Bottom NH4D) Bottom SiO4 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case II n C) Bottom PO4 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case II n B) Bottom NO3 Case II nCase II n Figure 28. Results from the Case II simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida o n 07 October 1999.

PAGE 114

99 Figure 29. Case II simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at mi dnight (A) and noon (B) along the Sarasota transect, 07 Oc tober 1999. 01020304050607080 -50 -40 -30 -20 -10 0 A)K. brevis(Tampa Transect)Case II 07 October 1999 Case II 07 October 1999 Sarasota TransectA)K. brevis(24:00 local hour) 01020304050607080 -50 -40 -30 -20 -10 0 B)K. brevis(12:00 local hour)Case II 07 October 1999 Sarasota Transect 17.2 g L chlmaxm-1a 5.2 g L chlmaxm-1a d e p t h ( m ) d e p t h ( m ) distance from shore (km)

PAGE 115

100 therefore atotal) were generally higher than those of Case I. This increased attenuation of the sub-surface light field effectively compressed the euphotic zone throughout the spatial domain, thereby shoaling the depth of satur ation intensity for all phytoplankton groups. While this had the general effect of reduc ing the overall volume of water in which new production could take place, it also had the effect of driving the scalar irradiance isolumes inshore (Fig. 30D), where the p hytoplankton competitors of K. brevis were more productive and more heavily grazed (Fig. 25E-F) compared to the Case I results. Assuming a C:chl ratio of 30.0 and a molar C:N rati o of 6.67 for K. brevis (Table 2), the total nitrogen demand within the K. brevis biomass (Fig. 25G) was either met or exceeded by the amount of recycled ammonium availab le as a result of zooplankton excretion (Fig. 25F) at all depths along the Saraso ta transect. Throughout the inner shelf, the depth-averaged accumulation of excreted ammoniu m (Fig. 30B) was greater than the average nitrogen content of the K. brevis biomass over the entire water column (Fig. 30A). Thus, it was evident that the nitrogen deman d of K. brevis in nearshore waters could be met by the excreted ammonium without requi ring additional sources of DIN. Despite these additions of recycled nitrogen, the D IN:DIP ratios in the nearbottom waters of the west Florida shelf indicated t hat DIN-limitation was still in effect (Figs. 27E, 30C). However, in nearshore waters whe re ammonium excretion (Fig. 25F) and DIN stocks (Fig. 25H) were maximal, the overlyi ng water column was not deep enough (nor was the absorption due to CDOM signific ant enough) to prevent photoinhibitive light intensities for K. brevis (Fig. 24H). Thus, light inhibition was still the greatest barrier to K. brevis growth inshore of the Sarasota 18 m isobath.

PAGE 116

101 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case II nCase II n B) Excreted NH4C) Bottom DIN:DIP Case II nCase II n D) Bottom E (PAR)oA) K. brevisPON Figure 30. Results from the Case II simulation of the: A) particulate organic nitroge n (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) nearbottom DIN:DIP ratios; and D) maximum near-bottom s calar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis.

PAGE 117

102 3.2.2.3. Ecological Model Results: Case III – “In creased Shading for K. brevis ” While the salinity fields used to estimate absorpti on due to CDOM (Eqs. 39 – 41) were estimated by the coupled physical model, these contributions were most likely an under-estimation of the true CDOM attenuation, as t hese salinity fields were as much as 4.0 psu higher than those measured from CTD casts n ear the coast during 06 – 10 September 1999 (Fig. 31). In order to compensate f or this under-estimation of aCDOM, the Case III simulation increased the concentration of non-biog enic suspended matter inshore of the 30m isobath (Eqs. 79 – 81). While t he magnitude of diatom production was relatively unchanged in surface waters (compare d to the Case II simulation), the geographic range of new diatom production extended well beyond the shelf-break (Fig. 32A). The increased shading of coastal waters (due to the simulated increase of suspended sediments) now increased near-bottom popu lations of K. brevis, particularly near Sarasota and Tampa Bay (Fig 32B). Again, near-bottom nutrient concentrations on the o uter shelf (Fig. 33) were largely unchanged compared to the Case II simulation (Fig. 27), despite the increased production of K. brevis biomass in the near-bottom waters (Fig. 32B). Sin ce the depthaveraged accumulation of algal biomass in Case III had increased from 1.10 to 1.67 mg chl-a L-1 at the Sarasota 10 m isobath (Table 10), nearshore stocks of inorganic nutrients were further reduced by ~0.3 mmol -3NO kg-1 and <0.01 mmol -3 4PO kg-1 in the nearbottom waters offshore Sarasota (Fig. 34). While the maximum surface concentrations of K. brevis remained at ~17.2 mg chl-a L-1 above the 30 m isobath, the increase of non-biogen ic suspended material inshore

PAGE 118

103 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -84-83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) Modeled SurfaceSalinity B) Modeled BottomSalinity D) Observed BottomSalinity C) Observed SurfaceSalinity Figure 31. Surface and near-bottom salinity fields as computed by the coupled physical model (A-B) or observed during CTD casts (C-D) at E COHAB stations during the period of red tide development (06 – 10 September 1999).

PAGE 119

104 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Near-surface chl-am( g/L) B) Near-bottom chl-am( g/L)Case III Diatoms Microflagellates Dinoflagellates K. brevis Case III Diatoms Microflagellates Dinoflagellates K. brevis Figure 32. Case III simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999.

PAGE 120

105 Figure 33. Results from the Case III simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelf-break on 07 October 1999. -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Bottom NH4 Case III n -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case III nB) Bottom NO3 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case III nC) Bottom PO4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case III nD) Bottom SiO4 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 E) Bottom DIN:DIP Case III n 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32

PAGE 121

106 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case III n C) Bottom PO4 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 B) Bottom NO3 Case III n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) Bottom NH4 Case III n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case III n D) Bottom SiO4 Figure 34. Results from the Case III simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida o n 07 October 1999.

PAGE 122

107 yielded a secondary surface amount of 3.4 mg chl-a L-1 at the 18 m isobath (Fig. 35). During the simulated day, this population escaped p hoto-inhibitive light intensities, while maintaining modest growth, such that a surface conc entration of 1 mg chl-a L-1 of red tide could propagate to within ~9 km of the coast (Fig. 35A). The use of Eqs. 79 – 81 to simulate coastal turbidi ty in Case III ultimately led to a complimentary increase in non-biogenic particulat es as the thickness of the water column decreased near the coast. Thus, K. brevis cells near the coast were shaded in the shallow, turbid waters while those in the deep, cle arer waters of the mid-shelf were able to migrate to a depth of saturation intensity. As a result, the maximum near-bottom light intensities were never photo-inhibitive for K. brevis (Fig. 36C-D) regardless of water depth or proximity to the coast. Thus, K. brevis cells that were concentrated near the bottom in order to avoid light inhibition in the sh allow, nearshore waters were also advected inshore in the bottom Ekman layer during p eriods of coastal upwelling (Fig. 10). As K. brevis cells were advected onshore, they were exposed to greater concentrations of excreted ammonium located between the 10 – 20 m isobaths (Figs. 37B, 38B) which relaxed the state of nitrogen-limit ation under saturating light intensities at the coast (Fig. 38D), thereby increasing K. brevis production at the bottom in nearshore waters (Fig. 36A-B). Similar to Case II, the excretion of ammonium from the nearshore grazer population was still more than sufficient to fuel the nitrogen demand of K. brevis (Figs. 37C, 38A), but again not sufficient to rever se the apparent nitrogen limitation for the total phytoplankton community, as indicated by the near-bottom DIN:DIP ratios (Fig. 38C).

PAGE 123

108 A)K. brevis(Tampa Transect)Case II 07 October 1999 01020304050607080 -50 -40 -30 -20 -10 0 Case III 07 October 1999 Sarasota TransectA)K. brevis(24:00 local hour) 01020304050607080 -50 -40 -30 -20 -10 0 B)K. brevis(12:00 local hour)Case III 07 October 1999 Sarasota Transect distance from shore (km) d e p t h ( m ) d e p t h ( m ) Figure 35. Case III simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at midnight (A) and noon (B) along the Sarasota transe ct, 07 October 1999.

PAGE 124

109 01020304050607080 -50 -40 -30 -20 -10 0 010203040506070 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 ( g chlL )biomassm-1A) New K. brevisa( g chlL )biomassm-1B) New K. brevisaC) Limiting FactorK. brevisfor D) Limiting FactorK. brevisfor 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 ( g chlL )biomassm-1F) New K. brevisa 01020304050607080 -50 -40 -30 -20 -10 0 G) Limiting FactorK. brevisfor 01020304050607080 -50 -40 -30 -20 -10 0 H) Limiting FactorK. brevisfor 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation 01020304050607080 -50 -40 -30 -20 -10 0 Phosphorus Limitation Nitrogen Limitation Light Limitation Case III n rnCase III n rnCase IV n rn( g chlL )biomassm-1E) New K. brevisaCase IV n rnCase IV n rnnCase IV n rnnCase III n rnnCase III n rnn Figure 36. Depth profiles of total K. brevis accumulation due to new production and the growth limitation conditions for K. brevis along the Tampa and Sarasota transect lines, from simulation Cases III – IV at 12:00 local hour on 07 October 1999. Contour l ines in panels C, D, G, and H indicate K. brevis standing stock (mg chl-a L-1).

PAGE 125

110 Figure 37. Depth profiles of total diatom accumula tion (mg chl-a L-1) due to new production, total ammonium accumulation (mmol N kg-1) due to zooplankton excretion, total K. brevis accumulation of particulate organic nitrogen (mmol N kg-1) due to new production, and standing stock of dissolved inorgan ic nitrogen (mmol N kg-1) from simulation Cases III – IV at 12:00 local hour on 07 October 1999. 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 01020304050607080 -50 -40 -30 -20 -10 0 A) New Diatom BiomassB) Excreted NH4Case III n rnnCase III n rnnC) New PONK. brevisCase III n rnnD) Total DINCase III n rnn 01020304050607080 -50 -40 -30 -20 -10 0 E) New Diatom BiomassCase IV n rnn 01020304050607080 -50 -40 -30 -20 -10 0 F) Excreted NH4Case IV n rnn 01020304050607080 -50 -40 -30 -20 -10 0 G) New PONK. brevisCase IV n rnn 01020304050607080 -50 -40 -30 -20 -10 0 H) Total DINCase IV n rnn

PAGE 126

111 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) K. brevisPON Case III nCase III nCase III nCase III n B) Excreted NH4D) Bottom E (PAR)oC) Bottom DIN:DIP Figure 38. Results from the Case III simulation of the: A) particulate organic nitroge n (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) nearbottom DIN:DIP ratios; and D) maximum near-bottom s calar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis.

PAGE 127

112 3.2.2.4. Ecological Model Results: Case IV – “Inc reased Shading/Increased Grazing” While maintaining the Case III conditions of increased turbidity inshore of the 3 0 m isobath, the Case IV simulation used a simplified grazer community (Eq. 61) with just two generalized grazers (metazoa and protozoa), usi ng mean carbon ingestion rates (Eqs. 62 – 64) and assuming a uniform vertical distributi on of the grazer biomass. While the Case III diatom fraction (96%) of the total grazed algal bi omass (2.449 mg chl-a L-1) was 2.35 mg chl-a L-1 at the 10 m isobath of the Sarasota transect by 07 October 1999 (Table 10), diatom growth was only 90% of the total algal growth (1.666 mg chl-a L-1), or 1.5 mg chl-a L-1; therefore, Case III grazing losses to the diatom group were ~156% of d iatom growth. Within the Case IV simulation, the diatom fraction (93%) of the total grazed algal biomass (1.787 mg chl-a L-1) was only 1.66 mg chl-a L-1. However, when compared to diatom growth (74% of 1.231 mg chl-a L-1, or 0.91 mg chl-a L-1), grazing losses to the diatoms in the Case IV simulation were ~182% of diatom growth. Thus, the simplified grazer community of the Case IV simulation represented an increase in grazing stre ss on the most dominant algal group in the simulation, th e diatoms. Despite the increased grazing stress, the uniform v ertical distribution of the Case IV grazers precluded any vertical migration of the me tazoa (i.e. the Group I grazers in Cases I – III). Thus, vertically-migrating phytoplankton (such as the non-toxic dinoflagellates) mitigated grazing losses by concen trating in sub-surface layers where the metazoan grazers could not aggregate, thereby incre asing the dinoflagellate biomass throughout the spatial domain (Fig. 39) compared to Case III (Fig. 32). While this increased grazing stress eliminated the microflagel late group, the range of both the diatom and K. brevis biomass in the surface waters was restricted to th e shelf break near

PAGE 128

113 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case IV Diatoms Microflagellates Dinoflagellates K. brevis B) Near-bottom chl-am( g/L) -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 Case IV Diatoms Microflagellates Dinoflagellates K. brevis A) Near-surface chl-am( g/L) Figure 39. Case IV simulation results of the: A) near-surface and B) near-bottom phytoplankton biomass (mg chl-a L-1) at 12:00 local hour on 07 October 1999.

PAGE 129

114 De Soto Canyon (Fig. 39A). While K. brevis was not subjected to grazing losses in Case IV, their reduced presence in the near-bottom waters (Fig. 39B) was due to increased competition with the vertically-migrating non-toxic dinoflagellates (Table 10). Near-bottom nutrient concentrations on the outer sh elf (Fig. 40) were similar to those in the Case III simulation (Fig. 33), although more of the near-bo ttom waters of the outer shelf were now phosphorus-limited. While the reduction in near-bottom nitrate was negligible (<0.2 mmol -3NO kg-1) on the inner shelf between Sarasota and Ft. Myers Florida (Fig. 41), available stocks of ammonium, ph osphate, and silicate were indistinguishable from those in Case III. Since total grazing losses to the diatom biomass we re 182% of total diatom growth in Case IV, this resulted in the swift removal of diatom chlo rophyll stocks along the Sarasota transect by 07 October 1999 (Fig. 37E) thereby reducing overall diatom production to just ~33% of the total diatom product ion simulated in Case III (Fig. 37A). This ultimately led to a reduction in the diatom biomass available for inges tion by the metazoan grazers; thus, total ammonium excretion in Case IV was reduced to just 55% of the total ammonium excretion by the Case III zooplankton grazers (Fig. 37B). In nearshore waters where initial chlorophyll stock s were maximal (Fig. 7), increased grazing stress on the diatoms in Case IV led to the reduction of diatom production to only 74% of the total algal productio n at the 10 m isobath of the Sarasota transect (Table 10); in Case III, diatom production was as much as 90% of the total algal production here. Thus, increased grazing stress on the diatoms ultimately resulted in reduced ammonium excretion (Fig. 37F), thereby limi ting K. brevis production in nitrogen-poor waters (Fig. 37H).

PAGE 130

115 Figure 40. Results from the Case IV simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); D) silicate (mmol 4SiO kg-1); and E) DIN:DIP ratios above the 100 m isobath at the shelf-break on 07 October 1999. -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 A) Bottom NH4 Case IV n Case IV B) Bottom NO3 Case IV C) Bottom PO4 Case IV D) Bottom SiO4 n n n -92-91-90-89-88-87-86-85-84-83-82-81 25 26 27 28 29 30 31 32 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Case IV nE) Bottom DIN:DIP

PAGE 131

116 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 B) Bottom NO3 Case IV n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case IV n C) Bottom PO4A) Bottom NH4D) Bottom SiO4 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case IV n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case IV n Figure 41. Results from the Case IV simulation of near-bottom: A) ammonium (mmol +4NH kg-1); B) nitrate (mmol -3NO kg-1); C) phosphate (mmol -3 4PO kg-1); and D) silicate (mmol 4SiO kg-1) between Tampa Bay and Charlotte Harbor, Florida o n 07 October 1999.

PAGE 132

117 Since initial chlorophyll stocks were lower offshor e than at the coast (Fig. 7), the increased grazing stress in Case IV affected diatom production at the Sarasota 20 m isobath more significantly than it did at the coast (Fig. 37E). Therefore, the difference in K. brevis biomass at the Sarasota 30 m isobath (Fig. 42A) wa s due to the reduction of nearshore sources of excreted ammonium (Fig. 37F) t hat reduced the growth of K. brevis in offshore waters. As a result of reduced ammonium excretion in nearsh ore waters (Fig. 37F), the effects of DIN-limitation were more significant her e as well, as indicated by the lower DIN stocks (Fig. 37H) and the lower DIN:DIP ratios (Fig. 43C). Regardless, the nitrogen-demand within the K. brevis biomass was still met by the excretion of recycled ammonium along the Sarasota transect (Fig. 37G-H) a nd throughout the inner shelf between Tampa Bay and Charlotte Harbor (Fig. 43A-B)

PAGE 133

118 A)K. brevis(Tampa Transect)Case II 07 October 1999 01020304050607080 -50 -40 -30 -20 -10 0 Case IV 07 October 1999 Sarasota TransectA)K. brevis(24:00 local hour) 01020304050607080 -50 -40 -30 -20 -10 0 B)K. brevis(12:00 local hour)Case IV 07 October 1999 Sarasota Transect Figure 42. Case IV simulation results, indicating the computed vertic al distribution of the K. brevis biomass (mg chl-a L-1) as a consequence of diel vertical migration at midnight (A) and noon (B) along the Sarasota transe ct, 07 October 1999.

PAGE 134

119 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 A) K. brevisPON Case IV n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 B) Excreted NH4 Case IV n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 C) Bottom DIN:DIP Case IV n -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 Case IV n D) Bottom E (PAR)o Figure 43. Results from the Case IV simulation of the: A) particulate organic nitroge n (mmol at-N kg-1) within the depth-averaged K. brevis biomass; B) depth-averaged accumulation of ammonium (mmol +4NH kg-1) due to zooplankton excretion; C) nearbottom DIN:DIP ratios; and D) maximum near-bottom s calar irradiance (mE m-2 s-1) at 12:00 local hour on 07 October 1999. Hatched areas indicate regions of saturation intensity (45 – 65 mE m-2 s-1) for K. brevis.

PAGE 135

120 Chapter 4: Discussion The red tide organism K. brevis has long been known as an oceanic species indigenous to oligotrophic waters (Steidinger & Had dad 1981, Tester & Steidinger 1997). Since oscillatory winds, in conjunction with water column stratification, ultimately result in an asymmetrical upwelling response on the WFS (W eisberg et al. 2001, He & Weisberg 2003, Weisberg & He 2003, Weisberg et al. 2004), this upwelling paradigm presents a cross-shelf conduit for material exchang e between the midand inner-shelf. Thus, an inoculum of K. brevis, having grown into a significant red tide at the m id-shelf (Walsh et al. 2006), could be delivered to the coas t if the bloom were constrained to the bottom Ekman layer during the period of coastal upw elling. Since it has been well demonstrated (Weisberg et al. 2000, He & Weisberg 2 003, Weisberg & He 2003, Weisberg et al. 2004) that the unique coastline and isobath geometries of the west Florida shelf tend to focus upwelling in the region between Tampa Bay and Charlotte Harbor, one would expect the nearshore delivery of red tide s to be focused in this region as well. In fact, a compilation of historical data of K. brevis cell counts indicate that the greatest likelihood of a coastal red tide existed within thi s upwelling-favorable region, near Sarasota and Venice, Florida (Fig. 44). If the red tide found near Sarasota on 05 – 07 Oc tober 1999 (Fig. 2C) was indeed initiated by the near-bottom inoculum that was samp led ~46 km offshore some 37 days earlier (Fig. 2A), the average daily cross-shelf tr ansport of the developing red tide must have exceeded 1.25 km day-1, or 1.45 cm sec-1. WFS-POM estimates of the cross-shelf transport indicate that during the presumed month o f bloom development, onshore crossshelf transport rates were greatest offshore Ft. My ers, Florida (Fig. 10C), where K. brevis

PAGE 136

121 -83.5-83-82.5-82-81.5 26 26.5 27 27.5 28 28.5 n n n 1,000,000 5,000,000 5,000,000 10,000,00010,000,000 20,000,00020,000,000 40,000,00040,000,000 60,000,000Cumulative Counts (cells L )1954 1997K. brevis-1S a r a s o t a V e n i c e Figure 44. K. brevis counts (cells L-1) from Florida Marine Research Institute (FMRI) sampling stations prior to 1998 (cumulative from 19 54-1997). Numbers in red indicate the total number of samples collected from each sta tion.

PAGE 137

122 first appears (Fig 2B). As these flowfields were c oupled to the biological model, the red tide inoculum at the 30 m isobath did indeed transi t inshore along the Sarasota transect (Fig. 45) throughout September 1999. During daylig ht hours, the inoculum remained in the near-bottom waters as a consequence of the mode led “stop response” in order to avoid photo-inhibitive light intensities. However, cross-shelf transport along the Sarasota transect (Fig. 10B) was not sufficient without a li ght-shading mechanism to allow red tide growth and maintenance in nearshore waters. Once the shading effects from CDOM and lithogenic p articulates were introduced in Case III, the simulated red tide then extended into coastal waters above the 10 – 15 m isobath (Fig. 35). Here, only 1 – 2 mg chl-a L-1 of the simulated red tide was realized because dead fish were not included in the model as a secondary source of nutrient supply during the maintenance of larger red tides. Although the photochemical degradation of CDOM is significant in these shallow coastal waters, previous simulations (Jolliff et al. 2003) indicate that the photo-degradation of CDOM does not release a significant amount of nitrogen for phytop lankton growth on the west Florida shelf; thus, CDOM is important mainly as an attenua ting compound. Within the Case I (“No Refuge for K. brevis ”) simulation, the estimated grazing losses far exceeded production, resulting in a shel fwide reduction of the phytoplankton biomass throughout the entire model domain, except at the shelfbreak in the northeastern Gulf of Mexico (Fig. 19). Contrary to the 1 – 3 mg chl-a L-1 measured on the west Florida shelf (Fig. 14A) during the period of red t ide development (07 September 1999) or the ~27.6 mg chl-a L-1 (Fig. 16A) at the height of the red tide (07 Octob er 1999), the

PAGE 138

123 Figure 45. Weekly distributions of K. brevis above the 30 m, 20 m, and 10 m isobaths of the Sarasota transect during August – December 1999 The sampling interval was ~7 days ( www.floridamarine.org ), but only the occasions where K. brevis were present (>1000 cells L-1) are shown. -84.5-84-83.5-83-82.5-82 26 26.5 27 27.5 S a r a s o t a t r a n s e c t l i n e30m 10m MOTE ECOHAB -30 -20 -10 0 B) 20 m Isobath 0 5000 10000 15000 33,000 cells L -1AUG SEPOCTNOVDEC -30 -20 -10 0 C) 10 m Isobath 0 200000 400000 600000 800000 1000000 1200000 1400000 1600000 1800000 2000000 1.9 x 10 cells L -1 6 -30 -20 -10 0 0 5000 10000 15000 20000 25000 30000A) 30 m Isobath

PAGE 139

124 Case I simulation failed to predict any appreciable chlor ophyll-a on the inner shelf between Tampa Bay and Charlotte Harbor, Florida (Fi g. 19). Since K. brevis cells were afforded no refuge from grazers in the Case I simulation, the use of published ingestion rates fo r the variety of zooplankton species contained within the grazing scheme (Tables 6 – 8) indicated that the few species of metazoan zooplankton which ingested K. brevis did so as a consequence of reduced electivity (i.e. the Group I grazers) or non-selective opportunism (i.e. the Group III grazers). As a result of this grazing stress, only 3% of the new phytoplankton growth was attributed to K. brevis above the 30 m isobath along the Sarasota transect (Table 10), thereby preventing the growth and maintenance of th e red tide (Fig. 24B). Although Lester (2005) had previously observed smal l (16,000 cells L-1) patches of K. brevis which were capable of growth rates which exceeded grazing losses on the west Florida shelf, only those red tides which were early in the initiation phase were subject to termination as a result of zooplankton g razing, even if K. brevis were growing at a realized growth rate of 0.2 day-1. Ultimately, Lester (2005) suggested that the methods which estimated grazing losses merely as fr actional reductions of the K. brevis biomass were not reliable predictors of red tide ma intenance when K. brevis was subjected to grazing stress; rather, it was necessa ry to consider the initial size and realized growth rate of the red tide as well. Therefore, the swift termination of the red tide in the Case I simulation indicated that either the modeled grazing losses to K. brevis were over-estimated or the light/nutrient conditions were unfavorable for red tide maintenance. While the low DIN:DIP ratios on the inner shelf indicated that DI N-limitation (Fig. 22C) was in effect,

PAGE 140

125 the average concentration of ammonium excreted by t he zooplankton grazers was 0.345 mmol +4NH kg-1 (Table 10), an amount which could have supported a n average of 0.94 mg chl-a L-1 of red tide at the Sarasota 30 m isobath (assuming a C:N molar ratio of 6.67 and a C:chl ratio of 30). In fact, in addition to high grazing losses, K. brevis growth in the Case I simulation was limited by light stress (Fig. 24D) rather than by nutrient availability. Since the maximum surface intensities frequently ex ceeded 500 mE m-2 s-1 by 07 October 1999, photo-inhibition was in effect for mu ch of the overlying water column due to insufficient shading by CDOM (estimated by Eqs. 34 – 38) in the Case I simulation (Fig. 24C-D). As a shade-adapted species, K. brevis has been shown to reach saturation at very low scalar irradiances, from 45–65 mE m-2 s-1 (Shanley & Vargo 1993). Consequently, photo-inhibitory effects swiftly redu ced the photosynthetic efficiency of K. brevis beyond this saturation intensity, thereby reducing the net realized growth rate. Since the maximum near-bottom scalar irradiance at solar noon was within the optimal range (45 – 65 mE m-2 s-1) for the red tide inoculum (Fig. 22D), the vertica llymigrating K. brevis cells would able to swim down to escape photo-inhi bitive intensities during periods of the most intense insolation, as l ong as the depth of the water column exceeded 40 m (Fig. 24C-D). As a result, K. brevis cells located on the mid-shelf could swim below these photo-inhibitive light intensities to escape light stress, while those K. brevis cells advected inshore would be unable to avoid li ght stress for much of the day. Of course, this presupposes a daily vertical migrat ion behavior, like that used within the model, where K. brevis cells will always swim up when light intensities a re sub-saturating and down when they are super-saturat ing. However, several studies

PAGE 141

126 indicate that the vertical migration behavior of K. brevis can be rather complex, where cells may exhibit occasional phototaxis and geotaxi s (Heil 1986, Kamykowski et al. 1998, Kerfoot et al. 2003) which vary in strength w ith time of day (Kamykowski et al. 1998). While these studies all suggest a vertical migration pattern that would tend to concentrate K. brevis at the surface during daylight hours (Kamykowski e t al. 1998, Kerfoot et al. 2003) rather than at night, there is evidence which suggests that laboratory (Kamykowski et al. 1998) and natural (Heil 1986) po pulations of K. brevis do indeed form sub-surface maxima during daylight hours and, in some cases, persist in the surface at night (Heil 1986). While the true vertical migration behavior of K. brevis in natural waters remains enigmatic, the imposition of positive phototaxis, w ithout an associated “stop response” when light becomes super-saturating, is inconsisten t with the photophysiology of K. brevis. As a shade-adapted organism (Walsh et al. 2007), K. brevis cells typically experience super-saturation (i.e. photo-inhibition) when ambient light intensities a re in excess of 65 mE m-2 s-1 (Shanley & Vargo 1993). Since chronic photo-inhib ition typically results in photo-induced damage to some p art of photosystem I (PSI) and/or photosystem II (PSII), this would lead to a reducti on in the maximum quantum yield for stable charge separation at PSII ( fIIe) during periods of light stress (Evans et al. 200 1). In fact, Evans et al. (2001) found that laboratory cultures and natural populations of K. brevis exhibited minimal alterations to fIIe during overcast (~200 mmol quanta m-2 s-1) conditions, while fIIe declined to just ~15% of the maximum fII during sunny (~1500 mmol quanta m-2 s-1) conditions. Since these photoinhibitory response s usually involve some type of repair/replacement mechanism (Evans et al. 2001), the vertical migration of

PAGE 142

127 K. brevis into surface waters during the daylight hours woul d increase the likelihood of damage to PSII, lower photosynthetic efficiency, an d incur additional metabolic expenditures to affect de novo protein/pigment/enzyme synthesis to reverse the ef fects (Prasil et al. 1992). Hence, light stress was mini mized in the model by using the supersaturation “stop response” intimated by Baden and M ende (1978). Therefore, light inhibition and grazing stress with in the Case I simulation were responsible for the failure to reproduce the signif icant red tide observed in nearshore waters during 05 – 07 October 1999 (Fig. 2C). Sinc e near-bottom irradiance (Fig. 22D) and overall grazer abundance (Fig. 8) were maximal near the coast, grazingand lightstress were prohibitive to the net production of K. brevis, thereby preventing the growth and maintenance of the 31 August 1999 near-bottom i noculum. While the Case I simulation failed to predict the 5.9 x 106 cells L-1 of K. brevis during 05 – 07 October 1999 (Fig. 2C), the swift termination of the red ti de is consistent with the conclusions reached by Lester (2005), whereby increased grazing stress may indeed present a mechanism for the termination of small red tides ea rly in the initiation phase. Since none of the non-selective ingestion rates wit hin the Group III zooplankton (Table 7) specifically tested K. brevis as a potential prey item, it was impossible to eliminate the possibility that in reality, K. brevis could be rejected by or otherwise incapacitate members of the Group III grazers (ther eby reducing the overall grazing stress placed upon the K. brevis biomass). Thus, the grazing scheme was modified i n the Case II (“Grazer Avoidance of K. brevis ”) simulation to afford protected status to K. brevis, rendering them “invisible” to zooplankton grazers w hile preserving the susceptibility of the remaining phytoplankton biomass to grazing loss es. This numerical experiment was

PAGE 143

128 thereby used to test the hypothesis that in the abs ence of grazing stress, the small red tide (33, 000 cells L-1 at the Sarasota 30 m isobath) measured near the bo ttom on 31 August 1999 (Fig. 2A) could indeed grow at a net realized growth rate sufficient to produce a large red tide (1.9 x 106 cells L-1 at the Sarasota 10 m isobath) within 37 days (Fig. 2C). Case II was also used to test the extent of CDOM-shading a s a consequence of estuarine efflux (Eqs. 39 – 41) rather than using the spatial ly-invariant DOC estimates (Darrow et al. 2003) employed in Case I (Eqs. 34 – 38). As a result, the Case II simulation indicated the growth and maintenance of the K. brevis biomass in nearshore waters of the west Florida sh elf (Fig. 26). Despite the widespread persistence of nitrogen-limitation with respect to the DIN:DIP ratios (Fig. 30C), the computed nitrate and phosphate concentrat ions (Fig. 28B-C) could have supported a great deal more phytoplankton productio n. In fact, model results suggest that while net production among the more palatable phyto plankton was limited by grazing losses (Table 10), the nearshore K. brevis biomass was instead limited primarily due to photo-inhibition (Fig. 24G-H). Consequently, any r eduction of the photo-inhibiting light field would expand the ecological space in which K. brevis could succeed, thereby increasing overall K. brevis production. Since the Case II simulation also contained an improved scheme (Eqs. 39 – 41) for estimating spectral absorption due to CDOM (aCDOM) using ambient salinity fields (Del Castillo et al. 2000), the nearshore sub-surfa ce light field was influenced more heavily by CDOM. As a result, the overlying water column was more effectively shaded by CDOM, where the near-bottom saturation (45 – 65 mE m-2 s-1) isolume had shifted to the 21 m isobath, only ~22 km offshore (Fig. 30D). Ultimately, this shoreward advance

PAGE 144

129 of the saturation isolume led to increased K. brevis production near the coast (Fig. 24EF), the dynamics of which were first elucidated by Walsh et al. (2003). These results are also consistent with the assert ion that red tides along the west Florida shelf originate in waters 18-74 km offshore (Steidinger 1973, Steidinger 1975, Steidinger & Haddad 1981) and are subsequently tran sported onshore. While even neutrally-buoyant phytoplankton experiencing light limitation/inhibition will exhibit maximal growth at depths where the scalar irradianc e approaches the saturation irradiance, K. brevis has been shown to exhibit a “stop response” (Baden & Mende 1978) at light intensities in excess of ~150 mE m-2 s-1 (Shanley 1985), thereby enabling a reversal of swimming direction in order to maintain position in waters nearest to the saturation intensity (Walsh et al. 2002). Thus, an y red tide developing offshore would occur in a sub-surface layer (most likely in the bo ttom Ekman layer), the inshore transport of which would be favored during periods of coastal upwelling. As a general consequence of this behavior, K. brevis cells migrate upward in the waning light of the afternoon and evening, becoming concentrated in the near-surface waters by the end of the day. Thus, reductions in the sub-surface light field due to increased attenuation due to CDOM or self-shading caused K. brevis to aggregate near the surface, particularly as nightfall approached. Consequently, the Case II simulation predicted that the 31 August 1999 nearbottom inoculum could indeed maintain position and growth within the bottom Ekman layer during the day, the bulk of which would becom e concentrated at the surface during the evening hours, reaching a maximum surface conce ntration of 17.2 mg chl-a L-1 by 07 October 1999 at 24:00 local hour (Fig. 29A). While the location of this surface

PAGE 145

130 aggregation was predicted incorrectly at the 30 m i sobath, its magnitude was similar to the surface concentrations of 19.3 mg chl-a L-1 (Fig. 16A) which were measured at the Sarasota 10 m isobath at 18:36 local hour. Although the increased light attenuation due to CDO M in the Case II simulation resulted in an inshore shift of the saturation isol ume in near-bottom waters (Fig. 30D), this shift was not significant enough to bring the calculated red tide near to the coast without the advection of the bottom Ekman layer, no r could it replicate the extent and range of the red tide witnessed there during 05 – 0 7 October 1999. Since the regressions that were used to estimate aCDOM (Eqs. 39 – 41) were dependent upon modeled salinit y fields that were as much as 4.0 psu higher than tho se observed in nearshore waters (Fig. 31) during the period of bloom development, the com puted light fields within the shallow, coastal waters were erroneously high. In an effort to rectify this presumed under-estim ation of light attenuation, the Case III (“Increased Shading for K. brevis ”) simulation applied an arbitrary increase in lithogenic particulates to those waters inshore of the 30m isobath, primarily to explore the effects of a hypothetical increase in coastal water turbidity. As a result, the Case III saturation isolume occupied a volume of near-bottom waters that extended all the way to the coast (Fig. 38D). While this inshore expansion of the saturation isolume had no effect on the accumulation of K. brevis at the surface near the 30 m isobath (17.2 mg chl-a L-1), it did lead to the shoreward expansion of a more favorable light climate, which extended the range of the red tide (3.4 mg chl-a L-1) inshore to the 10 m isobath (Fig. 35). Results from both the Case II and Case III simulations suggest that once a developing red tide comes within 40 – 50 km of the shore, the photo-inhibition of K.

PAGE 146

131 brevis growth becomes paramount in the shallow waters insh ore of the 30m isobath. Throughout the Case II simulation, K. brevis production was maximal above the Sarasota 30m isobath (Fig. 24F), where the maximum near-bott om scalar irradiance did not significantly exceed saturation intensity. This ef fectively limited the in situ growth of K. brevis at inshore locations along the Sarasota transect, preventing K. brevis from encroaching upon new water masses which may have be en advected alongshore toward Tampa (Fig. 24E) rather than offshore toward the sh elfbreak. As light stress was alleviated in the near-bottom waters at the Sarasot a 30 m isobath, the realized net growth rate of K. brevis was nitrogen-limited (Fig. 24H) due to the low DIN :DIP ratios in the near-bottom water of the west Florida shelf (Fig. 3 0C). Since the Case III simulation imposed conditions of increased attenua tion inshore of the 30m isobath, the conditions favorable for th e growth of the initial K. brevis inoculum were extended to the near-bottom waters of the Sarasota 20m isobath (Fig. 36B), where alongshore transport ultimately seeded the waters near the 20m isobath of the Tampa transect (Fig. 36A), allowing for an incr ease in the K. brevis biomass in the near-bottom waters where disparate cross-shelf tran sport effectively focused K. brevis fortuitously in a region where the effects of photo -inhibition were minimal (Fig. 36C-D). Both the Case II and Case III simulations suggest that red tide maintenance may rely upon two different light utilization strategies on the west Florida shelf; that is, clear waters of sufficient depth to allow K. brevis to swim below photo-inhibitive light intensities or shallow waters of sufficient turbidi ty to attenuate photo-inhibitive light intensities.

PAGE 147

132 Although the modeled concentrations of available ni trate far exceeded those measured on the west Florida shelf, nitrate concent rations within the Case II simulation (Fig. 28B) were barely diminished compared to the Case I simulation (Fig. 21B) despite the significant increase in K. brevis production (Fig. 24). The production of new K. brevis biomass while the nutrient levels remain relativel y unchanged is a phenomenon common to the west Florida shelf (Heil et al. 2002) where the sources of those nutrients required to maintain blooms for several months have yet to be identified (Vargo et al. 2001). While previous hypotheses range from DON in put from nitrogen fixation during Trichodesmium blooms (Lenes et al. 2001, Walsh & Steidinger 2001 ), remineralization of near-bottom diatom blooms fueled by shelf-break upw elling (Walsh et al. 2003), and estuarine flux of DIN/DIP (Vargo et al. 2003), simu lation results indicate that the dissolved inorganic nutrients excreted by zooplankt on grazers may represent a significant source of nitrogen in the form of ammonium (Fig. 30 B), far exceeding the calculated demand for nitrogen in order to maintain balanced g rowth within the K. brevis biomass (Fig. 30A). Since diatoms are generally light-adapted and exh ibit swift growth rates (Table 2), they are very effective competitors within nearshor e waters, ideally positioned to most effectively harvest the in situ nutrients proximal to the coast, particularly in r egions heavily influenced by estuarine effluents. However during periods of severe grazing stress, the nutrient content of the diatom biomass is quickly converted to ammonium by the zooplankton grazers, resulting in the rapid tur nover of nitrogen. This is particularly evident in the Case III simulation of cumulative diatom production, which was maximal at the points of estuarine outfall and in the nearb y coastal waters (Fig 37A, E). Of course,

PAGE 148

133 since diatoms dominated the fraction of phytoplankt on biomass which was susceptible to grazing, the highest estimates of excreted ammonium were collocated with waters exhibiting maximal diatom production (Fig. 37B, F). While the vertically-migrating K. brevis biomass was capable of scouring the water column for inorganic nutrients, resupply of n itrogen in offshore waters was somewhat limited by the reduced production of the d iatom biomass at the Sarasota 30 m isobath (Fig. 37A) as well as the reduced abundance of zooplankton grazers there (Fig. 8). Although this led to <0.1 mM DIN in waters occupied by the red tide developing at the Sarasota 30 m isobath (Fig. 37D), a larger and more persistent red tide would be favored if it were located closer to the coast, whe re zooplankton excretion of the metabolized diatom biomass provides a greater suppl y of ammonium to support new K. brevis production. The fact that red tides along the WFS are typically found shoreward of the 20 m isobath and rarely beyond the 30 m isob ath (Vargo, personal communication) supports this analysis. While the zooplankton excretion of ammonium from th e grazed diatom biomass represented an additional source of recycled nitrog en in the simulation cases, other sources of nitrogen exist which were not explicitly calculated in the ecological model. Most notably, Walsh et al. (2001, 2006) have advoca ted that red tide initiation and persistence on the west Florida shelf are a direct result of the aeolian delivery of iron-rich dust which alleviates iron-limitation within the co -occurring diazotroph biomass, thereby increasing the release of DON as a potential nitrog en source to fuel K. brevis production. Since the diazotroph T. erythraeum is capable of vertical migration via buoyancy cont rol (Lenes et al. 2001) and possesses a saturation inte nsity of 300 mE m-2 s-1 (Penta 2000),

PAGE 149

134 they would tend to aggregate higher in the water co lumn than the shade-adapted K. brevis. Thus, they may also provide a mechanism to incre ase biogenic particulate attenuation of the sub-surface light field (thereby reducing photo-inhibitive light intensities for K. brevis) in addition to providing atmospherically-fixed ni trogen as DON. While the lowN15dvalues associated with red tides are presumed to be indicative DON exudates from the diazotroph biomass subsequent to nitrogen fixation (Havens et al. 2002, Walsh et al. 2003), the Rayleigh distilla tion of N15 by the diatom biomass (prior to excretion by zooplankton grazers) may exp lain this as well. In regions where primary productivity is primarily supported by rege nerated ammonium, the isotopic value of both the regenerated ammonium and the PON should be close to that of oceanic nitrate, which is typically 5 – 7‰ (Mino et al. 200 2). However, in areas that are greatly affected by nitrogen fixation, lower N15d values should be expected because this process adds lighter nitrogen (-2 to 0‰) to the combined ni trogen pool (Mino et al. 2002). In fact, PON15d values from surface samples obtained at ECOHAB sta tions above the 10 – 50 m isobaths during the period of b loom initiation and development (August – September 1999) were 3.79 – 3.82‰ (Walsh et al. 2003), indicating a paucity of 15N in the ammonium and DON pools as a result of incr eased nitrogen fixation (and subsequent release DON) by T. erythraeum in the months preceding the October 1999 red tide (Lenes et al. 2001). However, PON15d values had increased to 4.97‰ during the October 1999 red tide (Fig. 2C), before falling to 3.03‰ by November 1999 (Walsh et al. 2006) when K. brevis had returned to background concentrations (Fig. 2D ). While Checkley and Miller (1989) found that the ammonium excreted by zooplankton was

PAGE 150

135 depleted in 15N (relative to an oceanic -3 15NOdof 5 – 7‰), PON15d values of 4.45‰ and 4.18‰ were found for diatom blooms near the mouth o f Tampa Bay in October 2000 (Walsh et al. 2003). Thus, if the diatom biomass i n September – October 1999 had PON15d values similar to these, K. brevis cells with a PON15d of 4.97% would be more indicative of a red tide maintained by the ammonium excreted by zooplankton that were primarily ingesting diatoms (Table 10). This dependence on the ammonium excreted by zooplan kton grazers was elucidated by the Case IV (“Increased Shading/Increased Grazing”) simulation, which initially sought to explore the necessity of using a complex grazing scheme (Eq. 58) featuring 13 different grazers (Tables 6 – 8). Ult imately, the Case IV simulation was less capable of simulating the true magnitude of the red tide observed during 05 – 07 October 1999, predicting maximum surface concentrations of 13.1 mg chl-a L-1 (Fig. 42) rather than the 17.2 mg chl-a L-1 predicted by Case III (Fig. 35). This was due in large part to the swift ingestion of initial chlorophyll-a stocks by the generalized metazoan grazers of Case IV. As the diatom biomass was rapidly grazed to back ground concentrations, subsequent production was severely limited (Fig. 37 E-F), thereby reducing the total ammonium excreted compared to the Case III simulation of zooplankton excretion. Without additional nitrogen sources from diazotroph exudates of DON or from the bacterial remineralization of organic nitrogen from dead fish associated with red tides, the excretion of ammonium from zooplankton g razers represents one of the most significant sources of recycled nitrogen necessary to support new K. brevis production. Although the equation used to estimate the ammonium excretion rate (Eq. 74) is linear, the equation governing the nutrient-limited carbon growth rate (Eq. 47) of phytoplankton

PAGE 151

136 is not. Therefore, the growth dynamics of the diat om biomass (as the primary source of this recycled nitrogen) provide the upstream limit that defines the extent of ammonium excretion, and ultimately, the nitrogen-limited gro wth dynamics of K. brevis. Since grazing losses were instantaneous in all nume rical simulations, all newlyproduced algal biomass was immediately subject to i ngestion (as was the standing stock); thus, in cases where ingestion rates were sufficien tly high, grazing losses could easily exceed new production, resulting in zero net produc tion and an attendant decrease in the algal standing stock. At the Sarasota 10 m isobath diatom production was only 74% of the 1.231 mg chl-a L-1 (or ~0.91 mg chl-a L-1) of the new algal biomass in the Case IV simulation, compared to 90% of 1.666 (or ~1.50 mg chl-a L-1) in Case III (Table 10). This ultimately led to an associated decrease in th e total ammonium excreted, which had declined to only 66% of the total ammonium excreted in Case III (Table 10). Although the generalized metazoan and protozoan gra zer community (Eqs. 61 – 64) utilized in Case IV had reduced the magnitude of K. brevis growth and maintenance on the west Florida shelf compared to Case III (Fig. 36), it did not change the fundamental dynamics which ultimately delivered an offshore inoculum of K. brevis to the coast, where increased surface concentrations i ndicated a significant red tide of 13.1 – 17.2 (Figs. 35, 42). However, the more complex gra zing scheme (Eq. 58), which included diverse rates of ingestion (Tables 7 – 9) and complex behaviors such as prey selection and diel vertical migration (Table 6), wa s regarded as the preferred method as it included true three-dimensional heterogeneity with regard to the complex interplay between predator/prey and offered the greatest pote ntial for future applications which

PAGE 152

137 could include simulations of zooplankton growth, fe cundity, and trophic transfers within a diverse grazer community. 4.1. Discussion Summary Although the conditions necessary for the initiatio n of red tides at the mid-shelf of the west Florida shelf have been well described (Wa lsh et al. 2007), the factors affecting bloom growth and maintenance near the coast are les s well understood. While the Case I – IV simulations sought to explore the role of light, n utrient, and grazing controls on the growth and maintenance of red tides, results from t he various numerical experiments indicated that all three may play a very significan t role. Regardless of condition, any relaxation of photo-inhibition within the shallow c oastal waters of the west Florida shelf resulted in the shoreward expansion of a growth-fav orable light climate, leading to increased K. brevis production in nearshore waters. In the absence of DON, the removal of photo-inhibit ive light stress resulted in nitrogen limitation due to the low DIN:DIP ratios i n the near-bottom waters. Therefore, K. brevis production was dependent upon other sources of nit rogen on the west Florida shelf, either in the form of new DON exudates relea sed by the nitrogen-fixing diazotroph T. erythraeum, recycled ammonium derived from the bacterial remi neralization of the dead fish associated with large red tides, or recyc led ammonium excreted by zooplankton grazers ingesting phytoplankton competitors of K. brevis. Among these, only the zooplankton excretion of ammonium was addressed; ho wever, simulation results indicated that the nitrogen demand within all new K. brevis production was met by the

PAGE 153

138 ammonium excreted by zooplankton grazers, provided that the grazers were avoiding K. brevis cells during the initiation and maintenance phases of the red tide. Of course, estuarine sources of nitrogen could also fuel red tide growth and maintenance in coastal waters, but these nutrients would be available to all phytoplankton competitors, not just K. brevis. Without significant shading by CDOM, nutrient-re plete waters will favor the growth of diatoms rather than a red tide (Walsh et al. 2003). Thus, significant grazing pressure in coastal waters will also lead to swift nutrient regeneration rates, thereby releasing the estuarine nutrients (r ecycled as zooplankton excreta) to the slower-growing K. brevis cells. However, if the coastal waters were signif icantly shaded by CDOM and/or suspended inorganic particulates, K. brevis success would be less sensitive to diatom turnover. Of course, this para digm presupposes some mechanism by which a K. brevis inoculum can be delivered to the coast. Ultimately, the nearshore delivery of K. brevis as a consequence of coastal upwelling was facilitated by the aggregation of K. brevis at the saturation isolume (45 – 65 mE m-2 s-1) in the bottom Ekman layer. While the transport o f estuarine water into the coastal zone leads to the development of thermal an d salinity fronts along which K. brevis blooms are frequently associated (Vargo et al. 200 3), the modeled “stop response” within the saturation isolume, conditions of selfand CDOM-shading, and the diel vertical migration behavior of K. brevis all led to an aggregation mechanism in the vertical dimension. Any increase in shading of the overlying water column, whether it was from increased CDOM concentrations, biogenic pa rticulates (such as a diazotroph bloom), or lithogenic particulates, ultimately redu ced photo-inhibitive light intensities and thereby removed light stress on K. brevis growth and maintenance.

PAGE 154

139 Chapter 5: Conclusions 1) Flowfields measured at moored ADCPs, observation s from AVHRR and TOPEX satellite imagery, and west Florida shelf circulati on models indicate that conditions of coastal upwelling existed during the period of bloom development, such that the K. brevis inoculum measured in the near-bottom waters at the Sarasota 30 m isobath on 31 August 1999 could have been maintained in growthfavorable conditions in the bottom Ekman layer and delivered near the coast by 05 – 07 October 1999. 2) Once a red tide was initiated by the diazotroph release of DON at the mid-shelf, the continued release of DON during red tide mainte nance was not necessary to supply K. brevis with sufficient nitrogen to reach the observed bio mass. In its maintenance phase, nitrogen demand within the red t ide could instead be supplied by the excretion of ammonium by zooplankton grazers which did not eat K. brevis but grazed all other phytoplankton competitors. 3) The vertical migration of K. brevis into the near-bottom Ekman layer was in response to photo-inhibitive light intensities in t he overlying water column. In order to maintain position in the saturation isolum e (45 – 65 mE m-2 s-1), K. brevis cells were constrained to the bottom Ekman layer du ring daylight hours.

PAGE 155

140 4) Confinement near the bottom of the water column reduced or eliminated light inhibition, thereby increasing the realized net gro wth rate of K. brevis cells while increasing the likelihood of onshore transport in t he bottom Ekman layer during periods of upwelling. 5) K. brevis cells delivered to the coast were capable of growt h maintenance in shallow water only when shaded by CDOM-rich or high ly turbid surface waters, thereby preventing photo-inhibition. 6) In situ growth and maintenance of K. brevis was sufficient to reach the observed proportions, but both the mechanisms of coastal upw elling and light-induced aggregation were necessary to reproduce the concent rations of K. brevis observed in nearshore waters during the 05 – 07 October 1999 red tide.

PAGE 156

141 References Alldredge, A.L., 1981. The impact of appendicularia n grazing on natural food concentrations in situ. Limnol. Oceanogr. 26 (2): 2 47-257. Atlas, E.L., Gordon, L.I., Hager, S.W. and Park, P. K. 1971. A practical manual for use of the Technicon AutoAnalyzer in seawater nutrient ana lysis (revised), A practical manual for use of the Technicon AutoAnalyzer in sea water nutrient analysis (revised). Oregon State University, Dept. of Oceano graphy, Corvallis, pp. 1-49. Baden, D.G. and Mende, T.J., 1978. Glucose transpor t and metabolism in Gymnodinium breve. Phytochemistry 17: 1553-1558. Bagoien, E., Miranda, A., Reguera, B. and Franco, J .M., 1996. Effects of two paralytic shellfish toxin producing dinoflagellates on the pe lagic harpacticoid copepod Euterpina acutifrons. Marine Biology 126 (3): 361-369. Besiktepe, S. and Dam, H.G., 2002. Coupling of inge stion and defecation as a function of diet in the calanoid copepod Acartia tonsa. Mar. Ecol. Prog. Ser. 229: 151-164. Bird, J.L., 1983. Relationships between particle-gr azing zooplankton and vertical phytoplankton distributions on the Texas continenta l shelf. Estuarine, Coastal and Shelf Science 16 (2): 131-144. Bissett, P., Carder, K.L., Walsh, J.J. and Dieterle D.A., 1999a. Carbon cycling in the upper waters of the Sargasso Sea: II. Numerical s imulation of the apparent and inherent optical properties. Deep Sea Research 46: 271-317. Bissett, P., Walsh, J.J., Dieterle, D.A. and Carder K.L., 1999b. Carbon cycling in the upper waters of the Sargasso Sea: I. Numerical sim ulation of differential carbon and nitrogen fluxes. Deep Sea Research 46: 205-269. Blumberg, A.F. and Mellor, G.L. 1987. A description of a three-dimensional coastal ocean circulation model. Page(s) 208-233 In: A description of a threedimensional coastal ocean circulation model, N. Hea ps (ed) AGU, Washington DC. Brown, C.M., McDonald-Brown, D.S., and Stanley, S.O ., 1975. Inorganic nitrogen metabolism in marine bacteria: Nitrate uptake and r eduction in a marine pseudomonad. Marine Biology 31(1): 7-13. Bott, A., 1989. A positive definite advection schem e obtained by nonlinear renormalization of the advective fluxes. Monthly We ather Review 117: 10061015.

PAGE 157

142 Bukata, R.P., Jerome, J.H., Kondratyev, K.Y. and Po zdnyakov, D.V. 1995. Optical properties and remote sensing of inland and coastal waters. CRC Press, Inc., Boca Raton. Calbet, A., Landry, M.R. and Scheinberg, R.D., 2000 Copepod grazing in a subtropical bay: species-specific responses to a midsummer inc rease in nanoplankton standing stock. Mar. Ecol. Prog. Ser. 193: 75-84. Cannon, J.A. 1990. Development and dispersal of red tides in the Port River, south Australia. Page(s) 110-116 In: Toxic Marine Phytoplankton, E. Graneli, B. Sundtrom, L. Edler and D. M. Anderson (eds). Elsevi er, New York, NY. Cannizzaro, J.P. HPLC absorption spectra for west Florida shelf phytoplankton, 07 July 1999. Personal communication. Cannizzaro, J.P., Carder, K.L., Chen, F.P., Walsh, J.J., Lee, Z., and Heil, C.A., 2003. A novel optical classification technique for detectio n of red tides in the Gulf of Mexico. Page(s) 282-284 In: Harmful Algae 2002: Proceedings from the Xth International Conference on Harmful Algae, K. A. St eidinger, C. R. Landsberg, C. R. Tomas and G. A. Vargo (eds). IOC of UNESCO, S t. Petersburg, FL. Carder, K.L., Steward, R.G., Harvey, G.R., and Ortn er, P.B., 1989. Marine humic and fulvic acids: their effects on remote sensing of oc ean chlorophyll. Limnol. Oceanogr. 34(1): 68-81. Carlson, C.A. and Ducklow, H.W., 1995. Dissolved or ganic carbon in the upper ocean of the Central Equatorial Pacific Ocean, 1992: Daily and finescale vertical variations. Deep Sea Research II 42 (2-3): 639-656. Carranza, A., Rosales, L. and Monreal, A., 1993. S uspended sediments in the southeastern Gulf of Mexico. Marine Geology 112 (14): 257-269. Cervetto, G., Pagano, M. and Gaudy, R., 1995. Feedi ng behavior and migrations in a natural population of the copepod Acartia tonsa. Hydrobiologia 300/301: 237248. Checkley, D.M. Jr. and Miller, C.A., 1989. Nitroge n isotope fractionation by oceanic zooplankton. Deep Sea Research 36(A): 1449-1456. Checkley, D.M. Jr., Uye, S., Dagg, M.J., Mullin, M. M., Omori, M., Onbe, T. and Zhu, M.Y., 1992. Diel variation of the zooplankton and i ts environment at neritic stations in the Inland Sea of Japan and the north-w est Gulf of Mexico. Journal of Plankton Research 14 (1): 1-40.

PAGE 158

143 Chew, F., 1956. A tentative method for the predicti on of the Florida red tide outbreaks. Bull. Mar. Sci. Gulf Carib. 6 (4): 292-304. Conover, S.A., 1956. Oceanography of Long Island So und, 1952-1954. IV. Phytoplankton. Bull. Bing. Oceanogr. Coll. XV: 62-1 12. Cotner, J.B., Ammerman, J.W., Peele, E.R., and Bent zen, E., 1997. Phosphorus-limited bacterioplankton growth in the Sargasso Sea. Aquat ic Microbial Ecology 13(2): 141-149. Dagg, M.J., 1995. Ingestion of phytoplankton by the microand mesozooplankton communities in a productive subtropical estuary. Jo urnal of Plankton Research 17 (4): 845-857. Darrow, B.P., Walsh, J.J., Vargo, G.A., Masserini J r., R.T., Fanning, K.A. and Zhang, J.Z., 2003. A simulation study of the growth of benth ic microalgae following the decline of a surface phytoplankton bloom. Continent al Shelf Research 23 (14-15): 1265-1283. Del Castillo, C.E., Gilbes, F., Coble, P.G. and Mul ler-Karger, F.E., 2000. On the dispersal of riverine dissolved organic matter over the west Florida shelf. Limnol. Oceanogr. 45: 1425-1432. Dolan, J.R. and Marrase, C., 1995. Planktonic cilia te distribution relative to a deep chlorophyll maximum: Catalin Sea, NW Mediterranean June 1993. Deep Sea Research 42 (11-12): 1985-1987. Dolan, J.R., Viddusi, F. and Claustre, H., 1999. Pl anktonic ciliates in the Mediterranean Sea: longitudinal trends. Deep Sea Research 46 (12 ): 2025-2039. Dragovich, A., J.A. Kelly, J. and Goodell, H.G., 19 68. Hydrological and biological characteristics of Florida's west coast tributaries Fishery Bulletin 66 (3): 463-477. Durbin, E.G., Garrahan, P.R. and Casas, M.C., 2000. Depth distribution of Calanus finmarchicus nauplii on the Georges Bank during 1995 and 1996. ICES Journal of Marine Science 57 (6): 1686-1693. Easter, R.C., 1993. Two modified forms of Bott's p ositive-definite numerical advection scheme. Monthly Weather Review 121: 297-304. EPA. 2006. Nutrient Database State Query Form for F lorida, Nutrient Database State Query Form for Florida. Evans, G.T., 1988. A framework for discussing seaso nal succession and coexistence of phytoplankton species. Limnol. Oceanogr. 33: 1027-1 036.

PAGE 159

144 Evans, T.J., Kirkpatrick, G.J., Millie, D.F., Chapm an, D.J., Schofield, O.M.E., 2001. Photophysiological responses of the toxic red-tide dinoflagellate Gymnodinium breve (Dinophyceae) under natural sunlight. J. Plankton Res. 23 (11): 1177-1193. Fernandez, D., Lopez-Urrutia, A., Fernandez, A., Ac una, J.L. and Harris, R., 2004. Retention efficiency of 0.2 to 6 mm particles by the appendicularians Oikopleura dioica and Fritillaria borealis. Mar. Ecol. Prog. Ser. 266: 89-101. Fraga, S., Reguera, B. and Bravo, I. 1990. Gymnodinium catenatum bloom formation in the Spanish rias. Page(s) 149-154 In: Toxic Marine Phytoplankton, E. Graneli, B. Sundtrom, L. Edler and D. M. Anderson (eds). Elsevi er, New York, NY. Fragopoulu, N., Siokuo-Frangou, I., Christou, E.D. and Mazzocchi, M.G., 2001. Patterns of vertical distribution of Pseudocalanidae and Par acalanidae (Copepoda) in pelagic waters (0 to 300 m) of the eastern Mediterr anean Sea. Crustaceana 74 (1): 49-68. Franks, P.J.S., 1992. Sink or swim: Accumulation o f biomass at fronts. Marine Ecology Progress Series 82: 1-12. Fulton, R.S.I., 1984. Distribution and community st ructure of estuarine copepods. Estuaries 7 (1): 38-50. Gallie, E.A. and Murtha, P.A., 1992. Specific absor ption and backscattering spectra for suspended minerals and chlorophyll-a in Chilko Lake British Columbia. Remote Sensing of the Environment 39 (2): 103-118. Gordon, H.R., 1989. Can the Lambert-Beer law be app lied to the diffuse attenuation coefficient of ocean water? Limnol. Oceanogr. 34 (8 ): 1389. Gordon, L.I., Jennings, J.C., Ross, A.A. and Krest, J.M. 1994. A suggested protocol for continuous flow automated analyses of seawater nutr ients (phosphate, nitrate, nitrite, and silicic acid) in the WOCE Hydrographic Program and the Joint Global Ocean Fluxes Study. Page(s) 1-120 In: WOCE Operations Manual, WHP. Off. Rep. WHPO 91-1, WOCE Rep. 68/91, U.S. World Ocean C irc. Exp. Off., College Station. Gregg, W.W. and Carder, K.L., 1990. A simple spectr al solar irradiance model for cloudless maritime atmospheres. Limnol. Oceanogr. 3 5 (8): 1657-1675. Guerrero, F., Rodriguez, V., Bautista, B. and Jimen ez-Gomez, F., 1988. Formation and loss of pheopigments in pelagic systems: Evaluation in laboratory experimental conditions. Scientia Marina 52 (3): 437-449.

PAGE 160

145 Guo, L., Coleman, C.H. and Santschi, P.H., 1994. Th e distribution of colloidal and dissolved organic carbon in the Gulf of Mexico. Mar ine Chemistry 45 (1-2): 105119. Haney, J.F., 1988. Diel patterns of zooplankton beh avior. Bull. Mar. Sci. 43 (3): 583-603. Hansen, G., Daugbjerg, N. and Henriksen, P., 2000. Comparative study of Gymnodinium mikimotoi and Gymnodinium aureolum, comb. nov. (=Gy rodinium aureolum) based on morphology, pigment composition, and molec ular data. Journal of Phycology 36: 394-410. Havens, J., Heil, C.A., Hollander, D., Vargo, G.A., Ault, D.N., Murasko, S. and Walsh, J.J. 2002. Isotopic constraints on nutrient sources supporting the 2001 Karenia brevis bloom. Page(s) 32-34 In: Harmful Algae 2002: Proceedings from the Xth International Conference on Harmful Algae, K. A. St eidinger, C. R. Landsberg, C. R. Tomas and G. A. Vargo (eds). IOC of UNESCO, S t. Petersburg, FL. He, R., 2001. WFS-POM model results for the autumn 1999 transition. Filename “wfsaut99.cdf”, 239 Mb. He, R. and Weisberg, R.H., 2002. West Florida shelf circulation and temperature budget for the 1999 spring transition. Continental Shelf R esearch 22 (5): 719-748. He, R. and Weisberg, R.H., 2003. West Florida shelf circulation and temperature budget for the 1998 fall transition. Continental Shelf Res earch 23: 777-800. Heil, C.A., Vargo, G.A., Spence, D.N., Neely, M.B., Merkt, R., Lester, K.M. and Walsh, J.J. 2002. Nutrient stoichiometry of a Gymnodinium breve bloom: What limits blooms in oligotrophic environments? Page(s) 165-16 8 In: Proceedings of the IX International Symposium on Harmful Algal Blooms, Ho bart, Australia, G. M. Hallegraeff (ed) U.N. Educ. Sci. and Cult. Org., Pa ris. Huntley, M., P. Sykes, S. Rohan, and V. Marin. 198 6. Chemically-mediated rejection of dinoflagellate prey by the copepods Calanus pacificus and Paracalanus parvus: mechanism, occurrence, and significance. Mar. Ecol Prog. Ser. 28: 105-120. Ikeda, T., 1970. Relationship between respiration r ate and body size in marine plankton animals as a function of the temperature of habitat Bull. Fac. Fish. Hokkaido Univ. 21: 91-112. JHU/APL, 2006. Johns Hopkins University (JHU) Appl ied Physics Laboratory (APL), AVHRR Imagery Ocean Remote Sensing Group, http://fermi.jhuapl.edu/avhrr/index.html

PAGE 161

146 Jassby, A.D. and Platt, T., 1976. Mathematical form ulation of the relationship between photosynthesis and light for phytoplankton. Limnol. Oceanogr. 21 (4): 540-547. Jeffrey, S.W., Sielicki, M. and Haxo, F.T., 1975. C hloroplast pigment patterns in dinoflagellates. Journal of Phycology 11: 74-84. Jeffrey, S.W. and Vesk, M. 1997. Introduction to ma rine phytoplankton and their pigment signatures. Page(s) 37-84 In: Phytoplankton pigments in oceanography: guidelines to modern methods, R. F. C. M. S.W. Jeff rey, and S.W. Wright (ed) UNESCO Publishing, Paris. Jeffrey, S.W., Wright, S.W. and Zapata, M., 1999. R ecent advances in HPLC pigment analysis of phytoplankton. Mar. Freshw. Res. 50: 87 9-896. Johnsen, G. and Sakshaug, E., 1993. Bio-optical cha racteristics and photo-adaptive responses in the toxic and bloom-forming dinoflagel lates Gyrodinium aureolum, Gymnodinium galatheanum, and two strains of Proroce ntrum minimum. Journal of Phycology 29: 627-642. Kamiyama, T., 1994. The impact of grazing by microz ooplankton in northern Hiroshima Bay, the Seto Inland Sea, Japan. Marine Biology 119 (1): 77-88. Kamiyama, T. and Arima, S., 2001. Feeding character istics of two tintinnid ciliate species on phytoplankton including harmful species: effects of prey size on ingestion rates and selectivity. J. Exp. Mar. Biol. Ecol. 257 (2): 281-296. Kamykowski, D., Milligan, E.J. and Reed, R.E., 1998 Relationships between geotaxis/phototaxis and diel vertical migration in autotrophic dinoflagellates. Journal of Plankton Research 20 (9): 1781-1796. Karentz, D. and Smayda, T.J., 1984. Temperature and seasonal occurrence patterns of 30 dominant phytoplankton species in Narragansett Bay over a 22-year period (19591980). Mar. Ecol. Prog. Ser. 18: 277-293. Kerfoot, J., Kirkpatrick, G., Lohrenz, S., Mahoney, K., Moline, M., and Schofield, O., 2003. Vertical migration of a Karenia brevis bloom: Implications for remote sensing of harmful algal blooms. Page(s) 279-281 In: Harmful Algae 2002: Proceedings from the Xth International Conference o n Harmful Algae, K. A. Steidinger, C. R. Landsberg, C. R. Tomas and G. A. Vargo (eds). IOC of UNESCO, St. Petersburg, FL. Kirchman, D.L., Suzuki, Y., Garside, C. and Ducklow H.W., 1991. High turnover rates of dissolved organic carbon during a spring phytopl ankton bloom. Nature 352 (6336): 612-614.

PAGE 162

147 Kirk, J.T.O. 1994. Light & Photosynthesis in Aquati c Ecosystems. Cambridge University Press, Cambridge, 509 pp. Knauss, J.A., 1997. Introduction to Physical Oceano graphy. Prentice Hall, Upper Saddle River, 309 pp. Kourafalou, V.H., Oey, L.-Y., Wang, J.D. and Lee, T .N., 1996. The fate of river discharge on the continental shelf 1. Modeling the river plume and the inner shelf coastal current. J. Geophys. Res. (C. Oceans) 101 ( C2): 3415-3434. Lenes, J.M., Darrow, B.P., Cattrall, C., Heil, C.A. Vargo, G.A., Callahan, M., Byrne, R.H., Prospero, J.M., Bates, D.E., Fanning, K.A. an d Walsh, J.J., 2001. Iron fertilization and the Trichodesmium response on the west Florida shelf. Limnology and Oceanography 46 (6): 1261-1277. Lester, K.M. 2005. Zooplankton of the west Florida shelf: Relationships with Karenia brevis blooms. Ph.D. Dissertation, University of South Fl orida, St. Petersburg, pp. 216. Li, Z. and Weisberg, R.H., 1999a. West Florida cont inental shelf response to upwellingfavorable wind forcing: Dynamics. Journal of Geoph ysical Research C. Oceans 104 (C10): 23427-23442. Li, Z. and Weisberg, R.H., 1999b. West Florida shel f response to upwelling-favorable wind forcing: Kinematics. Journal of Geophysical R esearch C. Oceans 104 (C6): 13507-13527. Liu, S. and Wang, W.X., 2002. Feeding and reproduct ive responses of marine copepods in South China Sea to toxic and nontoxic phytoplank ton. Marine Biology 140: 595-603. Loose, C.J. and Dawidowicz, P., 1994. Trade-offs in diel vertical migration by zooplankton: the costs of predator avoidance. Ecol ogy 75 (8): 2255-2263. Mackey, M.D., Mackey, D.J., Higgins, H.W. and Wrigh t, S.W., 1996. CHEMTAX a program for estimating class abundances from chemic al markers: application to HPLC measurements of phytoplankton. Mar. Ecol. Prog Ser. 144: 265-283. Margalef, R. 1968. Perspectives in ecological theor y. University of Chicago Press, Chicago, IL. Margalef, R., Estrada, M. and Blasco, D. 1979. Func tional morphology of organisms involved in redtides, as adapted to decaying turbul ence. Page(s) 89-94 In: Toxic Dinoflagellate Blooms, D. L. Taylor and H. H. Selig er (eds). Elsevier, New York, NY.

PAGE 163

148 Mazzocchi, M.G. and Paffenhofer, G.A., 1998. First observations on the biology of Clausocalanus furcatus (Copepoda, Calanoida). Journal of Plankton Researc h 20 (2): 331-342. Mellor, G.L. and Yamada, T., 1982. Development of a turbulence closure model for geophysical fluid problems. Review Geophysics 20: 8 51-875. Millie, D.F., Paerl, H.W. and Hurley, J.P., 1993. M icroalgal pigment assessments using high-performance liquid chromatography: a synopsis of organismal and ecological applications. Can. J. Fish. Aquat. Sci. 50: 2513-2527. Mino, Y., Saino, T., Suzuki, K., and Maranon, E., 2 002. Isotopic composition of suspended particulate nitrogen (susN15d) in surface waters of the Atlantic Ocean from 50N to 50S. Global Biogeochemical cycles 16 (4): 1059-1068. Mobley, C.D., 1994. Light and water: Radiative tra nsfer in natural waters. Academic Press, Inc., San Diego, 592 pp. Mobley, C.D., 2000. Hydrolight (v4.1) computer soft ware. Sequoia Scientific, Inc., Redmond, WA. Mopper, K., Zhou, X., Kieber, R.J., Sikorski, R.J. and Jones, R.D., 1991. Photochemical degradation of dissolved organic carbon and its imp act on the oceanic carbon cycle. Nature 353 (6339): 60-62. Morel, A., 1988. Optical modeling of the upper ocea n in relation to its biogenic matter content (Case I waters). Journal of Geophysical Res earch 93 (C9): 10,749-10,768. Morel, A., 1991. Light and marine photosynthesis: A spectral model with geochemical and climatological implications. Progress in Oceano graphy 26: 263. Narayanan, C. and Garvine, R.W., 2002. Large scale buoyancy driven circulation on the continental shelf. Dynamics of Atmospheres and Ocea ns 36 (1-3): 125-152. Nassogne, A., 1969. La coltura dei Copepodi in labo ratorio. Pubbl. Staz. Zool. Napoli 37 suppl.: 203-218. Ohtsuka, S., Bottger-Schnack, R., Okada, M. and Onb e, T., 1996. In situ feeding habits of Oncaea (Copepoda: Poecilostomatoida) from the upper 250 m of the central Red Sea, with special reference to consumption of appen dicularian houses. Bull. Plank. Soc. Japan 43 (2): 89-105. Omori, M. and Ikeda, T. 1984. Methods in Marine Zoo plankton Ecology. John Wiley & Sons, New York.

PAGE 164

149 Orlanski, I., 1976. A simple boundary condition for unbounded hyperbolic flows. Journal of Computational Physics 21: 251-269. Ornolfsdottir, E.B., Pinckney, J.L. and Tester, P.A ., 2003. Quantification of the relative abundance of the toxic dinoflagellate, Karenia brevis (Dinophyta), using unique photopigments. Journal of Phycology 39: 449-457. Pagano, M., Gaudy, R., Thibault, D. and Lochet, F., 1993. Vertical migrations and feeding rhythms of mesozooplanktonic organisms in t he Rhone River plume area (North-West Mediterranean Sea). Estuarine, Coastal and Shelf Science 37 (3): 251-269. Pagou, K. and Ignatiades, L. 1990. The peridoicity of Gymnodinium breve (Davis) in Saronicos Gulf, Aegean Sea. Page(s) 206-208 In: Toxic Marine Phytoplankton, E. Graneli, B. Sundtrom, L. Edler and D. M. Anderson ( eds). Elsevier, New York, NY. Parsons, T.R., Takahashi, M. and Hargrave, B. 1984. Biological Oceanographic Processes. Pergamon Press, Oxford. Peng, T.H., Takahashi, T. and Broecker, W.S., 1987. Seasonal variability of carbon dioxide, nutrients, and oxygen in northern North At lantic surface water: observations and model. Tellus 39B: 339-458. Pederson, B.A. CHEMTAX pigment ratios for west Flo rida shelf phytoplankton. Personal communication. Penta, B. 2000. Phytoplankton competition in the WF S: A simulation analysis with "red tide" implications. Ph.D. Dissertation, University of South Florida, St. Petersburg, FL, pp. 259. Pierce, R.H., Henry, M.S., Proffitt, L.S. and Hasbr ouck, P.A. 1990. Red tide toxin (brevitoxin) enrichment in marine aerosol. Page(s) 397-402 In: Toxic Marine Phytoplankton, E. Graneli, B. Sundtrom, L. Edler an d D. M. Anderson (eds). Elsevier, New York, NY. Pope, R.M. and Fry, E.S., 1997. Absorption spectrum (380-700nm) of pure water. II. Integrating cavity measurements. Applied Optics 36 (33): 8710-8723. Prasil, O., Adir, N., and Ohad, I., 1992. Dynamics of photosystem II: Mechanism of photoinhibition and recovery processes. Pages 295348 In: The Photosystems: Structure, Function, and Molecular Biology, J. Barb er (ed.). Elsevier, Amsterdam.

PAGE 165

150 Roach, P.J. 1976. Computational Fluid Dynamics. Her mosa Publications, Albuquerque, NM. Roff, J.C., Turner, J.T., Webber, M.K. and Hopcroft R.R., 1995. Bactivory by tropical copepod nauplii: extent and possible significance. Aquatic Microbial Ecology 9: 165-175. Rounsefell, G.A. and Dragovich, A., 1966. Correlati on between oceanographic factors and abundance of the Florida red tide (Gymnodinium breve Davis) during 19541961. Bulletin of Marine Science 16: 404-422. Sakamoto, Y., Lockey, R.F. and Krzanowski, J.J., 19 87. Shellfish and fish poisoning related to toxic dinoflagellates. South. Med. J. 80 : 866-870. Santos, B.A., 1992. Pastoreo de Paracalanus parvus (Claus, 1863) sobre dinoflagelados productores de mareas rojas en la plataforma argent ina. Bol. Inst. Esp. Oceanogr. 8 (2): 255-261. Satsmadjis, J. and Friligos, N., 1983. Red tide in Greek waters. Vie et Melieu 33 (2): 111117. Saunders, R.P. and Glenn, D.A. 1969. Memoirs of the Hourglass cruises, Memoirs of the Hourglass cruises. Marine Research Laboratory, Depa rtment of Natural Resources. SFWMD. 2006. South Florida Water Management Distric t's Water Quality (DBHYDRO) Data Browser, South Florida Water Management Distri ct's Water Quality (DBHYDRO) Data Browser. Shanley, E. 1985. Photoadaptation in the red tide d inoflagellate Ptychodiscus brevis. Ph.D. Dissertation, University of South Florida, St Petersburg, FL, pp. 122. Shanley, E. and Vargo, G.A. 1993. Cellular composit ion, growth, photosynthesis, and respiration rates of Gymnodinium breve under varying light levels. Page(s) 831836 In: Toxic Phytoplankton Blooms in the Sea, T. J. Smayd a and Y. Shimizu (eds). Elsevier, Amsterdam. Shimizu, Y., Chou, H. and Bando, H., 1986. Structur e of Brevitoxin A (GB-1 Toxin), the most potent toxin in the Florida red tide organism Gymnodinium breve (Ptychodiscus brevis). Journal of the American Chemical Society 108: 51 4-515. Shimizu, Y., Watanabe, N. and Wrensford, G. 1995. B iosynthesis of brevitoxins and heterotrophic metabolism in Gymnodinium breve. Page(s) 351-357 In: Harmful Marine Algal Blooms, P. Lassus, G. Arzul, E. Erard, P. Gentien and C. Marcaillou (eds). Lavoisier Intercept Ltd.

PAGE 166

151 Simon, J.L. and Dauer, D.M., 1972. A quantitative e valuation of red tide induced mass mortalities of benthic invertebrates in Tampa Bay, Florida. Environmental Letters 3: 229-234. Smayda, T.J., 1969. Some measurements of the sinkin g rate of fecal pellets. Limnol. Oceanogr. 14 (4): 621-625. Smayda, T.J. 1980. Phytoplankton species succession Page(s) 493-570 In: The Physiological Ecology of Phytoplankton, I. Morris ( ed) University of California Press. Smith, R.C. and Baker, K.S., 1981. Optical properti es of the clearest natural waters (200 800 nm). Applied Optics 20 (20): 177-184. Sommer, U. 1989. The role of competition for resour ces in phytoplankton succession. Page(s) 57-170 In: Plankton Ecology: Succession in Plankton Communit ies, U. Sommer (ed) Springer-Verlag, New York, NY. SooHoo, J.B. and Kiefer, D.A., 1982. Vertical distr ibution of phaeopigments 1. A simple grazing and photooxidative scheme for small particles. Deep Sea Research 29 (12A): 1539-1551. Steidinger, K.A. and Ingle, R.M., 1972. Observation s on the 1972 summer red tide in Tampa Bay, Florida. Environ. Letters 3: 271-278. Steidinger, K.A., 1973. Phytoplankton ecology: A c onceptual view based on eastern Gulf of Mexico research. CRC Crit. Rev. Microbiol. 3: 49-67. Steidinger, K.A. and Joyce Jr., E.A. 1973. Florida red tides, Florida red tides. State of Florida, Department of Natural Resources, pp. 26 pp Steidinger, K.A., 1975. Implications of dinoflagell ate life cycles on initiation of Gymnodinium breve red tides. Environmental Letters 9: 129-139. Steidinger, K.A. and Haddad, K.D., 1981. Biologic a nd hydrographic aspects of red tides. Bioscience 31 (11): 814-819. Steidinger, K.A., 1983. A re-evaluation of toxic di noflagellate biology and ecology. Progress in Phycology Research 2: 147-188. Stoecker, D.K. and Sanders, N.K., 1985. Differenti al grazing by Acartia tonsa on a dinoflagellate and a tintinnid. J. Plankton Res. 7 (1): 85-100.

PAGE 167

152 Sutton, T., Hopkins, T., Remsen, A. and Burghart, S ., 2001. Multisensor sampling of pelagic ecosystem variables in a coastal environmen t to estimate zooplankton grazing impact. Continental Shelf Research 21 (1): 69-87. Taft, W.H. and Martin, D.F., 1986. The potential fo r managing a Florida red tide. Journal of Environmental Science and Health A21 (2): 107-12 7. Tang, K.W., Chen, Q.C. and Wong, C.K., 1994. Diel v ertical migration and gut pigment rhythm of Paracalanus parvus, P. crassirostris, Acartia erythraea, and Eucalanus subcrassus (Copepoda, Calanoida) in Tolo Harbour, Hong Kong. Hydrobiologia 292/293: 389-396. Tester, P.A. and Steidinger, K.A., 1997. Gymnodinium breve red tide blooms: Initiation, transport, and consequences of surface circulation. Limnology and Oceanography 42 (5): 1039-1051. Thingstad, T.F., Li-Zweifel, U., and Rassoulzadegan F., 1998. P-limitation of heterotrophic bacteria and phytoplankton in the nor thwest Mediterranean. Limnol. Oceanogr. 43(1): 88-94. Turner, J.T., 1984. Zooplankton feeding ecology: C ontents of fecal pellets of the copepods Temora turbinata and T. stylifera from continental shelf and slope waters near the mouth of the Mississippi River. Mar ine Biology 82 (1): 73-83. Turner, J.T. and Tester, P.A., 1989. Zooplankton fe eding ecology: Copepod grazing during an expatriate red tide. Coastal and Estuarin e Studies 35: 359-374. Turner, J.T., Hopcroft, R.R., Lincoln, J.A., Huestr is, C.S., Tester, P.A. and Roff, J.C., 1998. Zooplankton feeding ecology: grazing by mari ne copepods and cladocerans upon phytoplankton and cyanobacteria fr om Kingston Harbour, Jamaica. Marine Ecology 19 (3): 195-208. USGS. 2006. Hydrologic Data Web Portal. http://hdwp .er.usgs.gov Uye, S., Imaizumi, K. and Matsuda, O., 1990. Size c omposition and estimated phosphorus regeneration rates of the copepod commun ity along the estuarineoffshore transect in the Inland Sea of Japan. Estua rine, Coastal and Shelf Science 31 (6): 851-863. Vargo, G.A., Heil, C.A., Spence, D.N., Neely, M.B., Merkt, R., Lester, K.M., Weisberg, R.H., Walsh, J.J. and Fanning, K.A. 2001. The hydro graphic regime, nutrient requirements, and transport of a Gymnodinium breve Davis red tide on the west Florida shelf. Page(s) 157-160 In: Proceedings of the IX International Symposium on Harmful Algal Blooms, Hobart, Australia, G. M. H allegraeff, S. I. Blackburn, C. Bolch and R. J. Lewis (eds). IOC of UNESCO, Pari s, France.

PAGE 168

153 Vargo, G.A., Heil, C.A., Ault, D.N., Neely, M.B., M urasko, S., Havens, J., Lester, K.M., Dixon, L.K., Merkt, R., Walsh, J.J., Weisberg, R.H. and Steidinger, K.A. 2003. Four Karenia brevis blooms: A comparative analysis Page(s) 14-16 In: Harmful Algae 2002: Proceedings from the Xth International Conference on Harmful Algae, K. A. Steidinger, C. R. Landsberg, C. R. Tom as and G. A. Vargo (eds). IOC of UNESCO, St. Petersburg, FL. Vinogradov, M.E. 1970. Vertical distribution of the oceanic zooplankton. Israel Program for Scientific Translations, Jerusalem. Walsh, J.J. and Dieterle, D.A., 1994. CO2 cycling in the coastal ocean. I A numerical analysis of the southeastern Bering Sea with applic ations to the Chukchi Sea and the northern Gulf of Mexico. Progress in Oceanograp hy 34: 335-392. Walsh, J.J., Penta, B., Dieterle, D.A. and Bissett, P., 2001. Predictive ecological modeling of harmful algal blooms. Human and Ecologi cal Risk Assessment 7 (5): 1369-1383. Walsh, J.J. and Steidinger, K.A., 2001. Saharan dus t and Florida red tides: The cyanophyte connection. Journal of Geophysical Resea rch C. Oceans 106 (C6): 335-392. Walsh, J.J., Haddad, K.D., Dieterle, D.A., Weisberg R.H., Li, Z., Yang, H., MullerKarger, F.E., Heil, C.A. and Bissett, P., 2002. A n umercial analysis of landfall of the 1979 red tide of Karenia brevis along the west coast of Florida. Continental Shelf Research 22 (1): 15-38. Walsh, J.J., Weisberg, R.H., Dieterle, D.A., He, R. Darrow, B.P., Jolliff, J.K., Lester, K.M., Vargo, G.A., Kirkpatrick, G.J., Fanning, K.A. Sutton, T.T., Jochens, A.E., Biggs, D.C., Nababan, B., Hu, C. and Muller-Karger, F.E., 2003. Phytoplankton response to intrusions of slope water on the west F lorida shelf: Models and observations. J. Geophys. Res. 108 (C6): 3190-3208. 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., Weisbe rg, R.H., Fanning, K.A., Muller-Karger, F.E., Shin, E., Steidinger, K.A., He il, C.A., Tomas, C.R., Prospero, J.S., Lee, T.N., Kirkpatrick, G.J., Whitl edge, T.E., Stockwell, D.A., Villareal, T.A., Jochens, A.E. and Bontempi, P.S., 2007. Red tides in the Gulf of Mexico: Where, when, and why. Journal of Geophysic al Research (submitted). Weisberg, R.H., Black, B.D. and Li, Z., 2000. An up welling case study of Florida's west coast. Journal of Geophysical Research 105 (C5): 11 459-11469.

PAGE 169

154 Weisberg, R.H., Li, Z. and Muller-Karger, F., 2001. West Florida self response to local wind forcing: April 1998. Journal of Geophysical Re search 106 (C12): 3123931262. Weisberg, R.H. and He, R., 2003. Local and deep-oce an forcing contributions to anomolous water properties on the West Florida Shel f. Journal of Geophysical Research 108 (C6), 3184, doi: 10.1029/2002JC001407. Weisberg, R.H., He, R., Kirkpatrick, G., Muller-Kar ger, F. and Walsh, J.J., 2004. Coastal ocean circulation influences on remotely sensed opt ical properties: A West Florida Shelf case study. Oceanography 17 (2): 68-7 5. Wilson, W.B. 1967. The suitability of seawater for the survival and growth of Gymnodinium breve Davis and some effects of phosphorus and nitrogen on its growth. Page(s) 1-42 In: Florida State University Professional Papers Serie s, Florida Board of Conservation, Tallahassee, FL. Wong, C.K., Chan, A.L.C. and Tang, K.W., 1992. Natu ral ingestion rates and grazing impact of the marine cladoceran Penilia avirostris Dana in Tolo Harbour, Hong Kong. Journal of Plankton Research 14 (12): 1757-17 65. Wright, S.W., Jeffrey, S.W., Mantoura, R.F.C., Llew ellyn, C.A., Bjornland, T., Repeta, D. and Welschmeyer, N., 1991. Improved HPLC method for the analysis of chlorophylls and carotenoids from marine phytoplank ton. Mar. Ecol. Prog. Ser. 77: 183-196. Wright, S.W. and van den Enden, R.L., 2000. Phytopl ankton community structure and stocks in the East Antarctic marginal ice zone (BRO KE survey, January-March 1996) determined by CHEMTAX analysis of HPLC pigmen t signatures. Deep Sea Research II 47: 2363-2400. Wyatt, T. 1990. Do algal blooms play homeostatic ro les? Page(s) 249-252 In: Toxic Marine Phytoplankton, E. Graneli, B. Sundtrom, L. E dler and D. M. Anderson (eds). Elsevier, New York, NY. Yang, H. and Weisberg, R.H., 1999. Response of the west Florida shelf circulation to climatological wind stress forcing. Journal of Geop hysical Research 104 (C3): 5301-5320. Yang, H., Weisberg, R.H., Niiler, P.P., Sturges, W and Johnson, W., 1999. Lagrangian circulation and forbidden zone on the West Florida Shelf. Cont. Shelf Res. 19: 1221-1245.

PAGE 170

155 Appendix

PAGE 171

156 Appendix A: Ecological Model Variables, Constants, and Coefficients Symbol Parameter Value Process Pi Carbon-specific phytoplankton biomass (mmol C kg-1) t Time (360 sec timestep-1) Tr Physical advective/diffusive transport mi* Carbon-specific realized net growth rate (day-1) gi Carbon-specific grazing rate (day-1) iw [-1.25 x 10-3, 0.0, 1.0, 1.0] Phytoplankton sinking/swimming rate (m hour-1) compE 6.0 Compensation irradiance (mE m-2 s-1) satE [190, 275, 150, 65] Saturation irradiance (mE m-2 s-1) inhibE [1940, 2810, 1530, 1530] Inhibition irradiance (mE m-2 s-1) i nitratek_ [1.05, 0.2, 1.8, 0.5] Half-saturation constant for nitrate (mmol m-3) i ammoniumk_ [1.5, 0.2, 0.9, 0.5] Half-saturation constant f or ammonium (mmol m-3) i phosphatek_ [0.5, 0.3, 0.2, 0.2] Half-saturation constant f or phosphate (mmol m-3) i silicatek_ [1.15, n.a., n.a., n.a.] Half-saturation consta nt for silicate (mmol m-3) mll_i Light-limited growth rate (day-1) mnl_i Nutrient-limited growth rate (day-1) mmax_i Maximum theoretical growth rate (day-1) mmaxT_i Temperature-dependent growth rate (day-1) T Temperature (C) fl Length of daily light cycle (hr) dj Julian day dj [0.0 – 1.0] Fractional orbital period of the earth Dsolar Solar declination (radians) LATref Reference latitude within the domain (28.4 394 N) p 3.14159265… Pi m'maxT_i Temperature-dependent hourly growth rate ( hr-1) Eo Total scalar irradiance (mE m-2 s-1) ai Photosynthetic efficiency (m2 (mmol quanta)-1) Fi 0.833 Maximum quantum yield (mmol C (mmol quanta)-1) a*i Carbon-specific spectral absorption coeff icient (m2 mmol C-1) Ed_z Spectral downwelling irradiance at depth z (mE m-2 s-1) Dz Path-length (m) Kd Diffuse attenuation coeff. for downwellin g plane irradiance (m-1) atotal Total spectral absorption (m-1) btotal Total spectral backscattering (m-1) md Average downwelling cosine at depth md 0Average downwelling cosine just beneath t he surface mdif 0Average cosine of the diffuse downwelling irradiance mdir 0Average cosine of the direct downwelling i rradiance Edif 0Diffuse spectral downwelling irradiance (mE m-2 s-1) Edir 0Direct spectral downwelling irradiance (mE m-2 s-1) Q Solar zenith angle (degrees from zenith) nw Index of refraction for seawater awater Spectral absorption due to seawater (m-1) aphyto Spectral absorption due to the total phyt oplankton biomass (m-1) aCDOM Spectral absorption due to CDOM (m-1) aCDOC_i Spectral absorption due to labile/relict C DOC (m-1) l Wavelength (nm) Model parameters for competition among diatoms, mic roflagellates, non-toxic dinoflagellates, and K. brevis [ i = d, f, n, b]

PAGE 172

157 Appendix A: (Continued) Symbol Parameter Value Process CDOCi Concentration of labile/relict CDOC (mg L-1) FRACi [0.8 relict, 0.2 labile] Fraction of total DOC w hich is labile/relict CDOC S Salinity (‰) water waterb b ~ Spectral scattering due to seawater (m-1) bphyto Spectral scattering due to biogenic parti culates (m-1) blith Spectral scattering due to lithogenic part iculates (m-1) SS Suspended sediment concentration (g m-3) NO3 Nitrate concentration (mmol m-3) NH4 + Ammonium concentration (mmol m-3) PO4 3Phosphate concentration (mmol m-3) SiO4 Silicate concentration (mmol m-3) Gnutrient Nutrient flux due to realized phytoplankto n growth (mmol kg-1) NIT Nutrient flux due to nitrification (mmol kg-1) nitk 0.10 Half-saturation constant for water co lumn nitrification (mmol kg-1) nitr 0.04 Nitrification rate (mmol kg-1 day-1) Ynutrient Nutrient flux due to estuarine discharge (mmol kg-1) R Daily riverine/estuarine streamflow (m3 s-1) SXnutrient Nutrient flux at the sediment-water interf ace (mmol kg-1) pvk 3.5 x 10-10 Mean bottom piston velocity (m s-1) AXnutrient Nutrient flux at the air-water interface (mmol kg-1) 2_CO xk 5.55 x 10-5 CO2 gas exchange coefficient (mmol m-2 mATM-1 s-1) 2PCOATM 377.38 Partial pressure of CO2 in the atmosphere (mATM) Dnutrient Nutrient flux due to fecal pellet reminera lization (mmol kg-1) i FECw_ [132, 36, 100, 100] Fecal pellet sinking rate (m day-1) Si P N CRtFEC, , [0.05, 0.05, 0.00, 0.01] Fecal pellet nutrient d issolution coefficient (day-1) cnutrient Nutrient flux due to zooplankton excretion (mmol kg-1) ZIRi Carbon-specific ingestion rate for grazer Z (mmol C eaten ind-1 s-1) ZEi [0.01, 1.0] Electivity coefficient for graze r Z ZPdiel [0.25 day, 0.75 night] Diel feeding periodicity for grazer Z VZ Volumetric abundance of grazer Z (ind m-3) AZ Areal abundance of grazer Z (ind m-2) wZ [0.5] Wave period of vertically migrating grazer Z ds Sigma layer thickness (m) jZ Phase shift of the vertical migration sin e function ZIRi Mean carbon-specific ingestion rate (mmol C eaten ind-1 s-1) kc(Z,I) Variable conversion factor to yield ZIRi in consistent units Zm,p Mean metazoan/protozoan grazer abundance (ind m-3) FECPi Carbon-specific fecal material from the gr azed Pi (mmol kg-1) D FECPi Fraction of fecal material lost due to diss olution (mmol kg-1) R’ Log-specific respiration rate of grazer Z (log [mL O2 mg-1 hr-1]) W Dry weight of grazer Z (mg) RQ [0.97] Respiratory quotient R Respiration rate of grazer Z (mg C hr-1) ERC/N/P Grazer excretion rate of inorganic C/N/P (m ass C/N/P ind-1 day-1) SIP Concentration of suspended inorganic part iculates (g m-3) asip Spectral absorption due to suspended inorg anic particulates (m-1) bsip Spectral backscattering due to suspended i norganic particulates (m-1) Model parameters for competition among diatoms, mic roflagellates, non-toxic dinoflagellates, and K. brevis [ i = d, f, n, b]

PAGE 173

About the Author Scott P. Milroy received a Bachelor’s Degree in Bio logy (Marine Emphasis) from Occidental College in Los Angeles, California in 19 91. After returning to academia in 1997, he received his Master’s Degree in Biology (M arine Emphasis) from Texas A&M University in Corpus Christi, Texas in 1999 studyin g coral reef ecology in Quintana Roo, Mexico. While pursuing his Doctorate Degree in Mar ine Science (Biological Oceanography Emphasis) at the University of South F lorida in St. Petersburg, Florida, he also taught several biology and oceanography course s as an Adjunct Instructor at St. Petersburg College.