|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Characterization of the Underwater Light Environment and Its Relevance to Seagrass Recovery and Sustainability in Tampa Bay, Florida by Christopher J. Anastasiou A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Major Professor: Kendall Carder Ph.D. Paula Coble, Ph.D. Penny Hall, Ph.D. Ma rk Luther, Ph.D. John Walsh, Ph.D. Date of Approval: November 10 2009 Keywords: light attenuation, seagrass, spectral, flow through, fluorescence, absorption, optical model, water clarity, water quality C opyright 2009 Christopher J. Anastasiou
Dedication To my parents who always encouraged my intense curiosity and wonder about the sea, and instilled in me the determination to pursue my dreams, and to never let anyt hing keep me from them. ...that which we are, we are; One equal temper of heroic hearts, Made weak by time and fate, but strong in will, To strive, to seek, to find, and not to yield. Alfred, Lord Tennyson
Acknowledgements I thank my major professor, Dr. Ken Carder, for his academic, professional, and moral support over these many years. I am truly honored to be counted as one of Dr. former students. I also thank my committee for all o f their support and dedication, specifical ly, Dr. Paula Coble for her insight and expertise in colored dissolved task, Dr. Mark Luther for his insight on residence times and hydrodynamic circulation in Tampa Bay and Dr. Penny Hall for her expertise in seagrass ecology. I thank Holly Greening and the Tampa Bay Estuary Program for their financial support, the Florida Department of Environmental Protection for in k ind services and for providing me the opportunity to complete this research, the Florida Fish and Wildlife Research Institute for in kind services, and the University of South Florida Ocean Optics Laboratory for the use of their optical sensors I am grateful to the following individuals for their techni cal support: Jennifer Kunzelman for the many hours spent in the field, Dave English for his support with the flow through system, Jennifer Cannizzaro for her laboratory support, Walt Avery and Roger Johansson for their institutional knowledge of Tampa Bay, Jim Ivey for his technical expertise and support, Keith Reynolds for his GIS support, and many others. I also wish to thank all my friends and family for their encouragement and support over the past 5 years. Finally, I am deeply indebted to my beautiful wife Wendy for her unconditional support encouragement, and unwavering confidence in my ability to finish this project.
Note to Reader The original of this document contains color that is necessary for understanding the data. The origi nal dissertation is on file with the USF library in Tampa, Florida.
i Table of Contents List of Tables ................................ ................................ ................................ ..................... iii List of Figures ................................ ................................ ................................ ...................... v List of Definit ions and Acronyms ................................ ................................ ...................... ix Abstract ................................ ................................ ................................ ................................ x Introduction ................................ ................................ ................................ .......................... 1 Chapter 1 Estimating Photosynthetically Available and Useable Radiation at the Seagrass Deep Edge in Tampa Bay, Florid a ................................ ............................ 7 1.1. Abstract ................................ ................................ ................................ ............ 7 1.2. Introduction ................................ ................................ ................................ ...... 9 1.2.1. Minimum light targets in Tampa Bay ................................ ............... 9 1.2.2. Wavelength specific light utilization for pho tosynthesis ............... 12 1.2.3. Measuring transparency in shallow waters ................................ ..... 15 1.3. Materials and Methods ................................ ................................ ................... 17 1.3.1. The Seagrass Management Area concept ................................ ....... 17 1.3.2. Measuring E d ( ), PAR( ), and PAR ................................ .............. 23 1.3.4. Mapping percent subsurface irradiance ................................ .......... 28 1.4. Results and Discussion ................................ ................................ .................. 33 1.4.1. K d ( ) and PAR( ) relationships across SMAs ............................... 33 1.4.2. PUR( ) relationships ................................ ................................ ...... 39 1.4.3. Mapping spectral light and depth targets ................................ ........ 43 1.5. Conclusions ................................ ................................ ................................ .... 51 Chapter 2 Relativ e Contribution and Magnitude of Phytoplankton, CDOM, and Detritus Absorption to the Total Absorption Coefficient in Shallow Seagrass Areas ................................ ................................ ................................ ....... 52 2.1. Abstract ................................ ................................ ................................ .......... 52 2.2. Introduction ................................ ................................ ................................ .... 54 2.2.1. The inherent optical properties ................................ ....................... 55 2.2.2. Seagrass light relationships in Tampa Bay ................................ ..... 60 2.2.3 Coupling the IOPs with the quasi IOPs and K d ( ) .......................... 62 2.3. Methods and Materials ................................ ................................ ................... 63
ii 2.3.1. Site Locations ................................ ................................ ................. 63 2.3.2. IOP field and laboratory methods ................................ ................... 66 2.3.3. Coupling the IOPs with th e quasi IOPs ................................ .......... 69 2.4. Results and Discussion ................................ ................................ .................. 72 2.4.1. IOP spatial and temporal patterns ................................ ................... 72 2.4.2. Relationship between the IOPs and the quasi IOPs ........................ 81 2.4.3. Modeling K d (480) for Tampa Bay ................................ .................. 85 2.5. Conclusions ................................ ................................ ................................ .... 86 Chapter 3 Synoptic Surveillance of the Underwater Light Field Using a Continuous Deck Mounted Flow Through System ................................ ............... 90 3.1. Abstract ................................ ................................ ................................ .......... 90 3.2. Introduction ................................ ................................ ................................ .... 92 3.2.1 The underway flow through system approach ................................ 93 3.2.2 Principles of in water fluorescence ................................ .................. 94 3.3. Materials and Methods ................................ ................................ ................... 95 3.3.1. Flow through system design and specification ............................... 95 3.3.2. Synoptic survey of Seagrass Management Area s ........................... 96 3 .3.3. Data management and analysis ................................ ..................... 100 3.4. Results and Discussion ................................ ................................ ................ 102 3.4.1. Inherent optical properties and fluorescence ................................ 102 3.4.2. CDOM and chlorophyll spatial var iability ................................ ... 113 3.5. Conclusions ................................ ................................ ................................ .. 116 Conclusions ................................ ................................ ................................ ...................... 119 References ................................ ................................ ................................ ........................ 121 About the Author ................................ ................................ ................................ ... End Page
iii List of Tables Tab le 1 1. Selected wavelength ranges of the color bands blue, green, and red and the pigmen ts that are represented by each ................................ .................... 15 Table 1 2. Fixed stations for each SMA where in water irradia nce measurements were collected ................................ ................................ ............................... 17 Table 1 3. Annual average light attenuation coefficient for nearshore and offshore stations for th e blue color band (400nm 490nm) ................................ ......... 33 Table 1 4. Monthly percent subsurface irradiance at bottom relative t o just below the water surface for each representative color band ................................ ... 36 Table 1 5. Proportion of the blue, green, and red color bands relative to total bottom PAR at the seagrass deep edge during w et and dry periods for each SMA ................................ ................................ ............................... 39 Table 1 6. PAR and PUR for each SMA under wet and dry conditions for each representative color band ................................ ................................ ............. 42 Table 1 7. Total measured depth, percent subsurface blue light, and percent subsurface PAR relative to surface conditions along the mapped seagrass deep edge. ................................ ................................ ...................... 45 Table 2 1. Common parameters associated with the inherent and appar ent optical properties of water ................................ ................................ ........................ 56 Table 2 2. CDOM source s and sinks in Tampa Bay summarizing inputs and outputs. ................................ ................................ ................................ ......... 61 Table 2 3. Mean standard deviation of the annual chlorophyll concentrations and the annual absorption coefficients at 440nm for chlorophyll, CDOM, detritus, and tot al absorption. ................................ ......................... 74 Table 2 4. Annual percent contribution to the total absorption coefficient by chlorophyll absorption, CDOM abso rption, and detrital absorption ............ 80
iv Table 2 5. Predictor equations f or based on chlorophyll a concentration for each Seagrass Management Area and for all areas combined ................ 81 Table 2 6. Average annual summary table for predicted , and st andard deviation fo r each Seagrass Management Area ................................ ................................ ........................ 86 Table 3 1. Payload description of the underway flow throu gh system used for this project ................................ ................................ ................................ ........... 93 Table 3 2. Regression equations, with for each instrument aboard the flow through system, used to convert raw fluorescence voltage to corresponding absorption coefficients for chlorophyll and CDOM and chlorophyll concentration ................................ ................................ ................................ .... 103 Table 3 3. Results of multiple regression analyses for establishing the relationship between CDOM fluorescence and the IOPs for shal low seagrass areas in Tampa Bay ................................ ................................ .................... 112 Table 3 4. Annual summary of CDOM fluorescence voltage and corresponding CDOM absorption for the K itchen Seagrass Management Area ................................ ................................ ...................... 113 Table 3 5. Annual summar y of chlorophyll concentrations calculated using chlorophyll fluorescence for the Kitchen Seagrass Management Area ................................ ................................ ...................... 11 5
v List of Figures Figure 1 1. Conceptual d iagram of light loss with depth ................................ ................... 10 Figure 1 2. Absorptance spectra and PUR( ) for two species of seagr ass Thalassia testudinum (Banks ex. Knig) and Halodule wrightii (Asch.), and bottom PAR( ) measured in Tampa Bay ................................ 13 Figure 1 3. Seagrass Management Area s of Tampa Bay ................................ .................. 18 Figure 1 4. Map showing the spatial extent of seagr ass coverage in the Kitchen SMA ................................ ................................ ................................ ............. 19 Figure 1 5. Map showing the spatial extent of seagrass coverage in the Wolf Branch SMA ................................ ................................ ................................ 20 Figure 1 6. Map showing the spatial extent o f seagrass cove rage in the Coffeepot Bayou SMA ................................ ................................ ................................ .. 22 Figure 1 7. Map showing the spatial extent of seagrass cove rage in the Coffeepot Bayou SMA ................................ ................................ ................................ .. 23 Figure 1 8. Spectral scans of PAR( ) with depth collected in August 2008 at Wolf Branch in eastern Tampa Bay. ................................ ................................ ..... 25 Figure 1 9. The light attenuation coefficient as a function of wavelength. Data from the Wolf Branch Seagrass Management Area collected in October 2008 ................................ ................................ ................................ 26 Figure 1 10. Concept of operations of the NASA Experimental Advanced Airborne Research LIDAR (EAARL) ................................ .......................... 29 Figure 1 11. Bathymetry of the Kitchen Seagrass Management Area ............................. 30 Figure 1 12. Bathymetry of the Wolf Branch Seagrass Management Area ..................... 31 Figure 1 13. Bathymetry of the Coffeepot Bayou Seagrass Management Area ............... 32 Figure 1 14. Annual average light loss with depth by color band for each Seagrass Management Area ................................ ................................ ........................ 35
vi Figure 1 15. Relationship between K d (blue) and the 30 day running average for rainfall for each Seagrass Management Area ................................ .............. 38 Figu re 1 16. PUR( ) curves for each of the four Seagrass Management Area s sampled in this study under wet conditions and dry conditions. .................. 41 Figure 1 17. Percent subsurface PAR reaching the bottom for Coffeepot Bayou calculated using the annual average and LIDAR bathymetry. ................................ ................................ ................................ ... 44 Figure 1 18. Percent subsurface blue light reaching the bottom for Coffeepot Bayou calculated using the annual average and LIDAR bathymetry. ................................ ................................ ................................ ... 46 Figure 1 19. Percent subsurface PAR reaching the bottom for Wolf Branch calculated using the annual average K d (PAR) and LIDAR bathymetry. ................................ ................................ ................................ ... 47 Figure 1 20. Percent subsurface blue light reaching the bottom for Wo lf Branch calculated using the annual average K d (blue) and LIDAR bathymetry. ................................ ................................ ................................ ... 48 Figure 1 21. Percent subsurface PAR reaching the bottom for the Kitchen calculated using the annual average and LIDAR bathymetry. ................................ ................................ ................................ ... 49 Figure 1 22. Percent subsurface blue light reaching the bottom for the Kitchen calculated using the annual average and LIDAR bathymetry ................................ ................................ ................................ .... 50 Figure 2 1. Typical phytoplankton, CDOM, and detrital absorption sp ectra measured in Tampa Bay, FL ................................ ................................ ........ 57 Figure 2 2. Tampa Bay nitrogen management strategy as it relates to light and seagra ss sustainability ................................ ................................ .................. 60 Figure 2 3. Modified Tampa Bay seagrass management strategy that includes spectral absorption by CDOM and detritus ................................ .................. 62 Figure 2 4. Seagrass Management Area s in Tampa Bay and the four areas used in this study. ................................ ................................ ................................ ..... 64
vii Figure 2 5. Mangroves dominate the shoreline in (a) the Kitchen and (b) Wolf B ranch Seagrass Management Area s ................................ ........................... 65 Figure 2 6. Average annual CDOM, detritus, and phytoplankton absorption spectra for ea ch Seagrass Management Area ................................ ............... 73 Figure 2 7. Total nitrogen and chlorophyll data from two long term monitoring stations near the Kitchen and Egmon t Key Seagrass Management Area s ................................ ................................ ................................ ............. 76 Figure 2 8. Month ly averages standard deviation of the mean for the inherent optical properties for the four Seagrass Management Area s ........................ 77 Figure 2 9. Historical monthly average rainfall and average rainfall for 2008 for St. Petersburg, FL ................................ ................................ ......................... 78 Figure 2 10. Scatter plot and best fit line of the 30 day running average for rainfall and (a) .and (b) fo r each Seagrass Management Area ................................ ................................ ........................ 79 Figure 2 11. Chlorophyll a concentration and for Tampa Bay, collected during this study, and data collected for the West Florida Shelf ................. 82 Figure 2 12. Relationship between turbidity and the scattering coefficients for blue light and red light fo r selected samples in Tampa Bay ................................ ................................ ................................ ... 83 Figure 2 13. PCU color at 345nm plotted against for samples taken .............. 84 Figure 2 14. Plot of measured against modeled ................................ .... 85 Figure 3 1. The deck mounted flow th rough system used in this study ............................ 96 Figure 3 2. Example survey track from June 2008 for the Wolf Branch Seagrass Management Area ................................ ................................ ........................ 99 Figure 3 3. Relationship between WetStar chloroph yll fluorescence and the phytoplankton absorption coefficients at 440nm and at 660nm ................................ ................................ ................................ ......... 103
viii Figure 3 4. Relationship between SeaTech chlorophyll fluorescence and the phytoplankton absorption coefficients at 440nm and at 660nm ................................ ................................ ................................ ......... 104 Figure 3 5. Relationship between WetStar chlorophyll fluorescence and chlorophyll concentration ................................ ........................... 105 Figure 3 6. Relationship between SeaTech chlorophyll fluorescence and chlorophyll concentration ................................ ........................... 106 Figure 3 7. Relationship between CDOM fluorescence and the CDOM absorption coefficient for the WetStar and ECO fluorometers ..................... 107 Figure 3 8. Wavelength dependence of the absorption coeff icient on the relationship between and for various wavelengths ............ 108 Figure 3 9. Plot of and for all sample data used in this study. .......... 109 Figure 3 10. Relationship betweeen CDOM fluorescence and CDOM absorption at 312nm for the (a) W etStar and (b) ECO fluorometers ........................... 110 Figure 3 11. Comparison at 440nm of the CDOM absorption coefficient with the (a) detritus (b) phytoplankton and (c) particulate absorption coefficients ................................ ...... 111 Figure 3 12. Survey results of (V) across the Kitchen Seagrass Management Area fo r (a) April and (b) August 2008 ................................ 114 F igure 3 13. Survey results of across the Kitchen Seagrass Management Area fo r (a) April and (b) August 2008 expressed in concentration ................................ ................................ ................................ .... 116
ix List of Definitions and Acronyms Symbol Definition Units Specific leaf a bsorptance nm 1 Spectral absorption coefficient of phytoplankton m 1 Spectral absorption coefficient of detritus m 1 Spectral a bsorption of CDOM m 1 Spectral absorption coefficient of particulate matter m 1 Total spectral absorption coefficient m 1 Total s cattering c oefficient m 1 Beam a ttenuation c oefficient m 1 CDOM Colored D issolved O r ganic M atter Chl a Chlorophyll a c oncentration g L 1 Spectral do wnwelling i rradiance W m 1 Spectral d ownwelling a ttenuation c oefficient m 1 LIDAR Light D etection and R anging PAR( ) Wavelength specific P hotosynthetically A cti ve R adiation mol m 2 s 1 PUR( ) Wavelength specific P hotosynthetically U seable R adiation mol m 2 s 1 SMA Seagrass Management Area CDOM f luorescence V Chlorophyll f luorescence V Wavelength specific t ransmittance V nm 1 VDC Voltage Direct Current V
x Characterization of the Underwater Light Environment and Its Relevance to Seagrass Re covery and Sustainability in Tampa Bay, Florida Christopher J Anastasiou Abstract The availability of light is a primary limiting factor for seagrass recovery and sustainability. Understanding not only the quantity but the quality of light reaching the bottom is an important component to successful seagrass management and the key focus of this study. This study explores the spectral properties of the sub surface light field in four shallow Seagrass Management Area s (SMA) in Tampa Bay. W avelength specific p hotosynthetically a ctive r adiation ( PAR( ) ) and the spectral light attenuation coefficient are used to estimate the percent blue, green, and red light remaining at the bottom relative to the surface. LIDAR B athymetry is combined with to produce high resolution maps of percent subsurface light along the seagrass deep edge. The absorptance spectra from two sea grass species together with PAR( ) is used to calculate the p hotosynthetically u seable r adiation (PUR( )) a term describing the actual wavelengths of light being used by the seagrass. Based on the average annual 32% 39% percent of PAR rea ched the bottom at the seagrass deep edge while only 14% 18% of b lue light reach ed bottom, suggest ing that seagrass may be blue light limited. Analysis of PUR( ) data further confirmed that seagrass are blue light limited. Each SMA wa s characterized in terms of the inherent optical properties (IOP) of absorption and scatter. Tampa Bay is considered a chlorophyll dominated estuary However in this study, colored dissolved organic matter (CDOM) was the major
xi absorber of blue light, accounting for 60% of t he total absorption. To infer past light conditions, the IOPs were related to parameters more commonly used in routine monitoring programs. T o estimate an empirically derived model u sing only the total absorption and scatter coefficients was used and resulted in a good fit between measured and modeled A deck mounted flow through system was used to survey each SMA for CDOM and chlorophyll a fluorescence among other properties A series of SMA specific predictor equations were empirically derived to relate raw fluorescence to the IOPs The Kitchen SMA was used as a case study. Survey results show a strong connection bet ween CDOM rich waters and the mangrove dominated shoreline.
1 Introduction Seagrass are extremely productive estuarine and coastal resources and are critical habitat for a number of fish, shrimp, and crab species (Zieman and Zieman 1989) These systems are experiencing world wide decline, due in large part, to coastal eutrophication (Duarte et al. 2007) A major priority of the resource management community in Tampa Bay is to protect and restore seagrass habitat to the greatest extent possible (T BNEP 1996) Seagrass have among the highest light requirements of any organism in the plant kingdom (Gallegos 1994) Light is the primary limiting factor for most seagrass and its availability is determined by physical, chemical, and biological conditions (Duarte 2002; Duarte et al. 20 07; Kirk 1994; Miller 1995) In Tampa Bay, resource management agencies tasked with protecting and restoring the s e resources have established a minimum light target of 20.5% of total light reaching the bottom as the primary metric to achieve this objectiv e (Dixon and Leverone 1995; T BNEP 1996; Tomasko and Lapointe 1991) Other p roposed minimum light requirements for Tamp a Bay and other systems throughout Florida are between 20% 40% depending on species location and method used (Dennison et al. 1993; Dixon and Leverone 1995; Fourqurean et al. 2003; Gallegos 1994; Kenworthy and Fonseca 1 996; Steward et al. 2005) Typically light is measured in terms of the photosynthetically active radiation (PAR). By definition, PAR is a broadband quantity in units of mol photons m 2 s 1 integrated across the visible spectrum ( 400nm 700nm ) (Mobley 1994) While it is true that a photon induces the same chemical change within a molecule of chlorophyll irrespective of the photon energy state, photons of different wavelengths are not equally likely to be absorbed by chlorophyll (Mobley 1994) While PAR is a relatively good indicator of light quantit y, it does not take into account the spectral properties of the light field nor does it provide any indication that the photons available for photosynthesis are actually being used by the seagrass. incom ing light field, it is important to express light quality in terms of the specific absorption characteristics of the light harvesting pigments found in seagrass. L ike all higher plants, seagrass are reliant mostly on light within the blue and red color ban ds,
2 though the presence of accessory pigments can increase the operational window beyo nd the blue and red wavelengths. The term photosynthetically useable radiation (PUR( )) has been used to quantify t he fraction of radiant energy that can be absorbed by a given plant for a given wavelength (Carder 1995; Morel 1978; Morel 1991) Operationally PUR( ) is the product of the PAR( ) and the seagrass leaf absorptance. As a result of absorption and scatter of the incoming solar flux, PAR ( ) diminishes with depth in an approximatel y exponential manner (Kirk 1994) The downwelling light attenuation coefficient (K d ( )) describes the loss of light with depth and can be calculated as the slope of the natural log arithm of PAR ( ) With K d ( ), the percent subsurface PAR( ) can be estimated along the seagrass deep edge. PUR ( ) can then be used to compare what is available for photosynthesis with what is actually being used by the seagrass at the deep edge. Blue light is a common limiting factor in marine waters and results from ab sorption by phytoplankton and CDOM (Hoge et al. 1993; Menon et al. 2006) and is likely the case in Tampa Bay, especially along the seagrass deep edge. Estimating the PUR( ) at the seagrass deep edge can provide valuable insight into seagrass light utilization in areas that may be blue light limited. Accessory light harvesting pigments, including chlorophylls b and c and photosyntheti c carotenoids, expand the light harvesting capabilities across the visible spectrum (K irk 1994; Malick 2004) and may be a primary mechanism for surviving blue light limited environments. If blue light is severely depleted along the deep edge light harvesting by the accessory pigments in the red and even green wavelengths may be only means of survival in these environments Technology in recent years has advanced to where it is now feasible to incorporate spectral light readings as part of routine monitoring programs. Using a geographic information system ( GIS ) approach, the percent subsurf ace PAR( ) can be calculated and mapped for a given seagrass area by specifying the K d ( ) and bathymetry. Values of the percent subsurface PAR( ) reaching the bottom along the deep edge can be extracted from the map product. Defining the contributions of w ater, CDOM phytoplankton and detritus to the optical properties of the water column is a fundamental objective of bio optical
3 oceanography (Kiefer and Soohoo 1982; Nelson and Robertson 1993; Prieur and Sathyendranath 1981; Smith and Baker 1978) Such information is also essential for estimating the concentration and composition of particulate and dissolved materi als using remote sensing and in situ optical techniques (Nelson and Robertson 1993) Blue light in coastal and estuarine waters can be limited through absorption by phytoplankton, CDOM and detritus (Hoge et al. 1993; Nelson and Robertson 1993) Understanding the relative contribution and magnitude of the se inherent optical properties (IOP) is critical to understand ing the root causes of seagrass light limitation. A bsorption by phytoplankton is typically expressed in terms of the phytoplankton absorption coefficient in units of m 1 (Kirk 1994; Mobley 1994) Similarly, CDOM and detritus absorption can be described in terms of their absorption coefficient s ( and respectively). In Tampa Bay, seagras s management has been predicated on the assumption that light attenuation is controlled largely by increased chlorophyll caused by excessive nitrogen loading (Cannizzaro 2004; Janicki Environmental 2001) This pa radigm was developed during a time when wastewater effluent and untreated stormwater were directly discharging into the bay (T BNEP 1996) Today, advances in wastewater treatment and stormwater management have resulted in si gnificant decreases in both nitrogen loads and chlorophyll concentrations This decrease in the relative contribution of chlorophyll in t he bay may have caused an increase in the relative importance of CDOM and detritus to blue light a bsorption. If CDOM is the dominant light attenuator, there may be little resource management agencies can do to improve light quality at the seagrass deep edge. However it is likely that phytoplankton absorption is still contributing a significant amount to the total absorpti on of blue light, and evidence from recent work suggests that the relationship between nitrogen and phytoplankton productivity though complex, does hold true in Tampa Bay and other CDOM rich estuaries along the west coast of Florida (Janicki et al. 2003; Janicki and Wade 1996; Pribble et al. 2001; Tomasko and Ott 2001) Most resource management agencies do not meas ure the inherent optical properties (IOP) such as , and but collect quasi inherent optical properties such as chlorophyll a concentration, color, and turbidity. Chlorophyll a
4 concentration is a relatively good proxy for phytoplankton absorption (Bricaud et al. 1995) though differences in photosynthetic efficiencies among phytoplankton and pigment packaging effects, can cause significant errors in this relationship (Bissett 1997; Kirk 1994) C olor is reported in p latinum c obalt u nits (PCU) and is visually determined according to the EPA appro ved method (EPA 140 A). Despite this crude method, the relationship between PCU color and can be relatively strong (Gallegos 2005) Turbidity is reported in Nephelometric t urbidity u nits (NTU) and, using the EPA approved Method l 80. 1, is based upon a comparison of the intensity of light scattered at an angle of 90 o by a sample with the intensity of light scattered by a standard reference suspension, at the same scattering angle ( USEPA 1999) Turbidity can be expressed as a function of the total sc attering coefficient and the backscattering coefficient among other variables (Gallegos and Kenworthy 1996) Because chlorophyll a concentration, PCU color, and turbidity have been collected on a monthly basis for as long as 35 years in Tampa Bay (E.P.C.H.C. 2007) and are still widely collected by regulatory and resource management agencies throughout the state of Florida, it is important to couple these parameters with the IOPs (Gallegos and Kenworthy 1996; Gallegos 2001; Gallegos 2005; Lee 1998) For resource managers, i s one of the most relevant metrics for assessing seagrass habitat suitability largely because together with depth, can be used to calculate the percent subsurface PAR( ). However, PAR is an apparent optical property (AOP) and contains error due to variations in the ambient conditions at the time of sampling. For example, variable cloud cover, sea state, time of day, and time of year all contribute to unaccounted variability in B ecause of the multiple scattering that takes place in n atural systems, and the inherent variability in the angular distribution of the light field, there is no analytical expression to directly calculate from the IOPs (Kirk 1981) An alternative approach has been to develop an empirical relationship between and the IOPs using Monte Carlo procedures (Gallegos 2001; Kirk 1981; Kirk 1984) This method has proven to be very accurate and holds for most turbid estuaries with an sc attering to absorption ratio less than 30 (Kirk 1994)
5 Obtaining adequate spatial and temporal data about the underwater light field, though critical to successful seagrass management, can be costly and t ime consuming to collect and analyze A useful approach to supplement discrete water samples and provide calibration of satellite and airborne remote sensing is the use of a continuous flow through system Many of these systems have been constructed but most have been deployed on larger researc h vessels in blue water or coastal ocean environments (Madden and Day 1992; Twardowski et al. 2005) One such system has been modified to operate aboard a small open hull boat in shallow seagrass areas This deck mounted flow through system has all of the necessary sensors to fully characterize the optical light environment. The absorption coefficients and are not directly measured but modeled using chlorophyll and CDOM fluorescence (Belzile et al. 2006; Ferrari and Dowell 1998; Ferrari 1996) The relationships between fluorescence and absorption are site specific and therefore must be calibrated to the specific survey area. A contouring program can be used to map , or any other flow through parameter producing a synoptic snapshot of a given survey area. A deck mounted flow through system is not only useful for seagrass management but for any shallow water operation where optical information i s desired. This dissertation is divided into th ree main parts. Chapter 1 : Estimating the photosynthetically available and useable radiation at the seagrass deep edge, focuses on the quantity and quality of light penetrating the bottom of selected Seagrass Management Area s in Tampa Bay to (1) evaluate t he appropriateness of the current minimum light target s (2) examine the relationship between bottom PAR( ) and PUR( ) along the seagrass deep edge, (3) use a GIS based model ing approach to estimate d percent subsurface PAR( ) along the seagrass deep edge and test the hypothesis that the seagrass deep edge is blue light limited, and (4) propose spectrally rele vant minimum light targets for PAR( ). Chapter 2 : Relative contribution and magnitude of phytoplankton, CDOM, and detritus absorption to the total absorption coefficients in shallow seagrass areas, (1) tests the hypothesis that CDOM is the major absorption component to the total absorption of blue light, (2 ) challenges the current seagrass management paradigm by comparing the
6 SMA specific IOPs with specific environmental variables, (3) establishes S MA specific predictor equations to model the IOPs based on the quasi IOPs, and ( 4 ) utilizes an empirically derived spectral attenuation model to relate the IOPs to Finally, C hapter 3 : Synoptic surveillance of the underwater light field using a continuous deck mounted flow through system (1) designs a framework for surveying the optical properties of the light field in very shallow seagrass areas, (2) dev elops SMA specific correlations between the IOPs and raw flow through data, and (3) applies this framework to survey the spatial and temporal distribution of CDOM and chlorophyll using the Kitchen SMA as a case study.
7 Chapter 1. Estimating Photosynthetic ally Available and Useable Radiation at the Seagrass Deep Edge in Tampa Bay, F lorida 1 .1. Abstract Seagrass are among the most productive habitats in the world and are a vital component to maintaining a healthy estuary. To properly manage this resource re quires both a solid understanding of light seagrass relationships and a framework by which to monitor these complex relationships. A major challenge in managing seagrass is setting appropriate minimum light targets and i n Tampa Bay, resource management ag encies have adopted a bay wide minimum light target of 20.5% of photosynthetically available radiation (PAR) This target was based on a single species growing under optimal conditions and may not be appropriate as a bay wide estimate. PAR is a broadband q uantity and does not take into consideration the spectral properties of the light field, nor does it consider the specific absorption characteristics of the seagrass themselves By multiplying PAR( ) with leaf absorptance (A L ( )) a measure of the photosyn thetically useable radiation (PUR( )) can be easily obtained. To address the need for better monitoring tools, a GIS based modeling approach was used to couple hi gh resolution bathymetry with the light attenuation coefficient to map the percent subsurface PAR( ) reaching the bottom along the seagrass deep edge. P ercent subsurface PAR( ) was also calculate d using direct measurements at the seagrass deep edge. PAR( ) was grouped into blue, green, and red color bands whose wavelengths were based on the specifi c absorption characteristics of the seagrass species Thalassia testudinum ( Banks ex Knig ) and Halodule wrighti i ( Asch .) both major species found in Tampa Bay. In all cases, the light field was dep leted of blue light accounting for as little as 5.3 perce nt of the total PAR at the bottom. Green light accounted for approximately half the total PAR a t bottom while red light accounted for about one third. Based on annual average light attenuation coefficients seagrass received 31 .7 38.9 percent of surface PAR and 13.6 18 1 percent of surface blue light along the deep edge In August, during the rainy season seagrass at the deep edge received as little as 2.51 percent of surface blue light while still receiving 17.5 percent of surface PAR, 19.1 percent of surface green light, and 26.8
8 percent of surface red light. Under the lowest light conditions measured in this study, seagrass were primarily dependent on red light and, to a lesser extent, on blue green and yellow light. Bottom PUR( ) at the deep edge was 13.0 mol m 2 s 1 for blue light and 66.0 mol m 2 s 1 and 56.3 mol m 2 s 1 for red and green light, respectively. The relatively high PUR( ) for the green wavelengths suggests that when blue light is limited, the accessory pigment s chlorophyll s b c and the carotenoids may be mo st important to maintaining photosynthesis and ultimately plant survival
9 1 .2 Introduction Seagrass are important primary producers in estuarine systems around the world and provide critical habitat an d food for many commercially and recreationally important fish. Seagrass also play a key role in biogeochemical processes, sediment stability, as well as many other functions (Bortone 2000; Hemminga an d Duarte 2000; Thayer et al. 1984) Because of the perceived importance of seagrass to maintaining a healthy estuary, resource management agencies have focused on restoring and protecting seagrass to the greatest extent possible. Light is the primary limi ting factor for most seagrass ecosystems. For this reason, most management plans attempt to set minimum light targets along the deep edge typically in terms of the photosynthetically available radiation (PAR). 1.2.1. Minimum light targets in Tampa Bay Whil e seagrass meadows are highly productive systems, they are vulnerable to light limitation due in part to their high light requirements (Abal et al. 1994; Major and Dunton 2002; Zimmerman 2003) Seagrass depth li mits are strongly related not only to the percent of incoming solar radiation reaching the bottom and the rate at which it is attenuated ctral quality (Figure 1 1) (Dennison 1987; Duarte 1991; Durako 2007; Nielsen et al. 2002) In most systems, including Tampa Bay, seagrass are light limited (Janicki Environmental 2001; Kenworthy and Fonseca 1996) For this reason, most management plans attempt to set minimum light targets to maximize seagrass coverage along the deep edge.
10 Figure 1 1 Conceptual dia gram of light loss with depth. The percent subsurface irradiance is defined as the percent of PAR( ) or PUR( ) reaching the bottom relative to just below the water surface. Source: Center for Environmental Science, University of Maryland O ver development and pollution in Tampa Bay has resulted in a decrease in seagrass coverage b y 75% between the years of 1950 and 1985. Since 1985, improvements in wastewater treatment and stormwater management have seen large improvements in water quality resulting in incr eased light penetration with a corresponding increase in seagrass coverage Despite these large increases, total seagrass coverage is still below that of the 1950s. The development of a minimum light target for seagrass is a major part of the restoration a nd management plan for seagrass in Tampa Bay. Though imperfect, it has provided a context by which site suitability can be easily measured. The Tampa Bay Estuary Program and its partners established a minimum light target for Tampa Bay of 20.5% of the tot al incoming PAR in units of ( TBNEP 1996) This target was largely based on work that focused on a single species g rowing in lower Tampa Bay under optimal conditions
11 (Dixon and Leve rone 1995) The goal of this early work was to determine the annual light regime along the deep edge of a stable Thalassia testudinum ( Banks ex Knig ) bed where light limitation was believed to be the limiting factor (Hall et al. 1991) Based on annual water column PAR at the maximum seagrass depth limits for this species, the average percent subsurface light reaching the bottom was determined to be 22.5% (Dixon and Leverone 1995) As Dixon and Leverone (1995) clearly indicate, this value must be used with caution as it is only representative of the light attenuation for the given conditions at those sample locations. The decision to ext rapolate the findings of Dixon and Leverone (1995) to include the entire bay and to reduce the established target from 22.5% to its current value of 20.5% was largely a policy decision (H. Greening, personal communication). Other researchers have proposed minimum light targets for other seagrass species including Halodule wrighti i (Asch.) and Syringodium fili f orme (Kutz) For example, estimates of between 24% 37% of subsurface PAR have been proposed for the Indian River Lagoon along the east coast of Flor ida (Kenworthy and Fo nseca 1996) In 2007, the Tampa Bay Estuary Program began an extensive re evaluation of this light target. This dissertation is a large part of that evaluation. To apply this minimum light target to the entire bay requires accurate delineation of the sea grass deep edge. Since 1988, the Southwest Florida Water Management District (SWFWMD) has been mapping the spatial extent of seagrass along the west coast of Florida, including Tampa Bay, using aerial photography collected roughly every two years. After th e images are georectified and orthorectified they are analyzed by certified photointerpreters who delineate seagrass polygons and classify them as either patchy (>25% of a polygon is unvegetated) or continuous (<25% is unvegetated) (Kurz 2002) S tringent quality contro l measures are used to establish the accur acy of identify ing each polygon with the correct seagrass classification A 90% accuracy rate is required for polygons greater than 0.4 hec tares in size (Kurz 2002) These maps are the basis for tracking long term seagrass coverage in Tampa Bay The 2006 maps are used here to estimate the e xtent of the mapped s eagrass deep edge within the study areas. For the purposes of this study, no distinction is made between the patchy and continuous coverage classifications. It is recognized that seagras s can grow beyond this mapped edge
12 but at densities that are too small to be detected It is estimated that t hese grasses represent a very small percentage of the total seagrass area. Therefore, the mapped edge is considered to be the seagrass deep edge for management purposes and is defined as such here. 1.2. 2 Wavelength specific light utilization for photosynthesis It is important to think of PAR in terms of its flux of quanta as opposed to its energy state because once a quantum has been absorbed by a plant cell, its contribution to photosynthesis is the same regardless of its wavelength specific energy (Kirk 1994; Mobley 1994) Of course a given photon must first be ab sorbed by one of the wavelength specific photosynthetic pigments. The usefulness of a given light field for photosynthesis is not simply a function of the total intensity of PAR, but how well the spectral distribution of PAR matches the abso rption spectrum of a given aquatic macrophyte or phytoplankton (Kirk 1994) Because PAR, by definiti on, is a broadband quantity it does not take into account the spectral variability of light reaching the bott om. Spectral PAR (PAR( )) provides much more information on the shape of the incoming light field than simply measuring PAR. However, n either PAR nor PAR( ) provide any indication that the photons a vailable are actually being absorbed Since seagrass are c the incoming light field it is important to express light qua lity in terms of the specific absorption characteristics of the light harvesting pigments found in seagrass. L ike all higher plants, seagrass are reliant mostly on l ight within the blue and red color bands, though the presence of accessory pigments can increase the operational window beyond the blue and red wavelengths (Figure 1 2 ).
13 Figure 1 2 A bsorptance spectr a and PUR( ) for two species of seagrass Thalassia testudinum ( Banks ex. Knig ) and Halodule wrightii ( Asch .), and bottom PAR( ) measured in Tampa Bay. PUR( ) is the product of the leaf absorptance and PAR( ).
14 Light harvesting p igments a ssociated with photosynthesis include the chlorophylls a b and c and the carotenoids (Kirk 1994) The c hlorophyll s are the primary pigments for light harvesting but the presence of carotenoids expands the absorbing capabilities into the near U V and blue green wavelengths (Kirk 1994) Of the chlorophylls, chlorophyll a is the primary light absorbing pigment with a primary absorption peak centered near 440nm and a secondary peak near 660nm (Figure 1 2 ). Seagrass do not posses antennae pigments capable of efficient harvesting of green light (Cummings and Zimmerman 2003) unlike some species of red and blue green algae that contain green light harvesting billiproteins (Kirk 1994) G reen light should not be thought of as being useless to seagrass however some of the carotenoids can extend the absorption range of seagrass well into the blue green region up to about 560nm (Kirk 1994) Photosynthetically useable radiation (PUR) is a spectrally integrated quantity, defined as the fraction of the radiant energy that can be absorbed by a given plant, in this case seagrass (Morel 1978; Morel 1991) Both PAR and PUR can be expressed in terms of their spectral quantities and are given the symbols PAR( ) and PUR( ), respectively. PUR( ) is calculated by multiplying PAR( ) by some dimensionless quantity that is proportional to the leaf absorption per wavelength (Kirk 1994) This dimensionless quantity is commonly defined by phytoplankton researchers as the ratio of the phytoplankton absorption coefficient to the maximum absorption coefficient, typically at 440nm (Morel 1978; Morel 1991) T ypical ly, seagrass rese archers express this dimensionless quantity in terms of the leaf absorptance (A L ) While pigment concentrations can vary significantly both within and among various seagrass species, leaf optical characteristics are quite similar because of the strong pac kage effect, partially due to the chloroplasts being limited to the leaf epidermis (Cummings and Zimmerman 2003; Durako 2007; Enriquez 2005) A comparison of the relative absorption curves of Thalassia testudinum ( Banks ex. Knig ) and Halodoule writtii ( Asch ) reveal s very little difference in the p eak absorption wavelengths (Figure 1 2 ). This does not mean that the absorption efficiencies are the same. Seagrass photoacclimate to changing irradiance levels by varying leaf pigment concentrations
15 (Abal et al. 1994; Cummings and Zimmerman 2003; Dennison and Alberte 1982; Zimmerman 2003) As a result, c hlorophyll content can vary significantly among different species and habitats in response to low light conditions (Cummings and Zimmerman 2003; Dennison and Alberte 1982; Herzka and Dunton 1997) This acclimation strategy has been found to be largely inefficient due in part to the strong package effect caused by the structural configuration of the leaf tissue restricting the chloroplasts to the epidermal layer (Cummings and Zimmerman 2003) For res ource managers, it is far too complicated to address light quality on a per nanometer basis. It is more practical both operationally and conceptually to combine wavelengths into broad color bands based on the absorption characteristics of the seagrass (Fig ure 1 2 ). In this study, three color regions or bands are defined (Table 1 1). Table 1 1 Selected wavelength ranges of the color bands blue, green, and red and the pigments that are represented by each. Only the red algae, b lue green algae, and the cryptophytes contain billiproteins that allow them to harvest green light. Blue Green Red Wavelength Range 400nm 490nm 490nm 600nm 640nm 690nm Pigments Chlorophyll/Carotenoid Protein Complex Billiproteins Chlorophylls Found in Seagrass Yes No Yes These regions correspond to the absorption peaks associated with the chlorophylls and the carotenoids. There is an inherent danger of over simplifying what is a very complex process by simply binning wavelengths and this issue is e xamined in some detail by comparing these bulk color bands with measurements of PUR( ). 1.2. 3 Measuring transparency in shallow waters Historically light measurements have been rather crude. One of the oldest techniques and still a very common method i s to use a Secchi disk. This approach has
16 several limitations when applied to seagrass management, the most common of which is that often in very shallow waters, Secchi depth is greater than bottom depth. In other words, if one sees the bottom, Secchi disk cannot be used and in most cases, when dealing with seagrass, the bottom is visible. A more quantitative approach commonly employed is to use an irradiance or quanta meter that measures the quantity of light as PAR. Because PAR is a single number, it give s no indication of the spectral quality of the light field. Further, it gives no indication of the amount of light that is useable by seagrass. L imitations in the current methods for measuring water clarity and estimating the attenuation coefficient have necessitated the development of an alternative approach to measuring light Of primary significance is this lack of spectral information. While PAR is a good bulk estimator of the subsurface light field, i solating specific PUR wavelengths allows a much mor e surgical approach to establishing relevant seagrass targets. Advances in technology have made it relatively easy to acquire spectral data in very shallow waters. P resent ed here is a framework for determining not only the quantity of light reaching the bo ttom but also the spectral shape of the underwater light field. With this information a light attenuation coefficient is calculated and the percent subsurface light reaching the bottom is determined A potential source of error in setting light and depth targets is that most light data have been and continue to be collected in deeper waters well beyond seagrass depth limits. The current method widely used for measuring PAR requires a minimum depth of 1.5m. The use of a Secchi disk is an alternative method commonly used in Tampa Bay for estimating water clarity but it is of no use in seagrass beds where Secchi depths are typically greater than the bottom. The method presented here allows for detailed measurements of spectral downwelling irradiance in water depths of less than 0.5m This method also employs a simple tool to determining percent subsurface irradiance at depth at any wavelength or range of wavelengths of interest.
17 1 .3. Materials and Methods 1 .3.1. The Seagrass Management Area concept As part o evaluation, Tampa Bay was subdivided into 30 individual Seagrass Management Area s (SMA) (Figure 1 3) (E PCHC 2007) For this study, four SMAs were selected based on a priori knowledge of the optical properties and historical seagrass coverage. Within each SMA two fixed stations were establis hed from which all in water irradiance measurements were collected. For each SMA except for Egmont Key, a nearshore and an offshore station were established (Table 1 2). Table 1 2 Fixed stations for each SMA where in water i rradiance measurements were collected. Meter marks correspond to the approximate distance from the shoreline. The seagrass Halodule wrightii disappeared from the offshore Wolf Branch and Kitchen sites mid way through the study. Depth is relative to MSL. S MA Strata Meter Mark Depth (m) Seagrass Species Coffeepot Bayou near shore offshore 100 900 0.92 1. 18 T halassia testudinum H alodule w rightii T halassia testudinum Wolf Branch near shore offshore 300 1100 0. 85 1. 50 T halassia testudinum H alodule w right ii H alodule w rightii (disappeared) Kitchen near shore offshore 600 1300 0. 88 1. 16 H alodule w rightii H alodule w rightii (disappeared) Egmont Key near shore 100 1.59 T halassia testudinum The nearshore sites were located within relatively healthy seagras s beds where light limitation was assumed not to be a factor. Originally t he offshore sites were to be located at the deepest extent of seagrass growth where light limitation was likely to be the primary limiting factor. In the Kitchen and Wolf Branch off shore areas, seagrass coverage was extremely sparse during the beginning of the study and had disappeared completely by the end of the study period. Because these SMAs were so shallow, the nearshore sites were actually located along the deep edge of the pe rsistent seagrass bed.
18 Figure 1 3 Seagrass Management Area s of Tampa Bay. T he Kitchen (SMA 5) is located in eastern Tampa Bay (Figure 1 3 ) a nd has an area of approximately 776ha. There have been s ignificant increases in seagrass coverage over the past two decades from approximately 40ha in 1996 to over 142ha in 2007 (HCEPC 2007). The area can be thought of as being hydraulically isolated with the Alafia Banks to the north, spoil islands 2D and 2E to the west, Port Sutton berths to the south, and the shoreline to the east boxing in the area (Figure 1 4) The shoreline is mostly mangrove with some salt marsh. The only direct freshwater inflow is from Bullfrog Creek, a 3km long tidal creek that drains mostly agriculture and s ome urban development Additonal freshwater inflow can come from the Alafia River just to the north. The nearshore and an offshore site were located approximately 600m and 1300m from the shoreline (Figure 1 4) T he offshore site had very sparse Halodule wr ightii ( Asch .) at the beginning of the study, but by October, what little grass there was had completely disappeared. Given the extremely shallow nature of this SMA, the nearshore
19 site actually represents the deep edge of the persistent seagrass bed. Presu mably, light condition s at this nearshore site are representative of the minimum light conditions necessary for seagrass persistence. Figure 1 4 Map showing the spatial extent of seagrass coverage in the Kitchen SMA. Cover age is ba sed on 2006 aerial photography. Wolf Branch (SMA 6) is located along the eastern shore of Middle Tampa Bay and is immediately to the south of the Kitchen (Figure 1 3 ). Wolf Branch is approximately 1554ha, roughly double the size of the Kitchen (F igure 1 3 ). Unlike the Kitchen, seagrass in Wolf Branch have been on a continual decline over the past twenty years, f rom 283ha in 1996 to 162ha in 2006 (HCEPC 2007). Both Wolf Branch and Kitchen are very rich in colored dissolved organic matter (CDOM) and routinely have
20 among the highest chlorophyll concentrations of any SMA. The shoreline is dominated by mangroves and some salt marsh communities but no major creeks or rivers, though several small tidal tributaries are located along the complex mangrove sh oreline. Within this expanse of mangroves are numerous mosquito ditches dug in the 1960s for mosquito and flood control. These ditches provide a direct conduit for surface runoff and may also be a significant conveyance for CDOM rich water. A nearshore and an offshore site were established approximately 300m and 1100m from the shoreline (Figure 1 5) Like the Kitchen, the offshore location at Wolf Branch contained very sparse Halodule wrightii ( Asch. ). at the beginning of the study and disappeared by mid st udy. Also like the Kitchen, the nearshore site is located near the deep edge of the persistent seagrass bed at 0.85m MSL. Figure 1 5 Map showing the spatial extent of seagrass coverage in the Wolf Branch SMA. Coverage is ba sed on 2006 aerial photography.
21 Coffeepot Bayou (SMA 18) is located along the western shore of Middle Tampa Bay and is approximate ly the same size as Wolf Branch Seagrass beds within the Coffeepot Bayou SMA have declined over the twenty years going from 2 43ha in 1997 to 162a in 2007 (HCEPC 2007). Coffeepot Bayou receives large amounts of storm water from the adjacent urban watershed often resulting in high chlorophyll concentrations greater than 30 g L 1 during the rainy season from July through September Because the bayou drains an urban watershed with little to no vegetation and because the seawall shoreline has no salt marsh or mangrove vegetation, Coffeepot Bayou is not thought to be a CDOM rich environment but rather more chlorophyll dominated. A nea rshore and an offshore site were located with the nearshore site located approximately 100m from the shoreline and the offshore site 900m from the shoreline (Figure 1 6) The nearshore site at Coffeepot Bayou was well inshore of the seagrass deep edge and it was assumed that seagrass growing here were not light limited. The offshore site was located near the deep edge of the persistent seagrass bed at a depth of 0.90m MSL.
22 Figure 1 6 Map showing the spatial extent of seagras s coverage in the Coffeepot Bayou SMA. Coverage is based on 2006 aerial photography. Egmont Key (SMA 11) is nearest to the Gulf of Mexico and is adjacent to a small 162ha island that is both a State Park an d a National Wildlife Refuge Given the unique c onditions that exist here, it was assumed that the light conditions would be markedly different from the other three SMAs. This is the smallest of the four SMAs included in this study and covers an area of approximately 518ha. Though small in area, seagras s here are quite healthy growing deep er than in most other areas in Tampa Bay. Over the past twenty years seagrass coverage has increased from 24ha in 1996 to 40ha in 2006 (HCEPC 1997). Seagrass here only extend to about 200m offshore beyond which depths
23 become too great to support seagrass. Given the small aerial extent of the seagrass beds here, only one station was established approximately 100m offshore at a depth of 1.6m MSL (Figure 1 7). Figure 1 7 Map showing the sp atial extent of seagrass coverage in the Coffeepot Bayou SMA. Coverage is based on 2006 aerial photography. 1.3.2. Measuring E d ( ), PAR ( ) and PAR A planar irradiance cosine collector (Hobi Labs, Inc., Bellevue, WA) was mounted onto a PVC measuring rod The cosine collector was then connected via a fiber optic cable to a portable spectrometer (HR2000, Ocean Optics Dunedin, FL). A field laptop PC (Panasonic Toughbook, Panasonic Corporation, New York, NY) running the Ocean Optics program OOI Base32
24 All data were stored on the PC. A planar irradiance cosine collector was chosen over a spherical sensor to remove any inherent bias caused by bottom reflectance. This makes it easier to compare different sites and provides a conservative estimate of E d ( ). Once onsite the boat was anchored using a hydraulic anchor pole instead of a traditional anchor to minimize sediment disturbance. Every effort was made to measure E d always on the sunny side of the boat and as far a measuring rod was kept perpendicular to the water surface to within 5 o of nadir. For each discrete depth, three consecutive scans were taken one second apart and averaged to create a composite scan. Multiple scans were taken to account for any variation caused by waves and movement of the sensor off nadir. Initial measurements were taken in the air just above the water surface followed by a surface reading approximately 0.01m below the water surface. Followin g the surface reading, scans were taken at 0.25m intervals. The maximum scan depth was 1.75m because of limitations in the fiber optic cable length. In reality, depths were never more than 1.50m so this limitation was not an issue. Because E d ( ) is an appa rent optical property, it is dependent on time of day, sun angle, sky conditions, and sea state. These factors are often overlooked in most monitoring programs and over long time periods become less significant. In order to mimic the type of data that woul d be collected during routine monitoring runs, only time of day was considered and an operational window between 1000 and 1400 standard time was set. Some bias toward an incoming or slack high tide was unavoidable giv en the extremely shallow depths in cert ain areas. At each discrete depth the three scans were averaged and then converted first, f rom raw digital counts to E d ( ) in units of W m 2 nm 1 and then, from E d ( ) to PAR( ) in units of photon flux ( mol m 2 s 1 ). A typical depth profile of PAR( ) i s shown in Figure 1 8
25 Eq. 1.1 Figure 1 8 Spectral scans of PAR( ) with depth collected in August 2008 at Wolf Branch in eastern Tampa Bay. The depressions located throughout the curve are called Fraunhofer lines and are caused by the a bsorption of light by the cooler gases in the sun's outer atmosphere at frequencies corresponding to the atomic transition frequencies of these gases. To obtain PAR, PAR( ) was grouped to the nearest nanometer using a linear interpolation procedu re in MATLAB (The MathWorks, Inc., Natick, MA). PAR was then calculated by integrating across the visible spectrum (400nm 700nm) using a trapezoidal integration routine in MATLAB (Kirk 1994; Mobley 1994) : w here is the wavelength (nm), h onstant and c is the speed of light. The loss of PAR( ) with depth for a given wavelength can be described by the diffuse light attenuation coefficient ( K d ( ) ): 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 400 450 500 550 600 650 700 PAR ( mol m 2 s 1 ) Wavelength (nm) 0.1m 0.25m 0.50m 0.75m 1.00m
26 Eq. 1.2 Eq. 1.3 Because K d ( ) is not c onstant with depth, it is more accurate to use the average K d ( ) over a depth interval from 0 to z (Kirk 2003) : T he symbology K d ( ) is used here to indicate the average attenuation coefficient a cross the depth interval 0 to z. A more accurate measure of K d ( ) is to take the slope of a linear regression line fitted to a plot of the natural logarithm of PAR ( ) with respect to depth (Kirk 1994) K d (PAR) is simply the slope of the natural l og of PAR with respect to depth. The same procedure can be used to determine the K d ( ) for PUR( ) instead of PAR( ) The light attenuation coefficient varies with wavelength (Figure 1 9 ) and therefore must only be compared to attenuation coefficients of th e same wavelengths. The wavelength specific nature of K d ( ) can be attributed to the concentration and composition of the constituent absorption and scatter. Figure 1 9 The light attenuati on coefficient as a function of wavelength. Data from the Wolf Branch Seagrass Management Area collected in October 2008.
27 Eq. 1.4 1.3.3. Estimating the percent subsurface PAR ( ) and PUR( ) Using K d ( ) calculated from measured PAR ( ) and PUR( ) at selected locati ons, the percent subsurface irradiance, expressed in terms of either PAR ( ) or PUR( ), can be determined at any depth ( d ) for any given K d ( ). For seagrass management purposes, determining the percent subsurface irradiance at the bottom along the seagrass deep edge, will ultimately determine minimum light targets. Total depth is defined as the total depth of the water column at mean sea level (MSL) This was chosen beca use it represents the average c ondition at a given location. The fraction of surface irr adiance is calculated as: where z is th e total depth in meters at MSL. Multiplying Eq. 1.4 by 100 gives the percent subsurface PAR( ). The percent PAR( ) for the blue, green, and red color bands is determined by using the K d ( ) integrated across the wavelength range for each color band. For calculating percent subsurface PAR, the K d ( ) integrated across the visible spectrum (400nm 700nm) is used. PUR( ) is calculated by multiplying the PAR( ) by the single leaf specific absorptance (A L ( ) ) Single lea f absorptance for a given species is expressed as where Tr ( ) is the transmittance across a single seagrass leaf and represents the ratio of the amount of light that passes through the seagrass leaf to the amount of light incident on the leaf sur face. A PUR( ) is calculated for each wavelength and binned to the nearest nanometer using the same MATLAB linear interpolation procedure as was used for calculating PAR( ). To calculate the PUR( ) for each of the three color regions, PUR( ) was summed acr oss the wavelengths corresponding to the blue, green, and red color bands (Table 1 1) PUR was calculated by summation across the visible spectrum. To determine A L ( ), h ealthy Thalassia testudinum ( Banks ex. Knig ) and Halodule wrightii (Asch.) leaves were harvested from Coffeepot Bayou and scanned onsite using an Ocean Optics DR2000 field spectrometer (Ocean Optics, Inc., Dunedin,
28 FL) and a fiber optic cable connected to a black mounting bracket. All scans were taken near solar noon. Before leaves were sca nned, they were carefully wiped clean of any epiphytes or other particulate matter. Leaves were then place d across the mounting bracket and the bitter end of the fiber optic cable was adjusted until making contact with the leaf surface. The apparatus was p ositioned to face the sun and the spectrometer integration time was adjusted to avoid saturation of the signal. Scans were collected using the same methodology for measuring E d ( ). Immediately after scanning the leaf, a scan of the incoming solar radiation was collected. This procedure was repeated four times for each species using different leaves each time. Results were compared to literature values to ensure consistency with other researchers (Cummings and Zimmerman 2003; Durako 2007; Zimmerman 2003) 1 .3. 4 Mapping percent subsurface irradiance A major limitation in understanding the seagrass light depth relationship is accurate bathyme tric data. Airborne laser bathymetry, also known as Light Detection and Ranging (LIDAR), can help overcome this limitation. LIDAR is a technique for measuring shallow waters using a pulsed laser beam from an airborne platform. This technology has been in u se since the mid 1960s (Hickman and Hogg 1969) At that time, laser technology was brand new and used primarily for anti submari ne warfare by the U.S. Navy. Today airborne LIDAR is routinely used for hydrographic surveys and has evolved into a n accurate operational technique In 2007, selected areas of Tampa Bay were mapped Advanced Airborne Research LID AR (EAARL) (Figure 1 10 )
29 Figure 1 10 C oncept of operations of the NASA Experimental Advanced Airborne Research LIDAR (EAARL) (Image source http://ngom.usgs.gov ) The EAARL system has a maximum measureable water depth of 26m and is determined by the strength of the bottom return signal and water clarity (Brock et al. 2002; Guenther et al. 2000) The minimum operating depth is 30cm and has a nominal depth accuracy of 4.0cm 1.0 (Brock et al. 2002) In Tampa Bay, the average maximum measureable water depth was approximately 2. 6 m MSL (Tyler et al. 2007) In depths greater than 0.5m, LIDAR data were cross checked with depth data collected using a ship borne acoustic system called the System for Accurate Nearshore Depth Surveys (SANDS) (Hansen et al. 2005) The Kitchen, Wolf Branch, and Coffeepot Bayou SMAs wer e mapped using this technique Bathymetry for the Kitchen SMA is very shallow with most of the area less than 0.5m MSL ( Figure 1 11). Depth penetr ation is a function of the optical properties of the water column and the bottom sediment composition (Brock et al. 2002; Guenther et al. 2000) The deep area toward the center of the image is the original Alafia River channel while the deep shaded areas toward the bottom of the SMA are old dredge holes.
30 Figure 1 11 Bathymetry of the Kitchen Seagrass Manag ement Area LIDAR system. Unshaded areas exceed the maximum measureable depth of 2.75m.
31 The W olf Branch SMA sits on a relatively flat shelf with a gradual slope terminatin g approximately 1500m from the shoreline (Figure 1 12). Figure 1 12 Bathymetry of the Wolf Branch Seagrass Management Area LIDAR system. Unshad ed areas exceed the maximum measureable depth of 3.50m. The Coffeepot Bayou SMA is characterized by a large shelf with a relatively sharp break approximately 1000m offshore (Figure 1 13). T he large unshaded area s to the southwest and northeast are dredge d areas greater than 4.00m All of the areas that exceed the maximum measurable depth also exceed the maximum seagrass depth limit.
32 Figure 1 13 Bathymetry of the Coffeepot Bayou Seagrass Management Area Bathymetry was co LIDAR system. Unshaded areas exceed the maximum measureable depth of 4.00. To calculate the percent PAR ( ) f or any location, only the total depth and K d ( ) are necessary For the Kitchen, Wolf B ranch, and Coffeepot Bayou SMAs a GIS based model ing approach was employed to merge the LIDAR bathymetry with site specific K d ( ) and using Eq. 1. 4 to calculate the percent PAR ( ) for each cell in the bathymetric grid. It was assumed that the K d ( ) was th e same throughout a given Seagrass Management Area This simplification was necessary to run the GIS model with the caveat that spatial differences in K d ( ) do exist but that this method is a good first cut in
33 the absence of high resolution attenuation inf ormation. In all cases the K d ( ) used for a given Seagrass Management Area was the average annual K d ( ) for both the nearshore and offshore sites. 1 .4 Results and Discussion 1.4.1. K d ( ) and PAR( ) relationships across SMAs For Coffeepot Bayou, the annua l average K d ( ) for blue light (K d (blue)) was greater at the seagrass deep edge, than at the nearshore site though not statistically significant (ANOVA; p > 0.10) (Table 1 3). This may be a function of differences in residence time and flushing rates in this SMA. The proximity of the nearshore station to relatively deep channels may increase the amount of water flow past this station. While the offshore site sits on the end of a large bar covered with seagrass which may act to impede water flow and increa se residence times. Table 1 3 Annual average light attenuation coefficient for nearshore and offshore stations for the blue color band (400nm 490nm). Stations depths are in parentheses and are relative to MSL. K d (blue) mol m 2 s 1 Standard Deviation Maximum Minimum Coffeepot Bayou Nearshore (0.924m) Offshore (1.18m) 1.37 1.65 0.477 0.526 2.13 (Jun) 2.34 (Aug) 0.959 (Apr) 0.814 (Apr) Kitchen Nearshore (0.880m) Offshore (1.16m) 2.42 1.63 0.833 0.322 3.47 (Dec ) 2.15 (Aug) 1.53 (Jun) 1.34 (Apr) Wolf Branch Nearshore (0.852m) Offshore (1.50m) 2.07 1.40 1.16 0.404 4.33 (Aug) 2.04 (Apr) 1.19 (Jun) 0.846 (Jun) Egmont Key Nearshore (1.60m) 0.917 0.200 1.28 (Aug) 0.708 (Dec)
34 The reverse pattern was ob served in the Kitchen and Wolf Branch SMAs where K d (blue) was significantly higher (ANOVA; p > 0.05) at the nearshore sites relative to the offshore sites (Table 1 3) and was most likely a function of shoreline morphology. T he relatively natural shoreline s of both Kitchen and Wolf Branch are heavily vegetated with mostly mangroves and some salt marsh. The increased attenuation at the nearshore sites is likely a result of increased loads of dissolved and particulate organic material from shore. For the Kitc hen, there is also direct discharge of organic rich waters from Bullfrog Creek and indirect discharge from the Alafia River just to the north. K d (blue) was lowest at Egmont Key where direct mi xing with Gulf of Mexico waters helps to buffer water originatin g from the upper parts of Tampa Bay. Graphically, it is easy to see that light loss with depth occurs at different rates for the different color bands and for PAR (Figure 1 1 4 ). Blue light attenuated much more rapidly than either green or red light in all cases.
35 Figure 1 14 Annual average light loss with depth by color band for each Seagrass Management Area The intersection of the vertical and horizontal lines represents the current minimum light target of 20.5% at the deep seagrass edge. Depths are in meters relative to MSL. On an annual average basis, the percent subsurface PAR for all four SMAs well exceeded the minimum light target of 20.5% (Figure 1 1 4 ). O ver the course of this study, the minimum light target was me t or exceeded for all but three sampling events (Table 1 4 ).
36 Table 1 4 Monthly percent subsurface irradiance at bottom relative to just below the water surface for each representative color band. Data were collected at the deep edge of the persistent seagrass bed. Total depth at each deep edge is in parentheses and is relative to MSL. No data were collected for the Kitchen in December due to technical difficulties. PAR April June August October December Ann ual Average Wolf Branch (0.85m) 40.8 62.0 17.5 53.0 49.7 40.9 Kitchen (0.88m) 46.8 52.0 26.5 34.1 39.8 Coffeepot Bayou (1.18m) 57.1 36.1 15.5 28.8 32.5 30.1 Egmont Key (1.60) 33.5 40.5 18.3 34.4 39.8 32.1 BLUE April June August October December Annu al Average Wolf Branch 19.5 36.4 2.51 27.7 30.6 17.1 Kitchen 23.4 26.0 5.25 16.1 17.7 Coffeepot Bayou 38.2 8.93 6.30 14.9 23.1 14.2 Egmont Key 24.1 29.7 13.1 22.4 32.4 23.3 GREEN April June August October December Annual Average Wolf Branch 48.0 71 .9 19.1 61.7 57.9 47.2 Kitchen 54.4 60.6 30.7 39.8 46.4 Coffeepot Bayou 67.2 26.9 18.6 34.7 38.8 32.8 Egmont Key 42.1 50.6 22.7 43.7 50.3 40.2 RED April June August October December Annual Average Wolf Branch 43.7 63.3 26.8 54.8 50.0 45.8 Kitchen 4 9.1 56.2 35.6 39.3 45.0 Coffeepot Bayou 51.9 23.0 16.9 28.3 29.3 26.6 Egmont Key 26.9 31.6 14.3 27.6 30.1 25.1 August percent subsurface PAR for Wolf Branch, Coffeepot Bayou, and Egmont Key were 17.5, 15.5, and 18.3, respectively while Kitchen was 2 6.5. This could lead to the incorrect conclusion that seagrass along the deep edge at the Kitchen site were the least light limited. However, the percent subsurface blue light for August tells a different story. In August, the Kitchen had the second lowest value at 5.25. Wolf Branch had the lowest value at 2.51 and Egmont Key had the greatest value at 13.3, followed by Coffeepot Bayou at 6.30. August percent subsurface irradiance for green and red light while lower than any other month, were greater than pe rcent subsurface blue light. This suggests that blue light was the limiting factor for seagrass during the month of August. T he minimum light target for Tampa Bay was based on annual average PAR (Dixon and
37 Levero ne 1995) Annual average percent subsurface PAR at the deep edge, during this study, ranged from 30.1 40.9. This range is 10 % 20 % higher than other estimates (Bortone 2000; Dennison and Alberte 1982; Duarte 1991; Kenworthy et al. 1993; Steward et al. 2005) though Kenworthy, et al (1993) reported light requirements as high as 37% in seagrass beds of Northeastern Saudi Arabia. Most estima tes do not take into consideration epiphyte load. Dixon (2000) reported that average annual e piphyte attenuation in Tampa Bay accounted for 32.0% 36.5% and while the minimum light target of 20.5% subsurface PAR may be appropriate for healthy seagrass wit h light epiphytic loads but where loads are moderate to heavy, the amount of light needed may be greater (Dixon 2000) While no attempts to quantify the epiphyte loads on the seagrass in this study were made, qualitative observations were taken and suggest that epiphyte loads were heavier in the Kitchen and Wolf Branch SMAs along eastern Tampa Bay and less so in Coffeepot Bayou and Egmont Key. Minimum light target estimates also do not take into account pulsed events such as turbidity plumes like those documented in the Gulf of Carpentaria in northern Australia (Longstaff and Dennison 1999) In Tampa Bay, especially in eastern Tampa Bay, pulsed events may be more likely to cause high colored disso lved organic matter (CDOM) conditions. Heavy rain events during summer months may also bring pulsed nutrient loads that can lead to increases in phytoplankton biomass. Most likely, the drastic decrease in both percent subsurface PAR and blue light in Augus t was the result of a pulsed rain event. The relatively moderate values in percent subsurface red light indicate that CDO M and/or detritus may have been the dominant light attenuators during this dark water event Whatever the cause, the sharp decrease in percent subsurface PAR and blue light indicates that this was not a localized event. Rainfall plays a major role in regulating the light field in Tampa Bay either in terms of direct runoff or increases in river discharge. H eavy rainfall during the 30 days preceding the August sampling was likely responsible for the anomalous low light conditions. For Kitchen and Wolf Branch, the rainfall amount recorded at Tampa International Airport during the 30 days prior to sampling was 29.5cm For Coffeepot Bayou and Egmont Key, rainfall amounts recorded at the St. Petersburg Airport were
38 24.3cm and 16.4cm, respectively. More rain fell prior to the August sampling than for any other sampling. When the 30 day average rainfall is plotted against the monthly light attenua tion coefficients for blue light, there is a correlation and this correlation appears to be stronger for the Kitchen and Wolf Branch than for Coffeepot Bayou or Egmont Key (Figure 1 1 5 ). Figure 1 15 Relationship between and the 30 day running average for rainfall for each Seagrass Management Area Rainfall data were taken from the nearest ASOS weather station. Rainfall data source: National Weather Service, NOAA. One anomalously high re ading occurred in December for the Kitchen and was most likely a function of wind driven sediment re suspension coupled with a very shallow measured depth (0.25m). A strong north wind following the passage of a cold front exacerbated the already shallow co nditions by pushing water offshore, further adding to the likelihood of wind driven re suspension and shallow depth as the likely cause of this anomaly.
39 Blue light made up only 5.3% of total PAR reaching the bottom at the seagrass deep edge for Wolf Branch during the relatively wet month of August whereas during a June dry period, blue light made up 13% of the total PAR (Table 1 5 ). Both green and red light were similar irrespective of wet or dry weather conditions. Green light made up about half of the to tal PAR at the bottom for all SMAs irrespective of rainfall. The percent of red light that made up PAR was less in wet periods relative to dry periods but only slightly (Table 1 5 ). Table 1 5 Proportion of the blue, green, a nd red color bands relative to total bottom PAR at the seagrass deep edge during wet and dry periods for each SMA. Rainfall is the total amount for the 30 days prior to sampling. Rainfall (cm) Blue (400nm 490nm) Green (490nm 600nm) Red (640nm 690nm) Co ffeepot Bayou Wet Dry 24.3 7.65 11 18 51 48 20 16 Kitchen Wet Dry 29.6 2.28 6.7 13 50 48 26 20 Wolf Branch Wet Dry 29.5 3.45 5.3 15 48 48 29 19 Egmont Key Wet Dry 16.4 2.08 18 22 51 47 14 15 Despite the fact that there appears to be plenty of light based on the percent subsurface PAR at bottom, it is probable that the se grasses are blue light limited For resource managers, relying only on PAR measurements without having any information about the spectral properties of the light field could lead to the wrong conclusions. 1 .4. 2 PUR( ) relationships While a photon may be available for photosynthesis, there is no guarantee that it will be used. The usefulness of a given photon is dictated by the plant of interest. In this ld for seagrass management
40 should focus on those usable wavelengths. In this study, the spectral regions for the usable color ranges of blue and red were estimated using spectral absorption ranges for seagrass light harvesting pigments found in the literat ure (Cummings and Zimmerman 2003; Durako 2007; Zimmerman 2003) While this is a good first approximation, a more sophisticated way to dete rmine spectral significance is through the use of PUR( ). The poorest light conditions occurred during the month of August across all SMAs and corresponded to a relatively rainy period. Under these worst case conditions PUR( ) approached zero at 440nm in creasing to near 1.0 mol m 2 s 1 at 490nm (Figure 1 1 6 ).
41 Figure 1 16 PUR( ) curves for each of the four Seagrass Management Area s sampled in this study under wet conditions and dry con ditions. The small rise in PUR( ) centered between 490nm and 500nm is likely absorption by the accessory chlorophyll b and c as well as the carotenoids. A second, more gradual rise, in PUR( ) between 550nm and 700nm is likely a function of chlorophyll c a nd perhaps to a lesser extent chlorophylls a and b There was a more pronounced rise in the
42 yellow to red region for the Kitchen (Figure 1 1 6 ). This same pattern was seen under the best light conditions though the magnitudes of PUR( ) was much greater (Tab le 1 6 ). Table 1 6 PAR and PUR for each SMA under wet and dry conditions for each representative color band. All values are in units of mol m 2 s 1 Coffeepot Bayou Kitchen Wolf Branch Egmont Key Wet Dry Wet Dry Wet Dry Wet Dry PAR (400nm 700nm) Blue (400nm 490nm) Green (490nm 600nm) Red (640nm 690nm) 250 28.1 127 49.4 1060 187 513 173 513 34.5 258 131 1320 177 629 266 255 13.4 123 74.4 1130 174 540 212 354 65.3 181 49.9 859 193 403 127 PUR (400nm 700nm) Blue (400nm 490nm) Green (490nm 600nm) Red (640nm 690nm) 161 27.2 58.3 43.9 680 181 236 154 348 33.4 118 117 858 171 289 236 179 13.0 56.3 66.0 734 168 248 188 223 63.1 83.2 44.3 566 186 185 113 The seagrass growing along the deep edge in Tampa Bay ar e blue light limited as evidenced by the sharp decrease in PUR( ) from 490nm to 400nm (Figure 1 1 6 ). Peak P U R( ) for the Kitchen and Wolf Branch SMAs was located in the red color region and not the blue region suggesting that these grasses are acclimated to absorbing red light CDOM rich water removes m ost of the blue light while much of the red light remains intact, although absorption due to water becomes significant at longer wavelengths. Average annual percent subsurface red light ranged from a maximu m of 45.8 at Wolf Branch to a minimum of 25.1 at Egmont Key (Table 1 4 ) suggesting an ample supply of red light even when light conditions are minimal. While l eaf absorptance is minimal at 550nm it is not zero. In fact, there is a significant amount of ab sorptance occurring in the green region (Figure 1 2). PUR( ) for the green region under the best light conditions during this study was as high as 289 mol m 2 s 1 measured in the Kitchen during a June dry spell (Table 1 6 ). Even under low light conditions a PUR( ) for the green region was 118 mol m 2 s 1 measured in Wolf Branch. It is evident that there is absorption taking place in the green color region and
43 further study is needed to isolate the pigments responsible and to understand the physiological mechanisms behind this apparent acclimation The light harvesting pigments largely responsible for absorption in the blue green region are the carotenoids but it is not clear what percentage are acting as photo protective pigments. Along the deep edge it i s doubtful that leaves have much in the way of photo protective pigments. One supposes that most production will be in the form of light harvesting pigments for photosynthesis. 1 .4. 3 Mapping s pectral light and depth targets Seagrass in Coffeepot Bayou are depth limited as evidenced by the relatively sharp shelf break at the offshore terminus (Figure 1 13 ). Most of the seagrass beds along the flat shelf behind this slope are classified on the 2006 seagrass map as continuous (Figure 1 6 ). Visually, these grasses appear to be in good health and are persistent. Depths along the shelf break are between 1. 2 5m 1.5 0 m MSL. T he annual average of 1.51 m 1 was used to map the percent subsurface PAR for the Coffeepot Bayou SMA (Figure 1 17).
44 Figure 1 17 Percent subsurface PAR reaching the bottom for Coffeepot Bayou calculated using the annual average K d (PAR) and LIDAR bathymetry.
45 Extracting the percent sub surface PAR reaching the bottom along the mapp ed seagrasss deep edge yielded an average of 31.7 6.7 (Table 1 7). Table 1 7 Total measured depth, percent subsurface blue light, and percent subsurface PAR relative to surface conditions along the mapped seagrass deep edge These data were extracted using a sub routine in GIS in which percent subsurface irradiance was collected for each pixel that fell along the seagrass deep edge. % Blue Light % Green Light % Red Light %PAR MSL Depth (m) Coffeepot Bayou Mean standar d deviation Median Max Min 13.6 3.5 13.8 24.1 4.5 48.5 5.1 48.9 87.15 26.26 30.7 7.3 31.5 43.6 13.6 31.7 6.7 32.7 43.6 15.5 1.21 0.30 1.14 2.05 0.82 Kitchen Mean standard deviation Median Max Min 17.3 2.7 17.4 29.6 11.0 44.5 3.2 44.8 57.15 36.26 40.7 2.72 40.3 51.8 35.0 38.7 3.25 38.9 51.8 30.3 0.87 0.08 0.86 1.09 0.60 Wolf Branch Mean standard deviation Median Max Min 18.1 4.9 18.0 27.8 18.0 45.1 5.6 45.4 55.0 34.4 40.7 5.7 40.9 50.9 29.8 38.9 5.8 39.1 49.2 28.1 0.97 0.15 0.95 1.29 0.72 By contrast, the percent of blue light reaching the bottom was only 13.6% (Figure 1 18). The percent of green and red light reaching the bottom was 48.5% and 30.7%, respectively (Table 1 7).
46 Figure 1 18 Percent subsurface blue light reaching the bottom for Coffeepot Bayou calculated using the annual average K d (blue) and LIDAR bathymetry. Using the annual average K d (PAR) of 0.985 m 1 for Wolf Branch yielded a percent subsurf ace PAR along the mapped seagrass deep edge of approximately 39% (Table 1 7) For Wolf Branch the 30% 40% range for subsurf ace PAR was located approximately 250m offshore (Figure 1 1 9 ) and corresponded to the mapped seagrass deep edge (Figure 1 5)
47 Figure 1 19 Percent subsurface PAR reaching the bottom for Wolf Branch calculated using the annual average K d (PAR) and LIDAR bathymetry. T he modeled percent subsurface blue light at bottom was 17% alon g the mapped seagras s deep edge using a K d (blue) of 1.78m 1 (Figure 1 20 ). Near the offshore sample station 1300m from the bank, is the approximate location of the minimum light target of 20.5% (Figure 1 20). Based on this target, there should be seagrass at this offshore st ation. While there were very sparse seagrass at the beginning of this study, after six months, the few shoots that were there had disappeared. The percent subsurface blue light along this same 20.5% PAR line was between 0% 10% further supporting the
48 hyp othesis that this is a blue light limited environment. Using only percent subsurface PAR without any information of the amount of blue light reaching the bottom would lead to the conclusion that this area is meeting its minimum light requirement. Fig ure 1 20 Percent subsurface blue light reaching the bottom for Wolf Branch calculated using the annual average K d (blue) and LIDAR bathymetry.
49 Spatial patterns in the Kitchen S MA were similar to those found at Wolf Branch (F igure 1 21). P ercent subsurface PAR at the deep edge was 38.9 % using an average K d (PAR) of 2.07m 1 Figure 1 21 Percent subsurface PAR reaching the bottom for the Kitchen calculated using the annual average K d (PAR) and L IDAR bathymetry.
50 For the Kitchen, where there were persistent seagrass (Figure 1 4), the percent subsurface blue light had a range of 20% 40% (Figure 1 22). Like Wolf Branch, at the beginning of the study, the offshore station had very sparse seagrass that disappeared during the rainy season. At this location, the percent subsurface blue light was approximately 10% (Figure 1 22) while the percent subsurface PAR was near 25%. Again this further suggests that these areas are blue light limited and that site suitability based solely on the minimum subsurface PAR c ould be misleading. Figure 1 22 Percent subsurface blue light reaching the bottom for the Kitchen calculated using the annual average K d (blue) and LIDAR bathymet ry.
51 1 .5. Conclusions In the context of the existing minimum light target of 20.5% PAR, the seagrass deep edge should not be light limited. Based on the results presented here, this target may be too low and should be increased to 30% or higher While th is may seem like more than enough light, it is important to remember that much of the photosynthetically useable blue light has been attenuated by the time the light reaches the bottom. Results presented here suggest that a target of 20% may actually be an appropriate blue light target. Because there were no significant differences in either K d (blue) or K d (PAR) for Coffeepot Bayou, Kitchen, and Wolf Branch, applying a bay wide target may be appropriate and the need to develop Seagrass Management Area specif ic targets unnecessary. EAARL system greatly enhances the ability to model the spatial distribution of light. Unfortunately such data are few and far between but none the less imp erative to accurate model development. The GIS based model presented here is an effective tool to quickly asse s s the status of the subsurface light field on an area wide basis and does not necessarily need such high resolution bathymetry What is necessary is an accurate understanding of the spatial and temporal variability in PAR ( ) and K d ( ). The framework presented here is relatively straightforward and can quantify the spectral properties of the subsurface light field in a way that is cost effective and can be readily integrated into existing water quality monitoring programs. Because this system is designed to work in very shallow waters makes it an ideal tool for seagrass applications. While understanding t he spectral characteristics of the light fiel d is a critical first step i t is not enough simply to know how much light is there at the bottom It is equally important to understand the underly ing causes of light attenuation and if these causes can be managed to facilitate improvements in water clari ty and ultimately to provide an environment suitable for seagrass recovery and growth.
52 Chapter 2. Relativ e C ontribution and M agnitude of Phytoplankton CDOM, and D etritus Absorption to the T otal A bsorption C oefficient in Shallow S eagrass Areas 2 1. Abst ract The quality of light plays a major role in limiting the distribution of seagrass. In sub tropical estuaries like Tampa Bay, seagrass are blue light limited along the deep edge. But it is not enough simply to know this. To effectively manage seagrass, the causes of blue light attenuation must also be u nderst oo d The t otal absorption of light is a function of phytoplankton, colored dissolved organic matter (CDOM) and detri tal material. Historically, Tampa Bay has been considered a chlorophyll dominated estuary and seagrass management efforts have focused on chlorophyll as the primary target for increasing light along the seagrass deep edge by reducing the total nitrogen load into the bay However, w hile bay wide chlorophyll concentrations have decrease d over the past decade there has been no major expansion of seagrass into deeper areas. Mounting evidence suggests that the major light attenuator in these shallow seagrass areas is not chlorophyll but CDOM T his hypothesis was tested in selected Seagrass Management Area s (SMA) by comparing the relative contribution and magnitude of the various components of the total absorption. Results confirmed that CDOM is the major absorption component, and at 440nm accounted for an average of 60% of the total absorpt ion. Detrital absorption at 440nm accounted for an additional 20% leaving only 20% o f the total absorption attributable to phytoplankton absorption The magnitude of the absorption coefficients varied as a function of the 30 day running average for tota l rainfall. The correlation between rainfall and was directly related to distance up the bay from the mouth. Rainfall and corre lated with rainfall for all SMAs, though the magnitude of increased with increasing distance up the bay from the mouth It is important to i nfer past light conditions in terms of the inherent optical properties (IOP). Predictor equations were developed relating chlorophyll a to turbidity to the scatter coefficient s and and PCU color to The correlati on coefficients were relatively strong ranging from 0.68 to 0.89. An empirically derived spectral attenuation model was used to relate the IOPs with the light attenuation
53 coefficient using the equation where is the cosine of the solar zenith angle, and is a coefficient that determines the relative effect of scattering on the total attenuation of irradiance. SMA specific and were used to generate modeled which agreed well with measured While this model was originally calibrated for the turbid waters of San Diego Harbor, it work ed fairly well in Tampa Bay.
54 2 .2 Introduction Chapter 1 explored the spectr al attenuation of light in four Seagrass Management Area s (SMA) in Tampa Bay and found that seagrass along the deep edge are blue light limited. This chapter explore s the root ca uses of blue light loss in shallow SMAs of Tampa Bay by comparing the relative contribution and magnitude of the inherent optical properties for each area. Because the inherent optical properties are independent of the understanding the radiance distribution of the underwater light field. An empirical optical model is employed to couple these inherent optical properties with the light attenuation coefficient (Kd( )), a term derived from the apparent optical properties. Light limitation is the fact or determining seagrass depth distribution in many subtropical estuaries like Tampa Bay. Seagrass provide critical habitat for many commercially, recreationally, and ecologically important species of fish (Bortone 2000; Hemminga and Duarte 2000; Hill 2002; Kenworthy et al. 1993; Zieman and Zieman 1989) Seagrasses in Florida provide juvenile nursery and adult feeding areas for red drum, spotted s eatrout, spot, silver perch, sheepshead, snook, shrimp, and the bay scallop (Zieman and Zieman 1989) Light quantity reaching the bottom is often expressed in terms of the photosynthetic ally available radiation (PAR ) (Morel 1978; Smith and Baker 1978) Photosynthetically useable radiation (PUR) is a spectrally integrated quantity, defined as the fraction of the radiant energy that can be abso rbed by the light harvesting pigments (Morel 1978; Morel 1991) Both PAR and PUR can be expressed in terms of their spectral quantities and are given the symbols PAR( ) and PUR( ), respectively. PUR( ) can be thought of as a measure of the quality of light relative to a given target species. Plant pigments typically absorb at blue green wavelengths whereas the maximum transparency of most Tampa Bay waters occurs at the greenish yellow wavelengths, consistent with the color of CDOM rich waters. Models based upon white light concepts such as PAR do not allow calculation of the photosynthetic ally useable radiation (PUR ) (Carder 1995; Morel 1978; Smith and Baker 1978) that seagrass require to thrive or even survive (Zimmerman 2003)
55 To effectively manage these seagrass systems it is not enough simply to know the amount of PAR or even the quality of light (PUR) reaching the bottom but also the causes of light loss. As light propagates down through the water column its attenuation is governed by a combination of absorption and scatter (Kirk 1994) The primary causes of water column absorption are phyt oplankton, colored dissolved organic matter (CDOM), and detrital material (Kirk 1994; Mobley 1994) 2.2. 1 The i nherent optical properties The term inherent optical property (IOP) is rooted i n radiative transfer theory with equations providing the theoretical framework for predicting and interpreting und erwater light fields in terms of the physical, chemical, and biological constituents of natural water bodies (Mobley 1994) These properties refer to those intrinsic properties of the aquatic medium wh ich are dependent solely on the radiance distribution and independent of the am bient conditions at the time of measurement (Kirk 1984; Mobley 1994; Preisendorfer 1961) The IOPs include the absorption coefficient s, the scattering coefficient s and the beam attenuation coefficient s (Table 2 1). By contra s t, the apparent optical properties (AOP) depend both on the radiance distribution of the aquatic medium and the ambient light conditions a t the time of sampling (Kirk 1984; Mobley 1994) Typical AOPs include downwelling irradiance, the diffuse attenuation coefficient and photosynthetically active radiation (PAR) (Table 2 1).
56 Eq. 2.1 Table 0 1 Common parameters associated with the inherent and apparent optical properties of water. Parameter Symbol Unit Inherent Optical Properties Absorption coefficient a( ) m 1 Scattering coefficient b( ) m 1 Beam attenuation coefficient c( ) m 1 Apparent Optical Properties Diffuse attenuation coefficient K d ( ) m 1 Downwelling irradiance E d ( ) W m 2 Photosynthetically active radiation PAR mol m 2 s 1 Absorption of light by constituents in the water column is the most common cause of light loss with depth. In hydrologic optics, the parameter most commonly used to describe absorption is the spectral absorption coefficient (Mobley 1994) The total spectral absorption coefficient for a giv en wavelength can be expressed as a function of its component parts Where is the total absorption coefficient, is the absorption by water, is the absorption by phytoplankton chlorophyll, is the absorption by CDOM, and is the absorption by detrital material. Of these components, can be treated as a constant because it is dependent on the molecular properties of pure water. The contribution of to is s mall in the blue and the green regions of th e visible spectrum but increase s exponentially above 550nm (Kirk 1994) (Figure 2 1) For photosynthetically useable blue light, is insignificant relative to , and F or example, (Morel and Prieur 1977; Pope and Fry 1997; Smith 1981) while typical absorption coefficients for phytoplankton, CDOM, and detritus in Tampa Bay range anywhere from 0.10 m 1 o greater than 1.0 m 1
57 Figure 2 1 Typical phytoplankton, CDOM, and detrital absorption spectra measured in Tampa Bay, FL. Pure water absorption spectrum is taken from Pope and Fry (1997). A bsorption by phytoplankton is accomplished primarily by p hotosynthetic and photoprotective pigments (Bricaud et al. 1995; Bricaud and Stramski 1990; Gordon 1983; Kirk 1994) While there are many varieties of l ight harvesting pigments found in phytoplankton each with unique absorption spectra, all phytoplankton contain chlorophyll a (Kirk 1994) In living phytoplankton cells, almost all of the chlorophyll and most of the carotenoids in the chloroplasts are complexed to proteins (Kirk 1994) There are two broad absorption peaks for The primary peak is in the blue region centered near 440nm with the s econdary peak located in the red region near 660nm (Figure 2 1). The spectral shape of phytoplankton absorption is a function of the absorption characteristics of chlorophyll a a nd the accessory pigments chlorophyll b and c the carotenoids and the billiproteins. B illiproteins are chloroplast pigments found in certain types of red and blue green algae (Kirk 1994; Rowan 1989) These pigments extend the absorption window into the yellow and green wavelengths. For exam ple, the absorption maxima for p hycoerythrin a billiprotein, lies between 490nm and 600nm and between
58 Eq. 2.2 490nm and 640nm for the billiprotein p hycocyanin (Kirk 1994) As higher plants, seagrass species do not have billiprotei ns and thus are not capable of harvesting much light in the green region. In many subtropical estuaries like Tampa Bay, can be a significant contributor to CDOM is operationally defined as that component of the total dissolved organic matter pool th at absorbs light over the vis ible and ultraviolet spectrum (Coble 200 7; Conmy 2008) There are many different names for CDOM in the literature including gelbstoff, gilvin, yellow matter, and chromophoric dissolved organic matter. The definition of CDOM can be further broken down into its component parts. Organic matter is any material that contains carbon and hydrogen and is of biological origin. An operational definition for dissolved is the mechanical separation of water samples using filtration, centrifugation, or other techniques to remove particles larger than some m inimum diameter (Coble 2007) Often 0.2 m is used as the operational cutoff between dissolved and particulate constituents (Hansell and Carlson 2002; Twardowski et al. 2004) T he term colored refers to the optical properties giving CDOM its characteristic yellow, or iced tea, color. CDOM absorbs and fluores c es in the ultraviolet to blue wavelengths. CDOM has been shown to play a major role in light attenuation, even in clear, open ocean waters like the Sargasso Sea (Siegel and Michaels 1996a) CDOM is the principal light absorbing constituent of the DOM pool in seawater (Blough and Vecchio 2002) Due to its complex nature, absorption spectra are broad and unstructured. CDOM absorption spectra decrease exponentially with increasing wavelength in the range 300nm 700nm. Absorption magnitudes vary significantly and increase with increasing proximity to terrestrial sources. Typically CDOM absorption increases along a continuum from open ocean to river water The absorption spectrum for can be approximated with the followi ng equation : w here is the CDOM absorption coefficient at wavelength is the CDOM absorption coefficient at some reference wavelength and is the spectral slope.
59 Eq. 2.3 Differences in can be indicative of CDOM origin such that lower slopes are typical of freshwater and coastal environments while steeper slopes are more indicative of offshore photobleached marine waters (Coble 2007; Hansell and Carlson 2002) Spectral slope varies depending o n the wav elength range used to calculate the slope (Coble 2007; Hansell and Carlson 2002; Stedmon et al. 2000; Stedmon 2003) Detrital material is esse ntially the non living particulate portion of the total organic matter pool that is left behind after a water sample has been passed through a 0.2 m filter. Because the chemical compounds that make up the detrital fraction are similar to those found in CDO M, the spectral shape of is very similar to that of though typically at much lower concentrations (Figure2 1) I n shallow coastal lagoons wind driven suspension of particulate matter can be a significant control on light attenuation (Lawson et al. 2007) Understanding the behavior of light requires knowledge of the scattering properties of the water (Kirk 1981) At the most fundamental level, scattering of light arises from interacti ons between photons and molecules (Mobley 1994) In natural waters, sc atter is dependent on particle size, phytoplankton species, particle mineralogy, detritus composition, and p article concentration (Coble 2007; Weidemann and Bannister 1986) T he total scattering coefficient can be calculated as: where is the total spectral absorption coefficient and is the sum of the component absorption coe fficients as stated in Eq (2 1) and is the total spectral beam attenuation coefficient in units of m 1 The probability of a photon travelling along a specific path length before interacting with some component within the water column, either through absorption or scattering is governed by (Kirk 1981). The beam attenuation coefficient i s calculated by in water measurement of the beam transmittance of a specific wavelength of light across a fixed pathlength.
60 2.2. 2 Seagrass light relationships in Tampa Bay Phytoplankton p roductivity is directly related to nutrient loads originating from terrestrial runoff, sediment re suspension, internal cycling, and direct atmospheric deposition. The m ost cited cause of seagrass decline in coastal systems including Tampa Bay and Charlotte Harbor, is anthropogenic nutrient enrichment (Janicki Environmental 2001; Tomasko et al. 1996) Pulsed events like hurricanes, regulated discharges from nutrient rich systems, or accidental nutrient releases can provide the catalyst for high phytoplankton productivity significantly increasing light limitation for seagrass Under the current Tampa Bay nitrogen managem ent strategy, seagrass management has been predicated on the assumption that phytoplankton chlorophyll is the major light attenuator limiting seagrass depth distribution (Figure 2 2 ) (Janicki Environmental 2001) Figure 2 2 Tampa Bay nitrogen management strategy as it relates to light and seagrass sustainability. This assumption was based on empirical relationships bet ween water quality and light attenuation from an impressive monthly data set collected over several decades by the Hillsborough County Environmental Protection Commission as part of their ambient monitoring program. However, almost all of the stations are located well offshore from existing seagrass beds. Results from a two year study in Old Tampa Bay concluded that light attenuation was greater in shallow seagrass beds immediately along the shoreline than f u rther offshore and that CDOM absorption has more of an impact than previously believed (Griffen and Greening 2004) the same basic conclusion in the Kitchen and Wolf Branch areas of the bay (Joha nsson,
61 personal com munication). In ocean water, CDOM is one of the strongest absorbing constituents often exceeding absorption by phytoplankton in the blue color region (Hoge et al. 1993) Not only can CDOM reduce the PAR and PUR of blue light, it also degrades the accuracy of chlorophyll concentration by satellite color sensors (Carder et al. 1989) CDOM can come from a variety of sources and sinks in Tampa Bay (Table 2 2). While t he primary source is via in situ biological production the major source in estuaries like Tamp a Bay is from freshwater inflow from rivers and streams such as the Hillsborough River and Alafia River. Along the immediate shoreline, mangrove swamps and salt marshes can also be an important source of CDOM. Groundwater could also be a significant sou rce of CDOM along coastal and estuarine areas where groundwater inputs exist either as discrete springs like Crystal Beach Spring off Pinellas County, or as diffuse groundwater discharge through sediments. CDOM sinks in estuaries like Tampa Bay are mostly thr ough direct export via tidal mixing and flushing. Photobleaching is the dominant process for the degradation of CDOM in shallow oceanic waters (Conmy et al. 2004; Siegel and Michaels 1996b; Warrior and Carder 2005) but in estuarine systems residence times are sufficiently short that CDOM is exported before major degradation can occur. Table 0 2 CDOM sources and sinks in Tampa Bay summarizing inputs and outputs. Sources Sinks Major Rivers Hillsborough, Manatee, Alafia Minor Rivers and Canals Bullfrog Creek Direct Runoff Stormwater Coastal Vegetation Mangrov es, Salt marsh In situ Biological Production Water column In situ Seagrass / Macroalgae Direct Groundwater Sediment flux Direct Export Tidal mixing and Flushing Photodegradation Biodegradation Detrital material can also play a major role in the absorption of light an d has been correlated with wind driven re suspension of organic particles (Lawson et al. 2007; Steward and Green 2007) If this is the case in the shallow seagrass areas of Tampa Bay, the current seagrass management paradigm should be modified to include the effects of detritus and CDOM absorption on the light field (Figure 2 3).
62 Figure 2 3 Modified Tampa Bay seagrass management strategy that includes spectral absorption by CDOM and detritus. 2.2.3 Coupling the IOPs with the quasi IOPs and K d ( ) While IOPs are considered to be the gold standard for understanding the optical properties of the water, most resource management agencies do no t collect these data and little historical IOP data exist. Today many agencies are beginning to appreciate the importance of understanding the optical properties of the water as it relates to resource management and have begun to collect IOP S tudy G roup and the Hillsborough County Environmental Protection Commission now routinely collect CDOM absorption as part of thei r monitoring programs (R. Johans son, personal communic ation). Nevertheless there is wealth of historical data th at could be use ful to infer historical IOPs. This can be accomplished using empirical relationships between the IOPs and historical quasi IOPs such as chlorophyll a concentration, water color in platinum cobalt units and turbidity in nephelometric turbidity units The t erm quasi IOP is used to indicate the close relationship s between these parameters and the various absorption and scattering coefficients.
63 The Apparent Optical Properties ( AOP ) define the light field within the water column and are parameters routinely mo nitored by limnologists and oceanographers (Kirk 1981) For resource managers, is the most relevan t metric that arises from the AOPs and is the fundamental metric for setting seagrass light and depth targets. To relate with the IOPs provides much more insight into the root causes of light loss with depth. Historically there has been a great deal research trying to couple the AOPs with the IOPs (Berwald et al. 1995; Berwald et al. 1998; Kirk 1984; Kirk 1994) Because of the multiple scattering that takes place in natural systems and the variability in the angular distribution of the light field, there is no analytical expression to directly calculate from the IOPs (Kirk 1981) The IOP s however, can be used to specify the probability of absorption and scatter occurring on an individual photon basis. This study was designed to investigate the relative contribution and magnitude of the IOPs for selected Seagrass Management Area s (SMA) in Tampa Bay. Based on the results presented in Chapter 1, s eagrass depth distribution in Tampa Bay is limited by blue light (400nm 490nm). To effectively manage seagrass along the deep edge, the root causes of blue light loss must be better understood. If C DOM is the dominant attenuator of blue light in these areas it further challenges the long standing seagrass management paradigm that phytoplankton chlorophyll is the primary light attenuator and that reductions in nitrogen loads will have a significant i mpact in reducing chlorophyll concentrations significantly enough to increase water clarity (Janicki Environmental 2001) Also e mpirical relationships between the IOPs and the quasi IOPs for these shallow water areas were constructed to couple the IOPs with parameters more commonly measured in resource management and regulatory programs. Finally the total absorption and scattering coefficients are coup led with the light attenuation coefficient at 480nm using the equation developed by Kirk (1981) and modified by Kirk (1991). 2 .3 Methods and Materials 2 .3.1. Site Locations Tampa Bay has been subdivided into 30 individual Seagrass Management Area s (SMA) (Figure 2 4 ) The SMA concept was proposed by the Tampa Bay Estuary
64 Program and is based on several factors such as historical seagrass patterns, water quality, sediment type, and watershed land use (E.P.C.H.C. 2007) A detailed description of each SMA may be found in Chapter 1. Figure 2 4 Seagrass Managemen t Area s in Tampa Bay and the four areas used in this study. The Kitchen SMA is located in eastern Tampa Bay and has a well developed mangrove shoreline (Figure 2 5). This is a CDOM and chlorophyll rich area characterized by average water depths of less th an 2.0m mean sea level (MSL). Direc t river discharge comes from B ullfrog Creek, a 3km long tidal creek that drains mostly agricu lture and some urban developed areas A second source of river water come s from the Alafia River just to the north of the Kitche n Discharge from these rivers is typically maximu m during the height of the rainy season during the months of August and September.
65 Figure 2 5 Mangroves dominate the shoreline in (a) the Kitchen and (b) Wolf Branch S eagrass Management Area s. The creek near the center of the Kitchen is Bullfrog Creek. The old mouth of the Alaf ia River can also be seen just above Bullfrog Creek. No major creeks flow into Wolf Branch. Image scale: 1:32,285. Wolf Branch is located along the eastern shore of Middle Tampa Bay and is immediately to the south of the Kitchen (Figure 2 4 ). Like the Kitchen, Wolf Branch also has a well developed mangrove shoreline (Figure 2 5) and is rich in both CDOM and, chlorophyll. Unlike the Kitche n there are n o significant river inputs into Wolf Branch though several small tidal tributaries are located throughout the area. Additionally, w ithin th e mangroves numerous mosquito ditches were dug in the 1960s for flood control. These ditches provide a direct conduit for surface runoff and may also be a significant conveyance for CD OM rich water. (a) ( b)
66 Coffeepot Bayou is loc ated along western Tampa Bay (Figure 2 4) S horeline morphology is very different from the Kitchen and Wolf Branch and is mostly seawall with no mangroves. The urban landuse adjacent to Coffeepot Bayou can deliver large amounts of stormwater runoff during wet weather. Because the re are no major sources of CDOM, blue light loss is not thought to be driven by CDOM absorption but by phytoplankton abs orption due to pulsed inputs of nutrient lad en urban runoff during storm events. The Egmont Key SMA is unique in that it is adjacent to a 162ha island that is both a State Park a nd a National Wildlife Refuge (Figure 2 4). The largely undeveloped island is located at the mouth of Tampa Bay. Given it s proximity to the Gulf of Mexico, CDOM and phytoplankton absorption should be minimal by comparison with the other three SMAs. 2 .3. 2 IOP f ield and laboratory methods For each SMA, a minimum of t en water sampl es were pseudo randomly selected every other month during 2008. For each SMA, a 25m by 25m grid was placed on top of the SMA map and e ach grid cell was assigned a number. Ten randomly generated numbers were selected and placed on the SMA map. Sites locatio ns were adjusted if the randomly generated plot resulted in sites clustered too close together. In some cases, especially in the Kitchen where water depths can be very shallow, station locations were moved to deeper water to allow boat access. The minimum operating depth during this study was 0.20m. Sampling protocol followed Florida Department of Environmental Protection standard operating procedures ( http://www.floridadep.org/labs/qa/sops.htm ) Samp les were collected approximately 0.10m below the surface using 1L brown plastic bottles and were immediately placed in a cooler of ice. Because some of the shallowest sites were so difficult to reach, attempts were made to sampl e on an incoming or a slack high tide Laboratory analyses were performed following the methods as described in Cannizzaro (2004). The particulate absorption coefficient (a p ( )) is the sum of the phytoplankton absorption coefficient (a ( )) and the detritus absorption coefficient
67 Eq. 2.4 Eq. 2.5 Eq. 2.6 (a d ( )). A bsorption spectra for a p ( ) were determined using the quantitative filter technique (Kiefer and Soohoo 1982; Yentsch 1962) Water samples were filtered slowly through 2.5cm GF/F (Whatmans) filters usi ng vacuum pressure s less than 15 in Hg. The volume of water filtered depend ed on the concent ration of the pigmented particles in the sample. Filters were wr apped in aluminum foil and 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 pl aced 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 2.4cm diameter glass plates in a custom made light box. Prior to each transmission scan, th e 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 tra nsmittance of the sample filter ( T sample ) and the reference filter ( T reference were each measured in triplicate using a custom made, 512 channel spectroradiometer with an effective range of 350 nm 850nm Particulate absorption wa s calculated as w here OD( ) is the optical density, is the path length amplification factor or beta factor, and the number one in the denominator is the geometric pathlength equivalent to the volume of water filtered divided by the clearance area of the filter (Butler 1962) OD( ) is the optical density and is calculated as The beta factor ( ) is an empirical formulation defined as the ratio of optical to geometric pathlength that corrects for multiple scattering inside the filter. To correct for pathlength amplification, was determined from published work (Bricaud and Stramski 1990; Nelson and Robertson 1993) using the equation
68 Eq. 2.7 Absorption at 750nm was subtracted from the entire spect r a to correct for either residual scattering caused by non uniformity in wetness between the sample and re ference filters or stray light. Phytoplankton pigments were extracte d 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 conce ntrations were determined on the filtrate using a Turner 10 AU 005 fluorometer (Holm Hansen and Rieman 1978) Following extraction, the sample filter was rinsed with Milli Q water to remov e any excess methanol and to rehydrate the filter. Transmittance spectra were measured on this filter and the reference filter. The a bsorption spectra for detrital material and non methanol extractable pigments (e.g. phycobiliproteins ) were calculated usin g Eq. 2 4. Lastly, a was calculated by subtraction using the following equation: For CDOM absorption, wa ter samples were filtered within four hours of collection first through a 0.45 m GF/F filter and then through a 0.2 m nylon membrane filter F iltrates were stored in clean 125mL amber bottles a t 10 o C and processed within one month of sample collection Prior to measurement, samples were thawed overnight at 6 o C and re filtered through a 0.2 m syringe filters to re move any particles that may have formed during the freezing and thawing process es Absorbance measurements were then made using the same spectrophotometer a s was used for particulate absorption (Cannizzaro et al. 2009) The beam attenuation coefficient ( for 480nm and 660nm was calculated using transmittan ce measurements collected in the field using two C star beam transmissometers (Wet Labs, Inc., Philomath, OR) Each transmissometer was housed in a flow chamber as part of a larger flow through system and sample water was pumped
69 Eq. 2.8 Eq. 2.9 into the tank via a small pump. Beam attenuation for a given wavelength was calculated from raw transmittance using the following equation: where is the pathlength between the light source and the detector and is transmittance and is calculated as: where is the measured output signal, is the manufacturer supplied clean water offset. 2.3.3. Coupling the IOPs with the quasi IOPs Historically, and still common ly in use b y the State of Florida chlorophyll a concentration s are determined using a spectrophotome ter in accordance with Standard Methods 10200H (Ea ton et al. 2005; F LDEP 2009) M any environmental laboratories and regulatory agencies have switched to the EPA approved f luorometric m ethod 445.0 which is considered to be a more accurate method (Arar and Collins 1997; Eaton et al. 2005) In the present study, chlorophyll a concentrations were determined based upon this fluorometric method. Samples for chlorophyll a concentration were collected at all water sample locations using one liter amber plastic bottle s Samples were placed on ice Optics Laboratory for analysis. Using regression techniques, an empirically derived equation was determined to all ow direct conversion of chlorophyll a with a (440). Additional water samples were collected at selected stations and analyzed for turbidity and color using methods commonly used for regulatory purposes. These samples were collected using 250mL clear plast ic bottles, placed on ice, and shipped to
70 the Florida Department of Environmental Protection Central Laboratory Facility in Tallahassee, FL for analysis. For regulatory purposes in Florida, water color is reported in units of Platinum Cobalt Units (PCU) an d have historically been analyzed using an EPA approved method ( EPA 140 A ) in which a color wheel is used to visually compare a given sample against a known PCU color scale (Eaton et al. 2005) This method is crude with a resolution of only 5 PCU and a minimum detection limit of 5 PCU. Despite the existence of better methods, it is worthwhile to relate to PCU col or given the thirty year record of color in Tampa Bay (Conmy 2008) In addition, t his method is still widely in use by many agencies including the FDEP Central Laboratory Facility, the labo ratory For this reason, and to be consistent with historical determinations of PCU color, water samples collected in this study were analyzed for PCU color using this crude method To relate P CU co lor with a linear regression model was fitted to data collected in this study Local and regional resource management agencies are now switching to the newer EPA method ( Method 110.3 ) that employs the use of a single wavelength spectrophotometer f or de termining PCU color at m uch greater resolution and with lower detection limits compared to the color wheel method (Eaton et a l. 2005) Over the past few years, t he HC EPC has been collecting side by side water samples from various parts of Tampa Bay and analyzing them for spectrophotometrically d etermined PCU color and for single wavelength CDOM absorption. The relationship betw een PCU color at 345nm and is examined for this larger Tampa Bay data set in the context of the current study. In turbid coastal waters, turbidity a quasi IOP, and the s catter coefficient, an IOP, are directly related and can be used interch angeably (Kirk 1994; Kirk 1991) Most regulatory agencies report t urbidity in Nephelometric Turbidity Units (NTU) using an EPA approved method (M ethod 180.1 ) (U U 1999) This method is based upon a comparison of the intensity of light scattered at an angle of 90 o by a sample with the intensity of light scatte red by a standard reference suspension at the same scattering angle. A primary standard suspension is used to calibrate the instrument. A secondary
71 Eq. 2.10 Eq. 2.11 Eq. 2.12 standard suspension is used as a daily calibration check and is monitored periodically for deterioration usi ng one of the primary standards. To relate this quasi IOP with the scatter coefficient, a subset of water samples were collected sent to the FDEP Central Laboratory Facility for turbidity analyses and compared to b(480) and b(660) as determined by Eq. 2. 3 A regression equation was fitted to these data allowing direct conversion of turbidity to the scatter coefficients which then can be used to estimate the light attenuation coefficients. To model an empirically derived relationship between and the total absorption and scattering coefficients is used. Kirk (1981) established this relationship between and the IOPs by expressing as a function of for monochromatic light using Monte Carlo simulation The values of over a range up to fit very closely to the following equation: where is a constant whose value is dependent on the volume scattering function used (Kirk 1981; Kirk 1984) While this has proven to be a ve ry robust relationship when the solar zenith angle is zero, increases as the direction of the incident light increases from vertical (Kirk 1984) Using Monte Carlo simulation, Kirk (1984) accounts for differences in by modifying Eq. 2 6 to the form: where is the cosine of and is a coefficient that determines the relative effec t of scattering on the total attenuation of irradiance (Gallegos 2001; Kirk 1991) is a linear function of and can be expressed as:
72 where and are numerical constants. Kirk (1991) determined and to be 0.425 and 0.19, respectively using the volume scattering f unction for San Diego Harbor and found that these values were applicab le to most coastal waters where Gallegos (2001) reproduced the values for and found by Kirk (1984) by conducting 432 model runs allowing a to vary between 0.5 and 4.0 m 1 and b to vary between 0.5 and 40 m 1 encompassing a range of b:a between 0.5 and 20. Although Eq. 2.11 is entirely empirical, it has proven over the years to be highly accurate and applicable to a wide range of solar angles and ratios (Gallegos et al. 1990; Kirk 1981; Kirk 1984; Kirk 1994; Kirk 1991) V alues for and determi ned by Kirk (1991) were used in the present study to estimate the using and To calculate the solar zenith angle the position and time of day for each sample station was input into the RADTRAN computer program (Sandia Nat ional Laboratories) (Weiner et al. 2008) 2 .4. Results and Discussion 2 .4.1. IOP s patial and temporal patterns On an average annual basis, absorption coefficients for CDOM, detritus and phytoplankton were higher in the Kitchen than in any other area followed by Wolf Branc h, Coffeepot Bayou and Egmont Key. Average annual absorption coefficients for CDOM, detritus, and phytoplankton were greatest in the Kitchen, followed by Wolf Branch, then Coffeepot Bayou, and finally Egmont Key (Figure 2 6 ).
73 Figure 2 6 Average annual CDOM, detritus, and phytoplankton absorption spectra for each Seagrass Management Area Note the different scales along the y axes. This pattern was inversely proportional to the di stance from the mouth of the bay with the Kitchen being furthest from the mouth and Egmont Key at the mouth. At 440nm in the photosynthetically useable blue region (400nm 490nm) the a verage annual for the Kitchen was more than four times that of E gmont Key and almo st twice that of Coffeepot Bayou (Table 2 3 ).
74 Table 0 3 Mean standard deviation of the annual chlorophyll concentrations and the annual absorption coefficients at 440nm for chlorophyll, C DOM, detritus, and total absorption. Chl a (mg L 1 ) a (440) (m 1 ) a g (440) (m 1 ) a d (440) (m 1 ) a t (440) (m 1 ) Kitchen 9.73 6.62 0.2655 0.129 0.8268 0.459 0.3269 0.131 1.4192 0.568 Wolf Branch 7.40 3.87 0.2233 0.085 0.6509 0.396 0.2315 0.07 1.1057 0.491 Coffeepot Bayou 5.45 3.66 0.1736 0.088 0.4893 0.109 0.1207 0.057 0.7836 0.229 Egmont Key 3.09 1.49 0.1052 0.039 0.2013 0.044 0.0774 0.030 0.3839 0.087 The standard deviation about the mean average annual was greatest in the Kitchen suggesting that this ar ea is subject to CDOM pulses driven largely by r ainfall Most CDOM sources in the Kitchen are locally derived and include a well developed mangrove shoreline, direct inputs from Bullfrog Creek, and secondary inputs from the Alafia River (Figure 2 5 a ). Simi lar to Kitchen, Wolf Branch also has a well developed mangrove shoreline but unlike the Kitchen, has no major creeks (Figure 2 5 b). By contrast, the shoreline along Coffeepot Bayou is contained by a seawall. There are no major freshwater rivers and landu se is urban. Those few wetland systems in connection to the bay. This results in an overall lower magnitude of CDOM absorption relative to that of Wolf Branch and Kitchen. An often used and relatively simple method for comparing the characteristics of CDOM from various locations is to compare their s pectral slopes (Branco 2005; Coble 2007; Hansell and Carlson 2002; Malick 2004; Minor et al. 2006; Steinberg et al. 2004; Vanderbloemen 2006) Spectral slope did not change much over the course of the study with a nnual average slopes ranging between 0.0190 to 0.0194. These values are con sistent with those found in other parts of Tampa Bay. The spatial and seasonal distributions of CDOM in Tampa Bay indicate that the two larges t rivers, the Alafia River near the Kitchen and Hillsborough River further to the north are dominant CDOM sources to most of the bay (Chen et al. 2007)
75 As with detritus absorption was greatest at the Kitchen and Wolf Branch relative to Coffeepot Bayou and Egmont Key (Table 2 2). In addition to shoreline morphology and inputs from Bullfrog Creek and the Alafia River, detrital absorption is a lso a function of re suspension of organic matter (Lawson et al. 2007) This is especially evident in the Kitchen where extremely shallow water and organic rich sediments are common. As with and average annual was also highest in the Kitchen and followed the same pattern of decreasing absorption with proxi mity to the mouth of the bay. This pattern may be driven by higher nitrogen concentrations in the Kitchen and Wolf Branch resulting in higher phytoplankton productivity. This hypothesis is supported by long term total nitrogen and chlorophyll data from fix ed monitoring stations near the Kitchen and Egmont Key (Figure 2 7).
76 Figure 2 7 Total nitrogen and chlorophyll data from two long term monitoring stations near the Kitchen and Egmont Ke y Seagrass Management Area s. Data are reported as averages standard deviation for the period 2004 2007. The Kitchen and Egmont Key were significantly different for both total nitrogen and chlorophyll a concentration (ANOVA; p<0.01). Source: Hillsboroug h County Environmental Protection Commission. Because samples were collected every other month, the temporal resolution was rather coarse. However seasonality in and for both the Kitchen and Wolf Branch is evident with maximum absorption occurring in August and minimum absorption occurring in December (Figure 2 8) Tampa Bay has distinct rainy and dry seasons and m aximum absorption coincid e d with the rainy season typically between the months of June through September although the 30 days prior to the June sampling were very dry with rainfall totals of less than 3.0cm
77 Figure 2 8 Monthly averages standard deviation of the mean for the inherent optical properties for the four Seagrass Management Area s. Rainfall totals at the St. Petersburg station for 2008 was 117cm and was slightly below the average of 126cm. Also, p eak rainfall occurred in July rather than the typical peak in August (Figure 2 9 ).
78 Figure 2 9 Historical monthly average rainfall and average rainfall for 2008 for St. Petersburg, FL. Source: National Weather Service. Both and were compared w ith average rainfall for the 30 days preceding a given sample event. There was a strong correlation between and rainfall for Kitchen (p<0.01; r 2 =0.81) and Wolf Branch (p<0.01; r 2 =0.92) but not for Co ffeepot Bayou (p>0.05; r 2 =0.12) or Egmont Key (p>0.05; r 2 =0.16) (Figure 2 10a). T he slopes of the best lines ranged from 0.032 in the Kitchen to 2.23 for Egmont Key decreasing with increasing proximity to the mouth of the bay (Figure 2 10a). The correlati on between and rainfall was significant (p<0.05) for all four SMAs and exhibited similar patterns (Figure 2 10b). The slopes of the best fit lines from the Kitchen (r 2 =0.64) Wolf Branch (r 2 =0.88) and Coffeepot Bayou (r 2 =0.65) ranged from 6.36 to 6.98 s uggesting that while the magnitudes of varied among SMAs, the response to rainfall was similar Egmont Key also exhibited a significant correlation between and rainfall (p<0.05; r 2 =0.67) but at a slightly different sl ope of 4.79 relative to the other three SMAs due to the mixing of bay water with the Gulf of Mexico.
79 Figure 2 10 Scatter plot and best fit line of the 30 day running average for rainfall and (a) .and (b) for each Seagrass Management Area Simple linear regressions were carried out for each curve with each the resultant listed to the right of the best fit line. Previous research has established a correlation between total nitrogen loads and rainfall in Tampa Bay (Janicki et al. 2003; Pribble et al. 2001) This suggests that the response to rainfall may be associated with increases in nitrogen loads during high flow periods Analysis of monthly HCEPC water quality data from 2004 2007 showed a significant differences between the two fixed stations closest to the Kitchen and
80 Egmont Key for total nitrogen and chlorophyll a concentrations (Figure 2 7) suggesting that the current nitrogen management paradigm may be appropriate for managing chlorophyll concentrations in the bay However, the effect of reducing chlorophyll con centrations on increasing the amount of blue light at depth in seagrass areas is a function of the relative contribution of to Overall, the relative contributions of , and to on an annual average basis were very similar for the Kitchen, Wolf Branch, Coffeepot Bayou, and Egmont Key. In all cases, represented approximately 61% of while only accounted for 20% of with accounting for the remaining 19% (Table 2 4 ). Absorption due to water represented less than 0.1% of the total absorption at 440nm and therefore was considered to be negligible. Table 0 4 Annual percent contribution to the total absorption coefficient by chlorophyll absorption, CDOM absorption, and detrital absorptio n. a (440) a g (440) a d (440) Wolf Branch 20.2% 58.9% 20.9% Kitchen 18.7% 58.3% 23.0% Coffeepot Bayou 22.2% 62.4% 15.4% Egmont Key 20.8% 63.8% 15.3% Traditionally the Tampa Bay model assumes that phytoplankton absorption is the primary cause of light attenu ation. This assumption is part of a larger model describing a n increase in nutrient availability leading to increased phytoplankton productivity and thus reducing the available light for seagrass (Cloern 2001; Janicki Environmental 2001) Most of these conceptual models are based on wat er quality information collected at stations not representat ive of the conditions in shallow waters (Lawson et al. 2007) In Tampa Bay, the data most u tilized for model development are from fixed stations located throughout the deeper waters Virtually none of these st ations are located anywhere near seagrass areas where the optical properties can d iffer markedly from offshore areas. The results presente d here strongly suggest that the existing model needs to be modified to
81 account for the dominance of CDOM as the major attenuator of light At the end of the day, may be the only component that can be effectively managed through nitrogen reductions and is much more problematic to manage for CDOM or detritus Nevertheless, resource managers must take into account reductions in although these reductions will a ffect only 20% of the total absorption. 2 .4. 2 Relationship between the IOPs and the quasi IOPs A multiple regression model was used to determine the contribution of and the contribution, if any, of to chloroph yll a concentration. Model results indicated that both and were significant at the 95% confidence level with an r 2 =0.87.Standard error for and was 0.378 and 1.35, respectively. Using only as the independent variable yielded similar results with an r 2 =0.86 and a standard error of 1.15. To predict from chlorophyll a concentration, a simple linear regression model was applied to the entire data set and to each SMA. In all cases, th e predictor equations explained between 68 % and 89 % of the variation (Table 2 5). Table 0 5 Predictor equations for based on chlorophyll a concentration for each Seagrass Management Area and fo r all areas combined. Wolf Branch Kitchen Coffeepot Bayou Egmont Key All Areas The weakest correlation was at Egmont Key where lowest chlorophyll a and values were reported Average chlorophyll a concentration for Egmont Key was 3.25 g L 1 ranging from 1.0 5 g L 1 to 6.44 g L 1 Average for Egmont Key was
82 0.1140m 1 ranging from 0.0571m 1 to 0.1812m 1 These low values may explain why the r 2 for Egmont Key was low er than for the other SMAs The chlorophyll a relationship for the areas in this study were consistent with relationships found by other researchers along the West Florida Shelf (Cannizzaro et al. 2008; Cannizzaro et al. 2004) (Figure 2 1 1 ). As expected, Tampa Bay data fall on the upper end of the West Florida Shelf curve and are representative of the nearsh ore end member s (Figure 2 1 1 ). Figure 2 11 Chlorophyll a concentration and for Tampa Bay, collected during this study, and data collected for the West Florida Shelf (Cannizzaro et al. 2004) Inset shows chlorophyll a concentration and for data collected in this study from each of the four SMAs.
83 Turbidity is a simple and very common water quality parameter measured using a nephelometric turbidimeter. Essentially, a beam of light is directed along the axis of a cylindrical g lass cell containing the sample. Light is scattered from the beam and a photomultiplier is positioned at a scattering angle of 90 o (Kirk 1994) The measurement provided is in N ephelometric T urbidity U nits (NTU) and is a measure relative to a know n standard. Turbidity meters do not provide a direct estimate of any fundamental scattering property of the water and measurements using these devices can be thought of as quasi inherent optical properties. Nevertheless, in waters with moderate to high tu rbidity due to inorganic particles, turbidity should approximate the scattering coefficient such that a linear relationship should bear out. Turbidity data were collected for a subset of sites across each of the four Seagrass Management Area s and when plot ted against the total scattering coefficient s for 480nm and 660nm yielded a moderately strong relationship (Figure 2 1 2 ). Figure 2 12 Relationship between turbidity and the scattering co efficients for blue light and red light for selected samples in Tampa Bay.
84 PCU color and were weakly correlated mostly because of the coarse resolution of color measurements (Figure 2 1 3 ). Color was reported in 5 PCU inc Bay Study Group has conducted extensive side by side comparisons of PCU color and measurements and has demonstrated an almost one to one relationship between them (Figure 2 1 4 ). The City determine s color by measuring absorbance at 345nm and then using a platinum cobalt standard, convert ing absorbance to color in PCU. This provides a much more robust measure of color and regulatory agency laboratories are slowly mak ing the change to the more quantitative method of measuring color. Figure 2 13 PCU color at 345nm plotted against for samples taken throughout Inset shows the relationship between color and for samples collected during this study. The stair step pattern is a result of PCU col or being reported in 5 PCU increments.
85 2.4.3. Modeling K d (480) for Tampa Bay In this study, the attenuation coefficient at 480nm was calculated using Eq. 2.10 and the values for and as determined by Kirk (1991) for selected stations Measured versus modeled fit well against a 1:1 line with scatter being evenly distributed on either side of the line (Figure 2 1 4 ) Figure 2 14 Plot of measured against modeled Th e ratio of for this study ranged from 1.74 to 34.4. With the exception of the single 34.4 value, all other values were within the recommended maximum value of 30 (Kirk 1984) R esiduals range d from 0.002 to 0.913 and the sum of the absolute values of the difference between measured and modeled values was 7.03 An optimization program adjusted the coefficients and but did not yield an improvement in the term Using Eq. 2 10, estimates of the for all water samples were calculated and summarized in Ta ble 2 6
86 Table 0 6 Average annual summary table for predicted , and b standard deviation for each Seagrass Management Area N Wolf Branch 1.16 0.432 0.67 8 0.357 4.0 1 1.23 6.64 2.38 60 Kitchen 1.47 0.429 0.8 30 0.324 5.9 5 2.16 7.83 2.96 57 Coffeepot Bayou 0.761 0.223 0.439 0.138 2.5 5 0.962 5.77 1.32 60 Egmont Key 0.459 0.099 0.22 4 0.051 2. 5 5 1.05 11.928 5.73 56 As expected t he Kitchen had the highest annual average modeled and highest standard deviation while Egmont Key had the lowest annual average modeled and the lowest standard deviation. Despite the rel atively small sample size in this study, the model yielded good results. Nevertheless, fu rther work is needed to refine the coefficients for the various conditions found in Tampa Bay and beyond. 2 .5. Conclusions The results in Chapter 1 of this study indi cate d that seagrass in Tampa Bay are largely blue light limited, especially along the deep edges. Management decisions designed to improve the light environment for seagrass growth should focus on the blue wavelengths which are largely dominated by CDOM ab sorption. The current management paradigm does not address CDOM but focuses exclusively on light attenuation by phytoplankton absorption. Historically management of nutrient load reductions has been successful in reducing chlorophyll concentrations in the bay, primarily through the increased level of wastewater treatment a nd improvements in stormwater management While ther e is still room for improvement it is unlikely that the large increases in seagrass coverage, as seen in the 1980s and 1990s will occ ur given the already significant reductions in water column chlorophyll and the large fraction of total blue light absorption due to CDOM Based on the results from this study, CDOM wa s the dominant blue light absorption component accounting for as much as 80% of the total absorption. This supports the conclusions of Chen, et al. (2007) who found on a bay wide basis, was five times higher than in June and ten times higher in August. In the present study, t he relative dominance of CDOM to was consistent across a ll four
87 SMA s even at Egmont Key which was expected to have a much great er percent contribution from given its proximity to the Gulf of Mexico and relative distance from any major CDOM sources The contribution of to was not in significant and in some cases exceeded Given the shallow nature of the seagrass beds in Tampa Bay wind driven resuspension of organic material may be a primary cause of the relatively high contribution of to especially in areas where major sources of detritus from river inflo ws are minimal. Lawson, et al. (2007) found bottom stresses from wind driven waves was the dominant predictor of light attenuation in Hogs Bay, Virginia, a shallow coastal bay off the U. S. mid Atlantic coast. Lawson, et al. (2007) also points out that the se wind driven forces are episodic and often missed due to fair weather monitoring or inappropriate sample site locations. While relative contribution s were similar across SMAs differences in the magnitude of CDOM were largely a function of proximity to t he Gulf of Mexico. Temporal variability in CDOM absorption was greatest in t he Kitchen and Wolf Branch and was largely a response to pulsed events. There was a strong response to rainfall in both the Kitchen and Wolf Branch. For seagrass, the timing of the se pulsed events may be critical to their survival. Typically the highest rainfall occurs in the summer rainy season with peak rainfall in August and September during the height of the growing season. Anecdotal evidence of increases in seagrass coverage o ne year following the winter El Nio of 1997/1998 supports the hypothesis that the timing of high C DOM pulses is extremely critical A better understanding of the impacts of the magnitude and timing of pulsed events is critical to successfully managing sea grass resources, especially in the face of sea level rise and global climate change. While there may be little resource management agencies can do to manage CDOM inputs to the bay, it is important to differentiate between anthropogenic and natural sources of light attenuation, especially in the face of growing regulatory pressure to implement numerical criteria for water bodies deemed impaired. One potential management action could be to remove the mosquito ditches adjacent to the Kitchen and Wolf Branch SM As. These ditches could be increasing the conveyance of CDOM to the
88 bay by providing a direct conduit However, more research is necessary to determine if hydrologic restoration of these ditches would in fact significantly change the CDOM load into these a reas There could be unintended consequences of such management actions. For example, r eductions in CDOM could potentially result in increased phytoplankton productivity due to less shading thus offsetting any benefit to removing the ditches. Typically w ith high CDOM come high nutrients resulting in higher chlorophyll concentrations and phytoplankton absorption. The pattern of chlorophyll and phytoplankton absorption seen in this study supports this contention. Kitchen and Wolf Branch, the areas with the highest CDOM also had the highest chlorophyll concentrations and phytoplankton absorption coefficients. Comparing nutrient and chlorophyll data from two long term monitoring stations, one near Egmont Key and the other near the Kitchen, shows a strong corre lation between increased total nitrogen and increased chlorophyll concentration. This is good news to resource management agencies spending millions of dollars to reduce nutrient loads into the bay By regressing chlorophyll a color, and turbidity with t he IOPs, a series of predictor equations were d eveloped and will be very useful in estimating SMA specific absorption and scatter coefficients Linking the IOPs with the quasi IOP s provides a means to infer past optical conditions using historical data T h e relationship between and chlorophyll concentration yielded simi lar regression equations across all four areas and is further evidence that while these shallow seagrass areas may be unique relative to the rest of the bay, it is likely that many of the 30 bay wide SM As can be merged together to create a simpler framework for management purposes Of course this will need to be verified for other SMAs and over longer time periods. The relationships determined in this study were based on data collected in 2008, a normal year for ra infall It is unknown whether these relationships w ill hold under varying conditions such as during a prolonged La Nia or El Nio or in the aftermath of a major hurricane. Finally, the utility of using an empirically derived model to estimate the light attenuation coefficient at 480nm was demonstrated by using absorption and scatter coefficients of the same wavelength. This model is very useful in that it links the AOPs with the IOPs and allows much more insight into the underlying causes of at tenuation.
89 For resource management agencies tasked with setting light targets, this model is very powerful in that based targets can be set on the basis of the IOPs that are by definition intrinsic to the optical properties of the water and independent of the ambient light conditions at the time of sampling. This drastically increases the operational tempo o f a monitoring program because it relieves the constraints normally associated with measuring AOPs, such as sky cover and sea state. Of course validation of this model must be incorporated into any monitoring program using derived from measured using a spectral light monitoring system like the one described in Chapter 1.
90 Chapter 2 Synoptic S urveillance of the U nderwater L ight F ield U sing a C ontinuous D eck M ounted F low T hrough System 3 .1 Abstract G reater spatial and temporal information a bout the shallow water light environment is critical to u nderstanding seagrass ecology. The spatial and temporal variability inherent in shallow seagrass areas makes it difficult to accurately assess conditions using discrete measurements. Remote sensing o f the water column can also be problematic due to interference from bottom reflectance. The use of a flow through system designed to operate in water depths of less than 2.5m would provide relevant optical data at spatial resolutions not possible using oth er techniques. The flow through system used in this study was modified to operate in a small open hulled boat in depths as shallow as 0.25m. The payload included chlorophyll and colored dissolved organic matter (CDOM) fluorometers that were used to estima te the chlorophyll a concentration, as well as the phytoplankton and CDOM absorption coefficients. To accomplish this, discrete water samples were collected during each survey and analyzed in the laboratory. Empirical equations were derived to calculate and from raw chlo rophyll and CDOM fluorescence voltage measurements. Relationships between and approximated a linear fit with correlation coefficients ranging from 0.80 to 0.84. A second order polynomial equatio n best fit the relationship between and with correlation coefficients ranging between 0.88 and 0.92. This non linearity was observed across all surveyed areas using two different models of fluorometers and was a function of the inherent non linearity within the CDOM absorption spectra. Regressing against approximated a linear fit (p<0.01; r 2 =0. 91). The potential for interference s from and the detrital absorption coefficient on were explored using multiple regression procedures but resulted in no improvement over the non linear predictor equations. The Kitchen Seagrass Management Area a 776 h ectare area located along the shoreline of eastern Tampa Bay, i s presented as a case study to demonstrate the utility of using a flow through system approach to detect spatial and temporal differences in
91 CDOM and chlorophyll a concentration CDOM was concentrated along the immediate shoreline. This CDOM rich water mas s only extended out approximately 600m from the shoreline and covered all of the seagrass growing within this area. The complex patterns in CDOM and chlorophyll a demonstrate the need to adequately characterize the variability in shallow seagrass areas. Th ere is a danger in relying too heavily on discrete water samples to infer conditions. The nearest long term water quality monitoring station is well outside the Kitchen yet data from this station are routinely used to evaluate the light environment over th e seagrass. Without the use of a flow through system, is evident management decisions designed to protect and restore seagrass could be based on erroneous conclusions.
92 3 .2 Introduction The need for more and better optical data is a constant challen ge, especially for r esource management agencies in Florida who are being t asked to develop transparency standards for protecting seagrass communities. Over the past two decades, there has been a n impressive amount of re search investigating the properti es o f the underwater light field and its effects on seagrass survival (Abal et al. 1994; Biber et al. 2008; Cummings and Zimmerman 2003; Dennison 1987; Dixon 2000; Enriquez 2005; Enriquez et al. 1992; Greening 2004; Kenworthy and Fonseca 1996; Miller and Mcpherson 1995; Ralph et al. 2007; Zimmerman 2003) All of these studies underscore the need for more spatial and temporal data, a common problem among budget conscious government agencies as well. Fortunately, technological advancements in hardware and data processing software have made it possible to sample large areas repeatedly without being cost prohibitive. Until recently, the lack of rugged and portably field instruments, such as spectrophotometers necess ary to accurately characterize optical measurements like CDOM absorption and fluorescence, played a large part in contributing to the paucity of optical information (Hoge et al. 1993) The use of a deck mounted flow through system to monitor optical condi tions while underway promises to greatly enhance the way transparency data are being collected and analyzed. Flow through systems have been in use for some time (Madden and Day 1992) but to date, none have been used in very shallow seagrass beds like those found in Ta mpa Bay, where depths rarely exceed 4m and are often less than 0.25m. Traditional methods of measuring in water optical properties in Tampa Bay require a minimum depth of a t least 1.5m (Dixon and Leverone 1995; E.P.C.H.C. 2007; TBNEP 1996) which can create data bias potentially resulting in erroneous conclusions. To assist environment al resource management agencies in developing better monitoring techniques, a method of using a deck mounted flow through system in shallow seagrass areas is presented.
93 3.2.1 The underway flow through system approach A n underway flow through system orig inally designed to be operated a t sea, was employed in very shallow water s to characterize the optical properties of selected Seagrass Management Area s in Tampa Bay, FL. The payload included three chlorophyll fluorometers, two colored dissolved organic mat ter (CDOM) fluorometers, two backscatter meters, two transmissometers a conductivity temperature (CT) probe, and an onboard GPS (Table 3 1) While there are newer instruments on the market today, the payload used aboard this system included all the necess ary hardware to collect optical data commonly used in seagrass and water quality management. Table 2 1 Payload description of the underway flow through system used for this project. The system in its current configuration includes redundant instrumentation for data validation and to account for operational differences among units. Parameter Unit Orientation CDOM Fluorescence Wet Labs WET Star Fluorometer W et Labs ECO Fluorometer Inline In tank Chlorophy ll Fluorescence Wet Labs WET Star Fluorometer Sea Tech Fluorometer Wet Labs BB2F Inline In tank In tank Red & Blue Backscatter Wet Labs BB2F HOBI Labs HyroScat 2 In tank In tank Red & Blue Transmittance Wet Labs C star Transmissometer In tank Conductivi ty / Salinity Falmouth Scientific CTD In Tank Temperature Falmouth Scientific CTD In Tank The concept of underway monitoring has been in around for some time (Buzzelli et al. 200 3; Ensign and Paerl 2006; Madden and Day 1992; Paerl et al. 2009) In this study the concept is applied to modeling the inherent optical properties (IOP) of shallow waters within discrete Seagrass Management Area s (SMA). Thirty SMAs have been designated b y the Tampa Bay Estuary Program and its partners based on a priori information related to historical seagrass coverage, water quality, hydrodynamics, and shoreline morphology (E.P.C.H.C. 2007) Of special interest in Tampa Bay is the dominance of CDOM as a major absorber of blue light. Results from Chapter 1 indicated that seagrass in much of Tampa Bay are
94 blue light limited. In Chapter 2, CDOM absorption accounted for as much as 60% of the tot al absorption of blue light in the SMAs studied. In addition, phytoplankton absorption made up an additional 20% of the total absorption of blue light In terms of the magnitude of CDOM and phytoplankton absorption, the Kitchen, located in eastern Tampa Bay was the most CDOM rich area with the highest CDOM absorption occurring during the rainy season when inputs from river discharge and runoff are more prevalent. Chlorophyll absorption was also greatest in the Kitchen. Likely CDOM sources include a well established mangrove shoreline and direct river inputs of CDOM rich waters from nearby streams. The flow through system was used to survey CDOM absorpti on in the Kitchen and other SMAs every other month for one year. The CDOM absorption coefficient at 440nm was empirically derived using raw CDOM fluorescence and mesured from discrete water samples collected during each survey. The phytoplankton absorption coefficient at 440nm and the chlorophyll a concentration were al s0 determined using empirical regression techniques. Surveys were designed using a the the dots approa quantify differences between routine monitoring stations. 3.2.2 Principles of in water fluorescence While phytoplankton cells absorb light in the photosynthetically useable blue and red regions of the visible spectrum, they can also emit light. About 1% of the ambient light absorbed by phytoplankton cells is re emitted as fluorescence in the red region centered around 685nm (Kirk 1994) Fluorescence can be induced b y exciting cells with a light source of known intensity at a wavelength centered at 460nm and measuring the intensity of the emitted light at a wavelength centered at 695nm. Most in water fluorometers use an LED as the excitation source Some fluorometers use a halogen light source as does the SeaTech unit aboard the flow through system used in this study Al l others aboard the flow through use d LED light s ources CDOM fluoresce s in the blue wavelengths when excited by light in the ultraviolet wavelengths Peak fluorescence is
95 commonly found at 250nm and 350nm depending on its source (Coble 2007) Highest fluorescence efficiencies are found in fresh terrestrial and deep marine waters while lowest efficiencies are typically found in offshore surface waters due in large part to photodegradation. Depending on the excitation (Ex ) and emission (Em) peak wavelengths, specific components can be identified. Eight general types of fluorescence peaks have been identified. For example, an Ex/Em pair of 260nm/400nm 460nm indicates the presence of humic like material and an Ex/Em pair of 275nm/305nm indicates a protein (tyrosine) like component suggesting an autochthonous carbon source (Coble 2007) The excitation/emission wavelengths for the fluorometers aboard the flow through system varied somewhat. 3 .3 Materials and Methods 3 .3.1 Flow through system desig n and specification An underway flow through s ystem allows continuous recording of optical data and affords the ability to detect spatial patterns not observable through traditional methods. The flow through system payload included a conductivity/temperature sensor, blue and red transmissometers, chlo rophyll fluorometers, CDOM fluorometers, and blue and red backscatterometers (Table 3 1). All underway instruments were mounted on a black metal frame which placed inside a large tank whose walls were painted black (Figure 3 1). Water was pumped through th e chamber using a small pump with an average discharge rate of 38 liters per minute and was powered by an external 12V marine battery. At this discharge rate residence time in the tank for a given parcel of water was approximately 3 minutes. The flow thr ough system was designed with built in redundancy to provide a backup in case of system failure and to cross check measurements from units with different design specifications and operational characteristics (Cannizzaro et al. 2009) a and CDOM fluorescence sensors provided a measur e of any high frequency changes that may not have been detected by sensors in the tank. The advantage to operating the instruments in a closed tank rather than under ambient light conditions is that it eliminates the potential for contamination by solar st imulated
96 fluoresced photons (Stramski et al. 2008) In order to minimize the chance that light emitted by one sensor is detected by another sensor, sensors were oriented within the chamber so that light emitted from one instrument would not directly enter the field of view of ano ther instrument (Cannizzaro et al. 2009) The scatterometers were arranged inside the tank to maximize the distance between the instrument s and the chamber walls reducing the likelihood of interference from reflected light off the tank walls. This is mainly a concern when operating in very clear waters and was not a concern in this study. Figure 2 1 The deck mounted flow through system used in this study 3.3.2. Synoptic survey of Seagrass Management Area s Synoptic surveys of the optical properties were conducted in selected Seagrass Management Area s in Tampa Bay, FL. Prior to leaving the laboratory on the morning of a survey, a full system s test was conducted on all instruments. All cables and instruments were visually inspected and the system was connected to a ruggedized field laptop computer (Panasonic Toug hbook CF 29). Except for the ECO fluorometer and the HydroScat 2 that were connected directly into the laptop via a seria l to USB switch, all other instruments were routed through the CTD via an analog to digital converter to the laptop for logging. In add ition to logging all flow through data while underway, the laptop also provided positional information from a built in GPS unit. All positional data were recorded in the WGS 1984 geographic coordinate system. The laptop also had navigational software (Fuga wi ENC Ver4.5, Northport Systems, Inc., Ontario, Canada)
97 that allowed the overlay of the predetermined survey track onto a digital nautical chart. Optics were wiped clean of moisture and debris and then allowed to sit in air for approximately 10 minutes. A ir measurements, in raw counts, were written down and compared to calibrated values to ensure they were within 5% Once the system passed its pre cruise check, all cables were carefully disconnected and the system was prepared for transport to the field s ite. Once onsite, the system was carefully loaded onto the boat and placed into the tank. Depending on the vess el used, the tank typically sat roughly between the center console and the bow. All cables were carefully reconnected and the system was turned on. A similar pre cruise check was conducted in the boat while still onshore to make certain all systems were operating This was a very important par t of the pre deployment process. Once the instruments passed their final pre cruise check, they were set to begin logging. It was not necessary to stop logging data until the survey was complete unless there were problems with the system that necessitated immediate action. A detailed field log was kept throughout the survey for mission reconstruction during t he data evaluation phase of the project Once the flow through system was running and actively logging, the pump was connected to the tank and the pump intake affixed to the gunwale with the nose pointing in the direction of the flow path (ie. toward the b ow) to minimize turbulence and air bubbles while underway. A small centrifugal wash down pump (Water Puppy) with neoprene impellers to minimize damage to phytoplankton cells was used, though ideally a diaphragm pump is preferred. This pump operated off o f a sta ndard 12V marine battery and had a discharge rate of 10gpm. At this flow rate, it took approximately three minutes to fill the tank. The hose leading from the pump to the tank was attached near the top of the tank This was done primarily to allow s ediments, or other large particles, to settle to the bottom of the tank instead of being repeatedly re suspended by the incoming flow as would happen if the hose were located near the bottom of the tank. Placing the inflow hose near the top also had the a dded advantage of allowing any air bubbles to escape out the top of the tank since the tank lid was not airtight An inline y valve was inserted just before the tank to divert some of the flow to the in line CDOM and chlorophyll instruments. To minimize tu rbulence, flow
98 through the in line fluorometers was adjusted to approximately 0.5gpm using the y valve. Water leaving the tank drained out of a hole near the tank base opposite the inflow. A hose was connected to the drain and allowed to flow over the gunw ale but away from the pump intake. cla m p to a flexible PVC hose attached to the pump. To keep the nose of the intake pointing at a downward angle of approximately 20 o from the water sur face, and to minimize yaw, t he intake was weighted b y two 10lb dive weights The intake was adjusted on the fly. This was accomplished by attaching two lines from the intake to the boat. The forward or bow line was attached from the base of the intake pipe to a forward cleat. This line was used to adjust the pitch of the intake and was typically not adjusted once set. The aft or stern line was attached from the upper e nd of the intake and wrapped around an aft cleat. This line was used to adjust the depth of the intake and typically was handled by an operator for making quick depth adjustments. Intake depth was usually between 0.25m and 0.50m and was located just below At the intake nose, a mesh screen was attached to help exclude large floating p articles from entering the tank The screen was also designed to be a failsafe during those inevitable ti mes when the intake dragged bottom in very shallow waters. When this occurred, the survey typically was stopped and the tank inspected for sedimentation. If it was determined that a large amount of sedimen t entered the tank, th e flow through system was lif ted out of the tank, the tank was drained, and then rinsed with seawater. The entire process t ook several minutes and was documented in the underway log for data evaluation in the post processing stage Survey areas in Tampa Bay varied from 3 km 2 12km 2 and depending on the number of legs, survey times lasted between two and three hours. Maximum underway speed was approximately 5kts, above which excessive turbulence at the intake caused erratic backscatter measurements. Maximum speed was also constrained by depth along the shallow margins. While the Seagrass Management Area s were well delineated, the extent of the survey track was dependent on the desired spatial resolution. All s urvey
99 routes were delineated and plotted prior to starting the survey but oc casionally needed to be edited in the field if depth conditions were too shallow or other circumstances made it impossible to follow the pre plotted track. Track routes were based on a number of factors and in no small part to trial and error. A minimum of ten water sample locations per survey were selected for IOP determination s to allow for conversion of raw fluorescence data into absorption coefficients and chlorophyll concentrations While these locations were initially chosen at random, some were adjus ted to ensure that a wide cross section of optical conditions were captured (Figure 3 2). Occasionally it was necessary to move a sample location due to shallow water conditions preventing boat access. Figure 2 2 Example survey track from June 2008 for the Wolf Branch Seagrass Management Area This set the initial survey framework by which a connect the dots approach was applied (Figure 3 2). A secondary constraint was the addition of additional field mea surements collected on the same cruise by other researchers, making it necessary to tailor track routes to accommodate these other sampling schemes. A third important constraint was depth. Early track routes were mostly based on depth data from nautical ch arts and limited local knowledge of the area, but as the project progressed, survey tracks were based more on actual depths sometimes determined the hard way by running aground
100 and being forced to walk the boat off the bar Over the course of a three hour survey, conditions did not remain static and by the time the boat reache d the far end of the survey track the optical properties may have changed This was an inherent bias when surveying large areas and must be kept in mind when analyzing spatial pattern s over relatively large areas. Tid al stage, wind conditions, and currents were documented during each survey and were useful during post mission reconstruction. Survey tracks were typically set up in a serpentine pattern (Figure 3 2) essentially creating a series of transects perpendicular to the shorel ine that could be parsed out of the dataset to be analyzed as a standalone product. Each transect took approximately 30 min utes to complete. Sea state was also a major concern and potential source of error. Because the flow through system was originally designed to be on a large research vessel with plenty of cabin space, some design modifications were made to ruggedize the s ystem for use on a small open deck boat. A secondary wiring box was added to the sy stem housing any excess cabl e the serial to USB converter box and the instrument batteries. While this box was not completely waterproof it kept the electronics relatively dry. Another modification was to the tank locking mechanism which would no longer lid. This generally is a major problem when the seas are flat. However, during pre deployment trials, whenever the sea state was greater than one foot, the tank lid would not stay closed and the flow through system would experience excess ive sloshing. The addition of adjustable straps eliminated this problem by securing the lid while still allowing some water to overflow the tank thus acting like a large de bubbler. 3 .3. 3 Data management and analysis All raw data w ere logged onto a hard drive mounted onboard the field laptop. Most of the instruments, with the exception of the ECO fluorometer, lo g g ed internally acting as a backup. The CTD acted as the main backup data logger for the in line fluorometers, the transmissometer s and the CTD i tself. At a data logging rate of half second int ervals, a typical survey r esult ed in about 30,000 rows of data with as many as 30 parameters per row. This large quantity of data required a sophisticated data management strategy. The foundation of this stra tegy was based on four basic data levels Each increasing level represent s a more refined data product. Depending on the analysis
101 Eq. 2.1 need, data from multiple levels may be used but generally analyses took place on Level 3 data. Level 0 is the most basic leve l and represents raw data taken directly from the instruments. All data were logged by default at 0.5sec intervals unless otherwise specified. All raw data from the CTD and associated instruments were in units of digital counts from 0 to 4095. Geographical position data from the onboard laptop GPS was stored in a separate data file by the navigational software (Fugawi ENC). All data stored as Level 1 data were binned to the nearest minute using a simple binning routine CTD GPS position, and backscatter d ata were stored in separate data files All Level 1 CTD data remained in digital counts while BB2F and HS2 data were reported as the various scatter coefficients. Conversion of scatter data from raw counts occurred onboard the instruments using calibration files resident on the instruments. Data files containing raw signal counts were also saved and afforded the opportunity to apply locally derived constants to the scatter conversions. At Level 2, optical fluorescence and transmittance data were converted into voltages and prepared for post mission reconstruction in which data were reviewed against the field logs. All instruments ha d an effective output range of 0 5 VDC. Counts were converted into voltages using the simple equation: Detailed notes entered into the field log we re an invaluable tool when post processing because, as stated earlier, once the instruments were set to log prior to starting the actual survey, they did not stop logging until th e survey w as complete. On a typical mission, breaks in the survey track occurred at least half a dozen times for various reasons such as if the boat stopped at a fixed location to collect other data or when the boat ran aground in very shallow water and ha d to be pushed back into deeper waters. Data not part of the survey w ere simply parsed from the Level 2 dataset. Once all erroneous data were removed, GPS position data were merged with the CTD and backscatter datasets.
102 Data at Level 3 represented the fi nal data products used for analysis. At this level, data were converted from voltages to relevant units. For the transmissometers, t he conversion from voltage to beam attenuation was a two step process. The first was to convert voltage into transmittance a nd then transmittance into beam attenuation as described in Chapter 2. Conversion of fluorescence voltage to absorption coefficients was done empirically using absorption coefficient data from se Each survey mission included at least ten water samples taken back to the laboratory and analyzed for absorption by CDOM, phytoplankton, detritus as well as for chlorophyll concentration. The resultant regression equations were used to convert the rest of the survey data into relevant units. 3 .4 Results and Discussion Determinations of , and chlorophyll a concentration, for the SMAs sampled in this study, were made from raw CDOM and chlorophyll a flu orescence using a combination of linear, non linear, and multiple regression methods. The spatial variability in CDOM and chlorophyll a was mapped for the Kitchen SMA as a case study. 3 .4.1. I nherent optical propert ies and fluorescence C hlorophyll a fluor escence and generally followed a linear fit across all SMAs For IOP and chlorophyll a determinations, data from all stations were analyzed together to maximize the effective fluorescence range s All chlorophyll fluorescence data were co mpared for each instrument and while the response curves were very similar, there were some differences in the slopes, and y intercept s (Table 3 2).
103 Table 2 2 Regression equations, with fo r each instrument aboard the flow through system, used to convert raw fluorescence voltage to corresponding absorption coefficients for chlorophyll and CDOM and chlorophyll concentration WetStar Chl SeaTech Chl WetStar ECO WetStar Chl SeaTech Chl The regression model between and for the WetStar fluorometer resulted in a good fit with and a y intercept of 0.0353 (Figure 3 3 ) For the SeaTech fluorometer the resultant best fit line produced an and a y intercept of 0.0384 (Figure 3. 4 ). Figure 2 3 Relationship between WetStar chlorophyll fluorescence and the phytoplankton absorption coefficient s at 440nm and at 660nm. y = 0.2392x 0.0353 R = 0.8373 y = 0.0701x 0.0224 R = 0.8539 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 a ( ) Fluorometer Volts 440nm 660nm
104 Figure 2 4 Relationship between SeaTe ch chlorophyll fluorescence and the phytoplankton absorption coefficient s at 440nm and at 660nm. Using the absorption coefficient at 660nm yielded similar results for both WetStar and SeaTech fluorometers (Figure 3 4) Residuals increased as fluorescence increased and interference factors became more prevalent. M easuring in situ fluorescence can sometimes be more art than science and several confounding factors have to be considered whe n utilizing these types of relationships. For example, in waters with high phytoplankton biomass, pigment packaging effects can cause chlorophyll fluorescence variability (Bissett 1997) Another factor to consider is p otential error due to CDOM and detritus In this study both CDOM and detritus absorption co varied with chlorophyll fluorescence (ANOVA p<0.01) This significance does not of course imply a cause and effect relationship but the slightly better fit between chlorophyll fluorescence and phytoplankton absorption at 660nm for both WetStar (Figure 3 3 ) and SeaTech (Figure 3 4 ) i nstruments suggests that there is a causal relationship For both instruments the residuals increased with increasing absorption. This is to be expected as increases in fluorescence may be caused y = 0.1306x + 0.0384 R = 0.8021 y = 0.0383x 0.0013 R = 0.8403 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 a ( ) Fluorometer Volts 440nm 660nm
105 by interference from CDOM and detrital absorption as well as scatter from both o rganic and inorganic particles at high concentrations. Chlorophyll fluorescence plotted against chlorophyll concentration yielded a good fit though some non linearity was visually evident in the Wet Star fluorometer when fluo re scence w as less than 1.0V (Figure 3 5). Figure 2 5 Relationship between WetStar chlorophyll fluorescence and chlorophyll concentration ( g L 1 ). For the SeaTech fluorometer the relationship between fluorescence and chlorophyll concentration was more linear although, like the WetStar fluorometer, scatter about the best fit line increased with increasing chlorophyll concentration (Figure 3 6). y = 11.423x 4.4332 R = 0.8791 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 Chlorophyll a ( g L 1 ) Fluorometer Volts
106 Figure 2 6 Re lationship between SeaTech chlorophyll fluorescence and chlorophyll concentration ( g L 1 ). Also, the y intercept for the resultant linear regression line was closer to the origin for the SeaTech fluoromete r relative to the WetStar unit (Table 3 2) Total scatter can also affect fluorescence b y increasing the effective pathlength of th e light source. However, when scatter was included in the multiple regression model, it was not significant (ANOVA, p>0.05). Measured from both the ECO and WetStar fluorometers were plotted against and result ed in very similar yet non linear response curve s (Figure 3 7 ). y = 6.2856x 1.0473 R = 0.8787 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Chlorophyll a ( g L 1 ) Fluorometer Volts
107 Figure 2 7 Relationship between CDOM fluorescence and the CDOM absorptio n coefficient for the WetStar and ECO fluorometers. A second order polynomial is fitted to each of the curves. (a) WetStar (b) ECO
108 P redictor equations for both WetStar and ECO fluorometers that be s t explained the variability in the data were second order polynomial s of the for m c with correlation coefficients of 0.88 and 0.92, respectively (Figure 3 7). The observed non linearity between and was dependent on the wavelength of the CDOM absorption coefficient. The degree of non linea rity between and decreased with decreasing wavelength approaching a straight line at (Figure 3 8 ) The differences between and can be seen in Figure 3 9 Figure 2 8 Wavelength dependence of the absorption coefficient on the relationship between and for various wavelengths. For comparison purposes, was scaled to the m aximum absorption coefficient for a given wavelength by dividing by
109 Figure 2 9 Plot of and for all sample data used in this study. For the relationship between the CDOM absorption coefficient and fluorescence intensity approximate d a straight line for both the WetStar (Figure 3 10 a) and ECO (Figure 3 10 b) fluorometers with correlation coefficie nts of 0.91 and 0.89, respectively.
110 Figure 2 10 Relationship betweeen CDOM fluorescence and CDOM absorption at 312nm fo r the (a) WetStar and (b) ECO fluorometers. The possibility of interference with the CDOM fluorescence signal by water column constituents other than CDOM, such as phytoplankton and detritus, was explored (a) WetStar (b) ECO
111 using multiple regression techniques. In this stud y , and all covaried with (Figure 3 1 1 ) Figure 2 11 Comparison at 440nm of the CDOM absorption coefficient with the (a) detritus (b) phytoplankton and (c) particulate absorption coefficients.
112 While phytoplankton can contribute to the total CDOM pool, CDOM from terre strial sources far exceeded any contribution from phytoplankton. The correlation between and was c oincident on the fact that during wet periods, both CDOM concentration and phytoplankton productivity we re elevated. The same conclu sion can be inferred about detrital absorption and CDOM absorpt i on A m ultiple regression model using measured with the ECO fluorometer, as the dependent variable, and , and as in dependent variables resulted in strongly significant correlation s at the 95% probability level for and (p<0.01) and and (p<0.01) but resulted in only a weakly significant correlation b etween and (p<0.05 ) (Table 3 3) Similar results were found using the WetStar fluorometer except and were more strongly correlated (p<0.0 1 ) (Table 3 3) Multiple regression analyses were performed to investigate the dependence of on Chlorophyll fluorescence measured with the SeaTech fluorometer was not significantly correlated with (ANOVA; p>0.10) (Table 3 3). Because detritus absorption was significantly correlated with CDOM fluorescen ce a multiple regression model was constructed using only and as independent variables. For the ECO and WetStar fluorometers, model results yielded correlation coefficients of 0.87 and 0.88, respectively (Table 3 3). Table 2 3 Results of multiple regression analyses for establishing the relationship between CDOM fluorescence and the IOPs for shallow seagrass areas in Tampa Bay. ECO WetStar ECO WetStar ECO WetStar
113 3 .4. 2 CDOM and chlorophyll spatial variability Survey data for CD OM fluorescence and chlorophyll concentration were converted into contour plots for the Kitchen SMA using a Kriging method in SURFER 8.0 (Gold en Software) to examine the spatial pattern s of CDOM and chlorophyll fluorescence. Survey results showed an increa se in along the immediate shoreline for August and, to a lesser extent, for April (Figure 3 12). While area wide average doubled from April to August (Table 3 4) and this increase was concentrated close to the shoreline. Table 2 4 Annual summary of CDOM fluorescence voltage and corresponding CDOM absorption for the Kitchen Seagrass Management Area Volts Month Mean stdev Mean stdev Median Max imum Minimum April 1.32 0.240 0.713 0.185 0.656 1.40 0.530 June 1.15 0.206 0.600 0.139 0.578 1.17 0.406 August 2.14 0.534 1.46 0.630 1.28 4.21 0.650 October 1.36 0.114 0.515 0.083 0.504 0.772 0.387 December 1.13 0.245 0.586 0.161 0.539 1.03 0.359
114 Figure 2 12 Survey results of (V) across the Kitchen Seagrass Management Area for (a) April and (b) August 2008. In Chapter 2, it was shown that was positively correlated with rainfall The well established mangrove shoreline appears to be the major CDOM so urce in this case. Under very wet conditions, Bullfrog Creek and the Alafia River can contribute large high concentrations of CDOM to the Kitchen SMA. However, major incursions of high (b) ( a )
115 CDOM water into the Kitchen were not observed from flow through data co llected during this study. A slight CDOM plume was detected at the mouth of Bullfrog Creek during the August survey (Figure 3 12 a ). Presumably, rainfall totals were not great enough to cause significant amounts of CDOM rich river water to enter the Kitchen Total r ainfall 3 0 days prior to sampling was 29.6cm in August compared to only 6.63cm in April This difference was expressed as lower CDOM in April, though still increase d shore ward (Figure 3 12b). The CDOM rich water mass concentrate d over the existing seagrass beds. Hi gh CDOM concentrations may be beneficial to seagrass growing in average depths of less than 0.5m, where UV exposure could be lethal witho ut this protective CDOM layer While CDOM decreases with increasing distance away from the shoreline, any increase in light penetration is offset by increases in average depth. C hlorophyll concentrations displayed a typical seasonal pattern with peak concentrations in August and minimum concentrations during the dry season (Table 3 5 ). Table 2 5 Annual summary of chlorophyll concentrations ( g L 1 ) calculated using chlorophyll fluorescence for the Kitchen Seagrass Management Area Month Mean stdev Median Maximum Minimum April 5.93 1.81 5.69 10.1 2.84 June 11.8 5.08 10.1 30.6 5.04 August 16.3 9.74 11.8 52.1 6.29 October 6.54 1.78 6.09 11.0 4.14 December 6.95 1.72 7.20 10.4 3.69 Unlike CDOM, no obvious spatial pattern s for April August or any other month sampled were readily evident. P hytoplankton dynamics are quite complex and involve a number of factors such as hydrodynamics, tide, season, and nutrient availability and therefore it is not surprising that a spatial pattern was not observed There wa s h owever, a more s ubtle pattern that can be observed for the August survey and to a lesser extent for the April survey as well and that is the presence of small pockets of relatively high chlorophyll concentration located throughout the area (Figure 3 1 3 ) This was visual ly confirmed by sampling personnel during the actual surveys and therefore likely not an artifact of the contouring technique.
116 F igure 2 13 Survey results of across the Kitchen Seagrass Management Area for (a) April and (b) August 2008 expressed in concentration 3 5 Conclusions Using a deck mounted flow through system in shallow seagrass beds like the ones found in Tampa Bay is a novel approach. In this chapter the utility of using such a system ( a ) ( b )
117 for monitoring the spatial variability of the optical environment wit hin a Seagrass Management Area was clearly demonstrated for the Kitchen Raw flow through data were collected in four SM As under different conditions and times of the year. Despite these differences, the correlations between the IOPs and raw fluorescence produced good results. Overall, correlations using all available data were stronger than for individual surveys. This sug gests that the underlying optical properties of these SMAs were similar. For example, CDOM slopes were similar across all SMAs indicating that CDOM may be originating from similar sources. Th at is not to say that this will always be the case across all SMA s or over all time periods, but it does suggest that at least in the SMAs studied here, predictor equations for calculating and are applicable across the SMAs. This work took place in 2008 which was a normal year for rainfall. It is possible that these equations would have to be re calibrated during periods of strong El Nio or La Nia events or in the after the passage of a tropical storm. Further survey work should be conducted in other SMAs and under varying rainfall conditions to better refine the equations presented here. The patterns in CDOM and chlorophyll demonstrate the inherent need to adequate ly characterize the variability in a complex area like the Kitchen. There is a danger in relying too heavily on discrete water samples to infer conditions in areas where conditions are markedly different For example, t he nearest long term water quality mo nitoring station is will outside of the Kitchen SMA yet data from this station are routinely used to evaluate light availability o ver the seagras s. Routine surveillance of SMAs should be part of a regional monitoring program in order to establish a base li ne understanding of CDOM and phytoplankton dynamics. Incorporating this method to an existing program like the Hillsborough County EPC the can be applied to establish optica l characteristics between stations. The flow through system used in this study was designed primarily for oceanographic research. Given the latest advancements in sensors and data processing, a flow through system specifically designed for shallow water en vironments could easily be constructed and would provide
118 an operational and cost effective tool for developing transparency criteria, as well as for other applications.
119 C onclusions Seagrass are important indicator s of estuarine health and, as such, thei r recovery and sustainability are of major importance to resource managers and regulatory interests. The work presented here demonstrated the importance of not only knowing the quantity of light reaching the bottom but also the quality of light. Light qua lity was defined in radiation at the seagrass deep edge. Using this method revealed that seagrass along the deep edge in Tampa Bay are blue light limited and that most of the light being utilized for photosynthesis is coming from the red and even the green wavelengths. Using a GIS based technique of calculating the percent subsurface light at the bottom relative to the surface revealed that on average seagrass are r eceiving 13.6% 18.1% of the blue light at just below the water surface. By comparison, 32% 39% of PAR originating at the surface reaches the bottom. This suggests that the current minimum light target for PAR of 20.5% is too low. The attenuation of bl ue light in the Seagrass Management Area s studied here is primarily caused by CDOM absorption accounting for 60% of the total absorption. Detritus absorption accounted for an additional 20% of the total, leaving only 20% of the total absorption attributabl e to phytoplankton. This challenges the current seagrass management paradigm that chlorophyll a is the dominant absorber of light and that nitrogen limitation will have a direct effect on transparency. While it is important to consider CDOM and detritus wh en setting seagrass restoration and management goals, there may be little that can be done to manage CDOM or detritus. Providing a linkage between the IOPs and parameters more typical of routine monitoring programs is important and adds value to existing datasets. In this study, Seagrass Management Area specific correlations were derived for the various absorption and scatter coefficients. Given the resource management and regulatory importance of the light attenuation coefficient to establishing transpar ency criteria for Florida estuaries, an empirically derived model originally developed using Monte Carlo simulation was used to calculate from the total absorption and scatter coefficients.
120 A major challenge in modeling the underwater light fi eld is capturing the spatial and temporal variability in the optical properties. In this study a deck mounted flow through system was used to survey the optical properties of shallow seagrass areas in Tampa Bay. Empirical relationships were derived between raw fluorescence and the IOPs. This system was very effective at providing a synoptic view of a given area using the Kitchen as a case study.
121 References Abal, E. G., N. Loneragan, P. Brown, C. J. Perry, J. W. Udy, and W. C. Denn ison. 1994. Physiological and morphological responses of the seagrass Zostera capricorni Ashers. to light intensity. Journal of Experimental Marine Biology and Ecology 178: 113 129. Arar, E. J., and G. B. Collins. 1997. In vitro determination of chlorophyl l a and pheophytin a in marine and freshwater algae by fluorescence, p. 22. In N. E. R. Laboratory [ed.]. U.S. Environmental Protection Agency. Belzile, C., C. S. Roesler, J. P. Christensen, N. Shakhova, and I. Semiletov. 2006. Fluorescence measured using the WETStar DOM fluorometer as a proxy for dissolved absorption. Estuarine, Coastal & Shelf Science 67: 441 449. Berwald, J., D. Stramski, C. D. Mobley, and D. A. Kiefer. 1995. Influences of absorption and scattering on vertical changes in the average cosi ne of the underwater light field. Limnology and Oceanography 40: 11. --. 1998. Effect of Raman scattering on the average cosine and diffuse attenuation coefficient of irradiance in the ocean. Limnology and Oceanography 43: 13. Biber, P., C. Gallegos, and W. Kenworthy. 2008. Calibration of a Bio optical Model in the North River, North Carolina (Albemarle Pamlico Sound): A Tool to Evaluate Water Quality Impacts on Seagrasses. Estuaries and Coasts 31: 177 191. Bissett, W. P., J. S. Patch, K. L. Carder, and Z. P. Lee. 1997. Pigment packaging and chl a specific absorption in high light oceanic waters. Limnology and Oceanography 42: 961 968. Blough, N. V., and R. D. Vecchio. 2002. Chromophoric DOM in the Coastal Environment. In D. A. Hansell and C. A. Carlson [ed s.], Biogeochemistry of Marine Dissolved Organic Matter. Academic Press. Bortone, S. A. 2000. Seagrasses : monitoring, ecology, physiology, and management. CRC Press.
122 Branco, A. B. A. J. N. K. 2005. The relative importance of chlorophyll and colored dissol ved organic matter (CDOM) to the prediction of the difuse attenuation coefficient in shallow estuaries. Estuaries 28: 643 652. Bricaud, A., M. Babin, A. Morel, and H. Claustre. 1995. Variability in the chlorophyll specific absorption coefficients of natura l phytoplankton: Analysis and parameterization. Journal of Geophysical Research 100: 13,321. Bricaud, A., and D. Stramski. 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. Brock, J. C., C. W. Wright, A. H. Sallenger, W. B. Krabill, and R. N. Swift. 2002. Basis and methods of NASA airborne topographic mapper LIDAR surveys for coastal studies. Journal of coast al research : JCR 18: 1 13. Butler, W. L. 1962. Absorption of light by turbid materials. Journal of the Optical Society of America 52: 292 299. Buzzelli, C. P., J. Ramus, and H. W. Paerl. 2003. Ferry based monitoring of surface water quality in North Carol ina estuaries. Estuaries 26: 975 984. Cannizzaro, J. P. 2004. Detection and quantification of Karenia brevis blooms on the West Florida shelf from remotely sensed ocean color imagery. University of South Florida. Cannizzaro, J. P., K. L. Carder, F. R. Chen C. A. Heil, and G. A. Vargo. 2008. 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 28: 137 158. Cannizzaro, J. P., K. L. Carder, F. R. Ch en, J. J. Walsh, Z. P. Lee, and C. Heil. 2004. A novel optical classification technique for detection of red tides in the Gulf of Mexico: Application to the 2001 2002 bloom event, p. 137 158. In F. F. a. W. C. C. a. I. O. C. o. UNESCO [ed.], Harmful Algae 2002. Cannizzaro, J. P., C. Hu, D. C. English, K. L. Carder, C. A. Heil, and F. E. Muller Karger. 2009. Detection of Karenia brevis blooms on the west Florida shelf using in situ backscattering and fluorescence data. Harmful Algae 8: 898 909.
123 Carder, K. L. R. G. Steward, G. R. Harvey, and P. B. Ortner. 1989. Marine humic and fulvic acids: Their effects on remote sensing of ocean chlorophyll. Limnology and Oceanography 34: 68 81. Carder, K. L., Z. P. Lee, J. Marra, R. G. Steward, and M. J. Perry. 1995. Calc ulated quantum yield of photosynthesis of phytoplankton in the Marine Light Mixed Layers (59 o N, 21 o W). Journal of Geophysical Research 100: 6655 6663. Chen, Z., C. Hu, R. N. Conmy, F. Muller Karger, and P. Swarzenski. 2007. Colored dissolve organic matte r in Tampa Bay, Florida. Marine Chemistry 104: 98 109. Cloern, J. E. 2001. Our evolving conceptual model of the coastal eutrophication problem. Marine ecology progress series 210: 223 253. Coble, P. G. 2007. Marine Optical Biogeochemistry: The Chemistry of Ocean Color. Chemical reviews 107: 17. Conmy, R. N. 2008. Temporal and Spatial Patterns in Optical Properties of Colored Dissertation. University of South Florida. Conmy, R. N., P. G. Coble, R. F. Chen, and G. B. Gardner. 2004. Optical properties of colored dissolved organic matter in the Northern Gulf of Mexico. Marine Chemistry 89: 127 144. Cummings, M. E., and R. C. Zimmerman. 2003. Light harvesting and the package effect in t he seagrasses Thalassia testudinum Banks ex Konig and Zostera marina L : optical constraints on photoacclimation. Aquatic botany 75: 261. Dennison, W., C. and others 1993. Assessing water quality with submersed aquatic vegetation. BioScience 43: 86 94. Den nison, W. C. 1987. Effects of light on seagrass photosynthesis, growth, and depth distribution. Aquatic Botany 27: 15 26. Dennison, W. C., and R. S. Alberte. 1982. Photosynthetic response of Zostera marina L. (eelgrass) to in situ manipulations of light in tensity. Oecologia 55: 137 144. Dixon, K. L., and J. R. Leverone. 1995. Light Requirements of Thalassia testudinum in Tampa Bay, Florida, p. 77. Mote Marine Laboratory.
124 Dixon, L. K. 2000. Establishing Light Requirements for the Seagrass Thalassia testudinu m : An example from Tampa Bay, Florida. In S. A. Bortone [ed.], Seagrasses: Monitoring, Ecology, Phsiology, and Management. Marine science series. CRC Press. Duarte, C. M. 1991. Seagrass depth limits. Aquatic Botany 40: 363 377. --. 2002. The future of sea grass meadows. Environmental Conservation 29: 15. Duarte, C. M., N. Marba, D. Krause Jensen, and M. Sanchez Camacho. 2007. Testing the Predictive Power of Seagrass Depth Limit Models. Estuaries and coasts : journal of the Estuarine Research Federation 30: 5. Durako, M. J. 2007. Leaf optical properties and photosynthetic leaf absorptances in several Australian seagrasses. Aquatic Botany 87: 7. EPCHC 2007. Tampa Bay Seagrass Management Area Determination, Status and Trends, and Water Quality Target Attainment Technical Memorandum, p. 84. In E. R. M. Division [ed.]. Eaton, A. D., L. S. Clescerl, E. W. Rice, A. E. Greenberg, and M. A. H. Franson [eds.]. 2005. Standard Methods for Examination of Water & Wastewater: 21st Edition. American Public Health Associatio n. Enriquez, S. 2005. Light absorption efficiency and the package effect in the leaves of the seagrass Thalassia testudinum. Marine ecology progress series 289: 10. Enriquez, S., S. Agusti, and C. M. Duarte. 1992. Light absorption by seagrass Posidonia oce anica leaves. Marine ecology progress series 86: 201 204. Ensign, S. H., and H. W. Paerl. 2006. Development of an unattended estuarine nutrient monitoring program using ferries as data collection platforms. Limnology and Oceanography: Methods 4: 399 405. F errari, G. M., and M. D. Dowell. 1998. CDOM absorption characteristics with relation to fluorescence and salinity in coastal areas of the southern Baltic Sea. Estuarine, Coastal & Shelf Science 47: 91 105. Ferrari, G. M., N. Hoepffner, and M. Mingazzini. 1 996. Optical properties of the water in a deltaic environment: Prospective tool to analyze stellite data in turbid waters. Remote Sensing of Environment 58: 69 80.
125 FDEP. 2009. Standard Operating Procedure for: Spectrophotometric Determination of Corrected and Uncorrected Chlorophyll a and Pheophytin, p. 14. In B. S. Bureau of Laboratories [ed.]. Fourqurean, J. W., N. Marba, and C. M. Duarte. 2003. Comment Elucidating seagrass population dynamics: Theory, constraints, and practice. Limnology and oceanograp hy 48: 4. Gallegos, and Kenworthy. 1996. Seagrass Depth Limits in the Indian River Lagoon (Florida, U S A ): Application of an Optical Water Quality Model. Estuarine, Coastal & Shelf Science 42: 22. Gallegos, C. L. 1994. Refining habitat requirements of su bmersed aquatic vegetation: Role of optical models. Estuaries 17: 187 199. --. 2001. Calculating optical water quality targets to restore and protect submersed aquatic vegetation: Overcoming problems in partitioning the diffuse attenuation coeficient for photosynthetically active radiation. Estuaries 24: 381 397. --. 2005. Optical water quality of a blackwater river estuary: The Lower St. Johns River, Florida, USA. Estuarine, Coastal, and Shelf Science 63: 57 72. Gallegos, C. L., D. L. Correll, and J. W. Pierce. 1990. Modeling spectral diffuse attenuation, absorption, and scattering coefficients in a turbid estuary. Limnology and Oceanography 35: 1486. Gordon, H. R., D. K. Clark, J. W. Brown, O. B. Brown, R. H. Evans, and W. W. Broenkow. 1983. Phytoplankto n pigment concentrations in the middle Atlantic Bight: Comparison of ship determinations and CSCZ estimates. Applied Optics 22: 20 36. Greening, H. 2004. Factors Influencing Seagrass Recovery in Feather Sound, Tampa Bay, Florida. Tampa Bay Estuary Program. Griffen, L., and H. Greening. 2004. Factors influencing seagrass recovery in Feather Sound, Tampa Bay, Florida. In T. B. E. Program [ed.]. Guenther, G. C., A. G. Cunningham, P. E. Larocque, and D. J. Reid. 2000. Meeting the accuracy challenge in airborne LIDAR bathymetry, p. 28. European Assiciation of Remote Sensing Laboratories.
126 Hall, M. O., D. A. Tomasko, and F. X. Courtney. 1991. Responses of Thalassia testudinum to in situ light reduction, p. 77 86. The Light Requirements of Seagrasses: Proceedings of a Workshop to Examine the Capability of Water Quality Criteria, Standards and Monitoring Programs to Protect Seagrasses. Hansell, D. A., and C. A. Carlson [eds.]. 2002. Biogeochemistry of Marine Dissolved Organic Matter. Elsevier Science & Technology Book s. Hansen, M., M. Crane, C. Hearn, and K. Yates. 2005. Gulf of Mexico Integrated Science Tampa Bay Study Watershed and Estuary Mapping. In U. S. G. Survey [ed.]. Hemminga, M. A., and C. M. Duarte. 2000. Seagrass Ecology. Cambridge University Press. Herzk a, S., and K. Dunton. 1997. Seasonal photosynthetic patterns in the seagrass Thalassia testudinum in the western Gulf of Mexico. Marine ecology progress series 152: 103 117. Hickman, G. D., and J. E. Hogg. 1969. Application of an airborne pulsed laser for near shore bathymetric measurements. Remote Sensing of Environment 1: 47 58. Hill, K. 2002. Seagrass Habitats. Smithsonian Marine Station at Fort Pierce. Hoge, F. E., A. Vodacek, and N. V. Blough. 1993. Inherent optical properties of the ocean: Retrieval o f the absorption coefficient of chromophoric dissolved organic matter from fluorescence measurements. Limnology and Oceanography 38: 1394 1402. Holm Hansen, O., and B. Rieman. 1978. Chlorophyll a determination: improvements in methodology. Oikos 30: 438 44 7. Janicki, A., R. Pribble, and K. Hackett. 2003. Nitrogen Loadings to Tampa Bay: Model Based Estimates of 1998 and 2010 Loads to Major Basins, and TN Load Reduction/Preclusion Apportionment. Tampa Bay Estuary Program. Janicki, A. and D. Wade 1996. Estim ating critical external nitrogen loads for the Tampa Bay Estuary: An empirically based approach to setting management targets Tampa Bay Estuary Program.
127 Janicki Environmental, I. 2001. Tampa Bay Estuary Program Model Evaluation and Update Model Evaluation and Update: Chlorophyll a Light Attenuation Relationship, p. 35. Tampa Bay Estuary Program. Kenworthy, W. J., M. J. Durako, S. M. R. Fatemy, H. Valavi, and G. W. Thayer. 1993. Ecology of Seagrasses in Northeastern Saudi Arabia One Year After the Gulf Wa r Oil Spill. Marine pollution bulletin 27: 213. Kenworthy, W. J., and M. S. Fonseca. 1996. Light requirements of seagrasses Halodule wrightii and Syringodium filiforme derived from the relationship between diffuse light attenuation and maximum depth distri bution. Estuaries 19: 740 750. Kiefer, D. A., and J. B. Soohoo. 1982. Spectral absorption by marine particles of coastal waters of Baja California. Limnology and Oceanography. Limnology and Oceanography 27: 492 499. Kirk, J. T. 1981. Estimation of the scat tering coefficient of natrual waters using underwater irradiance measurements. Journal of Marine and Freshwater Research 32: 533 539. --. 1984. Dependence of rel ationship between inherent and apparent optical properties of water on solar altitude. Limnology and Oceanography 29: 350 356. --. 1994. Light and photosynthesis in aquatic ecosystems, Second Edition ed. Cambridge University Press. Kirk, J. T. O. 1991. Vo lume scattering function, average cosines, and the underwater light field. Limnology and Oceanography 36: 455 467. --. 2003. The vertical attenuation of irradiance as a function of the optical properties of the water. Limnology and Oceanography 48: 9. Kis hino, M., M. Takahashi, N. Okami, and S. Ichimura. 1985. Estimation of the spectral absorption coefficients of phytoplankton in the sea. Bulletin of Marine Science 37: 634 642. Kurz, R. 2002. Seagrass Mapping: Accuracy Issues, p. 246. In H. S. Greening [ed .], Seagrass Management: It's Not just Nutrients! Tampa Bay Estuary Program.
128 Lawson, S. E., P. L. Wiberg, K. J. Mcglathery, and D. C. Fugate. 2007. Wind driven Sediment Suspension Controls Light Availability in a Shallow Coastal Lagoon. Estuaries and coast s : journal of the Estuarine Research Federation 30: 11. Lee, Z., K. L. Carder, C. D. Mobley, R. G. Steward, and J. S. Patch. 1998. Hyperspectral remote sensing for shallow waters: 1. A semianalytical model. Applied Optics 37: 6329 6338. Longstaff, B. J., and W. C. Dennison. 1999. Seagrass survival during pulsed turbidity events: The effects of light deprivation on the seagrasses Halodule pinifolia and Halophila ovalis Aquatic Botany 65: 105 121. Madden, C. J., and J. W. Day. 1992. An instrument system for high speed mapping of chlorophyll a and physico chemical variables in surface waters. Estuaries 15: 421 427. Major, K. M., and K. H. Dunton. 2002. Variations in light harvesting characteristics of the seagrass, Thalassia testudinum: evidence for photoaccl imation. Journal of Experimental Marine Biology and Ecology 275: 18. Malick, L. A. 2004. Light quality and phytoplankton viability. University of South Florida. Menon, H. B., A. Lotliker, and S. R. Nayak. 2006. Pre monsoon bio optical properties in estuari ne, coastal and Lakshadweep waters. Estuaries and Coasts 63: 211 223. Miller, R. L., and B. F. Mcpherson. 1995. Modeling Light Available to Seagrasses in Tampa Bay, Florida. Florida scientist 58: 116. Miller, R. L. A. B. F. M. 1995. Modeling photosynthetic ally active radiation in water of Tampa Bay, Florida, with emphasis on the geometry of incident irradiance. Estuarine, Coastal, and Shelf Science 40: 359 377. Minor, E. C., J. Pothen, B. J. Dalzell, H. Abdulla, and K. Mopper. 2006. Effects of salinity chan ges on the photodegredation and ultra violet visible absorbance of terrestrial dissolved organic matter. Limnology and Oceanography 51: 2181 2186. Mobley, C. D. 1994. Light and Water: Radiative Transfer in Natural Waters. Academic Press.
129 Morel, A. 1978. Av ailable, usable, and stored radient energy in relation to marine photosynthesis. Deep sea Research 25: 673 688. --. 1991. Light and marine photosynthesis: a spectral model with geochemical and climatological implications. Progress in Oceanography 26: 263 306. Morel, A., and L. Prieur. 1977. Analysis of variations in ocean color. Limnology and Oceanography 22: 709 722. Nelson, N. B., and C. Y. Robertson. 1993. Detrital spectral absorption: laboratory studies of visible light effects on phytodetritus absorpt ion, bacterial spectral signal, and comparison to field measurements. Journal of Marine Research 51: 181 207. Nielsen, S. L., K. Sand Jensen, J. Borum, and O. Geertz Hansen. 2002. Depth Colonization of Eelgrass (Zostera marina) and Macroalgae as Determined by Water Transparency in Danish Coastal Waters. Estuaries 25: 8. Paerl, H. W., K. L. Rossignol, R. Guajardo, N. S. Hall, A. R. Joyner, and B. L. Peierls. 2009. FerryMon: Ferry based monitoring and assessment of human and climatically driven environmental change in the Albemarle Pamlico Sound system. Environmental science & technology 43: 7609 7613. Pope, R. M., and E. S. Fry. 1997. Absorption spectrum (380nm 700nm) of pure water. II. Integrating cavity measurements. Applied Optics 36: 8710 8723. Preisendor fer, R. W. 1961. Application of radiative transfer theory to light measurements in the sea. IUGG Monographs 10: 11 29. Pribble, R., A. Janicki, H. Zarbock, S. Janicki, and M. Winowitch. 2001. Estimates of Total Nitrogen, Total Phosphorus, Total Suspended S olids, and Biochemical Oxygen Demand Loadings to Tampa Bay. Tampa Bay Estuary Program. Prieur, L., and S. Sathyendranath. 1981. An optical classification of coastal and oceanic waters based on the specific spectral absorption curves of phytoplankton pigmen ts, dissolved organic matter, and other particulate materials. Limnology and Oceanography 26: 671 689. Ralph, P. J., M. J. Durako, S. Enriquez, C. J. Collier, and M. A. Doblin. 2007. Impact of light limitation of seagrass. Journal of Experimental Marine Bi ology and Ecology 350: 176 193.
130 Roesler, C. S., M. J. Perry, and K. L. Carder. 1989. Modeling in situ phytoplankton absorption from total absorption spectra in productive inland marine waters. Limnology and Oceanography 34: 1510 1523. Rowan, K. S. 1989. Ph otosynthetic Pigments of Algae. Cambridge University Press. Siegel, D. A., and A. F. Michaels. 1996a. Quantification of non algal light attenuation in the Sargasso Sea: Implications for biogeochemistry and remote sensing. Deep sea research. Part II, Topica l studies in oceanography 43: 26. --. 1996b. Quantification of non algal light attenuation in the Sargasso Sea: Implications for biogeochemistry and remote sensing. Deep Sea Research 43: 321 345. Smith, R. C., and K. S. Baker. 1978. The bio optical state of ocean waters and remote sensing. Limnology and Oceanography 23: 247 259. Smith, R. C. A. K. S. B. 1981. Optical properties of the clearest natural waters (200 800 nm). Applied Optics 20: 177 184. Stedmon, C. A., S. Markager, and H. Kaas. 2000. Optical P roperties and Signatures of Chromophoric Dissolved Organic Matter (CDOM) in Danish Coastal Waters. Estuarine, Coastal & Shelf Science 51: 267 278. Stedmon, C. A., Stiig Markager, Rasmus Bro. 2003. Tracing dissolved organic matter in aquatic environments us ing a new approach to fluorescence spectroscopy. Marine Chemistry 82: 239 254. Steinberg, D. K., N. B. Nelson, C. A. Carlson, and A. C. Prusak. 2004. RESEARCH ARTICLES Production of chromophoric dissolved organic matter (CDOM) in the open ocean by zoopla nkton and the colonial cyanobacterium Trichodesmium spp. Marine ecology progress series 267: 12. Steward, J. S., and W. C. Green. 2007. Setting Load Limits for Nutrients and Suspended Solids Based upon Seagrass Depth limit Targets. Estuaries and Coasts 30: 657 670. Steward, J. S., R. W. Virnstein, L. J. Morris, and E. F. Lowe. 2005. Setting seagrass depth, coverage, and light targets for the Indian River Lagoon. Estuaries and Coasts : 923 935.
131 Stramski, D. and others 2008. Relationships between the surface c oncentration of particulate organic carbon and optical properties in the eastern South Pacific and eastern Atlantic Oceans. Biogeosciences 5: 171 201. T BNEP 1996. Charting the Course: The Comprehensive Conservation and Management Plan for Tampa Bay, p. 27 4. Tampa Bay National Estuary Program. Thayer, G. W., K. A. Bjorndal, J. C. Ogden, S. L. Williams, and J. C. Zieman. 1984. Role of Larger Herbivores in Seagrass Communities. Estuaries 7: 351 376. Tomasko, D. A., C. J. Dawes, and M. O. Hall. 1996. The effec ts of anthropogenic nutrient enrichment on Turtle Grass ( Thalassia testudinum ) in Sarasota Bay, Florida. Estuaries 19: 448 456. Tomasko, D. A., and B. E. Lapointe. 1991. Productivity and biomass of Thalassia testudinum as related to water column nutrient a vailability and epiphytic levels: field observations and experimental studies. Marine ecology progress series 75: 9 16. Tomasko, D. A. B., D. L. and J. A. Ott. 2001. Assessment of present and future nitrogen loads, water quality, and seagrass ( Thalassia testudinum ) depth distribution in Lemon Bay, Florida. Estuaries 24: 926 938. Twardowski, M. S., E. Boss, J. M. Sullivan, and P. L. Donaghay. 2004. Modeling the spectral shape of absorption by chromophoric dissolve organic matter. Marine Chemistry 89: 69 88 Twardowski, M. S., M. R. Lewis, A. H. Barnard, and J. R. Zaneveld. 2005. In water instrumentation and platforms for ocean color remote sensing applications, p. 345. In R. Miller, C. E. D. Castillo and B. A. McKee [eds.], Remote Sensing of Coastal Aquatic Environments. Springer. Tyler, D. and others 2007. Topobathymetric data for Tampa Bay, Florida: U.S. Geological Survey Open File Report 2007 1051 (revised) U SEPA 1999. Guidance Manual for Compliance with the Interim Enhanced Surface Water Treatment Rule : Turbidity Provisions. Office of Water.
132 Vanderbloemen, L. A. 2006. Satellite analysis of temporal and spatial chlorophyll patterns on the West Florida shelf (1997 2003). University of South Florida. Warrior, H., and K. L. Carder. 2005. Production of hyper saline pools in shallow basins by evaporation. Geophysical Research Letters 32: 5. Weidemann, A. D., and T. T. Bannister. 1986. Absorption and scattering coefficients in Irondequoit Bay. Limnology and Oceanography 31: 567 583. Weiner, R. F., D. M. Osborn, D. Hinojosa, T. J. Heames, J. Penisten, and D. Orcutt. 2008. RADCAT 2.3 Users Guide. In S. N. Laboratories [ed.]. Sandia National Laboratories. Yentsch, C. S. 1962. Measurement of visible light absorption by particulate matter in the ocean. Limnology and O ceanography 7: 207 217. Zieman, R. C., and R. T. Zieman. 1989. The ecology of the seagrass meadows of the west coast of Florida: A community profile. In U. S. F. a. W. Service. [ed.]. Zimmerman, R. C. 2003. A biooptical model of irradiance distribution and photosynthesis in seagrass canopies. Limnology and Oceanography 48: 568 585.
About the Author Chris Anastasiou started his PhD at the University of South Florida (USF) College of Marine Science (CMS) in 2004 while working full time wi th the Southwest Florida Water Management District as an environmental scientist. In 2005, Chris began working for the Florida Department of Environmental Protection where he developed new techniques and t echnologies for coastal zone and estuarine resource m anagement Before starting his PhD also at USF in Tampa. I n 2004 Chris also received his commission as a Nav y Reserve Officer within the Naval Meteorology and Oceanography Command. He is currently a Lieutenant and the E xecutive O fficer of the Naval Meteorology and Oceanography Reserve Activity aboard Naval Station Mayport, FL.
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 2200409Ka 4500
controlfield tag 001 002068400
007 cr mnu|||uuuuu
008 100413s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003148
Anastasiou, Christopher J.
Characterization of the underwater light environment and its relevance to seagrass recovery and sustainability in Tampa Bay, Florida
h [electronic resource] /
by Christopher J. Anastasiou.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 132 pages.
Dissertation (Ph.D.)--University of South Florida, 2009.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Kendall Carder, Ph.D.
ABSTRACT: The availability of light is a primary limiting factor for seagrass recovery and sustainability. Understanding not only the quantity but the quality of light reaching the bottom is an important component to successful seagrass management and the key focus of this study. This study explores the spectral properties of the sub-surface light field in four shallow Seagrass Management Areas (SMA) in Tampa Bay. Wavelength-specific photosynthetically active radiation (PAR()) and the spectral light attenuation coefficient (K[subfield d]()) are used to estimate the percent blue, green, and red light remaining at the bottom relative to the surface. LIDAR Bathymetry is combined with K[subfield d]() to produce high-resolution maps of percent subsurface light along the seagrass deep edge.The absorptance spectra from two seagrass species together with PAR() is used to calculate the photosynthetically useable radiation (PUR()), a term describing the actual wavelengths of light being used by the seagrass. Based on the average annual K[subfield d](), 32% 39% percent of PAR reached the bottom at the seagrass deep edge, while only 14% 18% of blue light reached bottom, suggesting that seagrass may be blue-light limited. Analysis of PUR() data further confirmed that seagrass are blue-light limited. Each SMA was characterized in terms of the inherent optical properties (IOP) of absorption and scatter. Tampa Bay is considered a chlorophyll-dominated estuary. However, in this study, colored dissolved organic matter (CDOM) was the major absorber of blue light, accounting for 60% of the total absorption. To infer past light conditions, the IOPs were related to parameters more commonly used in routine monitoring programs.To estimate K[subfield d]() an empirically-derived model using only the total absorption and scatter coefficients was used and resulted in a good fit between measured K[subfield d](480) and modeled K[subfield d](480). A deck-mounted flow-through system was used to survey each SMA for CDOM and chlorophyll a fluorescence, among other properties. A series of SMA-specific predictor equations were empirically derived to relate raw fluorescence to the IOPs. The Kitchen SMA was used as a case study. Survey results show a strong connection between CDOM-rich waters and the mangrove-dominated shoreline.
x Marine Science
t USF Electronic Theses and Dissertations.