USF Libraries
USF Digital Collections

Detection and quantification of Karenia brevis blooms on the West Florida shelf from remotely sensed ocean color imagery

MISSING IMAGE

Material Information

Title:
Detection and quantification of Karenia brevis blooms on the West Florida shelf from remotely sensed ocean color imagery
Physical Description:
Book
Language:
English
Creator:
Cannizzaro, Jennifer P
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
remote sensing
red tides
phytoplankton
light absorption
backscatter
Dissertations, Academic -- Marine Science -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Karenia brevis, a toxic dinoflagellate species that blooms regularly in the Gulf of Mexico, frequently causes widespread ecological and economic damage and can pose a serious threat to human health. Satellite-based ocean color imagery may provide a means for detecting and monitoring blooms, providing early alerts to coastal communities. However, a technique for discriminating K. brevis from other chlorophyll-containing algae is required. Between 1999 and 2001, a large bio-optical data set consisting of spectral measurements of remote-sensing reflectance (Rrs(lambda)), absorption (a(lambda)), and backscattering (bb(lambda)) along with chlorophyll a concentrations was collected on the central west Florida shelf (WFS) as part of the Ecology and Oceanography of Harmful Algal Blooms (ECOHAB) and Hyperspectral Coastal Ocean Dynamics Experiment (HyCODE) programs. Reflectance model simulations indicate that cellular pigmentation is not responsible for the factor of 3 to 4 decrease in Rrs(lambda) observed in waters containing greater than 10 4 cells 1 -1 of K. brevis. Instead, particulate backscattering coefficients measured inside K. brevis blooms are responsible for this decreased reflectivity as they were significantly lower than values measured in high-chlorophyll (> 1 mg m -3), diatom-dominated waters containing fewer than 10 4 cells 1 -1 of K. brevis. A paucity of high-backscattering detritus present in K. brevis blooms caused by decreased grazing pressure perhaps due to cellular toxicity along with a general inability of K. brevis to out compete diatoms and bloom in high-nutrient, high-backscattering estuarine waters may explain this low backscattering. A classification technique for detecting high-chlorophyll, low-backscattering K. brevis blooms is developed. In addition, a method for quantifying chlorophyll concentrations in positively flagged pixels using fluorescence line height (FLH) data obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) is introduced. Both techniques are successfully applied to Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and MODIS data acquired in late August 2001 and validated using in situ K. brevis cell concentrations.
Thesis:
Thesis (M.S.)--University of South Florida, 2004.
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 Jennifer P Cannizzaro.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 81 pages.

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 - 001469355
oclc - 55520569
notis - AJR1109
usfldc doi - E14-SFE0000257
usfldc handle - e14.257
System ID:
SFS0024952:00001


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 001469355
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 040524s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000257
035
(OCoLC)55520569
9
AJR1109
b SE
SFE0000257
040
FHM
c FHM
090
GC11.2
1 100
Cannizzaro, Jennifer P.
0 245
Detection and quantification of Karenia brevis blooms on the West Florida shelf from remotely sensed ocean color imagery
h [electronic resource] /
by Jennifer P Cannizzaro.
260
[Tampa, Fla.] :
University of South Florida,
2004.
502
Thesis (M.S.)--University of South Florida, 2004.
504
Includes bibliographical references.
516
Text (Electronic thesis) 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 81 pages.
520
ABSTRACT: Karenia brevis, a toxic dinoflagellate species that blooms regularly in the Gulf of Mexico, frequently causes widespread ecological and economic damage and can pose a serious threat to human health. Satellite-based ocean color imagery may provide a means for detecting and monitoring blooms, providing early alerts to coastal communities. However, a technique for discriminating K. brevis from other chlorophyll-containing algae is required. Between 1999 and 2001, a large bio-optical data set consisting of spectral measurements of remote-sensing reflectance (Rrs(lambda)), absorption (a(lambda)), and backscattering (bb(lambda)) along with chlorophyll a concentrations was collected on the central west Florida shelf (WFS) as part of the Ecology and Oceanography of Harmful Algal Blooms (ECOHAB) and Hyperspectral Coastal Ocean Dynamics Experiment (HyCODE) programs. Reflectance model simulations indicate that cellular pigmentation is not responsible for the factor of 3 to 4 decrease in Rrs(lambda) observed in waters containing greater than 10 4 cells 1 -1 of K. brevis. Instead, particulate backscattering coefficients measured inside K. brevis blooms are responsible for this decreased reflectivity as they were significantly lower than values measured in high-chlorophyll (> 1 mg m -3), diatom-dominated waters containing fewer than 10 4 cells 1 -1 of K. brevis. A paucity of high-backscattering detritus present in K. brevis blooms caused by decreased grazing pressure perhaps due to cellular toxicity along with a general inability of K. brevis to out compete diatoms and bloom in high-nutrient, high-backscattering estuarine waters may explain this low backscattering. A classification technique for detecting high-chlorophyll, low-backscattering K. brevis blooms is developed. In addition, a method for quantifying chlorophyll concentrations in positively flagged pixels using fluorescence line height (FLH) data obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) is introduced. Both techniques are successfully applied to Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and MODIS data acquired in late August 2001 and validated using in situ K. brevis cell concentrations.
590
Adviser: Carder, Kendall L.
653
remote sensing.
red tides.
phytoplankton.
light absorption.
backscatter.
690
Dissertations, Academic
z USF
x Marine Science
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.257



PAGE 1

Detection and Quantification of Karenia brevis Blooms on the West Florida Shelf from Remotely Sensed Ocean Color Imagery by Jennifer P. Cannizzaro A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science College of Marine Science University of South Florida Major Professor: Kendall L. Carder, Ph.D. Gabriel A. Vargo, Ph.D. John J. Walsh, Ph.D. Date of Approval: March 29, 2004 Keywords: Backscatter, Light Absorption, Phytoplankton, Red Tides, Remote Sensing Copyright 2004, Jennifer P. Cannizzaro

PAGE 2

ACKNOWLEDGEMENTS There are many people I would like to thank who have supported me greatly in one way or another during my many years working in the College of Marine Science. First and foremost, I would like to thank my advisor Dr. Kendall Carder for sharing with me his passion for ocean optics, providing me the freedom to explore my own interests, and patiently offering encouragement. I would also like to thank my committee members, Dr. Gabriel Vargo and Dr. John Walsh, for their support and guidance. Their classes were instrumental in providing me with the biological framework necessary to interpret my optical findings. My husband and son, I thank for providing balance to my life outside of the lab, so that I could put things into perspective. My fellow lab mates were also very supportive and deserve many thanks. Bob Steward not only provided me with computer and instrument support for many years, but also provided me an open ear for listening to my thoughts and ideas. My friend David English has helped me in countless ways. He taught me how to operate the in situ optical sensors, helped with data collection and processing, and supplied helpful editorial comments. Bob Chen kindly taught me the basics of image processing and Dan Otis also helped with data collection and processing. Karenia brevis cell counts presented in this thesis were kindly provided by Dr. Gabriel Vargo. His research group, led by Dr. Cindy Heil, generously provided our group with ship space and oftentimes patiently waited while we collected our shipboard data. High performance liquid chromatography samples were processed by Brad Pederson and Barb Berg (Mote Marine Laboratory, Sarasota, FL) under the supervision

PAGE 3

of Dr. Gary Kirkpatrick. SeaWiFS imagery was collected at the University of South Florida by Frank Mller-Karger and is presented courtesy of Orbimage and NASA. Financial support for this work wa s provided by NASA (NAS5-31716) and ONR (N00014-97-1-0006 and N00014-96-1-5013) funding.

PAGE 4

TABLE OF CONTENTS LIST OF TABLES..............................................................................................................ii LIST OF FIGURES...........................................................................................................iii ABSTRACT.......................................................................................................................vi 1. INTRODUCTION ........................................................................................................1 1.1 Background and Motivation.............................................................................1 1.2 Prior Bio-optical Studies of K. brevis...............................................................5 1.3 Objectives and Approach..................................................................................8 2. METHODS..................................................................................................................10 2.1 Shipboard Data...............................................................................................10 2.1.1 Discrete surface measurements........................................................11 2.1.2 Underway surface measurements.....................................................15 2.1.3 Remote-sensing reflectance measurements......................................16 2.2 Satellite Data...................................................................................................17 3. THEORY.....................................................................................................................18 4. RESULTS....................................................................................................................22 4.1 Shipboard Remote-sensing Reflectance Data.................................................22 4.2 Reflectance Model..........................................................................................24 4.2.1 Model parameterization....................................................................25 4.2.1.1 Phytoplankton absorption spectra....................................25 4.2.1.2 Detrital absorption spectra...............................................29 4.2.1.3 Gelbstoff absorption spectra.............................................33 4.2.1.4 Particulate backscattering spectra...................................36 4.2.2 Model application.............................................................................41 4.3 Classifying and Quantifying K. brevis Blooms from Space...........................44 4.3.1 Development and application of a classification technique.............44 4.3.2 Development and application of a quantification technique............48 5. DISCUSSION..............................................................................................................54 6. CONCLUSIONS..........................................................................................................61 LIST OF REFERENCES...................................................................................................62 i

PAGE 5

LIST OF TABLES Table 1 Symbol definitions.......................................................................................6 Table 2 Cruise data summary..................................................................................11 Table 3 Gelbstoff absorption slope coefficients, Sg................................................36 ii

PAGE 6

LIST OF FIGURES Figure 1. Map of the west Florida shelf study area...................................................10 Figure 2. Examples of phytoplankton, detrital and gelbstoff absorption spectra and the absorption spectra due to pure water (Pope and Fry, 1997)..........................................................................................................19 Figure 3. Example of particulate backscattering spectra and the backscattering spectra due to pure water (Morel, 1974)............................20 Figure 4. Median remote-sensing reflectance spectra measured on the WFS between 1999 and 2001 for K. brevis cell concentrations a) less than 104 cells l-1 and b) greater than 104 cells l-1, for various concentrations of chlorophyll (mg m-3).....................................................23 Figure 5. Phytoplankton absorption spectra, aph(), measured on the WFS between 1999 and 2001.............................................................................26 Figure 6. Relationship between phytoplankton absorption at 443nm, aph(443), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1).........................................................26 Figure 7. Spectral values of a) A(), b) B() and c) the coefficient of determination, r2, obtained by fitting a power law function to log-transformed aph() versus chlorophyll a concentration data for K. brevis concentrations less than (solid) and greater than (dashed) 104 cells l-1..................................................................................................28 Figure 8. Detrital absorption spectra, ad(), measured on the WFS between 1999 and 2001............................................................................................31 Figure 9. Relationship between detrital absorption at 443nm, ad(443), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1).....................................................................................................31 Figure 10. Contribution of detrital absorption to total particulate absorption at 443nm as a function of chlorophyll a concentration for various concentrations of K. brevis (cells l-1).........................................................32 iii

PAGE 7

Figure 11. Relationship between the spectral slope for detrital absorption, Sd, and chlorophyll a concentration for various concentrations of K. brevis (cells l-1)..........................................................................................32 Figure 12. Gelbstoff absorption spectra, ag(), measured on the WFS between 1999 and 2001............................................................................................34 Figure 13. Relationship between gelbstoff absorption at 400nm, ag(400), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1).....................................................................................................34 Figure 14. Relationship between the spectral slope for gelbstoff absorption, Sg, and chlorophyll a concentration for various concentrations of K. brevis (cells l-1)......................................................................................36 Figure 15. Relationship between particulate backscattering at 550nm, bbp(550), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1).........................................................37 Figure 16. Relative composition of the main accessory pigments for a) low-chlorophyll (< 0.2 mg m-3), non-K. brevis bloom waters, b) high-chlorophyll (>1.0 mg m-3), non-K. brevis bloom waters, and c) high-chlorophyll (>1.0 mg m-3), K. brevis bloom waters..........................39 Figure 17. Relationship between the particulate backscattering spectral shape parameter, and chlorophyll a concentration for various concentrations of K. brevis (cells l-1).........................................................41 Figure 18. Modeled remote-sensing reflectance coefficients at a) 443nm, b) 488nm, and c) 551nm calculated using a semi-analytical Rrs() model and plotted as a function of chlorophyll a concentration................43 Figure 19. Measured versus modeled chlorophyll a concentrations derived from shipboard Rrs() data collected on the WFS between 1999 and 2001 for various concentrations of K. brevis (cells l-1).......................46 Figure 20. Measured versus modeled bbp(550) derived from shipboard Rrs() data collected on the WFS between 2000 and 2001 for various concentrations of K. brevis (cells l-1).........................................................46 Figure 21. Relationship between modeled bbp(550) and modeled chlorophyll a concentration data derived from shipboard Rrs() data collected on the WFS between 1999 and 2001. Symbol size increases with increasing concentrations of K. brevis (cells l-1)........................................47 iv

PAGE 8

Figure 22. SeaWiFS chlorophyll concentrations (mg m-3) derived using the Carder et al. (1999) semi-analytical algorithm for the WFS on 30 August 2001...............................................................................................49 Figure 23. SeaWiFS data (30 August 2001) classified (shaded area) for K. brevis blooms according to the criteria: Chl > 1.5 mg m-3 and bbp(550) < the Morel (1988) expression. Shipboard K. brevis cell concentrations (cells l-1) collected 28-31 August 2001, are superimposed on top of the classified image.............................................49 Figure 24. Definition of fluorescence line height (FLH)............................................51 Figure 25. Relationship between measured chlorophyll a concentrations and fluorescence line height data (W m-2 m-1 sr-1) derived from shipboard Rrs() data collected on the WFS between 1999 and 2001 for various concentrations of K. brevis (cells l-1)..............................52 Figure 26. Chlorophyll a concentrations (mg m-3) derived from MODIS fluorescence line height data of the WFS (30 August 2001) for regions positively flagged for K. brevis.....................................................53 v

PAGE 9

Detection and Quantification of Karenia brevis Blooms on the West Florida Shelf from Remotely Sensed Ocean Color Imagery Jennifer P. Cannizzaro ABSTRACT Karenia brevis, a toxic dinoflagellate species that blooms regularly in the Gulf of Mexico, frequently causes widespread ecological and economic damage and can pose a serious threat to human health. Satellite-based ocean color imagery may provide a means for detecting and monitoring blooms, providing early alerts to coastal communities. However, a technique for discriminating K. brevis from other chlorophyll-containing algae is required. Between 1999 and 2001, a large bio-optical data set consisting of spectral measurements of remote-sensing reflectance (Rrs()), absorption (a()), and backscattering (bb()) along with chlorophyll a concentrations was collected on the central west Florida shelf (WFS) as part of the Ecology and Oceanography of Harmful Algal Blooms (ECOHAB) and Hyperspectral Coastal Ocean Dynamics Experiment (HyCODE) programs. Reflectance model simulations indicate that cellular pigmentation is not responsible for the factor of 3 to 4 decrease in Rrs() observed in waters containing greater than 104 cells l-1 of K. brevis. Instead, particulate backscattering coefficients measured inside K. brevis blooms are responsible for this decreased reflectivity as they were significantly lower than values measured in high-chlorophyll (>1 mg m-3), diatom-dominated waters containing fewer than 104 cells l-1 of K. brevis. A paucity of highvi

PAGE 10

backscattering detritus present in K. brevis blooms caused by decreased grazing pressure perhaps due to cellular toxicity along with a general inability of K. brevis to out compete diatoms and bloom in high-nutrient, high-backscattering estuarine waters may explain this low backscattering. A classification technique for detecting high-chlorophyll, low-backscattering K. brevis blooms is developed. In addition, a method for quantifying chlorophyll concentrations in positively flagged pixels using fluorescence line height (FLH) data obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) is introduced. Both techniques are successfully applied to Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and MODIS data acquired in late August 2001 and validated using in situ K. brevis cell concentrations. vii

PAGE 11

1. INTRODUCTION 1.1 Background and Motivation Harmful algal blooms of the toxic dinoflagellate, Karenia brevis (formerly Gymnodinium breve and Ptychodiscus brevis), occur regularly in the Gulf of Mexico, typically in late summer and fall (Steidinger et al., 1998). While background concentrations (1-103 cells l-1) are commonly found throughout the Gulf of Mexico (Steidinger, 1975), populations can increase above background levels to bloom proportions provided that physical, chemical, and biological conditions are suitable (Steidinger et al., 1998). K. brevis cells produce brevetoxins that accumulate in filter-feeding shellfish (i.e. oysters, clams, etc.) and when ingested by humans can cause Neurotoxic Shellfish Poisoning (Hemmert, 1975). Consequently, shellfish beds are ordered closed when K. brevis cell concentrations greater than 5000 cells l-1 are reported, making commercial shellfish industries extremely vulnerable to K. brevis blooms. Brevetoxins also cause bird, fish, and marine mammal mortalities (Landsberg and Steidinger, 1998) and can irritate human eyes and respiratory systems once they become airborne in sea spray (Hemmert, 1975; Asai et al., 1982). Fish kills typically occur when K. brevis cell concentrations exceed 105 cells l-1 (Steidinger et al., 1998). As a result, tourism industries are also vulnerable and have incurred millions of dollars in lost revenue due to bloom events (Habas and Gilbert, 1974). 1

PAGE 12

The west Florida shelf (WFS) extends ~200km west of the Florida coastline to the shelf break (~200m deep). It is ~700km long and is bordered to the north by the Florida Panhandle and to the south by the Florida Keys. Although K. brevis blooms have been reported in coastal waters throughout the Gulf of Mexico and as far north as North Carolina (Tester et al., 1991), they occur most frequently along the west Florida coast between Clearwater (~28oN) and Sanibel (~26.5oN) (Steidinger et al., 1998). Thus, the central WFS is an excellent region for studying the optical properties of K. brevis blooms. In general, background pigment concentrations (chlorophyll a and phaeopigments) on the western edge of the continental shelf are typically low (<0.25 mg m-3) (Mller-Karger et al., 1991) because nutrient concentrations are low (Masserini and Fanning, 2000). Low concentrations of phytoplankton and gelbstoff (or colored dissolved organic matter) that attenuate blue light cause the ocean to appear blue. Phytoplankton that dominate these oligotrophic waters include prokaryotes, prymnesiophytes and pelagophytes (Qian et al., 2003). In contrast, coastal waters typically appear various shades of green since blue light is attenuated by the high concentrations of phytoplankton and gelbstoff that are present. Gelbstoff originates mainly from the multiple rivers discharging into the Gulf coast of Florida. Diatoms generally dominate these low-salinity, high-nutrient waters (Qian et al., 2003) since they are fast growers and exhibit high nutrient-uptake efficiencies (Smayda, 1997). Increases in chlorophyll a concentrations above background levels (~0.25 mg m-3) also occur on the WFS during spring and summer due to phytoplankton other than K. brevis. Peak discharges by northwest Florida rivers (e.g. Apalachicola, Suwanee, etc.) 2

PAGE 13

during spring are responsible for mid-shelf phytoplankton blooms that discolor the water green and span from one week to two months (Gilbes et al., 1996). These plumes typically originate in the northeastern Gulf of Mexico and due to a combination of wind and buoyancy forcing (He and Weisburg, 2002) extend southeast parallel to the Florida coastline. CZCS imagery indicates that pigment concentrations inside these plumes are 0.5-5.0 mg m-3 (Gilbes et al., 1996). Succession from diatoms closest to the river source to cryptophytes has been postulated to occur in these plumes (Gilbes et al., 2002). Summertime southeasterly extensions of the Mississippi River also increase chlorophyll concentrations above background levels in offshore waters of the outer WFS. These low salinity, surface plumes contain a diverse array of eukaryotic (chromophytes and chlorophytes) and prokaryotic (Synechococcus) algae (Wawrik et al., 2003). Based on historical data, K. brevis blooms typically initiate in nutrient-poor waters located between 18 and 74km offshore (Steidinger, 1975; Steidinger and Haddad, 1981; Tester and Steidinger, 1997). The nature of the nitrogen supply (e.g. upwelling, riverine, etc.) that supports K. brevis blooms remains a topic of much debate. Walsh and Steidinger (2001), for instance, contend that wet deposition of Saharan aerosols in the eastern Gulf of Mexico may alleviate the iron limitation of diazotrophic cyanophytes (i.e. Trichodesmium spp.) which in turn fuel the nitrogen economy of K. brevis. Wind and currents can then transport blooms inshore to coastal waters. Here they are often maintained along physical fronts as they approach near monospecificity and are fueled by additional nutrient sources (Vargo et al., 2000). This is where the greatest ecological and economic damage occurs as well as the greatest hazard to human health. 3

PAGE 14

Current coastal management strategies employed by state agencies for detecting and monitoring K. brevis blooms in the Gulf of Mexico rely primarily on the collection of discrete samples from mainly nearshore waters after blooms have been transported inshore. These discrete measurements (e.g. microscopic cell counts, toxin and pigment analyses, and chlorophyll concentrations) are can be labor intensive and the resultant temporal and spatial resolution is poor. Consequently, discrete sample data collected by coastal monitoring programs are susceptible to considerable bias and are often untimely in terms of providing early alerts to state and local officials. To mitigate the human health risks and negative economic impacts associated with K. brevis blooms, an accurate system with high spatial and temporal resolution is required that can differentiate K. brevis blooms from both high-chlorophyll, coastal waters and the seasonal, episodic plume events mentioned previously. Since K. brevis blooms discolor oceanic surface waters (Carder and Steward, 1985; Vargo et al., 1987; Tester et al., 1998), satellite-based ocean color sensors may provide a robust tool for remotely detecting and monitoring harmful algal blooms (Cullen et al., 1997; Schofield et al., 1999). Ocean color sensors measure the amount of light reflected from the upper ocean at specific wavebands and provide daily coverage (~1km) of the Gulf of Mexico. Algal biomass, measured as the concentration of chlorophyll a (Chl), can be accurately estimated from ocean color data (Gordon et al., 1983; O'Reilly et al., 1998; Carder et al., 1999; Carder et al., 2003). However, since all phytoplankton contain chlorophyll a and since other types of algae (e.g. diatoms, cryptophytes, etc.) bloom regularly in the Gulf of 4

PAGE 15

Mexico (Lambert et al., 1999; Gilbes et al., 2002; Qian et al., 2003), a method for discriminating between K. brevis and other major bloom-forming algae is required. While algal quantity (i.e. Chl) can be estimated relatively accurately from satellite-based ocean color data, identifying algal quality or the type of algae present is a more challenging task (Garver et al., 1994). Since K. brevis blooms, defined in this study as waters containing greater than 104 cells l-1, occur in mixed algal populations at low bloom levels and achieve near-monospecificity only at high bloom levels (~106 cells l-1), detecting K. brevis blooms early is problematic. Discriminating among algal groups requires systemic deviations in spectral absorption, a(), and/or backscattering, bb(), coefficients (see Table 1 for symbol definitions) since the fraction of light entering the ocean that is reflected, or the remote-sensing reflectance, Rrs(), is dominated by the ratio of bb() to a() (Morel and Prieur, 1977). 1.2 Prior Bio-optical Studies of K. brevis The bio-optical properties of K. brevis have been the focus of numerous studies during the past several decades. Bjornland and Liaaen-Jensen (1989) documented the pigment composition of K. brevis (=Ptychodiscus brevis) from cultures. While most dinoflagellate species utilize peridinin as the major light-harvesting accessory pigment, K. brevis utilizes the carotenoid fucoxanthin instead. In an earlier study, Jeffrey et al. (1975) hypothesized that an endosymbiotic relationship between heterotrophic colorless flagellates and autotrophic, fucoxanthin-containing chrysophytes could possibly explain this evolutionary difference. More recent studies have revealed the presence of a minor carotenoid, gyroxanthin-diester, that is unique to the genus Karenia in the relatively 5

PAGE 16

warm Gulf of Mexico waters. (Millie et al., 1995; Millie et al., 1997). Since concentrations of this pigment exhibit a strong correlation with both chlorophyll a and cell concentrations, gyroxanthin-diester may serve as a bio-marker for natural populations. Table 1. Symbol definitions Symbols Description Units a Absorption coefficient (=aw+aph+ad+ag) m-1 aw Absorption coefficient of pure water m-1 ap Absorption coefficient of particulates (=aph+ad) m-1 aph Absorption coefficient of phytoplankton m-1 ad Absorption coefficient of detritus m-1 ag Absorption coefficient of gelbstoff m-1 A Empirical shape coefficient for power law function (y=AxB) b b Backscattering coefficient m-1 bbw Backscattering coefficient of pure water m-1 bbp Backscattering coefficient of particulates m-1 bp Scattering coefficient of particles m-1 B Empirical slope coefficient for power law function (y=AxB) c Attenuation coefficient m-1 Chl Chlorophyll a concentration mg m-3 Ed Downwelling irradiance W m-2 nm-1 f Water-to-air divergence factor FLH Fluorescence line height W m-2 m-1 sr-1 LG Radiance reflected from a 10% diffuse reflector or gray card Lsky Downwelling sky radiance W m-2 nm-1 sr-1 Lu Upwelling radiance W m-2 nm-1 sr-1 Lw Water-leaving radiance W m-2 nm-1 sr-1 n Refractive index of seawater nLw Normalized water-leaving radiance W m-2 nm-1 sr-1 Q Upwelling irradiance-to-radiance ratio sr-1 r Fresnel reflectance RG Reflectance of a 10% diffuse reflector or gray card Rrs Above surface remote-sensing reflectance sr-1 rrs Subsurface remote-sensing reflectance sr-1 Sd Spectral slope for detrital absorption spectra m-1 Sg Spectral slope for gelbstoff absorption spectra m-1 t Transmittance across the air-sea interface Angstrom exponent describing spectral shape of bbp() 6

PAGE 17

Comparisons of absorption properties between phytoplankton cultures indicate that the absorption spectra of K. brevis differ only slightly from other dinoflagellates (Jeffrey, 1980) and diatoms (Millie et al., 1997). This is true most likely because the main accessory pigments found in K. brevis (fucoxanthin, 19-acylofucoxanthins, and chlorophyll c) co-occur in other algal groups (e.g. prymnesiophytes and chrysophytes). In addition, the purported bio-marker pigment, gyroxanthin-diester, typically comprises less than 5% of the total pigment concentration and absorbs light in the same spectral region as the other carotenoids (Millie et al., 1995). Despite the fact that only weak differences in absorption spectra have been observed, a method to identify in situ populations of K. brevis was developed based on pigment absorption spectra (Millie et al., 1997). This technique has been successfully applied to natural populations in the Eastern Gulf of Mexico (Kirkpatrick et al., 2000; Kirkpatrick et al., 2002). However, this technique is only amenable to in situ platforms (i.e. ships, moorings and autonomous underwater vehicles). Satellite-based ocean color sensors cannot be used to identify K. brevis using this method since pigment absorption is difficult to discern from remotely sensed ocean color data (Garver et al., 1994). Furthermore, the need for hyperspectral data also eliminates satellite ocean color sensors as a possible platform for this method since current sensors typically contain fewer than ten visible wavebands. In the past, most bio-optical studies of K. brevis have focused primarily on absorption due to cellular pigmentation (Millie et al., 1995; Millie et al., 1997; Lohrenz et al., 1999; Kirkpatrick et al., 2000). As a result, the effects of backscattering have largely been ignored. This is probably true because backscattering spectral dependency is much 7

PAGE 18

weaker than absorption. Also, commercially available instrumentation for measuring backscattering has only recently become available (Maffione and Dana, 1997). Modeling efforts (Carder and Steward, 1985; Roesler and McLeroy-Etheridge, 1998; Mahoney, 2003) and recent field measurements (Kerfoot et al., 2002; Cannizzaro et al., 2004), though, indicate that K. brevis blooms exhibit relatively low backscattering. While the influence of backscattering variability on Rrs() was previously examined to determine how diel vertical migration of K. brevis populations modifies the optical properties of oceanic surface waters (Kerfoot et al., 2002; Mahoney, 2003), classification strategies for identifying K. brevis blooms were never explored. Therefore, prior to this study, satellite-based ocean color data have been used to quantify biomass, primary productivity, and areal extent for confirmed K. brevis blooms only (Vargo et al., 1987; Tester and Steidinger, 1997; Tester et al., 1998; Walsh et al., 2002). Blooms verified by microscopic analysis were identified in Coastal Zone Color Scanner (CZCS) imagery during most of these studies, typically based on increases in chlorophyll concentrations above background levels. 1.3 Objectives and Approach The objective of this study is to develop techniques to classify and quantify K. brevis blooms based on shipboard data that can be applied to satellite-based ocean color data. A large, multi-year/multi-season, comprehensive database of Chl, absorption, backscattering, and remote-sensing reflectance data collected on the WFS as part of the Ecology and Oceanography of Harmful Algal Blooms (ECOHAB) and Hyperspectral Coastal Ocean Dynamics Experiment (HyCODE) programs is used for this purpose. 8

PAGE 19

Absorption and backscattering spectra are decomposed into the sum of their optically significant particulate, dissolved, and molecular constituents in order to examine how changes in these parameters influence Rrs(), and hence ocean color. Remote-sensing reflectance spectra are modeled for waters containing greater than and less than 104 cells l-1 of K. brevis to determine which optical parameter(s) is (are) responsible for deviations observed in shipboard Rrs() measurements. A novel technique for classifying K. brevis blooms from Rrs() is introduced based on the results of these model simulations along with a method for quantifying Chl in extreme blooms (>100 mg m-3). The techniques are applied to Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and Moderate Resolution Imaging Spectroradiometer (MODIS) data collected in late August 2001, and validated using in situ data. 9

PAGE 20

2. METHODS 2.1 Shipboard Data Field data were collected on the WFS between March 1999 and October 2001 during eighteen monthly survey cruises as part of the ECOHAB program and one HyCODE cruise (Fig. 1, Table 1). Vertical profiles of chlorophyll fluorescence were measured at each station and the data were kindly provided by Dr. Gabriel Vargo (USF). Figure 1. Map of the west Florida shelf study area. 10

PAGE 21

Table 2. Cruise data summary. Cruise Dates Surface Discretea Surface Underwayb Ship EH0399 03/01/99-03-04/99 + R/V Suncoaster EH0799 07/05/99-07/08/99 + R/V Suncoaster EH0999 09/07/99-09/10/99 + R/V Suncoaster EH1199 11/06/99-11/08/99 + R/V Suncoaster EH0100 01/11/00-01/14/00 + R/V Suncoaster EH0300 03/01/00-03/04/00 + R/V Suncoaster EH0800 08/02/00-08/05/00 + + R/V Suncoaster EH0900 09/06/00-09/08/00 + R/V Suncoaster EH1000 10/04/00-10/06/00 + + R/V Suncoaster EH1100 11/07/00-11/10/00 + + R/V Suncoaster EB0201 02/06/01-02/07/01 + + R/V Bellows EH0401 04/03/01-04/06/01 + + R/V Suncoaster EH0601 06/05/01-06/08/01 + + R/V Suncoaster EH0701 06/30/01-07/03/01 + + R/V Suncoaster EH0801 08/01/01-08/01/01 + + R/V Suncoaster EH0901 08/28/01-08/31/01 + + R/V Suncoaster EB0901 08/29/01-08/30/01 + + R/V Bellows EB1001 10/01/01-10/02/01 + R/V Bellows HY1001 10/04/01-10/04/01 + R/V Subchaser a Discrete measurements include K. brevis cell concentrations, HPLC, Chl, ap(), ad(), ag(), and Rrs(). b Underway measurements include salinity, bb(), and c(532). 2.1.1. Discrete surface measurements Discrete surface samples were collected with an 8-liter Niskin bottle. K. brevis enumeration data were provided by Dr. Gabriel Vargo (USF). Cell count samples were collected by siphoning 15 ml of water into clean scintillation vials and were counted within 30 minutes of collection. Vials were gently inverted, and then five replicates of 100-200 l of sample were pipetted directly into glass well slides. Cells were counted live using a dissecting microscope at 40x magnification with K. brevis cells identified by morphological features and characteristic swimming motion. Average cell counts were calculated from the replicate samples. 11

PAGE 22

Chorophyll (a, b and c) and carotenoid pigment concentrations were measured using high-performance liquid chromatography (HPLC) (Wright et al., 1991) by the research group of Dr. Gary Kirkpatrick (Mote Marine Laboratory, Sarasota, FL). Pigments were separated with a C-18 Hypersil reverse-phase column and identified using a photodiode array UV-VIS detector (Shimadzu). Pigment absorbance spectra from standard microalgae cultures were used to identify and quantify the pigments. Absorption spectra due to all particles (phytoplankton and detritus), ap(), and detritus, ad(), were determined using the quantitative filter technique (Yentsch, 1962; Kiefer and SooHoo, 1982). Seawater was filtered slowly through 2.5cm GF/F (Whatmans) filters immediately following collection using vacuum pressure less than 15 in Hg. The volume of water filtered varied between ~0.05 and 6.0 liters depending on the concentration of pigmented particles in the sample. Filters were placed in tissue-tek holders that were wrapped in aluminum foil and then stored in liquid nitrogen for less than one week prior to being processed. Filters were allowed to thaw slowly at room temperature for 5-10 minutes prior to being placed in a dark petri dish and re-hydrated with a drop of Milli-Q water. The sample filter and a reference filter wetted with Milli Q water were placed on individual glass plates (diameter=2.4cm) in a custom-made transmissometer box. Prior to each transmission scan, the filters were slid one at a time over a tungsten-halogen light source that shone through a blue long-pass filter and a quartz glass diffuser. The transmittance of the sample filter, Tsample(), and the reference filter, Treference(), were measured three times each using a custom made, 512-channel spectroradiometer (~350-850nm). Optical densities, OD(), were calculated as 12

PAGE 23

TTlogODsamplereference10 (1) Particulate absorption spectra were calculated as lOD2.3pa (2) where 2.3 is a factor for converting from log10 to loge, l is the geometric pathlength equivalent to the volume of seawater filtered divided by the clearance area of the filter, and is the pathlength amplification factor or beta factor (Butler, 1962). The beta factor is an empirical formulation defined as the ratio of optical to geometric pathlength that corrects for multiple scattering inside the filter. In this study, an average of two published beta factor formulations (Bricaud and Stramski, 1990; Nelson and Robertson, 1993) 0.5OD0.61.0 (3) was chosen to correct for pathlength amplification. Spectra with OD(675) less than 0.04 (~30% of samples) were omitted from this study in order to minimize artifacts due to uncertainty in the factor (Mitchell and Kiefer, 1988; Bricaud and Stramski, 1990; Cleveland and Weidemann, 1993; Nelson and Robertson, 1993; Moore et al., 1995; Lohrenz, 2000). Absorption at 750nm was subtracted from the entire specta to correct for either residual scattering caused by non-uniformity in wetness between the sample and reference filters or stray light. Phytoplankton pigments were extracted from the sample filter with ~40-60ml of hot 100% methanol for 10-15 minutes in the dark (Kishino et al., 1985; Roesler et al., 1989). Fluorometric chlorophyll and pheaopigment concentrations were determined on 13

PAGE 24

the filtrate using a Turner 10-AU-005 fluorometer according to the methods of Holm-Hansen et al. (1965). Following extraction, the sample filter was rinsed with a few drops of Milli Q water to remove excess methanol and to rehydrate the filter. Transmittance spectra were measured again on this filter and the reference filter. Absorption spectra due to detrital particles (and non-methanol extractable pigments (e.g. phycobiliproteins)) were then calculated using Eqs. (1-3). Lastly, spectral absorption due to phytoplankton pigments, aph(), were calculated as follows: )(a)(a)(adpph (4) Gelbstoff absorption spectra, ag(), were measured on filtered seawater samples obtained using pre-rinsed 0.2m nylon membrane filters. Gelbstoff is the dissolved pool of colored compounds also commonly referred to as colored dissolved organic matter (CDOM). Filtered seawater samples were stored in clean 4 oz amber bottles at -30oC for less than three weeks prior to being processed. In order to prevent particle formation due to flocculation, samples were thawed slowly (~24 hours) at 0oC. Samples were then refiltered and scanned in 10cm quartz cells from 200-800nm using a Perkin-Elmer Lambda 18 spectrophotometer and referenced to Milli Q water. Gelbstoff absorbance spectra, Ag(), were converted to gelbstoff absorption spectra as follows: lgA2.3ga (5) after Eq. (2). Assuming gelbstoff does not absorb red light significantly, ag(600) values were subtracted from the entire spectra for weakly absorbing samples (Ag(400) < 0.0015) to remove any residual instrument offset and to prevent negative absorption in regions of 14

PAGE 25

low signal-to-noise. Gelbstoff absorption values at 680nm were subtracted from ag() for spectra with Ag(400) > 0.0015, since absorption at 600nm was not negligible. A temperature/salinity artifact centered at 732nm (Trabjerg and Hojerslev, 1996) prevented 750nm from being used to bias the curves as was used for the particulate and detrital absorption spectra. 2.1.2 Underway surface measurements Underway measurements of salinity, backscattering, and beam attenuation were measured with a CTD (Falmouth-Scientific Instruments), Hydroscat-2 (HOBI Labs), and C-Star transmissometer (WET Labs), respectively. The instruments were mounted on a metal frame that was placed in a closed, black-walled chamber (0.5 m3). Seawater was pumped by the ship flow-through system from a depth of ~2m through the chamber, which provided a ~5 minute residence time for a typical sample. Hydroscat-2 measurement, calibration, and data processing information are described in Maffione and Dana (1997). Backscattering coefficients at wavelengths other than those measured (488 and 676nm) were calculated by fitting a spectral power function to the measured wavebands and then interpolating to the desired wavelength. Particulate backscattering coefficients, bbp(), were derived from total backscattering by subtracting backscattering due to pure water (Morel, 1974). Particulate scattering at 532nm, bp(532), was obtained by subtracting ap(532) and ag(532) from the beam attenuation coefficient at 532nm, c(532), once the attenuation of pure water was removed. 15

PAGE 26

2.1.3 Remote-sensing reflectance measurements Measurements for calculating remote-sensing reflectance, Rrs(), were collected during daylight hours only when solar zenith angles were less than ~60o (Lee et al., 1996). All measurements were made from the bow of the ship to avoid ship shadow and wake bubbles. A custom made, hand-held 512-channel spectroradiometer (~350-850 nm) equipped with a 10o field-of-view was used to measure upwelling radiance, Lu(), at 30o from nadir and downwelling sky radiance, Lsky(), at 30o from zenith, both in the same plane (azimuth = 90o). The water-leaving radiance, Lw(), was then calculated by subtracting from Lu() that portion of the skylight and solar glint reflected into the sensor as follows: dskyuwELirLL (6) where r(i) is the Fresnel reflectance for zenith angle, i, and Ed() is a solar glint correction based on the assumption that Lw(750) is zero. Most glint is avoided by selecting glint-free areas and angles. The total downwelling irradiance, Ed(), was determined from spectral measurements of the light reflected from a standard diffuse reflector (Spectralon) or gray card, LG(), as follows: GGdRLE (7) where RG() is the reflectance of the diffuse reflector (~10%). Remote-sensing reflectance spectra were then calculated as dwrsELR (8) 16

PAGE 27

2.2 Satellite Data SeaWiFS data were processed using SeaWiFS Data Analysis System (SeaDAS) software (version 4.3). The Carder et al. (1999) semi-analytical algorithm with the global parameters table was applied to SeaWiFS data using Interactive Data Language (IDL) software (version 5.3). SeaWiFS data were classified using Environment for Visualizing Images (ENVI) software (version 3.4). MODIS Terra Collection 4 data were obtained from the NASA Goddard Distributed Active Archive Center (DAAC). 17

PAGE 28

3. THEORY Remote-sensing reflectance is defined as the ratio of water-leaving radiance to downwelling irradiance, measured just above the sea surface (Eq. (8)). For optically deep, vertically homogeneous waters, Rrs() is dependent on the absorption and backscattering properties of seawater and the angular distribution of light within the ocean. Using radiative transfer theory (Gordon et al., 1988; Mobley, 1994), Rrs() can be expressed as )(b)(a)(b)(Qfnt)(Rbb22rs (9) where t is the transmittance across the air-sea interface, n is the index of refraction of seawater, f is an empirical factor that is a function of the solar zenith angle, and Q() is the upwelling irradiance-to-radiance ratio. By making approximations for these latter terms (Lee et al., 1998), Rrs() can be related to the subsurface remote-sensing reflectance, rrs(), as follows: ))(r5.11()(r5.0)(Rrsrsrs (10) Based on model simulations (Lee et al., 1999), rrs() for optically deep waters is u)u170.0084.0()(rrs (11) with )(b)(a)(bubb (12) 18

PAGE 29

The absorption coefficient can be examined more thoroughly by decomposing it into the sum of its components: ) (a) (a) (a) (a) (ag d ph w (13) where the subscripts w, ph, d, and g refer to water, phytoplankton, detritus, and gelbstoff, respectively (e.g. Fig. 2). Similarly, the backscattering coefficient can be expanded as ) (b) (b) (bbp bw b (14) where the subscripts w and p refer to water and particles (phytoplankton and detritus), respectively (e.g. Fig. 3). Figure 2. Examples of phytoplankton, detrital and gelbstoff absorption spectra and the absorption spectra due to pure water (Pope and Fry, 1997). 19

PAGE 30

Figure 3. Example of particulate backscattering spectra and the backscattering spectra due to pure water (Morel, 1974). Since aw( ) (Pope and Fry, 1997)and bbw( ) (Morel, 1974) are constant and known, Rrs( ) can be calculated using a semi-analytical remote-sensing reflectance model formulated by combining Eqs. (10-14), provided that aph( ), ad( ), ag( ), and bbp( ) are known. In the following section, relationships are generated between Chl and aph( ), ad( ), ag( ), and bbp( ) so that a library of Rrs( ) curves for various chlorophyll concentrations can be modeled for waters containing fewer than 104 cells l-1 of K. brevis. Similar relationships derived for waters containing greater than 104 cells l-1 of K. brevis were then substituted into the model one at a time. Deviations in Rrs( ) as a function of chlorophyll concentrations are examined in order to understand which parameter(s) 20

PAGE 31

is(are) responsible for the differences in shipboard Rrs() observed between K. brevis bloom and non-bloom waters. Although modeling studies indicate that K. brevis diel vertical migration can result in minor changes in the magnitude and shape of Rrs() spectra (Kerfoot et al., 2002; Mahoney, 2003), a homogenous water column was assumed in this model since vertical profiles of chlorophyll fluorescence indicate that the water column was generally well mixed, except during an extreme bloom event (Chl = 130 mg m-3). 21

PAGE 32

4. RESULTS 4.1 Shipboard Remote-sensing Reflectance Data Median remote-sensing reflectance spectra measured between 1999 and 2001 in waters containing fewer than 104 cells l-1 of K. brevis (Fig. 4a) are indicative of typical non-K. brevis bloom conditions on the WFS. Oligotrophic, offshore waters with chlorophyll concentrations less than 0.2 mg m-3 visually appear blue since blue light (~400-450nm) is reflected strongly. Recalling that Rrs() is proportional to bb() and inversely proportional to a() (Eq. (9)), blue reflectance values are relatively high in these waters because absorption dominates backscattering in this spectral region (Figs. 2,3). Also, the absolute magnitude of absorption is relatively low, leaving more blue-rich backscattering due to water molecules (Fig. 3) and small particles. Transitioning from offshore to coastal waters, increases in absorption due to increased concentrations of chlorophyll and non-biogenous material (e.g. detritus and gelbstoff) cause blue reflectance values to decrease (Fig. 4a). Meanwhile, reflectance values at green wavelengths (~500-600nm) increase making waters visually appear green. This occurs because the relative contribution of particulate backscattering to Rrs() increases since contributions by absorption in this spectral region are minimal (Fig. 2). Peak reflectance values, in turn, shift from 400nm to ~570nm with increasing particle concentrations while transitioning from low-chlorophyll, offshore waters to high-chlorophyll, coastal waters. 22

PAGE 33

Figure 4. Median remote-sensing reflectance spectra measured on the WFS between 1999 and 2001 for K. brevis cell concentrations a) less than 104 cells l-1 and b) greater than 104 cells l-1, for various concentrations of chlorophyll (mg m-3). 23

PAGE 34

For high-chlorophyll (1-10 mg m-3) waters containing greater than 104 cells l-1 of K. brevis, remote-sensing reflectance values are ~3 to 4 times lower compared to non-K. brevis bloom waters for wavelengths less than 600nm (Fig. 4b). This decrease in Rrs() observed in K. brevis blooms makes the water appear darker since the green reflectance peak at ~570nm is less prominent. Indeed, when K. brevis cell concentrations exceeded 104 cells l-1, instead of appearing green, the waters visually appeared olive green. Increases in Chl beyond 10 mg m-3 are accompanied by further decreases in Rrs() (Fig. 4b), especially below 540nm, causing the water to appear dark brown or black in color. Although a red reflectance peak (~685 to 700nm) due to chlorophyll a fluorescence becomes increasingly significant with increasing Chl (Fig. 4b), K. brevis blooms do not appear as red in color visually as they do radiometrically because the color receptors of the human eye are only slightly sensitive to this portion of the visible spectrum (Commission Internationale de LEclairage). 4.2 Reflectance Model In order to understand how deviations in absorption and backscattering between K. brevis bloom and non-bloom waters influence Rrs() and in particular what causes reflectance values in K. brevis blooms to be ~3 to 4 times lower than in non-K. brevis bloom waters, a sensitivity analysis was performed using modeled data. Remote-sensing reflectance spectra were modeled using a semi-analytical model developed by combining Eqs. (10-14) and parameterized using the relationships developed as follows: 24

PAGE 35

4.2.1 Model parameterization 4.2.1.1 Phytoplankton absorption spectra Phytoplankton absorption spectra are composed of overlapping absorption spectra due to individual pigments belonging to three main groups: chlorophylls, carotenoids, and phycobiliproteins (Bidigare et al., 1989; Hoepffner and Sathyendranath, 1991). Two major absorption peaks at ~440nm and 675nm, mainly due to chlorophyll a, are typical of most aph() curves (Fig. 5). Accessory pigments absorb light between ~450 and 550nm, producing absorption shoulders and occasional peaks (Fig. 5 inset). Along with pigment variability, aph() are also influenced by pigment packaging (Bricaud et al., 1983; Nelson et al., 1993; Allali et al., 1997; Stuart et al., 1998). Essentially, increased cell size and intracellular pigment concentration causes self-shading, which leads to decreased absorption per unit chlorophyll, manifested as a flattening of absorption peaks (Morel and Bricaud, 1981). The relationship between Chl and aph() can be expressed by a power function (e.g. Prieur and Sathyendranath, 1981) )(BphChl)(A)(a (15) where A() and B() are constants derived empirically from log-transformed data. Surface chlorophyll concentrations observed on the WFS between 1999 and 2001 span more than two orders of magnitude, ranging from ~0.1 to 20 mg m-3 (Fig. 6). Phytoplankton absorption coefficients at 443nm, aph(443), range from ~0.01 to 0.70 m-1 25

PAGE 36

Figure 5. Phytoplankton absorption spectra, aph( ), measured on the WFS between 1999 and 2001. Figure 6. Relationship between phytoplankton absorption at 443nm, aph(443), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1). 26

PAGE 37

and are highly correlated with Chl (r2=0.98, n=290). Compared to a function derived by Bricaud et al. (1998) for a global dataset, the WFS exhibits slightly higher phytoplankton absorption at 443nm per unit Chl. Increased concentrations of photoprotective carotenoids relative to Chl and/or decreased pigment packaging on the WFS may explain these differences. Both conditions are typical of high-light, subtropical oceanic waters (e.g. the WFS) compared to the global dataset examined by Bricaud et al. (1998) (Stuart et al., 1998). Partitioning the dataset based on K. brevis cell concentration, the following relationships were determined on log-transformed data 7550 phChl057.0)443(a for <104 cells l-1 (r2=0.95, n=260) (16a) 7720 phChl051.0)443(a for >104 cells l-1 (r2=0.93, n=30) (16b) Using Students t-test, the differen ce between the slope coefficients, B( ), derived for K. brevis populations less than and greater than 104 cells l-1 was insignificant ( =0.05) at 443nm. Values for A( ) and B( ) were determined at all wavelengths between 400 and 700nm (Fig. 7) to parameterize the reflectance model. K. brevis bloom and non-K. brevis bloom waters exhibit similar spectral shape coefficients, A( ), for wavelengths greater than ~500nm (Fig. 7a). A(443) is ~11% lower, however, in K. brevis blooms. Since K. brevis cells are relatively large (20-40 m) and often low-light adapted (Shanley and Vargo, 1993), decreased phytoplankton absorption per unit Chl may be caused by increased pigment packaging and/or decreased concentrations of photoprotective carotenoids relative to Chl, respectively. 27

PAGE 38

28 Figure 7. Spectral values of a) A( ), b) B( ) and c) the coefficient of determination, r2, obtained by fitting a power law function to log-transformed aph( ) versus chlorophyll a concentration data for K. brevis concentrations less than (solid) and greater than (dashed) 104 cells l-1.

PAGE 39

Slope coefficients derived for waters containing less than and greater than 104 cells l-1 of K. brevis are similar for wavelengths less than 500nm and greater than 590nm (Fig. 7b). For wavelengths between 500 and 590nm, K. brevis blooms exhibit lower slopes than non-K. brevis bloom waters. In fact, B() values derived for wavelengths between 520 and 575nm were significantly different (=0.05), perhaps due to increased concentrations of phycobiliproteins and/or decreased pigment packaging in non-K. brevis bloom waters. Slopes derived for wavelengths less than 520nm and greater than 575nm were not significantly different (=0.05). Correlation coefficients for wavelengths less than 500nm are slightly lower in waters containing greater than 104 cells l-1 of K. brevis (r2=~0.93, n=30) compared to waters containing fewer than 104 cells l-1 of K. brevis (r2=~0.95, n=260) (Fig. 7c). Coefficients of correlation are maximal for both groups at 675nm because non-chlorophyll a pigments absorb light weakly at this wavelength. Since the total concentration of accessory pigments typically co-varies strongly with Chl in oceanic waters (Trees et al., 2000), the relationships between Chl and aph() for wavelengths between 490 and 550nm also exhibit high correlation coefficients (r2>0.93) even though chlorophyll a does not absorb light in this region. 4.2.1.2 Detrital absorption spectra Detrital absorption spectra exhibit exponentially decreasing absorption with increasing wavelength (Fig. 8) (Yentsch, 1962; Roesler et al., 1989) and can be modeled as 29

PAGE 40

)) (Sexp()(a) (a0 d 0d d (17) where 0 is a reference wavelength (443nm is used here) and Sd is the spectral slope for detrital absorption. For waters containing relatively high concentrations of phycobiliproteins, residual absorption peaks are present between ~450-550nm (Fig. 8 inset) since these pigments are non-methanol soluble. Modeled detrital absorption spectra for these stations were generated using Eq. (17) to eliminate these peaks and used instead of the measured spectra in Eq. (4) to derive aph( ). Detrital absorption at 443nm, ad(443), ranges from ~0.001 to 0.1 m-1 for surface stations measured on the WFS between 1999 and 2001 (Fig. 9) and constitutes a small and variable fraction (~5 55%) of the total particulate absorption at 443nm (Fig. 10). This latter range is consistent with values reported for tropical to subpolar waters (Cleveland, 1995; Bricaud et al., 1998). Trends between Chl and ad(443)/ap(443), however, were not observed. Since the sources and sinks of this non-living particulate pool of colored compounds (i.e. fragments of microorganisms, fecal pellets, terrigenous material, etc.) are often independent of phytoplankton in coastal regions such as the WFS, a weaker positive correlation exists between Chl and ad(443) (r2=0.88, n=290) (Fig. 9) compared to between Chl and aph(443) (Fig. 6). Partitioning the data set based on K. brevis concentration, the following relationships were developed from log-transformed data: 0301 dChl014.0)443(a for <104 cells l-1 (r2=0.78, n=260) (18a) 5810 dChl014.0)443(a for >104 cells l-1 (r2=0.36, n=30) (18b) 30

PAGE 41

Figure 8. Detrital absorption spectra, ad( ), measured on the WFS between 1999 and 2001. Figure 9. Relationship between detrital absorption at 443nm, ad(443), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1). 31

PAGE 42

Figure 10. Contribution of detrital absorption to total particulate absorption at 443nm as a function of chlorophyll a concentration for various concentrations of K. brevis (cells l-1). 32 Figure 11. Relationship between the spectral slope for detrital absorption, Sd, and chlorophyll a concentration for various concentrations of K. brevis (cells l-1).

PAGE 43

Waters containing fewer than 104 cells l-1 of K. brevis exhibit a steeper slope compared to bloom waters with greater than 104 cells l-1. Differences between these slopes are significantly different (=0.05), indicating that K. brevis blooms exhibit a paucity of detrital absorption on the WFS compared to similarly high-chlorophyll (> ~1 mg m-3), low K. brevis (<104 cells l-1) waters. Spectral slopes for detrital absorption generated between 400 and 650nm vary from ~0.009 to 0.015 nm (Fig. 11) and are consistent with values reported for other oceanic regions (Roesler et al., 1989; Bricaud et al., 1998). Average Sd values observed on the WFS between 1999 and 2001 for three Chl ranges (0.1-1, 1-10, and 10-100 mg m-3) are 0.011, 0.011, and 0.012 nm-1, respectively. Since detrital spectral slopes do not co-vary with Chl (r2=0.00, n=290), the mean detrital spectral slope, 0.012 nm-1, is used together with the relationships in Eq. (18) to generate ad() (Eq. (17)) for the reflectance model. 4.2.1.3 Gelbstoff absorption spectra Gelbstoff absorption spectra exhibit spectral features similar to the detrital pool of colored compounds, absorbing blue light strongly and showing decreased absorption with increased wavelength (Fig. 12). Thus, ag() can be modeled as ))(Sexp()(a)(a0g0gg (19) where Sg is the spectral slope for gelbstoff absorption, and the reference wavelength, 0, used for this parameter is 400nm. Since gelbstoff absorption is roughly two times greater 33

PAGE 44

Figure 12. Gelbstoff absorption spectra, ag( ), measured on the WFS between 1999 and 2001. Figure 13. Relationship between gelbstoff absorption at 400nm, ag(400), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1). 34

PAGE 45

at 400nm than at 443nm, a reference wavelength of 400nm offers greater signal-to-noise when a 10cm spectrophotometric cell is used for measurement. Gelbstoff absorption at 400nm, ag(400), ranges from ~0.01 to 3 m-1 for surface stations measured on the WFS from 1999 to 2001 (Fig. 13). This range is approximately an order of magnitude higher than ad(400) (Fig. 8). ag(400) is positively correlated with Chl (r2=0.92, n=385) (Fig. 13). Partitioning the data set based on K. brevis cell concentration, the following relationships were developed from log-transformed data: 8140 gChl142.0)400(a for <104 cells l-1 (r2=0.79, n=355) (20a) 8180 gChl156.0)400(a for >104 cells l-1 (r2=0.62, n=30) (20b) The difference between the slopes for these two groups is insignificant ( =0.05), indicating that gelbstoff absorption, unlike detrital absorption, is not unique in K. brevis blooms. Spectral slopes for gelbstoff absorption calculated between 350 and 450nm range from ~0.014 to 0.025 nm-1 (Fig. 14) and are generally higher than the slopes measured for detrital absorption. Data for wavelengths greater than 450nm were not considered due to signal-to-noise problems. The variability of Sg is high for Chl less than 1 mg m-3 with an average slope of 0.020 0.002 nm-1. For higher Chl waters (>1 mg m-3), the mean spectral slope (0.019 0.001 nm-1) is slightly lower and shows decreased variability. Similar Sg ranges and variability have been reported for other oceanic waters (Table 3) (Carder et al., 1989; Blough et al., 1993; Nelson and Guarda, 1995; Vodacek and Blough, 1997). Since Sg is only weakly correlated with Chl (r2=0.09, n=385), the mean gelbstoff 35

PAGE 46

spectral slope, 0.019 nm-1, is used together with the relationships in Eq. (20) to calculate ag( ) (Eq. (19)) for the reflectance model. Figure 14. Relationship between the spectral slope for gelbstoff absorption, Sg, and chlorophyll a concentration for various concentrations of K. brevis (cells l-1). Table 3. Gelbstoff absorption slope coefficients, Sg. Location Sg (nm-1) Reference Gulf of Mexico 0.011 0.017 Carder et al., 1989 Orinoco River plume 0.014 0.023 Blough et al., 1993 South Atlantic Bight 0.010 0.025 Nelson and Guarda, 1995 Mid-Atlantic Bight 0.010 0.033 Vodacek et al., 1997 West Florida Shelf 0.014 0.025 this study 4.2.1.4 Particulate backscattering spectra Particulate backscattering spectra can be fit to the form 36

PAGE 47

00bpbp)(b)(b (21) where 0 is the reference wavelength, and is the Angstrom exponent that describes the spectral shape of bbp(). A reference wavelength of 550nm was chosen to parameterize the reflectance model since changes in Rrs() at this wavelength are primarily due to backscattering. Absorption due to water is constant (Pope and Fry, 1997) and non-water absorption at this wavelength is relatively low. Particulate backscattering at 550nm, bbp(550), ranges from ~0.001 to 0.08 m-1 for surface stations measured on the WFS in 2000 and 2001 (Fig. 15). A positive correlation Figure 15. Relationship between particulate backscattering at 550nm, bbp(550), and chlorophyll a concentration for various concentrations of K. brevis (cells l-1). 37

PAGE 48

exists between Chl and bbp(550) (r2=0.85, n=194). For low chlorophyll concentrations (< ~0.5 mg m-3), bbp(550) is tightly coupled to Chl, while for higher chlorophyll concentrations increased variability is observed. Partitioning the data in Figure 15 based on K. brevis cell concentrations, the following relationships were generated from log-transformed data: 9770 bpChl0098.0)550(b for <104 cells l-1 (r2=0.87, n=180) (22a) 1800 bpChl0051.0)550(b for >104 cells l-1 (r2=0.08, n=14) (22b) Statistical analysis shows that the slope derived in K. brevis blooms is significantly lower ( =0.05) than the slope derived in non-K. brevis bloom waters. The expression obtained for bloom waters (Eq. 22b) is very similar to a relationship generated for modeled Case 1 data (Morel, 1988). Case 1 waters have detrita l and gelbstoff absorption coefficients that co-vary with phytoplankton, while such a correlation does not exist in Case 2 waters (Preisendorfer, 1961). Comparing the Chl and bbp(550) data collected on the WFS in 2000 and 2001 to this Case 1 expression (Morel, 1988), the WFS can be separated into three bio-optically unique provinces: 1. Data collected from oligotrophic waters typically west of the 30 m isobath (i.e. away from terrigenous influences) exhibit low chlorophyll concentrations (< 0.2mg m-3), low bbp(550) coefficients, and K. brevis concentrations less than 104 cells l-1. These data generally overlay the Case 1 line and are dominated by prochlorophytes and cyanophytes since they exhibit relatively high concentrations of the marker pigments divinylchlorophyll a and zeaxanthin, respectively (Fig. 16a). 38

PAGE 49

Figure 16. Relative composition of the main accessory pigments for a) low-chlorophyll (< 0.2 mg m-3), non-K. brevis bloom waters, b) high-chlorophyll (>1.0 mg m-3), non-K. brevis bloom waters, and c) high-chlorophyll (>1.0 mg m-3), K. brevis bloom waters. 2. Shallow, estuarine waters, containing fewer than 104 cells l-1 of K. brevis, typically located just outside of Tampa Bay and Charlotte Harbor, exhibit high chlorophyll concentrations (> ~1.0 mg m-3) and high bbp(550) coefficients. These data deviate significantly from the Case 1 relationship and are instead consistent with Case 2 39

PAGE 50

waters containing large concentrations of high-backscattering particles. Relatively high concentrations of the marker pigment fucoxanthin (Fig. 16b) along with light microscopic identification (pers. comm. C. Heil) indicate that diatoms typically dominated these regions. 3. Regions containing K. brevis cell concentrations greater than 104 cells l-1 also have high chlorophyll concentrations (> ~1.0 mg m-3), but they exhibit lower bbp(550) coefficients compared to diatom-dominated, estuarine waters. Data from these regions generally overlay the Case 1 line and contain relatively high concentrations of the K. brevis diagnostic marker pigment, gyroxanthin-diester (Fig. 16c) (Millie et al., 1995). Unlike the spectral slopes for detrital and gelbstoff absorption, the Angstrom exponent for particulate backscattering decreases with increasing Chl (r2=0.53, n=194) (Fig. 17). A negative correlation is observed between and Chl, with no significant differences observed between waters containing less than or greater than 104 cells l-1 of K. brevis. Thus, the following relationship, modified from Lee et al. (2000), )Chl1(9.11.0 (23) was developed for incorporation into Eq. (21) along with the relationships in Eq. (22) to derive bbp() from Chl for the reflectance model. Equation (23) ensures that approaches asymptotes for low and high chlorophyll concentrations that are consistent with modeled data (Gordon et al., 1988). 40

PAGE 51

Figure 17. Relationship between the particulate backscattering spectral shape parameter, and chlorophyll a concentration for various concentrations of K. brevis (cells l-1). 4.2.2 Model application Relationships derived between Chl and aph(), ad(), ag(), and bbp() for waters containing fewer than 104 cells l-1 of K. brevis were used to create a library of non-K. brevis Rrs() spectra for a wide range of chlorophyll concentrations (0.05-20 mg m-3). Relationships generated from waters containing greater than 104 cells l-1 of K. brevis were then substituted one at a time into the model to determine which parameter(s) is(are) responsible for the decreased Rrs() observed in blooms (Fig. 4b) with two exceptions: 41

PAGE 52

1. The influence of ag() on Rrs() was not examined since the difference in slopes between Chl and ag(400) for waters containing less than and greater than 104 cells l-1 of K. brevis was insignificant (Fig. 13). 2. The relationship generated between Chl and bbp(550) in bloom waters was replaced by the modeled Case 1 expression (Morel, 1988) to provide particulate backscattering coefficients similar to those measured for low chlorophyll concentrations (Fig. 15). Rrs() values generated by substituting into the model expressions representing K. brevis blooms between Chl and the three significantly different optical parameters (aph(), ad(), and bbp()) were then compared to non-K. brevis values by examining Rrs() as a function of Chl at three wavelengths (Fig. 18). Substituting the aph() parameters, A() and B(), derived for K. brevis bloom waters (Figs. 7a,b) into the model, reflectance values similar to non-K. brevis values were observed at 443 and 490nm (Figs. 18a,b). Bloom reflectance values at 555nm, though, were slightly higher than non-K. brevis values at higher chlorophyll concentrations (Fig. 18c). An increase in reflectance at 555nm occurs because the slope coefficient at this wavelength was significantly lower in bloom waters compared to non-K. brevis bloom waters (Fig. 7b). Increases in Rrs() were also observed when the relationship derived for K. brevis blooms between Chl and ad() (Eq. (18b)) was substituted into the reflectance model. Since ad() constitutes a very small portion of ap() (Fig. 10), only minor deviations in Rrs() were observed. 42

PAGE 53

Figure 18. Modeled remote-sensing reflectance coefficients at a) 443nm, b) 488nm, and c) 551nm calculated using a semi-analytical Rrs( ) model and plotted as a function of chlorophyll a concentration. 43

PAGE 54

Together, the results of these two model simulations indicate that the decreases observed in K. brevis bloom chlorophyll-specific phytoplankton absorption, a*ph() (= aph()/Chl), and chlorophyll-specific detrital absorption, a*d() (= ad()/Chl), would increase Rrs() only slightly in high-chlorophyll (> ~1 mg m-3) waters. This suggests that the 3to 4fold decrease in Rrs() observed in bloom waters (Fig. 4b) cannot be explained by deviations in absorption. Furthermore, since these deviations are relatively minor, they cannot be used as a basis for detecting K. brevis blooms from space using ocean color sensors. Substituting the Morel (1988) relationship between Chl and bbp(550) (Fig. 15) into the reflectance model to represent K. brevis blooms, decreases in Rrs() at 443, 488, and 551nm by a factor of 2 to 10 were observed relative to non-K. brevis bloom values for chlorophyll concentrations from 1 to 10 mg m-3, respectively (Fig. 18). Thus, the 3to 4-fold decrease in Rrs(550) observed in waters containing greater than 104 cells l-1 of K. brevis compared to those containing fewer than 104 cells l-1 of K. brevis can be attributed to decreases in particulate backscattering. 4.3 Classifying and Quantifying K. brevis Blooms from Space 4.3.1 Development and application of a classification technique Based on the results of the reflectance model simulations in the previous section, a technique is developed to classify waters containing greater than 104 cells l-1 of K. brevis using the differences observed in chlorophyll-specific particulate backscattering (Fig. 15). However, since Chl and bbp(550) cannot be measured directly by satellite-based ocean color sensors, a semi-analytical algorithm (Carder et al., 1999) developed for 44

PAGE 55

MODIS data is used to derive these values given Rrs() at 412, 443, 488, and 551nm. Assuming that reflectance values at SeaWiFS wavebands (412, 443, 490, and 555nm) are not significantly different from these MODIS wavebands, the classification technique developed here may be applied to both MODIS and SeaWiFS data. For Case 2 environments such as the WFS, chlorophyll concentrations determined semi-analytically have an advantage over empirical spectral ratio algorithms (Gordon et al., 1983; O'Reilly et al., 1998) since corrections for gelbstoff and detrital absorption that often do not co-vary with Chl are made. Gelbstoff and detritus can absorb enough light at blue wavelengths to increase empirically retrieved chlorophyll concentrations by as much as a factor of two (Hu, 2003). The semi-analytical algorithm separates the effects of phytoplankton absorption from gelbstoff and detrital absorption based on spectral differences at 412 and 443nm (Fig. 2). The root mean square error (rms) determined between measured and modeled log-transformed Chl data is 0.19 (r2=0.95, n=214) (Fig. 19). Particulate backscattering at 550nm can be calculated accurately from shipboard Rrs(551) using the expression bbp(550) = 2.058 Rrs(551) 0.000182 derived in Carder et al. (1999). The rms error determined between measured and modeled log-transformed bbp(550) data is 0.14 (r2=0.90, n=120) (Fig. 20). Calculating bbp(550) from Rrs(551), and plotting these values against the semi-analytically derived chlorophyll concentrations, Figure 21 shows how satellite-based ocean color data can be used to examine the Chl versus bbp(550) relationship. Note that the separation between clusters of points representing K. brevis blooms and non-bloom waters is even more distinct than was observed for the in situ measurements (Fig. 15). 45

PAGE 56

Figure 19. Measured versus modeled chlorophyll a concentrations derived from shipboard Rrs( ) data collected on the WFS between 1999 and 2001 for various concentrations of K. brevis (cells l-1). 46 Figure 20. Measured versus modeled bbp(550) derived from shipboard Rrs( ) data collected on the WFS between 2000 and 2001 for various concentrations of K. brevis (cells l-1).

PAGE 57

Figure 21. Relationship between modeled bbp(550) and modeled chlorophyll a concentration data derived from shipboard Rrs( ) data collected on the WFS between 1999 and 2001. Symbol size increases with increasing concentrations of K. brevis (cells l-1). Indeed, all observations with 1) chlorophyll concentrations greater than 1.5 mg m-3 and 2) bbp(550) values less than the modeled Case 1 relationship (Morel, 1988) contain greater than 104 cells l-1 of K. brevis. Only one observation with 14,000 cells l-1 of K. brevis is not discriminated by this approach, and no false positive values are observed. As a result, these two conditions shall serve as the initial criteria for classifying K. brevis blooms from space. From fall 2001 to spring 2002, the WFS experienced a major K. brevis bloom (Stumpf et al., 2003). The bloom first appeared in surface waters north of Charlotte 47

PAGE 58

Harbor in late August, concentrated just north and west of a low-salinity (~32.5 psu), estuarine plume dominated by diatoms (pers. comm.. C. Heil). Three weeks of weak but steady northerly winds may have upwelled the bloom from a deeper offshore initiation source. Surface cell counts as high as 7.5 x 106 cells l-1 with chlorophyll concentrations ~130 mg m-3 were observed during an ECOHAB cruise along with salinities greater than 36 psu and chlorophyll-specific particulate backscattering coefficients as low as 0.0006 m2 (mg Chl)-1. A SeaWiFS image acquired on 30 August 2001 shows a thin (~6 km) filament of high-chlorophyll water stretching ~65 km from just north of Charlotte Harbor offshore in a southerly direction (Fig. 22). Smaller regions of high chlorophyll concentrations (>10 mg m-3) are also observed outside of Tampa Bay. Deriving Chl and bbp(550) from SeaWiFS data using the Carder et al. (1999) semi-analytical algorithm and then classifying the data based on the criteria defined above, Figure 23 shows that regions with K. brevis concentrations greater than 104 cells l-1 are positively identified as K. brevis using this classification technique. Conversely, high-chlorophyll coastal waters and those observed outside of Tampa Bay that contained fewer than 103 cells l-1 of K. brevis are not flagged. These latter waters were instead dominated by diatoms (e.g. Rhizosolenia sp.) according to microscopic analyses (pers. comm.. C. Heil ). 4.3.2 Development and application of a quantification technique Since most remote-sensing algorithms for deriving Chl (Gordon et al., 1983; O'Reilly et al., 1998; Carder et al., 1999) rely on Rrs() values at or below 490nm, current 48

PAGE 59

Figure 22. SeaWiFS chlorophyll concentrations (mg m-3) derived using the Carder et al. (1999) semi-analytical algorithm for the WFS on 30 August 2001. Figure 23. SeaWiFS data (30 August 2001) classified (shaded area) for K. brevis blooms according to the criteria: Chl > 1.5 mg m-3 and bbp(550) < the Morel (1988) expression. Shipboard K. brevis cell concentrations (cells l-1) collected 28-31 August 2001, are superimposed on top of the classified image. 49

PAGE 60

space-based retrievals of accurate chlorophyll concentrations for extreme algal blooms (>100 mg m-3) may oftentimes be inaccurate due to signal-to-noise errors at these wavelengths (Fig. 4b). This problem is only exacerbated in K. brevis blooms as decreased particulate backscattering per unit Chl (Fig. 15) leads to even lower Rrs() values compared to non-bloom waters with similar chlorophyll concentrations (Figs. 4,18). Also, since the water-leaving radiance at 750 nm can no longer be considered negligible for use in atmospheric correction for extreme blooms (Siegel et al., 2000), low and even negative retrievals of Rrs() can occur at wavelengths below 490nm due to over-correction of aerosol radiance (Hu et al., 2000), leading to Chl retrieval inaccuracies. An alternative approach to standard algorithms is to quantify Chl in extreme blooms (>100 mg m-3) based on fluorescence since Rrs(685) increases with Chl due to chlorophyll fluorescence (Fig. 4). While it is well understood that fluorescence efficiency is quite variable (Kiefer, 1973), if a firm relationship can be developed between Chl and the height of the reflectance peak, then perhaps more accurate chlorophyll concentrations can be obtained using this method compared to standard approaches. Unlike SeaWiFS, MODIS contains wavebands designed to quantify this chlorophyll fluorescence peak. The MODIS fluorescence line height (FLH) data product is defined as )()())(nL)((nL)(nL)(nLFLH1314131515w13w13w14w (24) where nLw() is the water-leaving radiance normalized so that the sun is at zenith and the numerical subscripts refer to the MODIS band number (13=667nm, 14=678nm, and 50

PAGE 61

15=751nm) (Esaias et al., 1998). Essentially, FLH is the height of the fluorescence contribution at 678 nm above a point directly below on a baseline to be found if no fluorescence were present (e.g. Fig. 24). Note that the center fluorescence band could not be located at the nominal chlorophyll fluorescence peak wavelength at 685nm because of the atmospheric oxygen absorption line, and so a 678 nm waveband was used on MODIS as a replacement. Figure 24. Definition of fluorescence line height (FLH). Since shipboard measurements of nLw( ) were not collected, they were derived from measurements of Rrs( ) by multiplying by the extraterrestrial solar irradiance according to Gordon and Wang (1994). The relationship between Chl and FLH is shown in Figure 25. This diagram exhibits two separate trend lines: 1) a low slope for non-K. brevis bloom waters and 2) a higher slope for K. brevis blooms. 51

PAGE 62

The relationship generated for K. brevis blooms (Chl = 143.0*FLH1.66, r2=0.85, n=14) is quite linear in log-log coordinates for chlorophyll concentrations greater than ~1.5 mg m-3, even up to concentrations as high as 130 mg m-3. Thus, to the extent that K. brevis blooms can be identified using satellite-based ocean color data (e.g. Fig. 23), chlorophyll concentrations may be estimated from MODIS FLH data (Fig. 26). Figure 25. Relationship between measured chlorophyll a concentrations and fluorescence line height data (W m-2 m-1 sr-1) derived from shipboard Rrs( ) data collected on the WFS between 1999 and 2001 for various concentrations of K. brevis (cells l-1). 52

PAGE 63

Figure 26. Chlorophyll a concentrations (mg m-3) derived from MODIS fluorescence line height data of the WFS (30 August 2001) for regions positively flagged for K. brevis. 53

PAGE 64

5. DISCUSSION Shipboard remote-sensing reflectance values measured on the WFS between 1999 and 2001 were ~3 to 4 times lower on average within K. brevis blooms (>104 cells l-1) compared to in non-K. brevis bloom waters (<104 cells l-1). Model simulations indicate that this decrease in Rrs() cannot be attributed to deviations in absorption. These results refute the common perception that phytoplankton absorption (i.e. pigmentation) is responsible for the discoloration associated with K. brevis blooms. This is consistent with a previous study, which concluded that accurately identifying individual algal groups (i.e. diatoms, dinoflagellates, etc.) from space based on pigmentation is highly unlikely due to overlapping pigment absorption spectra and pigment packaging (Garver et al., 1994). Indeed, classifying K. brevis blooms based solely on differences in aph() would be difficult since the main accessory pigments belonging to K. brevis (fucoxanthin, 19-acylofucoxanthins, and chlorophyll c) co-occur in other algal groups (e.g. prymnesiophytes and chrysophytes). Furthermore, the only pigment unique to Karenia species in the Gulf of Mexico, gyroxanthin-diester, typically comprises less than 5% of the total pigment concentration and absorbs light in the same spectral region as the other carotenoids (Millie et al., 1995). Particulate backscattering coefficients measured during this study were significantly lower inside K. brevis blooms compared to in non-K. brevis bloom waters. 54

PAGE 65

For chlorophyll concentrations between ~1.5 and 20 mg m-3, bbp(550) values in K. brevis blooms were ~2.5 to 20 times lower than non-K. brevis bloom values, respectively. Model simulations indicate that this decrease in bbp() can explain the 3to 4fold decrease in Rrs() observed in K. brevis blooms. Thus, a technique was developed to classify K. brevis blooms from space based on the relationship between Chl and bbp(550). In addition, a second technique for quantifying extreme blooms with high chlorophyll concentrations was developed utilizing fluorescent line height data. The techniques developed in this study for classifying and quantifying K. brevis blooms from space have several advantages over methods that rely solely on the magnitude of Chl derived using standard empirical algorithms (Vargo et al., 1987; Tester and Steidinger, 1997; Tester et al., 1998; Walsh et al., 2002; Stumpf et al., 2003). The relationship between Chl and bbp(550) is used to discriminate high-chlorophyll waters containing greater than and less than 104 cells l-1 of K. brevis. Overestimations of Chl caused by high ratios of ag() to aph(), a common problem for standard empirical algorithms (Hu, 2003), are minimized by using a semi-analytical algorithm (Carder et al., 1999). Shallow water regions that exhibit reflectance signals contaminated by light reflected from the bottom are not mistakenly flagged as K. brevis blooms. Although elevated Rrs(551) values caused by bottom reflectance will lead to overestimations in Chl (Lee et al., 2001), particulate backscattering values derived from these same elevated Rrs(551) values will also be erroneously high, preventing these regions from mistakenly being flagged as K. brevis blooms. Lastly, signal-to-noise errors that adversely affect empirically-derived Chl estimates in extreme K. brevis blooms may be eliminated when MODIS FLH data are used to quantify Chl. 55

PAGE 66

Normally on the WFS, chlorophyll concentrations typically increase when moving from offshore to near shore waters due to increased nutrient availability from terrigenous sources (e.g. riverine input and bottom resuspension). Increased absorption due to gelbstoff and increased backscattering due to detrital particles typically accompany this increase in Chl (data not shown). Classification errors may occur if this trend is not observed. Model simulations indicate that false positive flagging of K. brevis blooms may occur if increases in Chl are accompanied by increases in gelbstoff absorption without similar increases in high-backscattering detritus. In this situation, decreased reflectance due to high gelbstoff absorption would not be offset by increased reflectance if backscattering were anomalously low. Subsequently, low Rrs(551) values would lead to erroneously low backscattering values, which would false positively be flagged for K. brevis. Although it is unknown how frequently this situation occurs, MODIS FLH data may possibly serve as a means to discriminate between positively flagged regions containing K. brevis from false positively flagged high gelbstoff absorption/low backscattering regions since gelbstoff does not fluoresce at 678nm. While false positive flagging may occur when ratios of ag() to bbp() are anomalously high, false negative flagging may occur due to subpixel variability. Since dense algal blooms exhibit spatial patchiness (Franks, 1997) and may not occupy entire SeaWiFS or MODIS pixels (~1 km2), extremely low reflectance values exhibited by spatially small K. brevis blooms may be overwhelmed by higher reflectance values from surrounding background populations of phytoplankton. Similar misclassifications may 56

PAGE 67

occur if K. brevis blooms spatially co-exist with high-backscattering Trichodesmium blooms (Subramaniam et al., 1999) as was suggested by Walsh and Steidinger (2001). In order to understand why backscattering per unit chlorophyll is relatively low in K. brevis blooms, the major factors responsible for controlling backscattering in oceanic waters, particle size distributions and particle composition, are considered. While neither of these factors was measured directly during this study, assumptions are made based on measurements of the backscattering ratio. The backscattering ratio, defined as the ratio of bbp() to bp(), is the proportion of light scattered by particles in the backward direction, where bp() is the scattering coefficient due to particles in all directions. Theoretical models indicate that backscattering ratios decrease in response to 1) decreases in the fraction of submicron particles (Ulloa et al., 1994) and 2) decreases in the refractive index of particles (Twardowski, 2001). Indices of refraction (discussed here relative to seawater) reflect particle composition with low values (1.02-1.07) typically associated with living cells and high values (1.14-1.26) typically associated with mineral particles (Twardowski, 2001). Since in situ backscattering ratios measured on the WFS in 2000 and 2001 (~0.6-2.3%; data not shown) are significantly greater than ratios measured for pure phytoplankton (typically < 0.1%) (Bricaud et al., 1983; Ahn et al., 1992), only a small fraction of particulate backscattering can be attributed to living algal cells. Indeed, K. brevis is an ineffective backscatterer since it is large (20-40 m) and exhibits a relatively low index of refraction (~1.05) (Mahoney, 2003). Instead, the primary source of particulate backscattering in oceanic waters is particles less than 1m (Morel and Ahn, 1991; Stramski and Kiefer, 1991). Thus, natural populations of phytoplankton on the 57

PAGE 68

WFS must contain significant concentrations of submicron detrital particles in order to account for the high backscattering ratios observed. More importantly, differences in the amount and composition of this submicron pool of detritus and not the phytoplankton themselves must be responsible for the relatively low backscattering coefficients observed in K. brevis blooms. For high-chlorophyll waters (> 1.5 mg m-3), K. brevis blooms exhibited lower in situ backscattering ratios (< ~1%) compared to high-chlorophyll, non-K. brevis bloom waters (> ~1%). This suggests that K. brevis blooms are associated with either a paucity of submicron particles and/or particle assemblages with lower bulk indices of refraction compared to non-bloom waters (Ulloa et al., 1994; Twardowski, 2001). Similar backscattering ratios were observed during a diel vertical migration experiment on the WFS in October 2001(Kerfoot et al., 2002). The following data collected indicate that both a paucity of detritus and a decreased bulk index of refraction were most likely associated with K. brevis blooms observed during this study. Reduced grazing pressure due to cellular toxicity along with a general inability of K. brevis to out compete diatoms and bloom in high-nutrient, high-backscattering estuarine waters are proposed mechanisms that may be responsible for the observed differences in backscattering ratios. Together they may explain the low backscattering per unit chlorophyll values observed in K. brevis blooms. Shipboard data demonstrate that chlorophyll-specific detrital absorption coefficients were significantly lower in K. brevis blooms compared to high-chlorophyll, non-K. brevis bloom waters (Fig. 9). Assuming that a positive correlation exists between detrital absorption and detrital concentration, K. brevis blooms exhibit relatively low 58

PAGE 69

detrital concentrations. Reduced grazing pressure may be responsible for this paucity of detritus. Lower pheaopigment-to-chlorophyll ratios measured in K. brevis blooms (0.10 0.04, n=26) compared to in high-chlorophyll, non-bloom waters (0.23 0.10, n=24) indicate that zooplankton grazing may have been lower inside blooms (Shuman and Lorenzen, 1975). While relationships between zooplankton and K. brevis are complex (Lester et al., 2003), previous studies have demonstrated that grazing pressure declines in both natural populations and cultures of K. brevis when blooms are monospecific possibly due to cellular toxicity (reviewed by Turner and Tester, 1997). Reduced concentrations of detrital substrates due to decreased grazing would cause microbial populations to decline, leading to further reductions in detritus due to decreased microbial decomposition. Not only may K. brevis blooms exhibit a paucity of detritus due to reduced grazing pressure as a result of cellular toxicity, they may also exhibit reduced detrital concentrations and contain a smaller proportion of particles with relatively high indices of refraction because they do not occur in estuarine waters. Estuarine waters typically contain high concentrations of terrigenous material with high indices of refraction (i.e. minerals), causing them to exhibit high backscatter coefficients. Indeed, high salinity values (>~35 psu) observed in K. brevis blooms indicate that blooms were not directly associated with low-salinity, estuarine waters (<~35 psu) from Tampa Bay and Charlotte Harbor. Instead, blooms were often encountered adjacent to freshwater plumes (Vargo et al., 2000). In general, phytoplankton groups with high growth rates and high nutrient uptake efficiencies (i.e. diatoms) tend to dominate algal populations in high-nutrient estuarine 59

PAGE 70

waters (Smayda, 1997). Since K. brevis cells grow relatively slowly, doubling only once every 3-5 days (Steidinger et al., 1998), they cannot out-compete these faster growing phytoplankton groups. Instead, K. brevis has evolved numerous physiological and behavioral strategies (i.e. cellular toxicity, diel vertical migration, mixotrophy, photoadaptation, etc. (Steidinger et al., 1998)) that allow it to succeed in nutrient-poor environments where particulate backscattering per unit chlorophyll is relatively low. As a result, a bridge between optics and ecology exists whereby the optical niche occupied by K. brevis is directly related to adaptational strategies evolved by K. brevis that allow it to succeed in low-nutrient environments. Furthermore, the results of this study demonstrate that the uniqueness of this optical niche may permit K. brevis blooms to be identified using space-based ocean color data. 60

PAGE 71

6. CONCLUSIONS K. brevis blooms are optically unique from a remote sensing standpoint not due to the cells themselves, but due to the environment inhabited by the cells. A paucity of detritus along with possible decreased concentrations of inorganic material are displayed by K. brevis blooms on the WFS relative to high-chlorophyll (> 1mg m-3), non-K. brevis bloom estuarine waters. These differences may be a product of the physiological and behavior strategies evolved by K. brevis that allow it to out-compete other phytoplankton groups to bloom in low-nutrient, oligotrophic environments. The anomalously low backscattering that results from these conditions may permit classification of K. brevis blooms using optical sensors located on a variety of remote (aircraft and satellite) and in situ (moorings and ships) platforms. Accurate quantification of extreme bloom events (>100 mg m-3) may also be possible using MODIS fluorescence line height data since a strong correlation exists between FLH and Chl. Together, the classification and quantification techniques developed in this study may be implemented into existing monitoring programs and ecological models upon further validation in order to better mitigate the harmful effects of K. brevis blooms on the WFS and throughout the Gulf of Mexico. 61

PAGE 72

LIST OF REFERENCES Ahn, Y., Bricaud, A., Morel, A., 1992. Light backscattering efficiency and related properties of some phytoplankters. Deep Sea Research 39, 1835-1855. Allali, K., Bricaud, A., Claustre, H., 1997. Spatial variations in the chlorophyll-specific absorption coefficient of phytoplankton and photosynthetically active pigments in the equatorial Pacific. Journal of Geophysical Research 102, 12,413-12,423. Asai, S., Krzanowski, J.J., Anderson, W.H., Martin, D.F., Polson, J.B., Lockey, R.F., Bukantz, S.C., Szentivanyi, A., 1982. Effects of the toxin of red tide, Ptychodiscus brevis, on canine tracheal smooth muscle: a possible new asthma-triggering mechanism. The Journal of Allergy and Clinical Immunology 69, 418-428. Bidigare, R.R., Morrow, J.H., Kiefer, D.A., 1989. Derivative analysis of spectral absorption by photosynthetic pigments in the western Sargasso Sea. Journal of Marine Research 47, 323-341. Bjornland, T., Liaaen-Jensen, S., 1989. Distribution patterns of carotenoids in relation to chromophyte phylogeny and systematics. In: Green, J.C., Leadbeater, B.S.C., Diver, W.L. (Eds.), The Chromophyte Algae: Problems and Perspectives. Clarendon Press, Oxford, pp. 37-61. Blough, N.V., Zafiriou, O.C., Bonilla, J., 1993. Optical absorption spectra of waters from the Orinoco River outflow: terrestrial input of colored organic matter to the Caribbean. Journal of Geophysical Research 98, 2271-2278. Bricaud, A., Morel, A., Babin, M., Allali, K., Claustre, H., 1998. Variations of light absorption by suspended particles with chlorophyll a concentration in oceanic (case 1) waters: analysis and implications for bio-optical models. Journal of Geophysical Research 103, 31,033-31,044. Bricaud, A., Morel, A., Prieur, L., 1983. Optical efficiency factors of some phytoplankters. Limnology and Oceanography 28, 816-832. Bricaud, A., Stramski, D., 1990. Spectral absorption coefficients of living phytoplankton and nonalgal biogenous matter: a comparison between the Peru upwelling area and the Sargasso Sea. Limnology and Oceanography 35, 562-582. 62

PAGE 73

Butler, W.L., 1962. Absorption of light by turbid materials. J. Opt. Soc. Am. 52, 292-299. Cannizzaro, J.P., Carder, K.L., Chen, F.R., Heil, C.A., Vargo, G.A., 2004. A novel technique for detection of the toxic dinoflagellate, Karenia brevis, in the Gulf of Mexico from remotely sensed ocean color data. Continental Shelf Research, accepted. Carder, K.L., Chen, F.R., Cannizzaro, J.P., Campbell, J.W., Mitchell, B.G., 2003. Performance of MODIS semi-analytic ocean color algorithm for chlorophyll-a. Advances in Space Research, accepted. Carder, K.L., Chen, F.R., Lee, Z.P., Hawes, S.K., Kamykowski, D., 1999. Semi-analytic Moderate-Resolution Imaging Spectrometer algorithms for chlorophyll a and absorption with bio-optical domains based on nitrate-depletion temperatures. Journal of Geophysical Research 104, 5403-5422. Carder, K.L., Steward, R.G., 1985. A remote-sensing reflectance model of a red-tide dinoflagellate off west Florida. Limnology and Oceanography 30, 286-298. Carder, K.L., Steward, R.G., Harvey, G.R., Ortner, R.B., 1989. Marine humic and fulvic acids: their effects on remote sensing of ocean chlorophyll. Limnology and Oceanography 34, 68-81. Cleveland, J.S., 1995. Regional models for phytoplankton absorption as a function of chlorophyll a concentration. Journal of Geophysical Research 100, 13,333-13,344. Cleveland, J.S., Weidemann, A.D., 1993. Quantifying absorption by aquatic particles: a multiple scattering correction for glass-fiber filters. Limnology and Oceanography 38, 1321-1327. Cullen, J.J., Ciotti, A.M., Davis, R.F., Lewis, M.R., 1997. Optical detection and assessment of algal blooms. Limnology and Oceanography 42, 1223-1239. Esaias, W., Abbott, M., Barton, I., Brown, O.B., Campbell, J.W., Carder, K.L., Clark, D.K., Evans, R.H., Hoge, F.E., Gordon, H.R., Balch, W.M., Letelier, R., Minnett, P.J., 1998. An overview of MODIS capabilities for ocean science observations. IEEE Trans. Geosci. Remote Sens. 36, 1250-1265. Franks, P.J.S., 1997. Spatial patterns in dense algal blooms. Limnology and Oceanography 42, 1297-1305. Garver, S.A., Siegel, D.A., Mitchell, B.G., 1994. Variability in near surface particulate absorption spectra: what can a satellite ocean color imager see? Limnology and Oceanography 39, 1349-1367. 63

PAGE 74

Gilbes, F., Muller-Karger, F.E., Castillo, C.E.D., 2002. New evidence for the West Florida Shelf Plume. Continental Shelf Research 22, 2479-2496. Gilbes, F., Tomas, C., Walsh, J.J., Muller-Karger, F.E., 1996. An episodic chlorophyll plume on the West Florida Shelf. Continental Shelf Research 16, 1201-1224. Gordon, H.R., Brown, O.B., Evans, R.H., Brown, J.W., Smith, R.C., Baker, K.S., Clark, D.K., 1988. A semianalytic radiance model of ocean color. Journal of Geophysical Research 93, 10,909-10,924. Gordon, H.R., Clark, D.K., Brown, J.W., Brown, O.B., Evans, R.H., Broenkow, W.W., 1983. Phytoplankton pigment concentrations in the Middle Atlantic Bight: comparison of ship determinations and CZCS estimates. Applied Optics 22, 20-36. Gordon, H.R., Wang, M., 1994. Influence of oceanic whitecaps on atmospheric correction of ocean-color sensors. Applied Optics 33, 7754-7763. Habas, E.J., Gilbert, C.K., 1974. The economic effects of the 1971 Florida red tide and the damage it presages for future occurrences. Environmental Letters 6, 139-147. He, R., Weisburg, R.H., 2002. West Florida shelf circulation and temperature budget for the 1999 spring transition. Continental Shelf Research 22, 719-748. Hemmert, W.H., 1975. The public health implications of Gymnodinium breve red tides, A review of the literature and recent events. In: LoCicero, V.R. (Eds.), Proceedings of the First International Conference on Toxic Dinoflagellate Blooms, pp. 489-497. Hoepffner, N., Sathyendranath, S., 1991. Effect of pigment composition on absorption properties of phytoplankton. Marine Ecology Progress Series 73, 11-23. Holm-Hansen, O., Lorenzen, C.J., Holmes, R.W., Strickland, J.D.H., 1965. Fluorometric determination of chlorophyll. J. Cons. Perm. Int. Explor. Mer 30, 3-15. Hu, C., Carder, K.L., Muller-Karger, F.E., 2000. Atmospheric correction of SeaWiFS imagery over turbid coastal waters: a practical method. Remote Sensing of Environment 74, 195-206. Hu, C., F.E. Muller-Karger, D.C. Biggs, K.L. Carder, B. Nababan, D. Nadeau, J. Vanderbloemen, 2003. Comparison of ship and satellite bio-optical measurements on the continental margin of the NE Gulf of Mexico. International Journal of Remote Sensing 24, 2597-2612. Jeffrey, S.W., 1980. Algal pigment systems. In: Falkowski, P.G. (Eds.), Primary Productivity in the Sea. Plenum Press, New York, pp. 33-58. 64

PAGE 75

Jeffrey, S.W., Sielicki, M., Haxo, F.T., 1975. Chloroplast pigment patterns in dinoflagellates. Journal of Phycology 11, 374-384. Kerfoot, J., G. Kirkpatrick, Lohrenz, G., Mahoney, K., Moline, M., Schofield, O., 2002. Vertical migration of Karenia brevis bloom: implications for remote sensing of harmful algal blooms. In: K.A. Steidinger, J.H.L., C.R. Tomas, G.A. Vargo (Eds.), Florida Fish and Wildlife Conservation Commission and Intergovernmental Oceanographic Commission of UNESCO, St. Pete Beach, FL, pp. 5. Kiefer, D., 1973. Fluorescence properties of natural phytoplankton populations. Marine Biology 22, 263-269. Kiefer, D.A., SooHoo, J.B., 1982. Spectral absorption by marine particles of coastal waters of Baja California. Limnology and Oceanography 27, 492-499. Kirkpatrick, G.J., Millie, D.F., Moline, M.A., Schofield, O., 2000. Optical discrimination of a phytoplankton species in natural mixed populations. Limnology and Oceanography 45, 467-471. Kirkpatrick, G.J., Schofield, O.M., Millie, D.F., Moline, M.A., 2002. In situ, autonomous optical detection and 3-D mapping of harmful algal blooms. In: K.A. Steidinger, J.H.L., C.R. Tomas, G.A. Vargo (Eds.), Florida Fish and Wildlife Conservation Commission and Intergovernmental Oceanographic Commission of UNESCO, St. Pete Beach, FL, pp. Kishino, M., Takahashi, M., Okami, N., Ichimura, S., 1985. Estimation of the spectral absorption coefficients of phytoplankton in the sea. Bulletin of Marine Science 37, 634-642. Lambert, C.D., Bianchi, T.S., Santschi, P.H., 1999. Cross-shelf changes in phytoplankton community composition in the Gulf of Mexico (Texas shelf/slope): the use of plant pigments as biomarkers. Continental Shelf Research 19, 1-21. Landsberg, J.H., Steidinger, K.A., 1998. A historical review of Gymnodium breve red tides implicated in mass mortalities of the manatee (Trichechus manatus latiriostris) in Florida, USA. In: Reguera, B., Blanco, J., Fernandez, M.L., Wyatt, T. (Eds.), Proceedings of the 8th International Conference on Harmful Algal Blooms, Vigo, Spain, pp. 97-100. Lee, Z., Carder, K.L., 2000. Band-ratio of spectral-curvature algorithms for satellite remote sensing? Applied Optics 39, 4377-4380. Lee, Z., Carder, K.L., Chen, R.F., Peacock, T.G., 2001. Properties of the water column and bottom derived from Airborne Visible Infrared Imaging Spectrometer (AVIRIS) data. Journal of Geophysical Research 106, 11,639-11,651. 65

PAGE 76

Lee, Z., Carder, K.L., Mobley, C.D., Steward, R.G., Patch, J.S., 1999. Hyperspectral remote sensing for shallow waters: 2. deriving bottom depths and water properties by optimization. Applied Optics 38, 3831-3843. Lee, Z.P., Carder, K.L., Mobley, C., Steward, R.G., Patch, J.S., 1998. Hyperspectral remote sensing for shallow waters: I. a semi-analytical model. Applied Optics 37, 6329-6338. Lee, Z.P., Carder, K.L., Steward, R.G., Peacock, T.G., Davis, C.O., Mueller, J.L., 1996. Remote-sensing reflectance and inherent optical properties of oceanic waters derived from above-water measurements. In: Ackleson, S.G., Frouin, R. (Eds.), Ocean Optics XIII, Halifax, Nova Scotia, Canada, pp. 160-166. Lester, K.M., Heil, C., Neely, M., Spence, D., Murasko, S., Milroy, S., Hopkins, T., Sutton, T., Burghart, S., Bohr, R., Remson, A., Vargo, G., Walsh, J., 2003. Zooplankton and Karenia brevis in the Gulf of Mexico. Continental Shelf Research, submitted. Lohrenz, S.E., 2000. A novel theoretical approach to correct for pathlength amplication and variable sample loading in measurements of particulate spectral absorption by the quantitative filter pad technique. Journal of Plankton Research 22, 639-657. Lohrenz, S.E., Fahnenstiel, G.L., Kirkpatrick, G.J., Carroll, C.L., Kelly, K.A., 1999. Microphotometric assessment of spectral absorption and its potential application for characterization of harmful algal species. Journal of Phycology 35, 1438-1446. Maffione, R.A., Dana, D.R., 1997. Instruments and methods for measuring the backward-scattering coefficient of ocean waters. Applied Optics 36, 6057-6067. Mahoney, K.L., 2003. Backscattering of light by Karenia brevis and implications for optical detection and monitoring. Ph.D, University of Southern Mississippi, Stennis Space Center. Masserini, R.T., Fanning, K.A., 2000. A sensor package for the simultaneous determination of nanomolar concentrations of nitrite, nitrate, and ammonia in seawater by fluorescence detection. Marine Chemistry 68, 323-333. Millie, D.F., Kirkpatrick, G.J., Vinyard, B.T., 1995. Relating photosynthetic pigments and in vivo optical density spectra to irradiance for the Florida red-tide dinoflagellate Gymnodinium breve. Marine Ecology Progress Series 120, 65-75. Millie, D.F., Schofield, O.M., Kirkpatrick, G.J., Johnsen, G., Tester, P.A., Vinyard, B.T., 1997. Detection of harmful algal blooms using photopigments and absorption signatures: A case study of the Florida red tide dinoflagellate, Gymnodinium breve. Limnology and Oceanography 42, 1240-1251. 66

PAGE 77

Mitchell, B.G., Kiefer, D.A., 1988. Chlorophyll a specific absorption and fluorescence excitation spectra for light-limited phytoplankton. Deep Sea Research 35, 639-663. Mobley, C.D. (1994). Light and water: radiative transfer in natural waters. San Diego, Academic Press. Moore, L.R., Goericke, R., Chisholm, S.W., 1995. Comparative physiology of Synechococcus and Prochlorococcus: influence of light and temperature on growth, pigments, fluorescence and absorptive properties. Marine Ecology Progress Series 116, 259-275. Morel, A., 1974. Optical properties of pure water and pure sea water. In: Jerlov, N.G., Nielsen, E.S. (Eds.), Optical Aspects of Oceanography. Academic Press, London, pp. 1-24. Morel, A., 1988. Optical modeling of the upper ocean in relation to its biogenous matter content (case I waters). Journal of Geophysical Research 93, 10,749-10,768. Morel, A., Ahn, Y.-H., 1991. Optics of heterotrophic nanoflagellates and ciliates: a tentative assessment of their scattering role in oceanic waters compared to those of bacteria and algal cells. Journal of Marine Research 49, 177-202. Morel, A., Bricaud, A., 1981. Theoretical results concerning light absorption in a discrete medium, and application to specific absorption of phytoplankton. Deep Sea Research 28, 1375-1393. Morel, A., Prieur, L., 1977. Analysis of variations in ocean color. Limnology and Oceanography 22, 709-722. Mller-Karger, F.E., Walsh, J.J., Evans, R.H., Meyers, M.B., 1991. On the seasonal phytoplankton concentration and sea surface temperature cycles of the Gulf of Mexico as determined by satellites. Journal of Geophysical Research 96, 12,645-12,665. Nelson, J.R., Guarda, S., 1995. Particulate and dissolved spectral absorption on the continental shelf of the southeastern United States. Journal of Geophysical Research 100, 8715-8732. Nelson, J.R., Robertson, C.Y., 1993. Detrital spectral absorption: laboratory studies of visible light effects on phytodetritus absorption, bacterial spectral signal, and comparison to field measurements. Journal of Marine Research 51, 181-207. Nelson, N.B., Prezelin, B.B., Bidigare, R.R., 1993. Phytoplankton light absorption and the package effect in California coastal waters. Marine Ecology Progress Series 94, 217-227. 67

PAGE 78

O'Reilly, J.E., Maritorena, S., Mitchell, B.G., Siegel, D.A., Carder, K.L., Garver, S.A., Kahru, M., McClain, C., 1998. Ocean color chlorophyll algorithms for SeaWiFS. Journal of Geophysical Research 103, 24,937-24,953. Pope, R., Fry, E., 1997. Absorption spectrum (380-700nm) of pure waters, II, Integrating cavity measurements. Applied Optics 36, 8710-8723. Preisendorfer, R.W., Application of radiative transfer theory to light measurements in the sea. Monogr. 10, pp.11-30, Intl. Union Geod. Geophys., Paris, 1961. Prieur, L., Sathyendranath, S., 1981. An optical classification of coastal and oceanic waters based on the specific spectral absorption curves of phytoplankton pigments, dissolved organic matter, and other particulate materials. Limnology and Oceanography 26, 671-689. Qian, Y., Jochens, A.E., II, M.C.K., Biggs, D.C., 2003. Spatial and temporal variability of phytoplankton biomass and community structure over the continental margin of the northeast Gulf of Mexico based on pigment analysis. Continental Shelf Research 23, 1-17. Roesler, C.S., McLeroy-Etheridge, S.L., 1998. Remote detection of harmful algal blooms. In: Ackleson, S., Campbell, J. (Eds.), Ocean Optics XIV, Kailua-Kona, HI, pp. 12. Roesler, C.S., Perry, M.J., Carder, K.L., 1989. Modeling in situ phytoplankton absorption from total absorption spectra in productive inland marine waters. Limnology and Oceanography 34, 1510-1523. Schofield, O., Grzymski, J., Bissett, W.P., Kirkpatick, G.J., Millie, D.F., Moline, M., Roesler, C.S., 1999. Optical monitoring and forecasting systems for Harmful Algal Blooms: Possibility or pipe dream? Journal of Phycology 35, 1477-1496. Shanley, E., Vargo, G.A., 1993. Cellular composition, growth, photosynthesis, and respiration rates of Gymnodinium breve under varying light levels. In: Smayda, T.J., Shimizu, Y. (Eds.), Toxic Phytoplankton Blooms in the Sea. Elsevier, New York, pp. 831-836. Shuman, F.R., Lorenzen, C.J., 1975. Quantitative degradation of chlorophyll by a marine herbivore. Limnology and Oceanography 20, 580-586. Siegel, D.A., Wang, M., Maritorena, S., Robinson, W., 2000. Atmospheric correction of satellite ocean color imagery: the black pixel assumption. Applied Optics 39, 3582-3591. Smayda, T.J., 1997. Harmful algal blooms: their ecophysiology and general relevance to phytoplankton blooms in the sea. Limnology and Oceanography 42, 1137-1153. 68

PAGE 79

Steidinger, K., 1975. Basic factors influencing red tides. In: LoCicero, V.R. (Eds.), Proceedings of the First International Conference on Toxic Dinoflagellate Blooms, pp. 153-162. Steidinger, K.A., Haddad, K., 1981. Biologic and hydrographic aspects of red tides. BioScience 31, 814-819. Steidinger, K.A., Vargo, G.A., Tester, P.A., Tomas, C.R., 1998. Bloom dynamics and physiology of Gymnodinium breve, with emphasis on the Gulf of Mexico. In: Anderson, E.M., Cembella, A.D., Hallengraff, G.M. (Eds.), Physiological Ecology of Harmful Algal Blooms. Springer-Verlag, New York, pp. 135-153. Stramski, D., Kiefer, D.A., 1991. Light scattering in microorganisms in the open ocean. Progress in Oceanography 28, 343-383. Stuart, V., Sathyendranath, S., Platt, T., Maas, H., Irwin, B.D., 1998. Pigments and species composition of natural phytoplankton populations: effect on the absorption spectra. Journal of Plankton Research 20, 187-217. Stumpf, R.P., Culver, M.E., Tester, P.A., Tomlinson, M., Kirkpatrick, G.J., Pederson, B.A., Truby, E., Ransibrahmanakul, V., Soracco, M., 2003. Monitoring Karenia brevis blooms in the Gulf of Mexico using satellite ocean color imagery and other data. Harmful Algae 2, 147-160. Subramaniam, A., Carpenter, E.J., Falkowski, P.G., 1999. Bio-optical properties of the marine diazotrophic cyanobacteria Trichodesmium spp. II. A reflectance model for remote sensing. Limnology and Oceanography 44, 618-627. Tester, P.A., Steidinger, K.A., 1997. Gymnodinium breve red tide blooms: initiation, transport, and consequences of surface circulation. Limnology and Oceanography 42, 1039-1051. Tester, P.A., Stumpf, R.P., Steidinger, K., 1998. Ocean color imagery: What is the minimum detection level for Gymnodinium breve blooms? In: Reguera, B., Blanco, J., Fernandez, M.L., Wyatt, T. (Eds.), Proceedings of the 8th International Conference on Harmful Algal Blooms, Vigo, Spain, pp. 3. Tester, P.A., Stumpf, R.P., Vukovich, F.M., Fowler, P.K., Turner, J.T., 1991. An expatriate red tide bloom: Transport, distribution, and persistence. Limnology and Oceanography 36, 1053-1061. Trabjerg, I., Hojerslev, N.K., 1996. Temperature influence on light absorption by fresh water and seawater in the visible and near-infrared spectrum. Applied Optics 35, 2653-2658. 69

PAGE 80

Trees, C.C., Clark, D.K., Bidigare, R.R., Ondrusek, M.E., Mueller, J.L., 2000. Accessory pigments versus chlorophyll a concentrations within the euphotic zone: A ubiquitous relationship. Limnology and Oceanography 45, 1130-1143. Turner, J.T., Tester, P.A., 1997. Toxic marine phytoplankton, zooplankton grazers, and pelagic food webs. Limnology and Oceanography 42, 1203-1214. Twardowski, M.S., E. Boss, J.B. Macdonald, W.S. Pegau, A.H. Barnard, and J.R.V. Zaneveld, 2001. A model for estimating bulk refractive index from the optical backscattering ratio and the implications for understanding particle composition in case I and case II waters. Journal of Geophysical Research 106, 14,129-14,142. Ulloa, O., Sathyendranath, S., Platt, T., 1994. Effect of the particle-size distribution on the backscattering ratio in seawater. Applied Optics 33, 7070-7077. Vargo, G.A., Carder, K.L., Gregg, W., Shanley, E., Heil, C., Steidinger, K.A., Haddad, K.D., 1987. The potential contribution of primary production by red tides to the west Florida shelf ecosystem. Limnology and Oceanography 32, 762-767. Vargo, G.A., Heil, C.A., Spence, D., Neely, M.B., Merkt, R., Lester, K., Weisburg, R.H., Walsh, J.J., Fanning, K., 2000. The hydrographic regime, nutrient requirements, and transport of a Gymnodinium breve Davis red tide on the West Florida Shelf. In: Hallegraff, G.M., Blackburn, S.I., Bolch, C.J., Lewis, R.J. (Eds.), Proceedings of the 9th International Conference on Harmful Algae, Hobart, Australia, pp. 157-160. Vodacek, A., Blough, N.V., 1997. Seasonal variation of CDOM and DOC in the Middle Atlantic Bight: terrestrial inputs and photooxidation. Limnology and Oceanography 42, 674-686. Walsh, J.J., Haddad, K.D., Dieterle, D.A., Weisburg, R.H., Li, Z., Yang, H., Muller-Karger, F.E., Heil, C.A., Bissett, W.P., 2002. A numerical analysis of landfall of the 1979 red tide of Karenia brevis along the west coast of Florida. Continental Shelf Research 22, 15-38. Walsh, J.J., Steidinger, K.A., 2001. Saharan dust and Florida red tides: the cyanophyte connection. Journals of Geophysical Research 106, 11,597-11,612. Wawrik, B., Paul, J.H., Campbell, L., Griffin, D., Houchin, L., Fuentes-Ortega, A., Muller-Karger, F., 2003. Vertical structure of the phytoplankton community associated with a coastal plume in the Gulf of Mexico. Marine Ecology Progress Series 251, 87-101. Wright, S.W., Jeffrey, S.W., Mantoura, R.F.C., Llewellyn, C.A., Bjornland, T., Repeta, D., Welschmeyer, N., 1991. Improved HPLC method for the analysis of chlorophylls and carotenoids from marine phytoplankton. Marine Ecology Progress Series 77, 183-196. 70

PAGE 81

Yentsch, C.S., 1962. Measurement of visible light absorption by particulate matter in the ocean. Limnology and Oceanography 7, 207-217. 71