CONTROL OF FLORIDA RED TIDES USING PHOSPHATIC CLAY

FINAL REPORT

Donald M. Anderson Principal Investigator

with

Mario R. Sengco, Aishao Li and Stace E. Beaulieu

WOODS HOLE OCEANOGRAPHIC INSTITUTION Woods Hole, MA 02543

Prepared for

FLORIDA INSTITUTE OF PHOSPHATE RESEARCH 1855 West Main Street

Bartow, Florida 33830 USA

Contract Manager: Steven G. Richardson FIPR Project Number: 99-03-138

July 2004

DISCLAIMER

The contents of this report are reproduced herein as received from the contractor. The report may have been edited as to format in conformance with the FIPR Style Manual. The opinions, findings and conclusions expressed herein are not necessarily those of the Florida Institute of Phosphate Research, nor does mention of company names or products constitute endorsement by the Florida Institute of Phosphate Research. © 2004, Florida Institute of Phosphate Research.

iii

PERSPECTIVE

The phosphate industry in Florida produces large quantities (about 30 million tons per year) of phosphatic clay as a byproduct of mining and washing of the phosphate ore or “matrix.” Following the washing step, the dilute clay slurry (initially about 3% solids) is pumped to large settling ponds, often about one square mile (about 250 ha) each, to begin the process of dewatering and consolidation. The phosphatic clays are not pure clays but also contain fine, unrecoverable particles of the phosphate mineral (apatite), plus clay-sized particles of dolomite, calcite, silica, etc., in addition to clay minerals. The phosphatic clays also contain small amounts of metals and radionuclides.

Rather than look upon the phosphatic clay as a waste, the Florida Institute of Phosphate Research considers the phosphatic clay a byproduct that potentially has other uses. In addition to the topic of this report on the use of phosphatic clay as an agent to help control red tide algae through the process of flocculation and settling, phosphatic clay has been and is being tested for making tile, bricks, aggregate, a component of cement and concrete, etc. It has also been shown to be a fertile agricultural soil.

Before the phosphatic clay can be recommended as an agent to help control red tide algae blooms, several questions need to be answered. Is the phosphatic clay effective in removing red tide organisms through flocculation? Is it practical and economical? Is it safe; are there any environmental problems that might come from clay treatment? Do the potential benefits outweigh the potential drawbacks? The research reported here is based mainly on laboratory studies, but results of a field test using 2m-diameter polyvinyl cylinders immersed in red-tide-affected waters near Sarasota Bay are also included.

A variety of phosphatic clay samples were very effective in removing Karenia brevis (the Florida red tide alga) from the water column when clay loading was 0.25 g/L or greater. Low level additions (e.g., 5 ppm) of poly aluminum chloride enhanced the removal ability of the phosphatic clay, especially at lower clay loading rates. Algal cell death depended on the cells remaining trapped in the settled flocs, so factors that might affect resuspension and escape of the algal cells, such as currents, were also studied. The study also includes a preliminary examination of potential environmental impacts, as well as engineering requirements and costs associated with treatment of a red tide affected area. Steven G. Richardson Reclamation Research Director

v

ABSTRACT

Red-tides of the dinoflagellate Karenia brevis are a serious and recurrent problem in Florida coastal waters, causing severe impacts on public health, marine life, and regional economies. This study evaluated the effectiveness, practicability, and possible impacts of clay dispersal as a means of controlling these phenomena. The method, used effectively in Japan and Korea to protect fish mariculture, is based on the co-flocculation between microalgae and mineral particles, leading to the formation of larger aggregates, which settle faster and further entrain cells during descent. In modified jar tests, phosphatic clays from varying stages in beneficiation, pond age, and geographic origin, all displayed > 80% removal efficiency (RE), with ≥ 0.25 g/L at 9,000 cells/mL. Phosphate rock only showed 10% RE. Lower clay amounts combined with 5 ppm of polyaluminum chloride (PAC) increased RE by one order of magnitude, while alum and two cationic polymers were less effective. Cell removal > 70% required a minimum of 1,000 cells/mL at ≥ 0.10 g/L. Post-treatment mortality depended on clay concentration, resuspension frequency, and the duration of clay/cell contact prior to first resuspension. Bulk chemical analysis showed elevated Cr content relative to standard sediments, while analysis of the supernatant revealed significant phosphate release (moderated by PAC), and possibly 210Pb. Flume experiments defined flow speeds that influence cell removal (settling) and erodibility (including the effect of consolidation over time and PAC addition). The primary logistical and engineering challenges of possible field implementation were associated with handling and processing of wet clays, and finding a suitable method of clay dispersal at larger scales. The highest costs were estimated in clay transport, followed by ship/crew hiring. In summary, this study demonstrated the efficacy of phosphatic clays to remove and kill Karenia brevis in laboratory experiments, the physicochemical, hydrodynamic and scale conditions that can influence cell removal and sedimentation, balanced by the possible impacts on seawater chemistry, sediment properties, and economic costs. These findings provide crucial information needed to assess this bloom control strategy, but also indicate a need to examine this concept further under well-defined field conditions.

vi

ACKNOWLEDGMENTS

We are grateful to John Keating and his coworkers at IMC Phosphates, Inc., (of Florida) for providing samples of phosphatic clay and technical support during this project. We also thank Cytec Industries (Charlotte, NC) and Ciba Specialty Chemical (Suffolk, VA) for providing samples of coagulants and flocculants. We appreciate the technical contributions provided by members of the Anderson lab at the Woods Hole Oceanographic Institution (WHOI), especially David M. Kulis and Judy Kleindinst. Radionuclide analysis was performed through the help of Alan Fleer from the Woods Hole Oceanographic Institution. Additional thanks go to C. DiBacco and D. Kulis for help with culture rooms, S. Faluotico, S. Fries, J. Savage, J. Sisson, and P. Traykovski for help with flume experiments, and J. Trowbridge for guidance in analyzing bottom stress in the flumes. The mesocosm experiments were conducted in collaboration with researchers at Mote Marine Laboratory: Richard Pierce, Michael Henry, and Jim Culter. Information for cost analysis and engineering requirements was provided by Bob Dahlquist and his team from J.M. Huber Co. (Macon, GA.).

Other funding sources for this research included START (Solutions to Avoid Red Tides), Sholley Foundation, Inc., Cove Point Foundation, United States Environmental Protection Agency, Woods Hole Summer Student Fellowships, the Ford Foundation Pre-Doctoral Fellowships, and the Woods Hole Oceanographic Institution Academic Programs Office.

vii

TABLE OF CONTENTS

PERSPECTIVE.................................................................................................................. iii

ABSTRACT.........................................................................................................................v ACKNOWLEDGMENTS ................................................................................................. vi EXECUTIVE SUMMARY .................................................................................................1 INTRODUCTION ...............................................................................................................5

Control Strategies.....................................................................................................6 Flocculation..............................................................................................................7 Previous Work on Clay Flocculation and Its Application .......................................8 Preliminary Research at Woods Hole ......................................................................9 Objectives and Project Framework........................................................................10

METHODOLOGY ............................................................................................................13

Phosphatic Clay Screening Against Karenia Brevis..............................................13 Cell Removal Using Phosphatic Clay and Flocculants..........................................16 Varying Cell Concentration and Water Column Height........................................16 Cell Viability and Growth......................................................................................17 Seawater Chemistry Following Clay Treatment....................................................19 Physical Behavior of Settled Clay/Cell Flocs........................................................20 Oxygen Demand at the Sediment ..........................................................................25

RESULTS ..........................................................................................................................29

Cell Removal Efficiency........................................................................................29 Coagulants and Flocculants ...................................................................................29 Cell Concentration and Water-Column Height......................................................34 Viability and Growth After Treatment ..................................................................40 Seawater Chemistry ...............................................................................................42 Physical Behavior of Settled Flocs ........................................................................50 Oxygen Demand in Mesocosm Experiments.........................................................56

CONCLUSIONS AND RECOMMENDATIONS ............................................................59

Clay Effectiveness .................................................................................................59 Cell Viability, Recovery and Growth ....................................................................62 Potential Impacts of Clay Dispersal.......................................................................63

viii

TABLE OF CONTENTS (CONT.)

Seawater Chemistry ...................................................................................63 Behavior of Clay Flocs on the Ocean Floor...............................................67 Oxygen Demand ........................................................................................67

ENGINEERING REQUIREMENTS AND TREATMENT COSTS ................................67

Phosphatic Clay Acquisition and Transport to Treatment Site..............................68 Preparation and Dispersal of Phosphatic Clay Slurry............................................71 Ship and Crew Hiring ............................................................................................75

REFERENCES ..................................................................................................................79 APPENDIX A: Flume Experiments .............................................................................. A-1

ix

LIST OF FIGURES Figure Page 1 Florida red tide organism, Karenia brevis, formerly called Gymno- dinium breve..................................................................................................5 2 Location of sampling sites where phosphatic clay was taken..........................15 3 Settling columns to determine cell removal with clay at several depth intervals.......................................................................................................18 4 Mesocosms (called limnocorrals) used in Florida field experiments with clay flocculation..................................................................................26 5 Removal efficiency of the Florida red tide organism, Karenia brevis, by several phosphatic clays.........................................................................30 6 Removal efficiency of the Florida red tide organism, Karenia brevis, by phosphatic clays and phosphate rock.....................................................31 7 Removal ability of Florida phosphatic clays against Karenia brevis in combination with selected flocculants and coagulants ...............................32 8 Removal ability of Florida phosphatic clay IMC-P2 against the red tide organism, Karenia brevis, with and without the flocculant poly- aluminum chloride (PAC)...........................................................................33 9 Removal efficiency of Karenia brevis at various cell concentrations treated with Florida phosphatic clay (IMC-P2) ..........................................35 10 Removal efficiency of Karenia brevis at various depths during clay treatment with IMC-P2 phosphatic clay .....................................................36 11 Removal efficiency of Karenia brevis using IMC-P2 phosphatic clays at various depths following the 2-hour incubation .....................................37 12 Viability and growth of Karenia brevis cells following treatment with IMC-P2 phosphatic clay .............................................................................41 13 Dynamics of biologically-significant inorganic nutrients in suspension with selected clays ......................................................................................49 14 Critical bed shear stress (τ0crit) for resuspension of clay/algal flocs in the 17-m flume..................................................................................................50

x

LIST OF FIGURES Figure Page 15 Optical backscatter (OBS) plots for racetrack flume runs ...............................52 16 Cell counts for racetrack flume runs................................................................55 17 Karenia brevis removal in mesocosm experiments using IMC-P4 phosphatic clay............................................................................................56 18 Biological oxygen demand, total settled solids and volatile settled solids from mesocosm experiments ............................................................57 19 Clay dispersal machine from South Korea ......................................................73 A-1 Flume Experiments ..............................................................................A-2 - A-4

xi

LIST OF TABLES

Table Page 1 Samples of Florida phosphatic clay tested for their ability to remove the Florida red tide organism, Karenia brevis, from suspension, through flocculation and settling ................................................................14 2 Summary of Woods Hole flume experiments..................................................23 3 Observed arrival times of clay pulse, sinking rates and predicted aggregate size..............................................................................................39 4 Recovery of Karenia brevis following clay treatment with Florida phosphatic clay............................................................................................40 5a Chemical analysis of several samples of Florida phosphatic clays from various sources in central Florida ......................................................43 5b Published analyses of Florida phosphatic clays and phosphorites from various sources............................................................................................44 6 Sediment standards provided by U.S. EPA Gulf Breeze Laboratory ..............45 7 Comparison of phosphatic clay and natural sediment quality standards .........46 8 Radioactivity measurements of Florida phosphatic clays................................47 9 Critical shear velocity (u*crit) for the resuspension of clay/algal flocs in the 17-m flume........................................................................................51 10 Critical shear velocities (u*crit) for resuspension of clay/algal flocs in the racetrack flume............................................................................................53 11 Clay deposition and algal removal efficiency (RE) in the racetrack flume.....54 12 Cost analysis of red tide treatment using phosphatic clays: estimated amount of clay required and freight charges...............................................69 13 Cost of hiring offshore supply vessels (OSVs) for clay transport and dispersal ......................................................................................................76 14 Total cost of treatment .....................................................................................77

1

EXECUTIVE SUMMARY

Harmful algal blooms (HABs), commonly known as red tides, are aquatic phenomena caused by the rapid growth and accumulation of certain microalgae, which can lead to marked discoloration of surface waters, and severe impacts on public health, commerce, and the environment. In Florida, blooms of the dinoflagellate, Karenia brevis (formerly Gymnodinium breve), have been a serious and recurrent problem, especially along the Gulf coast. Karenia blooms have caused shellfish toxicity, human illnesses from the consumption of contaminated shellfish and aerosolized toxins, mortalities of wild fish and manatees, and economic losses reaching tens of millions of dollars. Although the significance and consequences of these phenomena would seem to justify bloom control as a high priority topic for research, virtually no focused investigation has been undertaken in the United States for nearly 40 years (Boesch and others 1997). By contrast, countries such as Japan, China, South Korea and Australia, have actively pursued bloom control and mitigation with some notable successes. The studies presented in this report evaluated a promising method of dispersing clay minerals to physically remove the organisms from suspension, thereby minimizing or preventing their impacts. This strategy has been used effectively in Japan and South Korea to minimize the damage of local blooms on commercially-valuable fish mariculture. The method is based on the principle of mutual flocculation between algal cells and mineral particles, leading to the formation of larger aggregates, which rapidly settle and further entrain cells during descent. Clay minerals were deemed suitable for this purpose because they are naturally occurring particles that are commonly found in marine systems, abundant, readily available, relatively inexpensive and easy to use. Moreover, clay addition is thought to have a low probability of causing environmental damage. For these reasons, this research program was mounted to examine the efficacy, practicability and potential impacts of clay dispersal, specifically using Florida phosphatic clays, to control the Florida red tide.

In laboratory experiments and analyses, the goals of this project in the first year

were to examine the removal ability of several phosphatic clays from various sources in central Florida against Karenia brevis, to determine whether cell removal can be enhanced using chemical coagulants, to determine the effect of water column height and Karenia concentration on removal efficiency, to examine the effect of clay treatment on cell viability and growth, and to examine changes in water column chemistry. In the second year, experiments were conducted on the behavior of settled clays at various flow regimes, to determine the chemical behavior of flocs at the sediment surface, especially on oxygen demand, and to assess the practical and logistical requirements of mounting a field operation. Several technical and engineering challenges of larger-scale dispersal were identified. Finally, some preliminary calculations were made to estimate the cost of such an effort at various treatment scales.

Eleven samples of clay, including nine from IMC Phosphates, Inc., one from Nu

Gulf, Inc., and one from the Florida Institute of Phosphate Research (FIPR), consistently displayed high removal efficiencies (i.e., > 80%), at concentrations ≥ 0.25 g/L, and

2

constant cell concentrations of 9,000 cells/mL. Seven clays showed > 88% RE, despite differences in their geographic source, age, initial concentration, and the stage of the clays during the beneficiation and settling processes. Although high cell removal was also demonstrated by several clays at lower concentrations (e.g., 0.03-0.13 g/L), larger variability was observed (e.g., 2-94%). In comparison with phosphate rock, phosphatic clays were far more effective with RE’s > 70%, while cell removal with phosphate rock did not exceed 10% in this study. In tests to examine whether chemical coagulants or flocculants can further enhance the removal ability of phosphatic clay by increasing its “stickiness”, the addition of alum did not improve the removal ability of phosphatic clay (designated IMC-P2) relative to untreated clay, until alum concentration reached 1000 ppm. Similarly, RE did not improve when IMC-P2 was treated with two cationic polymers (Percol LT-7990 and LT-7991), relative to untreated clay, especially with flocculant concentrations greater than 10 ppm. The clay slurry was found to aggregate rapidly upon polymer addition even before dispersal into the cell suspension. By contrast, the use of polyaluminum chloride (PAC) vastly improved the removal ability of IMC-P2, especially at the lower range of clay concentration. At 0.01 g/L and 5 ppm of PAC, for example, 80% RE was achieved with 90% less clay compared to clay alone, while rapid (self-) flocculation of the clay slurry was not observed. Aside from the clay concentration, high cell removal also depended on having a sufficient number of cells in the water column. For example, RE’s > 70% were only observed when cell concentration was around 1,000 cells/mL (and clay concentration was ≥ 0.10 g/L). This finding suggested that the presence and incorporation of a certain number of cells in the system, and in the settling aggregates, are necessary to maximize the effectiveness of a given clay dosage, as larger particles project a larger surface area for particle encounters during descent. In 2.5-m settling columns, cell removal decreased with depth as the clay slurry was gradually diluted. Therefore, to maintain uniformly high cell removal across the entire water column, the depth (or volume) where the organisms occur must be accounted for in scaling up the treatment.

Clay treatment can lead to mortality of Karenia brevis, depending on the interplay

of clay concentration, the frequency of resuspension events (e.g., mixing), and the amount of time that the clays and cells are in contact prior to the first resuspension event. At lower clay concentrations (e.g., < 0.10 g/L), practically all of the settled cells survived, escaped and resumed vegetative growth, with or without resuspension. Above 0.50 g/L, high RE was also accompanied by extremely high cell mortality in the sedimented layer, even with frequent (daily) resuspension. Cell death was not instantaneous, but increased dramatically between 2.5 and 12 hrs after deposition. At intermediate concentrations (e.g., between 0.10 and 0.50 g/L), cell removal was expectedly high, but the amount of mortality was dependent on when the first resuspension event took place. At these clay amounts, the cells can survive in the floc layer for up to three days, but can only return to vegetative growth, if allowed to escape through physical resuspension. Regarding the possible effect of phosphatic clay on seawater chemistry, bulk measurements of several clay samples showed an elevated level amount of Cr (and possibly Cd and Ni). The levels of As, Hg, Pb and Zn were consistently similar to those found in reference sediments, natural sediments and the proposed sediment quality parameters for Florida. The amount of 266Ra, and 238U does not appear to be a cause for concern with regard to

3

release in natural waters, based on effluent standards, while the possible release of 210Pb may be an issue depending on the clay sample used. With respect to important algal nutrients, phosphatic clays do not release nitrate, ammonium and silicate. However, phosphate was released to seawater, as expected. The release of phosphate was controlled by the addition of 4 ppm PAC to the clay suspension. In a mesocosm study incorporating effects of settled clay on the bottom, the addition of phosphatic clay resulted in an increased total settled solids (TSS) by 1426%, volatile settled solids (VSS) by 527%, and the biological oxygen demand (BOD) rate by 31%. These findings suggest that clay addition may have implications on the relative nutritional value of the sediment for benthic organisms that feed on surface sediment.

At WHOI's Coastal Research Laboratory, we conducted two sets of flume

experiments in order to predict flow environments in which clay and algae would flocculate, settle, and accumulate in a layer on the sea floor. Flume runs were conducted with the dinoflagellate Heterocapsa triquetra and the phosphatic clay IMC-P4 and then repeated with the addition of PAC. The first set of experiments, conducted in a 17-m straight-channel flume, was designed to compare the erodibility of clay/algal flocs that settled in a still water column for different time periods. The second set of experiments was conducted in a racetrack flume with flow speed initially at 3, 10, or 20 cm s-1. Cell counts during this set of experiments allowed us for the first time to determine removal efficiencies in flow. Results from both sets of experiments indicated that consolidation, or dewatering, of the layer over time increased the critical shear stress for resuspension (i.e., decreased erodibility). Resuspension of a 2-mm thick layer that settled for 3 hours in relatively low flow speeds (≤ 3 cm s-1) would be expected at a flow speed of 15-16 cm s-1 in the field, as compared to 18-19 cm s-1 for a layer that accumulated in 9 or 24 hours. The addition of PAC increased the erodibility of the flocs (i.e., resuspension occurred at lower flow speeds). In the racetrack flume with the flow at 3 and 10 cm s-1, >80% of the cells were deposited in the floc layer that accumulated on the bed. However, removal efficiencies were considerably lower when the experiments were repeated with PAC. Overall, if the objective is for treated cells to settle and remain on the bottom (which is not always desirable), we predict that the application of phosphatic clay to remove algal cells from the water column in the field would be effective in environments with flow < 12 cm s-1, or about 1/4 kt. With higher flows, settling would be slower and the floc would be dispersed and diluted further afield, which would be a beneficial result in some cases. However, in the field other factors such as waves must also be considered in addition to mean flow in contributing to bottom shear stress. We do not advise the addition of PAC in flowing water in an attempt to reduce the clay loading.

Finally, an analysis of the logistical and engineering requirements of bringing this

method to larger-scale trials and field applications resulted in three major areas of need: (1) clay handling (e.g., acquisition, transport and storage), (2) preparation of clay slurries for dispersal, and (3) clay dispersal system or apparatus. A significant technical (and economic) concern was found with the handling and processing of phosphatic clays due to the fact that they are “wet” materials that cannot be dried due to the loss of effectiveness. Based on a preliminary cost analysis for mounting a treatment effort, the greatest expense was associated with the transport of clays from the source (or storage

4

sites) to the staging areas, presumably along the Gulf coast. This was then followed by the cost of hiring ships and workers who would be involved in actual dispersal and monitoring of bloom removal and water quality.

In summary, this investigation revealed the strong, innate ability of Florida

phosphatic clays to remove and kill the red tide-causing Karenia brevis from suspension in laboratory experiments, and the physicochemical, hydrodynamic and scale conditions which can influence cell removal and sedimentation. Compared to other clay minerals, phosphatic clays were among the most effective against this and other bloom-forming species, at clay concentrations significantly lower than those used in Asia. There are concerns, however, regarding the release of inorganic phosphorus and certain trace metals into the marine environment, changes in the sediment composition and oxygen demand, and the fate of settled materials in flow. Eventually, these and other possible impacts need to be addressed in controlled experiments in actual field settings. In the end, we believe these results are promising, in many respects, and merit further and careful investigation.

5

INTRODUCTION

For many years, red tides have been a recurrent and serious problem along the west coast of Florida. They are caused by the dinoflagellate Gymnodinium breve (now called Karenia brevis, Figure 1), a single-celled alga that can grow to extremely high densities sufficient to discolor surface waters, cause shellfish toxicity, and lead to mortalities of fish and other marine organisms (reviewed by Tester and Steidinger, 1997). In addition to massive fish kills that litter beaches with dead and rotting fish, and the frequent closure of shellfish beds contaminated with brevetoxins, K. brevis blooms have severely affected the Florida manatee population, killing 10% of the population during an outbreak in 1995-1996 that lasted over 18 months. A further impact of these phenomena comes from the aerosolization of the algal toxin, which leads to respiratory problems among those living near or visiting the beach. Estimates of the economic impacts of the Florida red tides vary considerably, but costs of up to $20 million are often cited for episodes that last several months. The impact from the 18-month red tide in 1995-96, which spanned two tourist seasons, was undoubtedly much higher.

Although the significance and recurrence of these and other red tide phenomena

would seem to justify bloom control as a high-priority topic for research, virtually no focused investigation on control has been undertaken in the United States for nearly 40 years (Boesch and others 1997). Instead, the primary objectives of past and ongoing research on Karenia brevis, and other harmful algal blooms (HABs) in the U.S., have been to obtain a fundamental understanding of the biological, chemical, physical and ecological processes underlying blooms and their impacts. Notably, the gap in research addressing control of marine species is in stark contrast to work done in the terrestrial environment where human efforts to control insects, diseases, and weed species are more common. The reasons for this lack of effort are many (Anderson 1997). However, the lack of research progress is not due to an absence of strategies worth exploring, but rather, to a preference on the part of the HAB community, in general, to pursue fundamental research to better understand bloom phenomena. The rationale for this one-sided approach has been that fundamental understanding will be essential for man to ever manage or mitigate blooms: that is, we cannot control what we do not understand.

Figure 1. Florida Red Tide Organism, Karenia brevis, Formerly Called Gymnodinium breve. (Photo by Fukuyo and Mote Marine Laboratory, Sarasota, FL).

6

However, understanding is a subjective concept, and after decades of research on HABs, most scientists still have pressing questions to answer and more details to resolve. Hence, research into proactive bloom control and mitigation have remained secondary or tertiary priorities. CONTROL STRATEGIES

Management approaches against red tides or HABs can be broadly categorized as either indirect or direct (Shirota 1989; Boesch and others 1997). Indirect strategies focus on prevention (e.g., reduce nutrient inputs to the ocean), and/or the minimization of bloom impacts (e.g., early warning, cell and toxin monitoring, harvesting bans). Most strategies currently in place fall under this category. By contrast, direct strategies target the causative organisms directly, and attempt to eliminate them from the water column, or minimize their proliferation. Some examples of direct control include the use of chemicals that kill HAB cells during blooms, the addition of clays and chemical flocculants to aggregate cells into larger, more-rapidly settling particles, the use of large-scale filters or screens to remove cells, and the addition of biological agents such as viruses, bacteria, or parasites that are lethal pathogens to HAB species.

The last reported effort to control a red tide directly in U.S. coastal waters took place in 1957 (Rounsefell and Evans 1958). During an outbreak along the Gulf coast of Florida, copper sulfate was added over a 16-mi2 area using crop-dusters. Karenia brevis concentrations decreased to near-zero in treated areas. However, the red tide reappeared in several locations within several weeks. The conclusion was made that until a cheaper and more effective means of control was discovered, the use of copper should only be considered in local situations to give short-term, temporary relief from airborne brevetoxin. It is now known that these blooms originate offshore, so the reappearance of bloom after the copper sulfate treatment was a matter of new water and cells being advected into the region, replacing the treated water mass. Since that time, however, only Japan, China and South Korea have pursued research on direct bloom control (Shirota 1989; Yu and others 1994a, 1994b, 1994c, 1995; Na and others 1996; Anderson 1997).

To date, the most promising strategy for controlling red tides in the field is the application of suspended clay particles over the bloom to flocculate and settle the algal cells (Anderson 1997; Boesch and others 1997). This method has been investigated and used effectively during red tide events in Japan (Shirota 1989) and South Korea (Na and others 1996; Bae and others 1998; Choi and others 1998; Choi and others 1999) to protect vital mariculture operations. Due to its effectiveness and practicability, clay dispersal has become a part of South Korea’s management scheme. Furthermore, clay minerals were deemed suitable for this purpose because they are naturally-occurring particles that are commonly found in marine systems, abundant, readily available, relatively inexpensive and easy to use. Moreover, clay addition is thought to have a low probability of causing environmental damage. For these reasons, we pursued the research program presented here to investigate the potential application of clay control in the

7

United States, focusing primarily on locally-available clays (such as phosphatic clay) against problematic HAB species in U.S. coastal waters.

FLOCCULATION

The principle behind this control strategy is the mutual flocculation between clay particles and algal cells, leading to the formation of larger, more-rapidly sinking agglomerates (or flocs), which settle to the ocean floor. According to physicochemical theory, aggregation can be divided into two sequential steps: transport and attachment (O'Melia and Tiller 1993; Elimelech and others 1995). Transport is a physical process that generates particle collisions and is controlled by the hydrodynamics of the system, and external forces such as gravity. The three main mechanisms are Brownian diffusion, fluid motion (laminar or turbulent), and differential sedimentation. In a system with flagellated organisms, like Karenia brevis blooms, particle collisions may also be generated by their swimming ability (Jackson and Lochmann 1993).

After collision, particle attachment occurs when the particles adhere to produce a floc. This process is generally controlled by the surface chemical properties of the particles, and chemical properties of the surrounding medium (e.g., pH, ionic strength, valence of the ions, polyelectrolytes). Clay minerals develop surface charge through the exchange of ions of lower valence in the crystal structure (i.e., isomorphic substitution), the specific adsorption of charged species from the medium, and the exchange of ions along the mineral surface. These charges, typically electronegative, are balanced by excess ions (i.e., counterions) from the medium to form the so-called double layer arrangement (Thomas and others 1999). The interaction of similarly-charged double layers leads to electrostatic repulsion and the poor attachment of colliding particles (i.e., stable suspension). As the concentration of counterions increases, the double layer is compressed, allowing attractive forces to dominate (e.g., van der Waals), and attachment to occur. Particle attachment may also take place through polymer bridging, a process that involves the adsorption of a long-chained molecules between two charged particles (Gregory 1987).

Freshwater algae were also found to have negative surface charges (Ives 1956; Geissler 1958) that arise from the hydrolysis of functional groups from the various organic molecules on the cell surface (e.g., amino acids, nucleic acids, proteins, lipids, sugars, and carbohydrates). Similar charge measurements for marine microalgae are lacking, although the same model has been inferred (Maruyama and others 1987; Yu and others 1994b). In a recent study, Sengco (2001) demonstrated that various marine alga clearly displayed negative surface charges, although the magnitudes of their electrophoretic mobilities were small. In seawater, the thickness of the double-layer is small (Stumm and Morgan 1996) and the stability of the cell suspension may be controlled by organic polymers (i.e., steric stabilization) (O'Melia and Tiller 1993).

8

Flocculation in a suspension composed of clay minerals and algal cells is complex. It consists of differently sized particles (i.e., heterodisperse) ranging from submicron (clays) to tens of microns for the largest algal species. Moreover, the particles have varying shapes, composition (i.e., inorganic vs. organic), surface properties, and behavior (i.e., passive vs. motile). Nevertheless, the same physicochemical concepts have been applied to clay-algal systems, although in a mostly qualitative and descriptive manner (Avnimelech and others 1982; Alldredge and Silver 1988; Shirota 1989). In the current model, transport and collisions are brought about by the same suite of mechanisms (Alldredge and Silver 1988). However, Brownian diffusion has often been excluded since its effectiveness diminishes in the size range of most algal species (Yu and others 1995). Avnimelech and others (1982) and Leslie and others (1984) proposed that the attachment of clay particles on the cell surface is mediated by surface-active organic polymers produced by the organisms. In essence, the clays (specific gravity about 2.6) act as mineral ballasts for the cells.

Aside from particle aggregation and settling, cell removal may also take place by the sweeping action of the aggregates as they fall through the water column. This process may be significant below the water surface where only larger agglomerates are present to interact with the algal cells. Collisions can occur through differential sedimentation. However, the entrainment of cells into the aggregate may not be governed necessarily by surface charges. Instead, cells that come in contact with the aggregate may be pushed downward or captured by the inertia of the particle, without attachment via surface-specific chemical interactions. PREVIOUS WORK ON CLAY FLOCCULATION AND ITS APPLICATION

Most of the early research on the use of clays to control red tides have taken place in Japan (Maruyama and others 1987; Shirota 1989), China (Yu and others 1994a, 1994b, 1994c, 1995), and South Korea (Kim 1997; Na and others 1996). Depending on the treatment used, removal values of 95 to 99% of the cells in cultures have been accomplished with small amounts of added clay (e.g., between 0.01 to 0.10 g/L). However, none of these reports described the effectiveness of clays against common HAB species in the U.S. In most studies, montmorillonites showed the greatest versatility against a wide variety of algal species (Maruyama and others 1987, Na and others 1996). Shirota (1989) attributed this ability to the high adsorptive (ion-exchange) capacity of this three-layered mineral. By contrast, Yu and others (1995) found that kaolinites were better overall than montmorillonites against several species experimentally, which they corroborated using theoretical estimates of electrostatics. The prevailing hypothesis is that clay surface charge and charge density, dictated by their mineralogy, is more important in determining its removal ability for a given species. Lastly, additional work by Yu and others (1994c, 1995) demonstrated that clay’s removal capacity can be further enhanced by modifying the surface charge by treatment with polyhydroxy aluminum chloride (PAC), a chemical flocculant. When PAC was added, more than 90% of the cells of the dinoflagellate Prorocentrum minimum were removed from suspension using clay at 0.10 g/L, compared to 2 g/L with no treatment--a 20-fold

9

improvement. Similar results were found from chemical insertions into the clay lattice (e.g., Mg2+) and other chemical treatments that increased the net positive charge of the clay minerals.

Focusing on algal properties, Yu and others (1994b) found that two diatom

species (Nitzschia pungens and Skeletonema costatum) were removed more readily than two dinoflagellates (Prorocentrum minimum and Noctiluca scintillans) by both kaolinite and montmorillonite. The authors explained that the higher removal of diatoms may be attributed to a higher “specific surface area” which is associated with their size and shape. The effect of cell concentration was not addressed. In addition, the higher removal of diatoms was linked to stronger adsorption to the cell surface due to higher amounts of organic matter which is typically associated with this group of marine phytoplankton. While there have been numerous studies of clay removal using individual species, there have been no previous reports of clay treatment of water masses containing a mixture of algal species. Cell mortality due to clay contact was reported by Shirota (1989) and Bae and others (1998), although the exact mechanism for cell death is not fully understood.

In field trials, the Japanese dispersed montmorillonite at the rate of 200 g/m2 in and around culture cages where fish were dying due to a Cochlodinium red tide (Shirota 1989). The bloom disappeared and no further fish mortality was observed. Despite the apparent success of the treatment, the method was not readily accepted due to the high cost of the clay (reported to be an exorbitant $200-300 per ton), its storage on-site, and the dispersion methods.

By far, the most frequent and extensive application of clay control occurs in South Korea. The earliest treatments began in 1996 and 1997 during which yellow loess was used to control an outbreak of Cochlodinium polykrikoides in the vicinity of a large fish farming industry (Na and others 1996; H.G. Kim, pers. comm.). Due to the urgency of the application, details of the operations are sketchy, but a total of 60,000 tons of clay were reported dispersed over 100 mi2 in 1996, by spraying water onto a pile of clay on a flat barge. The target loading was 400 g/m2, but no attempt was made to measure the actual concentration of clay in the water (H.G. Kim, pers. comm.). This crude procedure undoubtedly resulted in highly variable and inefficient loadings. Nevertheless, the operation was considered a major success, as it resulted in reduced C. polykrikoides concentrations near the surface, no adverse impacts on caged fish and other organisms, and a 99% decrease in fish mortalities compared to a similar red tide in 1995, which was untreated. Encouraged by the success, Korean officials stockpiled 100,000 tons of clay and used much of it in 1997 to control another bloom event, again with no reported adverse effects (H.G. Kim, pers. comm.). PRELIMINARY RESEARCH AT WOODS HOLE

Prior to this investigation, we began some exploratory research into the potential application of clay control against problematic species in the United States. First, a fluorescence-based method was developed for estimating cell concentration before and

10

after clay addition, to expedite the clay screening process (Sengco and others 2001). Clay samples with varying mineralogies and compositions, from several suppliers, were then tested against a number of different algal species in culture. This phase of the work clearly highlighted the substantial differences in removal abilities among clay types. Against Karenia brevis, for example, pure montmorillonites and one phosphatic clay (a smectite-rich, finely-divided clay) were the most effective, while kaolinite, attapulgite, zeolite and loess were moderately or poorly capable at removing this species. Phosphatic clay was the best clay tested by far, showing 20-50 times more effectiveness than the Chinese and Korean clay samples at equal loading rates (Yu and others 1994b, 1994c; Na and others 1996). As phosphatic clay would also be available locally where K. brevis blooms occur, further tests were conducted with this clay.

Preliminary results indicated that the clay did not remove algal species equally, which suggest that the clay could be selective for the red tide, leaving other species relatively unaffected. A similar species “selectivity” was observed by Korean workers in their field applications of clays near fish farms (Na and others 1996), and in a study of algal flocculation in freshwater lakes by Avnimelech and others (1982), and mesocosm tank experiments by Søballe and Threlkeld (1988). Treatment of a natural community with clay should thus leave some planktonic species relatively unaffected, which is environmentally desirable. The research presented in this report followed these encouraging observations. OBJECTIVES AND PROJECT FRAMEWORK

The overall goal of this project was to obtain the experimental data needed to evaluate whether clay flocculation would be a suitable control strategy against the Florida red tide. Several tasks focused on quantifying the effectiveness and the factors governing the removal capacity of several Florida phosphatic clays--one of the most promising clay samples from preliminary experiments. These tasks included the following: (1) to screen a number of phosphatic clays from various sources in central Florida for their removal ability against Karenia brevis, (2) to investigate the possible enhancement of cell removal using chemical coagulants, and (3) to explore the effects of water column height, cell density, and other system parameters on removal efficiency. These laboratory studies were performed during the first year of the project.

Secondly, experiments were conducted to examine changes in water column chemistry and the behavior of settled clays in the sediment to infer the possible impacts of dispersing phosphatic clay into the marine environment. The specific objectives were (1) to explore the fate of Karenia brevis cells in the settled flocs, (2) to investigate seawater chemistry following clay dispersal, emphasizing the release or removal of major plant nutrients, radioactive species, trace metals, and anthropogenic pollutants, (3) to determine the physical behavior of clay/cell flocs at the sediment surface, and (4) to determine the chemical behavior of flocs at the sediment surface, especially effects on oxygen demand. Most of the experiments, especially the chemical analyses, were conducted in the laboratory. They were performed in the second and third years of the

11

project. The experiments involved in studying the physical behavior of settled materials were conducted using two types of flume channels.

In the second year of the program, the practical requirements of mounting a field operation were analyzed and discussed. The technical and engineering challenges of larger-scale dispersal were identified. Calculations were also made to estimate the cost of such an effort at various scales. Overall, this program provided much of the critical information needed to evaluate the technical feasibility, environmental acceptability, and the logistical or economic practicality of this bloom control strategy. The results from this project have guided the direction of the second phase of our research program, which moves the technology of clay control from the laboratory to intermediate and pilot-scales.

13

METHODOLOGY

PHOSPHATIC CLAY SCREENING AGAINST KARENIA BREVIS

Karenia brevis (Wilson strain, CCMP718) were grown in batch cultures using modified f/2-Si medium under conditions described by Anderson and others (1999). Growth was monitored using in vivo cellular fluorescence (model 10-AU Fluorometer, Turner Designs) calibrated against direct microscope cell counts (Avnimelech and others 1982). Removal experiments were performed using cultures in early to mid-exponential growth. Cell concentrations of 7,000-10,000 cells/mL were routinely used.

Samples of phosphatic clay were obtained from two mining companies (IMC Phosphates and Nu-Gulf), and the Florida Institute for Phosphate Research (FIPR) (Table 1, Figure 2). Several samples were taken directly from the flow stream at the beneficiation plant, typically those with higher water content. Other samples were taken from settling ponds of varying ages and solid content. These were placed in plastic jars or buckets, and shipped to Woods Hole. All of the clay samples were kept in the dark at room temperature until used in experiments. The percent solid content of each sample was determined by drying a known mass of wet clay overnight in a laboratory oven (100˚C), then dividing the dry weight by the wet weight (Table 1). All of the samples provided were dispersed in freshwater. For removal experiments, the phosphatic clay suspensions were prepared in the laboratory by diluting the stock solution to the desired concentration using distilled/deionized water (MilliQ). For comparison with phosphatic clay, ground phosphate rock was also tested (provided by Dr. S. Richardson, FIPR). It was produced from +150 mesh material with ground pebble added to represent the run of mine production specifically for analysis and does not represent the normal grinding regime. The sample was a “check” sample and is typical of the rock processed on the date the sample was taken. Known amounts were weighed out and then dispersed in MilliQ water for the removal experiments.

Ten milliliters of K. brevis culture were placed in triplicate borosilicate test tubes (14 mm inner diameter). The initial cell concentration was determined from in vivo cell fluorescence. One milliliter of clay slurry was then added dropwise to the surface of the cell suspension using an air displacement pipet (11 mL final volume). Note that the working clay slurries were 11 times more concentrated than the final suspensions. In this set of experiments, the final clay loadings ranged from 0.01 to 4.0 g/L. One milliliter of MilliQ water was added to the controls in the same manner. The clay-cell suspensions were allowed to flocculate at 20˚C for 2.5 hours under quiescent conditions (Yu and others 1994b). Afterwards, the supernatant directly above the pellet (here defined as the upper 10 mL) was carefully transferred to a new test tube, mixed, and the number of remaining cells was estimated by fluorescence. The removal efficiency (% RE) was then calculated using the following equation: % RE = [1 - (final fluorescence ÷ final fluorescence of control)] * 100 (Eq. 1)

14

The final fluorescence of the control (i.e., 2.5 hours after the addition of DI water) was used to account for cell sinking. Removal efficiency was then plotted against clay concentration for each clay. Table 1. Samples of Florida Phosphatic Clay Tested for Their Ability to Remove

the Florida Red Tide Organism, Karenia brevis, from Suspension, Through Flocculation and Settling.

Company Code Site and Source % Solids Content

IMC Phosphates

IMC-P1 Kingsford Site, Beneficiation Plant 2.6

IMC-P2 Kingsford Site, Settling Pond 16.7 IMC-P3 Kingsford Site, Settling Pond 34.0 IMC-P4* Kingsford Site, Settling Pond 48.61 IMC-P5* Kingsford Site, Settling Pond 76.91 IMC-P6 Kingsford Site, Settling Pond 2.59 IMC-P7-1 Fort Green Site, Primary Core Overflow 4.62 IMC-P7-2 Fort Green Site, Overflow from Concrete

Tanks 5.56

IMC-P7-3 Fort Green Site 2.47 IMC-P8 Hopewell Site 1.25 IMC-P9 Four Corners Site 1.14

FIPR FIPR FIPR (Florida Institute of Phosphate Res.) 46.6 Nu-Gulf NU Nu-Gulf, Mulberry Corporation 39.9

*Not tested in the laboratory, used in mesocosm experiments. NOTE: The percent solid content of each sample was determined by drying known amounts of wet clay overnight at 100°C, then dividing the dry weight by the wet weight.

Figure 2. Location of Sampling Sites Where Phosphatic Clay Was Taken. (Map provided by IMC Phosphates Company.)

15

16

CELL REMOVAL USING PHOSPHATIC CLAY AND FLOCCULANTS

In these experiments, the ability of selected coagulants and flocculants to improve the removal capacity of phosphatic clays was tested. We hypothesized that the addition of these substances would increase cell removal by enhancing the adhesion between particles (i.e., “particle stickiness”). By improving the removal ability of phosphatic clays in this way, the amount of clay needed to treat the red tide efficiently would decrease, thereby reducing the inorganic load on the system, and minimizing the potential impact on the environment. These selected chemicals are commonly used to enhance the clarity of drinking water by promoting the rapid flocculation of very fine, slow-settling particles or colloids.

The coagulants/flocculants tested were alum, polyaluminum hydroxychloride (PAC, Superfloc® 9001, Cytec Industries, Inc., W. Paterson, NJ, USA), and two cationic flocculants (Percol LT-7990 and LT-7991, Allied Colloids, Inc., Suffolk, VA, USA). The latter two consist of polyquaternary amines, which were designed for potable water treatment (NSF Standard 60). They were also selected because they were reported to have low toxicity to fish (Allied Colloid, MSDS). Following the removal experiments above, increasing amounts of IMC-P2 (phosphatic clay) were mixed with 1, 10, 100, 1000 ppm of alum, LT-7990 and LT-7991 just prior to the screening protocol. In another set of experiments, 0.01 g/L – 0.09 g/L of IMC-P2 was treated with 5 ppm PAC.

VARYING CELL CONCENTRATION AND WATER COLUMN HEIGHT

According to theory, flocculation rate is dependent on particle concentration (O’Melia and Tiller 1993). In the previous experiments, clay concentration was varied while cell concentration was kept constant. In the following studies, the concentration of Karenia brevis was varied with a range of values that can be found in the field.

K. brevis were grown in batch cultures as previously described. When the cultures reached mid- to late-exponential growth, they were diluted with fresh f/2-Si media or concentrated by reverse filtration--a process that removes the culture media, leaving behind intact cells with minimal damage. This was done by placing the culture in a plastic Nalgene beaker, then carefully pushing a 4-cm (inner diameter) PVC-tube fitted with a 10-µm Millipore filter downwards, allowing the water to flow through the filters, which was then siphoned off from the inside of the PVC tube. Cell concentrations ranging from 200 to 13,000 cells/mL were treated with varying concentrations of IMC-P2 phosphatic clay following the same protocols for the removal experiments.

The following experiments focused in greater detail on cell removal at specific

depth intervals over time, and in larger volumes. To obtain enough culture for these experiments, K. brevis was grown in carboys as described in Anderson and others (1990). Autoclaved glass carboys were filled with 16 L of sterile-filtered (0.2-µm SuporCap cartridge filter, Gelman Sciences, Ann Arbor, MI, USA) seawater (0.2 pre-filtered Vineyard Sound water, salinity = 30). Each was then enriched with f/2 nutrients (without

17

Si) and inoculated with dense cultures of K. brevis. The carboys were bubbled gently. Removal experiments were conducted when the carboys reached mid-exponential growth with cell concentrations between 8,000 and 10,000 cells/mL.

Two settling columns were constructed from polycarbonate coreliners, typically used in deep-sea coring operations (Figure 3). They are 2.5-m tall with a 9-cm inner diameter. One end was sealed with a tight-fitting polycarbonate bottle which was held in place with a metal ring clamp and silicone sealant. A series of 1.1 cm holes were cut out along the length of the tube at 0.5-m intervals. The holes were sealed with 15-D silicone stoppers. To obtain water samples at each depth, a 16 gauge needle (1.5 inches) was plunged into each stopper and a 20 cc syringe was attached to each needle. The needles extended 3.5 cm from the inner wall towards the center of each column.

For the experiment, 12.7 L of culture was added to the settling column, corresponding to a height of 2 m with this design. 15 mL aliquots were taken at 0.25, 0.75, 1.25 and 1.75 m from the surface. K. brevis concentration was determined through in vivo fluorescence. Concurrently, clay concentration was estimated as an absorbance value (at 302 nm) using a spectrophotometer. The highest absorbance for this clay was found at this wavelength during control experiments. Afterwards, IMC-P2 was added to the surface of the settling column in 200 mL volumes using a squirt bottle. The final clay concentrations for each experiment were 80, 200, and 500 g/m2 loadings, which corresponded to 0.04, 0.10 and 0.25 g/L of clay, respectively. Water samples were taken at each depth every 5 min for the first 20 min after dispersal. Then, the intervals were increased to 10 min until 1 h after dispersal. Finally, samples were taken at 30 min intervals until the end of the experiment at 2 hrs. CELL VIABILITY AND GROWTH

The effects of phosphatic clay treatment (IMC-P2) on the viability and growth of Karenia brevis were studied using three strategies, two of which considered the role of physical resuspension on the ability of a treated cell to recover. In the first experiment, the clay-cell pellet was incubated in each tube for 2.5, 12, 24 and 48 h at 20°C following clay treatment (0, 0.03, 0.10, 0.50 g/L). After the remaining cell concentration in the supernatant had been determined (by fluorescence), the supernatant was returned to the original tube and mixed well with the sedimented clay and cells. One milliliter aliquots were treated with 2.5 µM (final) 5-chloromethylfluorescein diacetate (CMFDA) (Molecular Probes, OR, USA), a vital stain that only penetrates the cell membrane of live cells and reacts with esterases to produce a green signal under FITC fluorescence (450 to

18

Figure 3. Settling Columns to Determine Cell Removal with Clay at Several Depth

Intervals.

sampling port

2.5 meters

0.50 m

0.25 m clay slurry in freshwater layer

0.25 m

0.75 m

1.25 m

1.75 m

Settling Column Design

polycarbonate core liner

19

490 nm excitation, Zeiss axioscope). Samples were incubated at 20°C for 20 min in the dark. The first 400 cells encountered on the slide were counted, keeping track of living and dead (dying) cells. Dead cells were identified as intact cells without green FITC fluorescence, or easily discernible cell fragments that contain cytoplasmic material.

In a second experiment, the supernatant was carefully returned to the original tube immediately after the fluorescence measurement without disturbing the pellet. The tubes were incubated at 20°C, and cell growth was monitored over 150 h by making daily fluorescence measurements on the unmixed supernatant. The controls consisted of cultures treated with distilled/DI water (1 ml) or f/2 medium (1 ml).

For the third experiment, the goal was to determine whether potentially viable cells from the pellet can recover and grow if allowed to disaggregate from the pellet as a result of resuspension. After the supernatant had been removed, the clay-cell pellet was resuspended in fresh f/2 medium (10 ml final volume). The volume was distributed equally into 3 new test tubes and incubated for 24 h at 20°C. The cultures were then subjected to a mixing schedule in which the pellet was resuspended daily, every 2 d, or every 3 d. Resuspension was achieved by gently mixing the tube by hand until the material on the bottom was evenly dispersed. Direct fluorescence measurements of the supernatant were taken daily over 7 d using fluorescence prior to each resuspension procedure. SEAWATER CHEMISTRY FOLLOWING CLAY TREATMENT

An important area of concern regarding the potential use of clays to treat red tides in natural waters is its impact on water quality, particularly on the concentration of certain chemical constituents. Aside from the possibility of introducing harmful trace metals, anthropogenic pollutants, and radioactive elements into seawater, clay dispersal may also contribute important dissolved macronutrients that can stimulate algal growth. These studies determined and quantified the release or uptake of several dissolved chemical species due to clay addition.

Ten phosphatic clay samples were shipped to B&B Laboratory (TDI-Brooks International, College Station, TX, USA). The samples were kept at 6.9oC during shipment and stored frozen (-20oC) in the dark until processing. All ten samples were analyzed for trace metals (TM), including mercury, according to the methodology described by Lewis and others (submitted).

To analyze the release of radionuclides (266Ra, 238U and 210Pb) into the surrounding water, 0.25 g/L of phosphatic clay (IMC-P4) was suspended in filtered seawater. The clay was allowed to flocculate and settle for 2.5 hrs at 20oC under quiescent conditions. After incubation, the entire volume was filtered to separate the supernatant. The filtrate and filter were analyzed for γ-count using a Ge(Li) detector with multichannel analyzer in the Chemistry Department at WHOI.

20

To examine the dynamics of important algal nutrients, a working solution of IMC-P2 phosphatic clay was prepared by diluting the initial slurry with filtered Sargasso seawater (SSW) with very low background inorganic nutrient content. For comparison, two other clays were tested: Wyoming bentonite (WB-B, Wyo-ben, Inc) and polymer treated (cationic) kaolinite (H-DP, J.M. Huber Co, Macon GA, USA). For these experiments, the clays were added to triplicate, acid-washed (10% HCl), 50-ml plastic centrifuge tubes containing SSW to a final concentration of 0.25 g/L. In a parallel set of tubes, 4 ppm (final concentration) of polyaluminum hydroxychloride (PAC) was added to the seawater medium alone or in combination with IMC-P2 or H-DP. Controls consisted of clean SSW only. All of the suspensions were thoroughly mixed after the slurry was added and then allowed to flocculate and settle for 2.5 h at 20°C. This incubation time was shown to be sufficient for adsorption equilibrium between the minerals and the medium (Beschoten and Edzwald 1990). After the incubation, the samples were spun down in a bench top centrifuge at 8650 rpm for 5 min. The supernatant then filtered through 0.2 µm Whatman Puradisk syringe and then transferred to acid-washed sample vials, frozen and sent for analysis at the University of New Hampshire. The supernatants were analyzed for nitrate (NO3), nitrite (NO2), ammonium (NH4), phosphate (PO4), and silicate (SiO4) by flow injection using a Lachat QuikChem 8000 Automated Analyzer. Nitrite analysis was performed according to Lachat QuikChem Method 31-107-04-1-A (Lachat Instruments 1999). Ammonium method was analyzed using Lachat QuikChem method 31-107-06-1-A (Lachat Instruments, 1994). Determination of phosphate and silicate was based on Lachat QuikChem methods 31-115-01-3-A and 31-114-27-1-B respectively (Lachat Instruments 1998 and 1996). All samples were quantified against four working standards prepared from diluted stock standards to encompass the concentration ranges of the samples.

In another set of experiments, the removal capacity of each clay (with and without PAC) was determined by analyzing the change in nutrient content in a medium to which known amounts of inorganic nutrients were added. A solution was prepared containing 4.0 µM KNO3, 4.0 µM (NH4)2SO4, 2.0 µM KH2PO4, and 1.0 µM Na2SiO3 in SSW. IMC-P2 and H-DP were added to the spiked medium (with and without 4 ppm of PAC) in acid washed 50-mL tubes at a final concentration of 0.25 g/L. The tubes were mixed thoroughly and then the clays were allowed to flocculate and settle for 2.5 hrs. The tubes were centrifuged and the supernatants were processed and analyzed as before. PHYSICAL BEHAVIOR OF SETTLED CLAY/CELL FLOCS

In addition to the potential impacts on the natural environment described above, the dispersal of clay into the water column will create at least a temporary increase in turbidity and the concentration of suspended solids. This increase in suspended solids may have negative impacts on sensitive, filter-feeding marine animals, for example clams, other bivalves, and some crustaceans. Thus, before considering clay dispersal as a mitigation strategy for red tides, it is important to determine the flow environments in which clay/algal flocs would be expected to settle and accumulate on the sea floor. Once on the sea floor, we need to understand how the consolidation, or dewatering, of the layer

21

affects its erodibility upon increased stress from tidal currents or waves. In addition to making predictions for the mass balance of dispersed clay, the mechanisms for flocculation of clay with algae in flowing water must be addressed. It should be noted that all of the experiments described above were conducted in still water columns, thus limiting particle contacts to differential sedimentation and the movements of algal cells.

At WHOI's Coastal Research Laboratory, we conducted two sets of flume experiments to predict flow environments in which clay and algae would flocculate, settle, accumulate, and then be resuspended from the sea floor. All flume experiments, including exploratory experiments in which optimal clay loadings were tested, are listed in Appendix A (an EXCEL spreadsheet). Drawings and additional information about flow in the two flumes are provided by Fries and Trowbridge (2003). Flume runs were conducted with the nontoxic dinoflagellate Heterocapsa triquetra, similar in size to Karenia brevis, and the phosphatic clay IMC-P4 in 10-µm filtered seawater (concentrations listed in Table 2). Each experiment was repeated with the addition of 5 ppm polyaluminum hydroxychloride (PAC), diluted in deionized water from the stock solution.

The first set of experiments, conducted in a 17-m long, 60-cm wide straight-

channel flume, was designed to compare the erodibility of clay/algal flocs that settled in a still water column for different time periods. These experiments were intended to measure the critical bed shear stress (τ0crit) of deposited flocs (i.e., the threshold flow for resuspension of flocs from the bed). The experimental design was a 3 x 2 factorial design (3-, 9-, and 24-hr settling periods, with or without PAC) with three replicate flume runs for each of the 6 treatments (Table 2). At the beginning of each experiment, the flume was filled with 10-µm filtered seawater to a depth of 12 cm, flowing at ~3 cm s-1 (the slowest flow possible in the 17-m flume). A panel with a recessed 20 x 20 x 2-cm deep box was filled to ~2 mm below flush with sieved sand and placed in the flume at 13 m downstream. A plastic fence that extended above the water surface was placed around this test bed, and the dinoflagellates, PAC, and clay were added to the enclosed volume using a wide-mouthed pipette. A floc layer accumulated on the test bed and consolidated over the test period.

The analysis of critical shear stress began when the fence was removed, exposing the floc layer to the overlying flow. The test bed surface was illuminated from above with a light sheet (~2-cm wide at the bed), and a CCD video camera recorded a section of the flow directly above the test bed. As the flow speed was increased in the flume, the following visual criteria were used for determining thresholds for four stages of particle transport: initial motion (when flocs started rolling); bedload (when most flocs were rolling and some were saltating); wave resuspension (when flocs were periodically lifting from and not returning to the bed); and mean flow resuspension (when flocs were continuously lifting from and not returning to the bed). Video clips of these stages of particle transport may be viewed on the Internet at the following URL: http://www.whoi. edu/science/AOPE/cofdl/stace/FIPR/. Mean shear stress ( 0τ ) at each stage of particle transport was estimated by analyzing profiles of mean along-channel velocity, measured with a Laser Doppler Velocimeter (LDV). Flow profiles were four-min records at each

22

of ten points between 0.9 and 8.0 cm above bottom. Due to a standing wave in the flume, estimates of total bottom shear stress had to include wave-forced bottom stress )(~

0 tτ , calculated via a model for an oscillatory boundary layer. Critical shear stress then was estimated as ( ))(~max 000 tcrit τττ += , and critical shear velocity as ρτ critcritu 0* = (ρ is seawater density). Detailed methods for the analysis of bed shear stress in the 17-m flume are provided by Beaulieu (2003). Statistical comparisons of the mean values for τ0crit included a two-way ANOVA and planned multiple comparisons by the least significant difference (LSD) test. Homogeneity of variances was checked with the Hartley Fmax-test, and significance was tested at α = 0.05.

23

Table 2. Summary of Woods Hole Flume Experiments.

Experiment Description

Experiment Run #

Settling Time (Hr)

Flow Speed (cm s-1) During Settling

T (°C)

Salinity (ppt)

Heterocapsa Cells (ml-1x103)

PAC Conc. (µl l-1)

Areal Loading of Clay (g m-2)

Conc. of Clay (g l-1)

17-m Flume Clay 1 3 0 21.5 33 0.00 n/a 33.13 0.41 Clay 2 3 0 21.5 33 0.00 n/a 33.13 0.41 Clay 3 3 0 21.5 33 0.00 n/a 33.13 0.41 Clay 1 9 0 22 33 0.00 n/a 44.96 0.56 Clay 2 9 0 22 33 0.00 n/a 44.96 0.56 Clay 3 9 0 22 33 0.00 n/a 38.22 0.48 Clay 1 24 0 23 33 0.00 n/a 56.80 0.71 Clay 2 24 0 23 33 0.00 n/a 47.33 0.59 Clay 3 24 0 23 33 0.00 n/a 47.33 0.59 Clay + PAC 1 3 0 23.5 33 0.00 5 33.13 0.28 Clay + PAC 2 3 0 23.5 33 0.00 5 33.13 0.28 Clay + PAC 3 3 0 23.5 33 0.00 5 33.13 0.28 Clay + PAC 1 9 0 23.5 33 0.00 5 37.86 0.32 Clay + PAC 2 9 0 23.5 33.5 0.00 5 37.86 0.32 Clay + PAC 3 9 0 23.5 33.5 0.00 5 37.86 0.32 Clay + PAC 1 24 0 24 33.5 0.00 5 47.33 0.39 Clay + PAC 2 24 0 24 33.5 0.00 5 47.33 0.39 Clay + PAC 3 24 0 24 33.5 0.00 5 47.33 0.39 Racetrack Flume Clay 1 3,3,9 3 23 30 0.00 n/a 33.21 0.28 Clay + PAC 1 3,3,9 3 23 30 0.00 5 33.21 0.28 Clay 1 3,3,9 10 21 30 0.00 n/a 33.21 0.28 Clay + PAC 1 3,3,9 10 22 30 0.00 5 33.21 0.28 Clay 2* 3 20** 20.5 30 0.00 n/a 33.21 0.28 * Run #1 lost due to power outage. **Initially at 20 cm s-1 and then decreased by 2 cm s-1 every 3 hrs.

24

The second set of experiments, conducted in a racetrack flume with essentially a 17.1 m2 test bed, was designed to study the deposition and resuspension of clay/algal flocs in flowing water. The objectives were several-fold, including determining τ0crit for deposition of flocs, simulating a tidal cycle for multiple resuspension “events,” and estimating algal removal efficiencies after repeated resuspension. Although more realistic than the 17-m flume experiments, these flume runs were far more labor-intensive due to the large culture volumes necessary to seed the 2052 liters used in each run, and we were unable to replicate the experimental treatments. We conducted a total of six experiments, at three initial flow speeds (3, 10, or 20 cm s-1) with or without PAC (Table 2). One of the flume runs at 20 cm s-1 ended prematurely with a power outage.

Before each experiment the flume was filled to 12-cm depth with 10-µm filtered seawater, the paddle speed was set to 10 cm s-1, and two optical backscatter (OBS) turbidity sensors were placed in the flume at 3.4 and 8 cm above bottom. The OBS sensors were calibrated to expected clay concentrations. Then several carboys of algal culture (50 L) were added to the flume and allowed approximately one hour to mix freely. For the flume runs with PAC, 22.8 ml of stock PAC solution were diluted in 18 L of deionized water and sprayed at 5 L min-1 on the surface of the flowing flume water. After another 30 min, the paddle speed was set to 3 or 20 cm s-1 (or left at 10) and given another 30 min prior to the addition of clay. For a final concentration of 0.28 g (dry) L-1 in the flume, 1.2 kg (wet) of IMC-P4 clay was blended, strained through a 63-µm sieve, suspended in a total of 54 L of 10-µm filtered seawater, and sprayed at 5 L min-1 on the water surface. For the flume runs with flow initially at 3 or 10 cm s-1, the flocs settled for 3 hr prior to the first resuspension “event” in which paddle speed was stepped at 1 cm s-1 increments to 25 cm s-1. The second and third resuspension “events” occurred after 3- and 9-hr settling periods, respectively. For cell counts to determine removal efficiencies, water samples were withdrawn from the flume 30 min after the addition of cells, ~1 hr later (just prior to the addition of clay), and before the 1st 3hr, 2nd 3hr, and 9hr resuspension events. For the flume runs with flow initially at 20 cm s-1, paddle speed was stepped down by 2 cm s-1 every 3 hours.

Analysis of critical shear stress (τ0crit) for resuspension of flocs in the racetrack flume was far more complicated than described above for the 17-m flume. Because the turbidity in the flume during these experiments precluded the use of an LDV to measure flow speeds, thirteen flow profiles were recorded after the flume was cleaned and filled to 12-cm depth with clear seawater. Estimates of bottom shear stress, including both a mean and wave-induced component, were calibrated to paddle speed. The maximum slope of the OBS records (i.e., change in OBS value divided by change in time) was used to determine the threshold for resuspension of flocs. Critical shear stress was estimated from the paddle speed in the flume during this maximum input of suspended solids. In contrast, τ0crit for deposition of flocs was estimated by the minimum slope for the OBS record during the flume run in which flow was stepped down from 20 cm s-1. More detailed methods for the analysis of τ0crit in the racetrack flume are in preparation for a manuscript (Beaulieu and others, forthcoming).

25

OXYGEN DEMAND AT THE SEDIMENT

If this strategy is effective, clay flocculation of algal blooms would potentially transport a substantial amount of dissolved and particulate organic matter from the surface to the ocean floor in a relatively short period of time. Depending on local bottom boundary conditions, this flocculated material may accumulate close to the point of release, or remain in suspension and be further distributed laterally. One possible consequence of introducing organic-rich material would be to increase the biological oxygen demand (BOD) at the sediment-water interface as bacterial growth and productivity are stimulated. Increased bacterial decomposition could lead to reduction in dissolved oxygen concentration along the sediment-water interface which, in turn, could cause benthic infaunal mortalities due to hypoxia/anoxia, and a change in the redox potential of the sediment. This section analyzed the change in biological oxygen demand due to clay dispersal during a mesocosm experiment wherein phosphatic clay was used to treat an actual Karenia bloom in the field.

Experiments were conducted during a red tide outbreak in Florida in November 2001. Cell counts were between 5 x 105 to 1 x 106 cells/L at the time of the study. The mesocosms consisted of clear, polyvinyl limnocorrals with open ends (diameter = 2 m, height = 3 m) (Figure 4A). Using a small powerboat and divers, they were positioned at the study area located in Sarasota Bay adjacent to Mote Marine Laboratory’s dock in a depth of 3 m (Figure 4B-D). The mesocosms were spaced 2 m apart in one line parallel to the shore. Two mesocosms served as treatments and one was used as the control along with an unbounded (ambient) site 2-m away. The open bottoms were anchored from the sides with PVC pipes to prevent drift and to focus the aggregates over the test plots. Three, 1-L sediment traps were carefully placed 10 cm above the bottom within each mesocosm and at the ambient site to capture the flocs for analysis.

The initial cell concentration within the mesocosms, and at the ambient site, was measured by taking integrated water samples. A marked and weighted Tygon tubing was lowered in the center of the mesocosm at 0.5 m and 2.5 m from the surface. Four liters were withdrawn by vacuum pump and combined in samples bottles. Subsamples were then taken for direct microscope cell counts.

The target loading was a final concentration of 0.25 g/L. IMC-P4 phosphatic clay

was soaked overnight in filtered seawater. The softened clay was broken up by hand, then blended until a fine slurry was produced. The slurry was dispersed over the surface using a submersible pump with a garden hose and nozzle attachment. The flocculation and settling was allowed to proceed for 24 h.

26

Note: The limnocorral in the foreground shows turbidity at the surface following clay dispersal. Figure 4. Mesocosms (Called Limnocorrals) Used in Florida Field Experiments

with Clay Flocculation. (A) Limnocorral Fully Extended. (B) Schematic of Deployed Limnocorral and Experimental Set-up. (C) Overhead Schematic of Limnocorral Placement. (D) Photo of Deployed Limnocorral.

Integrated water samples were again taken at 1, 2 and 4 h after clay addition to

determine cell concentration as before. The sediment traps were retrieved after 24 h to

2 m

PVC stakes

bucket

sediment trap

3 m

floation ring

control (no clay)

clay treatment 2

clay treatment 1

control (ambient)

A B

27

allow the contents to fully settle. The contents were analyzed for biological oxygen demand (BOD), total settled solids (TSS), and volatile settled solids (VSS). BOD was determined by EPA method 405.1, by incubating a well-aerated sample at 20°C in a 300 mL BOD bottle. Intermediate dissolved oxygen readings were taken for these samples during the incubation period, so that the dissolved oxygen levels would not drop below the recommended minimum. Samples were re-aerated as needed. Total amount of oxygen consumed by the sample during the 5-d incubation period was computed as the BOD rate. Total suspended solids and volatile solids were determined by methods 2540 D (Standard Methods, 18th edition) and 2540 E (Standard Methods, 17th edition), respectively. A measured volume of well mixed sample was filtered using pre-weighed glass fiber filter and the residue retained on the filter was dried in an oven at 103-105°C to a constant weight. Then, the residue with the filter was ignited to a constant weight in a muffle furnace at a temperature of 550°C. Total settled solids and volatile settled solids were computed in mg/L using initial, dried and ignited weights, and the volume filtered.

29

RESULTS

CELL REMOVAL EFFICIENCY

Florida phosphatic clay exhibited a wide range of removal abilities at the lowest concentrations between 0.01 and 0.03 g/L (Figure 5). Cell removal efficiency (RE) ranged from 2% (FIPR) up to 89% (IMC-P2). Eight out of the eleven samples had removal efficiencies < 50%. When clay loading was between 0.10 to 0.125 g/L, the range of RE’s narrowed to between 52% (IMC-P9) and 94% (IMC-P2), with 7 out of 11 samples exceeding 70% RE. At 0.20 to 0.25 g/L, all but one clay (i.e., IMC-P9) produced RE’s >80%. Above 0.25 g/L, all of the clays removed > 80% of the suspended cells, with 7 clays exhibiting RE’s above 88%. Though the differences were slight, the most effective clay in this survey was IMC-P2 from the Kingsford site of IMC Phosphate, Inc. The least effective was IMC-P9 from the IMC’s Four Corners site. There was no difference between the samples collected from the settling ponds and clays from the beneficiation plant, especially when the clays were applied above 0.20 g/L.

By comparison, the removal ability of phosphatic clay far exceeded that of

phosphate rock (Figure 6). At comparable cell concentrations in the culture (e.g., 1,000 and 8,000 cells/mL), removal efficiency with phosphate rock did not exceed 10%. Cell removal with phosphatic clay was >70% at 0.10 g/L and 1,000 cells/mL. This increased to >87% with clay loadings ≥0.05 g/L and 8,000 cells/mL. COAGULANTS AND FLOCCULANTS

The results of these experiments were reported in Sengco and others (2001) and are shown in Figures 7 and 8. In general, the addition of alum did not improve the removal ability of IMC-P2 phosphatic clay relative to untreated clay. For example, cell removal with 0.01 g/L of clay did not increase until alum concentration was 1000 ppm. Likewise, there were no differences between the RE of untreated and alum-treated clays at the two higher clay concentrations, except when 1000 ppm of alum was added. However, RE decreased slightly when this amount of alum was used. In the absence of clay, cell loss was also observed, particularly at the higher alum concentrations.

The addition of ≥ 10 ppm of Percol LT-7990 and LT-7991, without clay, resulted in cell removal between 25 and 40%. Percol LT (low toxicity) are cationic polymers used in drinking water treatment (NSF Standard, February 1999). At the lowest clay concentration, the use of the two flocculants improved cell removal slightly relative to the untreated phosphatic clay: the application of LT-7990 increased RE four-fold from 10% (untreated) to 37% (1000 ppm),

Figure 5. Removal Efficiency of the Florida Red Tide Organism, Karenia brevis, by Several Phosphatic Clays.

0

10

20

30

40

50

60

70

80

90

100

IMC-P1 IMC-P2 IMC-P3 IMC-P6 IMC-P7-1 IMC-P7-2 IMC-P7-3 IMC-P8 IMC-P9 FIPR Nu-GULF

cell

rem

oval

eff

icie

ncy

(%)

30

Figure 6. Removal Efficiency of the Florida Red Tide Organism, Karenia brevis, by Phosphatic Clays and Phosphate Rock.

31

32

Figure 7. Removal Ability of Florida Phosphatic Clays Against Karenia brevis in Combination with Selected Flocculants and Coagulants. (Data from Sengco and others 2001.)

33

Figure 8. Removal Ability of Florida Phosphatic Clay IMC-P2 Against the Red

Tide Organism, Karenia brevis, With and Without the Flocculant Polyaluminum Chloride (PAC). (Data from Sengco and others 2001.)

while the addition of LT-7991 also increased RE four-fold from 8% (untreated) to 31% (100 ppm). At higher clay concentrations, the clay suspension flocculated immediately after the flocculant was added, before it could be added to the cultures. Vigorous shaking and vortexing of the clay slurry did not disrupt the aggregates. Furthermore, after the slurry was added to the culture, clay flocs formed quicker and were much greater in size macroscopically than untreated clays. These particles sank more rapidly to the bottom of the test tube. As a result, the use of > 10 ppm of LT-7990 decreased RE from 89 to 50% with 0.10 g/L of clay. At 50 g/L of clay, cell removal decreased from 90% to 44% when 100 ppm of LT-7990 was used, although it recovered to 80% when the flocculant concentration increased to 1000 ppm. Similarly, the use of > 10 ppm of LT-7991 reduced RE from ca. 87% to 29% when combined with 0.10 g/L of clay. The addition of 100 ppm of LT-7991 let to a drop in cell removal from 93% to 32% with 0.50 g/L, but it increased again to 68% when the flocculant concentration increased to 1000 ppm.

34

Finally, the addition of polyaluminum hydroxychloride (PAC) vastly improved the removal ability of phosphatic clay, especially at the lower range of clay concentration--the only chemical flocculant or coagulant to accomplish this in the present study (Figure 8). With 5 ppm of PAC, it took 90% less clay to achieve an 80% cell removal (cf. 0.01 g/L with and without PAC). A 15% RE was observed with PAC alone.

CELL CONCENTRATION AND WATER-COLUMN HEIGHT

The concentration of Karenia brevis in suspension influenced the magnitude of cell removal with phosphatic clay. In almost all cases, the maximum removal efficiency was reached with ≥ 0.10 g/L of clay (Figure 9). However, the magnitude of the plateau increased with increasing cell concentration. For instance, cell removal exceeded 70% only when the cell concentration was approximately 1000 cells/mL. In addition, 80% RE was achieved only when the cell concentration was ≥ 4400 cells/mL.

The magnitude of cell removal at various depths was dependent on the amount of clay added at the surface (Figure 10). Experiments took place in a 2 m water column wherein samples could be collected at 0.5-m intervals along the side of the settling column (see also Figure 3). At a loading rate of 80 g/m2 (corresponding to 0.04 g/L in this settling column), cell removal at 0.25 m from the surface was ca. 75%. RE decreased linearly with increasing water column depth, reaching only 24% at 1.75 m. The disappearance of cells near the surface occurred rapidly during the first 40 min of the experiment (Figure 10A), then tapered off as the initial, large pulse of clay passed through (Figure 10B). A similar disappearance pattern was observed at 0.75 and 1.25 m, although the amount of cell loss was less at these depths. Interestingly, a large pulse of cells was detected at 1.75 m between 30 and 40 min, which appeared to correspond with the arrival of cells entrained within the clay around 40 min. The pulse of clays as measured with the spectrophotometer was detected at 0.75, 1.25 and 1.75 m at about the same time.

At 200 g/m2 (corresponding to 0.10 g/L), cell concentration nearest the surface decreased rapidly between 20 and 50 min, tapering off after 60 min (Figure 10C). This dramatic loss of Karenia brevis also corresponded to the arrival of the clay pulse over the same time interval (Figure 10D). At deeper depths, the same magnitude of cell losses observed at 0.25 m occurred, but slightly later in time with increasing distance from the surface. Likewise, the arrival of the clay peaks at lower depths was displaced with time. The height of the peaks was only half the height of the initial peak at 0.25 m, suggesting dilution of the clays. At the end of the experiment, cell removal across the entire length of the water column was quite uniform and high, ranging between 72 and 83% (Figure 11).

Figure 9. Removal Efficiency of Karenia brevis at Various Cell Concentrations Treated with Florida Phosphatic Clay (IMC-P2).

35

-10

0

10

20

30

40

50

60

70

80

90

100

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

clay concentration (g/L)

cell

rem

oval

eff

icie

ncy

(%)

200 cells/mL

300 cells/mL

860 cells/mL

1,049 cells/mL

4,400 - 13,000 cells/mL

36

Figure 10. Removal Efficiency of Karenia brevis at Various Depths During Clay Treatment with IMC-P2 Phosphatic Clay.

37

Lastly, the highest cell removal across the water column was observed at a clay

loading of 500 g/m2 (or 0.25 g/L), ranging from 85 to 90% (Figure 11). As in the case of 200 g/m2, most of the cell loss occurred between 5 and 60 min of the experiment with a time delay in removal over time (Figure 10E). At lower depths (≥ 0.75 m), the cell concentration peaked initially as cells from above arrived, carried by the clays. Then, the cell concentration dropped steeply as the clay pulses passed through. With the large amount of clay added near the surface, the clay peak at 0.25 m was both high and broad (Figure 10F). The clay peak was diluted by half by the time it reached 0.75 m, and continued to decrease slightly at each subsequent depth.

Using the observed arrival times for when the clay signal was first detected (Table 3A), as well as the arrival of the peak (Table 3B), the sinking rates were estimated for each depth interval. Furthermore, these sinking rates were used to calculate the predicted particle (or aggregate) sizes using Stokes’ Law, assuming that the density of the aggregates is equivalent to (or dominated by) the density of the clay particle. Sinking rates ranged from 0.10 to 4.17 mm/s. The corresponding particle sizes were estimated to be between 11.7 to 73.7 µm. The highest sinking rates were typically found with

Figure 11. Removal Efficiency of Karenia brevis Using IMC-P2 Phosphatic Clays at

Various Depths Following the 2-Hour Incubation.

38

Karenia and cells combined, compared to clays alone. Two arrival times were reported (refer to Figure 10B, 10D and 10F): (A) first arrival is the time (in minutes) when the leading edge of the clay pulse was detected, while (B) peak arrival is the time when the maximal signal was detected at each depth interval. The corresponding sinking rates were calculated by dividing the distance from the surface by the arrival times. The expected particle sizes were calculated using Stokes’ Law, using the estimated sinking rates: u = 1/18 * ((ρp - ρw) / η) * d2 * g, where u = sinking rate in m/s; ρp = particle density (3000 kg m-3); ρw = seawater density (1026 kg m-3); η = fluid viscosity at 20°C (1.4 kg m-1 s-1); g = 9.8 m s-2; d = particle diameter (in m).

39

Table 3. Observed Arrival Times of Clay Pulse, Sinking Rates and Predicted Aggregate Size.

Clay Concentration A. First Arrival of Clay Pulse 80 g/m2 = 0.04 g/L 200 g/m2 = 0.25 g/L 500 g/m2 = 0.50 g/L

Treatment

Depth from

Surface (m)

Time Detected

(min)

Sinking Rate

(mm/s)

Agg. Size (µm)

Time Detected

(min)

Sinking Rate

(mm/s)

Agg. Size (µm)

Time Detected

(min)

Sinking Rate

(mm/s)

Agg. Size (µm)

Clays only 0.25 m 0.75 m 1.25 m 1.75 m

20 15 30 60

0.21 0.83 0.69 0.49

16.5 33.0 30.1 25.2

15 20 30 50

0.28 0.63 0.69 0.58

19.0 28.5 30.1 27.6

5 15 20 30

0.83 0.83 1.04 0.97

33.0 33.0 36.8 35.6

Karenia + Clay

0.25 m 0.75 m 1.25 m 1.75 m

12.5 17.5 17.5 17.5

0.33 0.71 1.19 1.67

20.8 30.5 39.4 46.6

2.5 5 5

10

1.67 2.50 4.17 2.92

46.6 57.1 73.7 61.7

1 7.5 7.5

17.5

4.17 1.67 2.78 1.67

73.7 46.6 60.2 46.6

Clay Concentration B. Peak Arrival of Clay

Pulse 80 g/m2 = 0.04 g/L 200 g/m2 = 0.25 g/L 500 g/m2 = 0.50 g/L

Treatment Depth from

Surface (m)

Time Detected

(min)

Sinking Rate

(mm/s)

Agg. Size (µm)

Time Detected

(min)

Sinking Rate

(mm/s)

Agg. Size (µm)

Time Detected

(min)

Sinking Rate

(mm/s)

Agg. Size (µm)

Clays only 0.25 m 0.75 m 1.25 m 1.75 m

40 60 90 120

0.10 0.21 0.23 0.24

11.7 16.5 17.4 17.8

30 30 90 120

0.14 0.42 0.23 0.24

13.5 23.3 17.4 17.8

20 40 60 75

0.21 0.31 0.35 0.39

16.5 20.2 21.3 22.5

Karenia + Clay

0.25 m 0.75 m 1.25 m 1.75 m

20 30 30 30

0.21 0.42 0.69 0.97

16.5 23.3 30.1 35.6

15 30 30 40

0.28 0.42 0.69 0.73

19.0 23.3 30.1 30.8

15 30 35 40

0.28 0.42 0.60 0.73

19.0 23.3 27.9 30.8

40

VIABILITY AND GROWTH AFTER TREATMENT

The results of these experiments were reported in Sengco and others (2001) and are shown in Figure 12 and Table 4. Resuspension through manual mixing took place at varying intervals (daily, every 2 days, every three days). Based on vital staining, the number of dead (or dying) Karenia brevis cells increased with both clay concentration and the time of exposure to phosphatic clay (Figure 12A). Immediately after addition (2.5 h), no dead cells were detected in any of the treatments, although the cells were clearly visible within the clay flocs. Mortality dramatically increased after 12 h, especially at 0.10 and 0.50 g/L. The number of dead cells was more difficult to count at higher clay loadings because most of the cells were lysed and became indistinguishable from the clay particles. Therefore, the mortality values at 0.50 g/L are likely to be underestimates of cell death. By 24 and 48 h, most of the removed cells were dead and lysed in the 0.50 g/L treatment. By comparison, the cell mortality was minimal in the control and 0.03 g/L clay treatment, although 46% of the cells were removed with that clay loading after 2.5 hrs.

Table 4. Recovery of Karenia brevis Following Clay Treatment with Florida Phosphatic Clay.

Treatment

(g L-1) Mean Slope

(Change in Cell Density t-1) n

Standard Error

0.00 43.33 (A) 21 2.14 0.01 47.48 (A) 21 4.50

0.10 (Daily) 127.74 (B) 7 6.91 0.10 (Every 2 Days) 124.12 (B) 7 5.54 0.10 (Every 3 Days) 90.57 (C) 7 2.55

0.20 248.7 (D) 21 14.54 0.50 9.72 (E) 21 1.45

Note: Data from Sengco and others 2001. Using Student’s t-test, treatments with the same letter did not differ significantly from

each other (p > 0.05), while treatments with different letters differed significantly.

The recovery of Karenia brevis after treatment was inversely related to the clay

dosage and disappeared completely once a certain threshold concentration was added (Figure 12B). First, there was no difference between the growth of the controls (distilled water added to f/2) and 0.03 g/L clay addition, although 67% of the cells had been sedimented by the clay compared to 0.03% for the controls (data not shown). Despite the similarity in ultimate cell yield at the end of the experiments, the culture treated with clay showed a 24-h delay in the recovery. This delay increased to 48 h when the clay concentration was doubled to 0.06 g/L. Recovery and growth ceased when the clay dosage exceeded 0.13 g/L in the unmixed cultures.

41

Figure 12. Viability and Growth of Karenia brevis Cells Following Treatment with

IMC-P2 Phosphatic Clay. (Data from Sengco and others, 2001.)

42

In the resuspension experiments, the fluorescence values were first plotted against time for each experiment, and the best-fit regression lines through the points were calculated (data not shown). The slope of these lines was then compared statistically using Student's t-test. No difference was found in the slope values among the 3 mixing schedules within each clay treatment, except for 0.10 g/L (Table 4). At this dosage, there was a significant difference between daily mixing and every 3-d mixing, and between 2- and 3-d mixing. Following the comparisons of mixing frequency within each clay treatment, the data showing no statistical differences were grouped together and a new regression line was determined. The slope values were compared among the different clay loadings (Table 4). There was no difference between the control and the lowest loading (0.01 g/L). At 0.20 g/L, the large slope value was observed which seemed to indicate that a large number of cells escaped from the floc and recovered. The lowest slope value was found at 0.50 g/L suggesting low cell recovery. At the intermediate amount of clay (0.10 g/L), the recovery seems to be determined by the frequency of mixing: cell recovery was greatest when the pellet was resuspended daily or within 2 d after treatment. Survival of the cells decreased when the pellet was mixed after the third day. SEAWATER CHEMISTRY

Data on the concentration of several trace metals from Florida phosphatic clays are presented in Table 5a. Measurements were made from the various clay samples at Woods Hole as well as samples from the U.S. EPA and IMC Phosphates. For comparison, some additional data from various sources were also obtained and presented, including values from the literature on Florida phosphorites (from Altschuler 1980) and in-house analysis performed at IMC Phosphates (Table 5b). Overall, our measurements for As, Cd, and Fe were lower than those in other reports, while our Cr value was higher. For Cu, Mn and Zn, we found higher amounts than those from other phosphatic clays measured by IMC Phosphates, but lower than those from Florida phosphorites and phosphorites from other regions. Hg, Ni, Pb and Se measurements fell within the range of the other reported values, while our values for Ag and Sb were less than those from the available data sets.

Compared to reference sediments used by the U.S. EPA laboratory in Gulf Breeze (FL) (Table 6), all of our measurements (i.e., means) were higher than those for the reference materials. After considering the variability in samples, however, only the measurements for Al, Cd, Cr and Mn were significantly different. When our values were also compared to those of “natural sediments” (Lewis and others, submitted; Table 7), we found Cd, Cr, Ni and Se elevated in phosphatic clays than in natural sediments.

Finally, using sediment quality parameters proposed by MacDonald and others

(1996), we found that our values for As, Cu, Hg, Pb and Zn were all below the Threshold Effects (TEL) and Probable Effects (PEL) Guideline Values. Ni was between the TEL and PEL, while Cd and Cr were both slightly above the PEL. There were no suggested guidelines for Ag, Fe, Mn, Sb, Se, Sn and Al.

43

Table 5a. Chemical Analysis of Several Samples of Florida Phosphatic Clays from Various Sources in Central Florida.

Code and Source Location

% Solids

Ag, ppm

Al%

As, ppm

Cd, ppm

Cu, ppm

Cr, ppm

Fe%

Hg, ppm

Mn, ppm

Ni, ppm

Pb, ppm

Se, ppm

Sn, ppm

Zn, ppm

Sb, ppm

IMC-P2, Kingsford Settling Pond* 16.7 0.20 8.40 3.60 5.40 265.0 16.60 2.80 0.100 146 39.6 24.2 4.500 1.00 83.7 nd

IMC-P4, Kingsford Settling Pond 48.61 0.17 7.97 3.03 4.75 250.1 14.16 2.53 0.072 139 32.6 21.5 5.999 0.36 69.0 0.50

IMC-P5, Kingsford Settling Pond 76.91 0.12 7.87 1.39 3.72 234.0 10.36 2.44 0.062 130 27.8 18.5 3.343 0.69 61.9 0.66

IMC-P6, Kingsford Settling Pond 2.59 0.14 5.35 1.29 3.83 133.6 15.51 2.02 0.044 280 38.7 13.4 1.312 1.51 61.4 0.78

IMC-P7-1, Fort Green Primary Core Overflow

4.62 0.11 6.71 4.46 5.62 172.0 13.17 2.28 0.071 147 23.7 18.5 5.744 1.45 54.4 1.20

IMC-P7-2, Fort Green, Overflow from Concrete Tanks

5.56 0.09 6.50 5.99 4.11 153.1 11.47 2.38 0.060 145 21.2 18.5 6.188 0.83 53.1 0.72

IMC-P7-3, Fort Green Site 2.47 0.11 7.25 2.63 8.40 222.2 15.94 2.07 0.079 112 27.8 20.3 5.435 1.02 56.0 1.19

IMC-P8, Hopewell Site 1.25 0.39 5.15 8.21 3.99 208.5 12.32 6.75 0.092 449 107.7 10.1 0.544 1.87 129.0 3.03

IMC-P9, Four Corners Site 1.14 0.12 7.31 2.96 5.71 175.2 14.61 2.21 0.085 124 25.3 21.4 15.788 1.26 63.7 1.03

FIPR, Florida Institute of Phosphate Research 46.6 0.23 5.51 1.15 3.22 104.3 19.85 1.74 0.087 288 42.2 20.2 1.112 0.65 53.0 0.88

NU, Nu-Gulf Mulberry Corp. 39.9 0.11 6.89 4.58 1.82 132.4 9.19 1.94 0.095 136 17.0 20.4 2.886 1.25 51.0 0.72

0.16 6.81 3.57 4.60 186.4 13.93 2.65 0.077 191 36.7 18.8 4.805 1.08 66.9 1.07 Average => S.D. => 0.09 1.10 2.16 1.70 53.1 3.06 1.39 0.017 105 24.9 3.9 4.185 0.44 22.6 0.73

*Data for this clay sample (IMC-P2) was published in Lewis and others (2003). NOTE: nd = not determined. Samples provided primarily by IMC Phosphates Co. (IMC-P), Nu-Gulf (NU) and Florida Institute of Phosphate Research (FIPR). All samples provided as wet slurry of varying water content, although analysis was done on dry clays.

44

Table 5b. Published Analyses of Florida Phosphatic Clays and Phosphorites from Various Sources.

Sample Ag As Cd Cr Cu Fe Hg Mn Ni Pb Se Zn Sb Reference Settling Solids, µg/g nd 6.8 <10.03 51.2 1.27 5958 0.126 59.67 <24.50 <102.5 8.64 26.10 nd 1 Settling Solids, µg/g nd 6.40 <10.0 56.93 0.92 5443 0.020 63.45 <24.5 <102.46 9.51 22.51 <98.04 2 Fort Green 1, mg/kg nd 4.20 28.0 119.0 7.1 9330 0.07 81.0 13.0 9 2.00 27.0 nd 3 Fort Green 2, mg/kg nd 2.00 1.7 51.0 3.2 5130 0.03 35.0 5.7 5 1.00 16.0 nd 3 Four Corners, mg/kg nd 7.70 3.5 nd 2.8 5270 nd 44.0 nd 10 nd 24.0 nd 3 Bone Valley Formation, Florida, ppm 1.4 12 16 60 13 nd 0.025 230 9 55 2.6 180 nd 4 Average from 18 Regions, ppm 2 23 18 125 75 nd 0.055 1230 53 50 4.6 195 nd 4 NOTE: nd = not determined. Reference numbers above refer to:

(1) PEDCo (1981). (2) PEDCo (1983). (3) Data provided by IMC Phosphates Co., in-house survey. (4) Altschuler ZS (1980).

45

Table 6. Sediment Standards Provided by U.S. EPA Gulf Breeze Laboratory.

Lab ID

Client ID

% Moist.

Ag, ppm

Al%

As, ppm

Cd, ppm

Cr, ppm

Cu, ppm

Fe%

Hg, ppm

Mn, ppm

Ni, ppm

Pb, ppm

Se, ppm

Sn, ppm

Zn, ppm

EPA-246 Mississippi Sound A 50 0.299 4.51 8.41 0.113 30.4 3.79 2.52 0.073 17.50 6.09 17.50 1.59 1.89 32.00

EPA-247 Mississippi Sound B 26 0.024 0.16 0.57 0.014 3.6 0.46 0.08 0.010 1.54 0.82 1.54 0.60 0.13 4.25

EPA-248 Mississippi Sound 01 57 0.401 6.34 10.01 0.149 44.8 13.50 3.43 0.075 25.00 19.80 25.00 5.24 2.76 96.80

EPA-249 Mississippi Sound 02 23 0.020 0.06 0.38 0.009 2.2 0.49 0.06 0.006 1.36 0.63 1.36 0.52 0.11 3.30

EPA-250 Old River A 60 0.229 3.63 10.17 0.242 33.5 16.00 2.05 0.081 13.20 11.20 13.20 6.59 1.86 67.50 EPA-251 Old River B 22 0.011 0.05 0.16 0.009 1.6 0.18 0.03 0.003 0.54 0.23 0.54 0.55 0.08 3.35 EPA-252 Old River 01 30 0.028 0.28 1.39 0.035 6.4 1.91 0.26 0.021 1.88 1.43 1.88 1.28 0.22 9.11 EPA-253 Old River 02 24 0.012 0.31 0.15 0.054 1.8 0.78 0.03 0.001 0.81 0.95 0.81 0.47 0.07 6.81 EPA-254 Bay La Launch A 29 0.041 0.45 1.60 0.024 8.2 1.46 0.34 0.010 2.26 1.60 2.26 1.00 0.31 7.52 EPA-255 Bay La Launch B 24 0.030 0.12 0.62 0.015 5.8 1.11 0.16 0.001 1.61 0.87 1.61 0.85 0.26 4.01 EPA-256 Bay La Launch 01 49 0.129 2.32 6.66 0.090 20.0 5.05 1.42 0.047 7.39 5.47 7.39 3.37 1.14 26.50 EPA-257 Bay La Launch 02 24 0.024 0.13 0.28 0.017 2.9 0.73 0.09 0.001 1.34 0.58 1.34 0.63 0.17 2.36 EPA-258 Withlacoochee

River A 29 0.078 0.07 0.70 0.286 5.5 0.82 0.23 0.007 2.00 1.21 2.00 1.09 0.12 3.44

EPA-259 Withlacoochee River B 45 0.031 0.45 0.89 0.119 10.6 2.71 0.49 0.020 5.21 2.80 5.21 2.87 0.24 11.80

EPA-260 Withlacoochee River 01 28 0.013 0.19 0.43 0.049 6.1 0.63 0.18 0.006 2.07 1.07 2.07 1.31 0.08 2.38

EPA-261 Withlacoochee River 02 46 0.034 0.73 1.29 0.127 16.0 3.35 0.79 0.026 7.49 4.04 7.49 3.40 0.27 10.70

EPA-262 Suwannee River A 23 0.010 0.10 0.33 0.074 4.3 0.00 0.22 0.001 1.05 0.82 1.05 0.69 0.06 6.32 EPA-263 Suwannee River B 47 0.045 0.70 0.84 0.325 8.4 1.76 0.50 0.030 4.89 2.53 4.89 1.97 0.42 19.80 EPA-264 Suwannee River 01 19 0.016 0.11 0.39 0.070 3.9 0.01 0.23 0.001 1.34 0.91 1.34 0.81 0.08 7.47 EPA-265 Suwannee River 02 77 0.124 2.07 2.14 0.897 25.1 4.03 1.17 0.057 12.80 5.57 12.80 4.45 1.31 39.10 EPA-266 Grand Lagoon A 31 0.015 0.08 1.06 0.050 9.4 1.87 0.13 0.001 1.35 1.00 1.35 0.97 0.20 6.11 EPA-267 Grand Lagoon B 26 0.001 0.07 0.75 0.011 5.0 0.45 0.07 0.000 0.43 0.41 0.43 0.62 0.10 2.52 EPA-268 Grand Lagoon 01 24 0.008 0.10 0.55 0.012 1.9 1.06 0.10 0.000 0.83 0.43 0.83 0.63 0.17 3.65 EPA-269 Grand Lagoon 02 19 0.006 0.03 0.43 0.008 3.9 0.25 0.06 0.000 0.37 0.31 0.37 0.67 0.07 3.04 EPA-270 Old River Seagrass 23 0.058 0.05 0.36 0.217 1.6 0.25 0.07 0.003 0.34 0.36 0.34 0.99 0.07 3.04 EPA-360 Sample #A&B 9 0.013 0.01 0.08 0.024 0.9 0.30 0.05 0.000 0.40 0.27 0.40 0.29 0.09 2.53 EPA-361 Sample #C 19 0.012 0.23 0.30 0.020 2.2 8.52 0.12 0.046 3.75 0.51 3.75 0.50 0.13 12.10 EPA-362 Sample #F 21 0.027 0.48 0.63 0.245 4.4 22.30 0.26 0.185 10.20 2.54 10.20 1.39 0.27 74.60-

0.062 0.85 1.84 0.118 9.7 3.35 0.54 0.025 4.61 2.66 4.61 1.62 0.45 16.82 Average => STD => 0.096 1.55 2.99 0.179 11.1 5.40 0.84 0.040 6.05 4.16 6.05 1.60 0.69 24.39

46

Table 7. Comparison of Phosphatic Clay and Natural Sediment Quality Standards. Concentrations Compared to Sediment Quality Guidelines Proposed for Florida Coastal Areas by MacDonald and Others 1996.

Trace Element

TEL* (µg/g)

PEL* (µg/g)

Phosphatic Clay This Study

(ppm)

Natural Sediments**

(ppm) Ag -- -- 0.16 0.2 As 7.2 41.6 3.57 12.5 Cd 0.7 4.2 4.60 0.3 Cr 52.3 160 186.4 96.3 Cu 18.7 108 13.93 22.7 Hg 0.13 0.7 0.077 0.3 Mn -- -- 191 589 Ni 15.9 42.8 36.7 26.9 Pb 30.2 112 18.8 33.1 Sb -- -- 1.07 Se -- -- 4.805 1.3 Sn -- -- 1.08 2.9 Zn 124 271 66.9 116.5

Al % -- -- 6.81 6.8 Fe % -- -- 2.65 4.8

*TEL = Threshold Effects, PEL = Probable Effects Guideline Values. **Published in Lewis and others (2003).

The amounts of 266Ra, 238U and 210Pb in four selected phosphatic clays are

presented in Table 8. First, we observed some variability among the samples. For example, samples from Nu-Gulf contained about one-third the amount of each element compared to samples from IMC Phosphates. Nu-Gulf samples also contained about half the amount of each element compared to samples from the Florida Institute of Phosphate Research (FIPR).

47

Table 8. Radioactivity Measurements of Florida Phosphatic Clays.

Source Sample Units 266Ra 238U 210Pb

Phosphatic Clay

IMC-P2 (IMC Phosphates Co.)

IMC-P4 (IMC Phosphates Co.)

FIPR (FL Institute of Phosphate Research) NU (Nu-Gulf, Inc.)

dpm/g dpm/g dpm/g dpm/g

58.95 (1.5) 72.70 (1.0) 47.32 (1.4) 15.44 (3.0)

70.44 (3.7) 80.56 (2.7) 49.97 (4.0) 16.67 (9.5)

76.27 (3.0) 73.99 (2.5) 46.35 (3.6) 15.62 (6.2)

Concentration in Sediment Concentration in Seawater

dpm/g dpm/L

1-3 0.10

na 2.5

2-125 0.1 Comparison

Values EPA Drinking Water StandardEPA Max. Effluent Standard

dpm/Ldpm/L

0.666 133.2

0.033 666

0.044 22.2

These values were then compared to concentrations in sediment, as well as to U.S.EPA guidelines for effluents. The amount of 266Ra in all phosphatic clay samples were higher than the concentration in sediment, while the amount of 210Pb fell within the range of values found in sediments. Assuming full release of the elements from the clay to the surrounding medium, the amounts carried in phosphatic clay for 266Ra, and 238U are lower than those for the effluent standards. For 210Pb, samples from Nu-Gulf were lower than the effluent standards, while those from IMC Phosphates and FIPR were above the standard by a factor of 2 or 3. However, these are the worst case scenarios. More tests are needed to determine the actual amounts of each element released into water by phosphatic clays.

The results of the nutrient chemistry are summarized in Figure 13. In terms of nitrate (Figure 13A), Wyoming bentonite (WB-B) released significant amounts of this nutrient, while phosphatic clay (IMC-P2) and cationic polymer-treated kaolinite (H-DP) did not. Likewise, polyaluminum chloride (PAC) added alone did not add nitrate to the medium. None of the clays tested (or PAC) added significant amounts of ammonium into the seawater (Figure 13B). However, phosphatic clay released significant amounts of phosphate into the media, as expected (Figure 13C). The addition of WB-B, H-DP and PAC did not significantly contribute to the phosphate concentration. Finally, WB-B significantly increased the concentration of silicate in the system (Figure 13D), while the two other clays and PAC did not.

In the second study, the experiments were repeated using seawater containing a known amount of each nutrient (i.e., “spiked”). Furthermore, the H-DP and IMC-P2 were also tested in combination with PAC. With respect to nitrate (40 µM spike), all of the clay and clay+PAC treatments significantly reduced the amount of dissolved nitrate (Figure 13A). However, the greatest reduction was observed with PAC-treated phosphatic clay, which saw change by two orders of magnitude (i.e., from 4 µM to 0.046 µM). While both untreated and PAC-treated H-DP decreased the amount of the nitrate

48

spike, there was no significant difference between the two. Similarly, all of the clays (with and without PAC) were able to decrease the amount of ammonium spike (4.5 µM) in the medium, mostly to about half the ammonium concentration (Figure 13B). Again, the greatest reduction was achieved by PAC-treated phosphatic clay, which produced a loss of one order of magnitude in ammonium concentration (i.e., from 4.5 µM to 0.45 µM).

With phosphate, 2 µM was added to the ambient seawater (Figure 13C). The first experiment quantified the amount of each nutrient released or absorbed by clay alone in nutrient-poor seawater. The second experiment determined the amount of each nutrient absorbed or released by clay alone or PAC-treated clay in a nutrient-spiked mixture. The highest reduction in dissolved phosphate was observed with the use of untreated and PAC-treated H-DP. H-DP alone decreased the amount of phosphate by a factor of 5.5, while H-DP with PAC combined decreased the amount by a factor of 9. The application of WB-B also saw a reduction in phosphate, but only by a factor of 2.5 in this experiment. The use of phosphatic clay alone did result in reduction of phosphate content by a factor of 1.2. However, when PAC was added to phosphatic clay, phosphate concentration decreased by a factor of 3.6 in the final suspension. The reduction in phosphate content produced by PAC-treated IMC-P2 was greater than that of WB-B bentonite without PAC.

A. Nitrate0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

seawater only

H-DP

IMC-P2

WB-B

PAC

nitrate + seawater only

nitrate + H-DP

nitrate + IMC-P2

nitrate + WB-B

nitrate + H-DP + PAC

nitrate + IMC-P2 + PAC

concentration (uM)

B. Ammonium0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

seawater only

H-DP

IMC-P2

WB-B

PAC

ammonium + seawater only

ammonium + H-DP

ammonium + IMC-P2

ammonium + WB-B

ammonium + H-DP + PAC

ammonium + IMC-P2 + PAC

concentration (uM)

C. Phosphate0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00

seawater only

H-DP

IMC-P2

WB-B

PAC

phosphate + seawater only

phosphate + H-DP

phosphate + IMC-P2

phosphate + WB-B

phosphate + H-DP + PAC

phosphate + IMC-P2 + PAC

concentration (uM)

D. Silicate0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

seawater only

H-DP

IMC-P2

WB-B

PAC

silicate + seawater only

silicate + H-DP

silicate + IMC-P2

silicate + WB-B

silicate + H-DP + PAC

silicate + IMC-P2 + PAC

concentration (uM)

Figure 13. Dynamics of Biologically-Significant Inorganic Nutrients in Suspension with Selected Clays. IMC-P2 = Florida Phosphatic Clay (IMC Phosphates Co.), H-DP = Cationic-Polymer-Treated Kaolinite (J.M. Huber Co.), and WB-B = Wyoming Bentonite (Wyo-Ben, Inc.). PAC = Polyaluminum Chloride.

49

50

Lastly, the amount of silicate added to the system was negligible (0.003 µM) in our experiment (Figure 13D). As in the previous experiment where no spike was added, WB-B contributed a significant amount of silicate into the surrounding medium, while H-DP and IMC-P2 did not. Moreover, the use of PAC with the two latter clays did not significantly add or remove silicate, but there was a high degree of variability associated with the measurements made for H-DP+PAC. PHYSICAL BEHAVIOR OF SETTLED FLOCS

In both sets of flume experiments, critical bed shear stress (τ0crit) for the resuspension of clay/algal flocs was influenced by the settling time period and by the presence/absence of PAC. For the first set of experiments in the 17-m flume, both factors (settling time and PAC) were significant in the two-way ANOVA (with no interaction effect). Multiple comparison of the mean values for τ0crit (plotted in Fig. 14) revealed the following, with treatments listed in order of τ0crit increasing to the right and nonsignificant differences connected with an underline:

3PAC 9PAC 3 24PAC 9 24

Note: Each symbol indicates mean ± std. devn. for three replicate flume runs. (Values reported in Table 8

as u*crit for mean flow resuspension.) (Source: Beaulieu 2003).

Figure 14. Critical Bed Shear Stress (τ0crit) for Resuspension of Clay/Algal Flocs in the 17-m Flume.

51

These results suggest that the most important change in consolidation for the 2-mm thick layer occurred between 3 and 9 hours. In general, initial motion occurred at about half the flow speed (25% the bed shear stress) for mean flow resuspension. Bedload occurred at about 40% and wave-induced resuspension at about 60% of the stress for mean flow resuspension. Critical shear velocities (u*crit), along with a typical quadratic drag coefficient for horizontal flow above a bottom boundary layer (CD = 0.0025) were used to estimate the approximate mean flow speeds that would be measured in the field during resuspension events (Table 9). In the absence of PAC, resuspension of a 2-mm thick layer that settled for 3 hours in relatively low flow speeds (≤ 3 cm s-1) would be expected at a flow speed of 15-16 cm s-1 in the field, as compared to 18-19 cm s-1 for a layer that accumulated in 9 or 24 hours. Table 9. Critical Shear Velocity (u*crit) for the Resuspension of Clay/Algal Flocs in

the 17-m Flume.

Experiment Description

Settling Time (Hr)

u*crit (cm s-1)

Approx. Flow (cm s-1)

Clay + Heterocapsa 3 0.82 ± 0.04* 16.4 Clay + Heterocapsa 9 0.89 ± 0.01 17.8 Clay + Heterocapsa 24 0.93 ± 0.03 18.6 Clay + PAC + Heterocapsa 3 0.76 ± 0.01 15.2 Clay + PAC + Heterocapsa 9 0.82 ± 0.03 16.4 Clay + PAC + Heterocapsa 24 0.84 ± 0.01 16.8

*Contains an outlier based on post-experiment observations of video. Note: Values are mean ± std. devn. for three replicate flume runs. Values for the approximate mean

horizontal flow above the bottom boundary layer in the field were calculated assuming a typical quadratic drag coefficient (CD = 0.0025).

Records of suspended mass during the second set of experiments in the racetrack flume are presented in Fig. 15. Critical shear velocities (u*crit) for the three resuspension events during experiments with flow initially at 3 or 10 cm s-1 are given in Table 9. The higher values for u*crit during the 10 cm s-1 runs are likely because only the flocs with greater settling velocity (i.e., the heavier flocs) were deposited while the rest remained in suspension (see Table 10 for the fraction of clay mass that deposited). During the flume run in which paddle speed was stepped down from 20 cm s-1, a layer began to accumulate at the bed with the paddle speed at 12 cm s-1, and critical shear stress for deposition was calculated as 0.034 Pa (corresponding to ~12 cm s-1 as a mean flow in the field). Cell counts during the racetrack flume experiments are plotted in Fig. 16. For the runs without PAC at 3 and 10 cm s-1, >80% of the cells were deposited in the floc layer that accumulated on the bed. However, removal efficiencies were considerably lower when the experiments were repeated with PAC (Table 11).

52

Note: The average calibration of the OBS to suspended mass was used to set the values for the right

axis (note slight offset from 0 due to gain set on OBS). Linear fits to 2-min subsets of data are superimposed on plots A, B, D, and E. A, B. Paddle speed 3 cm s-1. OBS at 3.4 cm above bottom. A. No PAC. B. With PAC. C, D. Paddle speed 10 cm s-1. C. No PAC. OBS at 6 cm a.b. Gaps indicate manual displacement of the OBS to 3.4 and 8 cm a.b. during resuspension events. D. With PAC. OBS at 8 cm a.b. E. Paddle speed 20 cm s-1. No PAC. OBS at 3.4 cm a.b. Paddle speed was decreased by 2 cm s-1 every three hours. Arrows indicate decreases to 12, 10, and 8 cm s-1. (Source: Beaulieu 2003.)

Figure 15. Optical Backscatter (OBS) Plots for Racetrack Flume Runs.

53

Table 10. Critical Shear Velocities (u*crit) for Resuspension of Clay/Algal Flocs in the Racetrack Flume.

u*crit (cm s-1)

for Mean Flow Resuspension Experiment Description

Flow Speed (cm s-1)

During Settling 1st 3 hr 2nd 3 hr 9 hr Clay + Heterocapsa 3 0.76 0.76 0.76 Clay + PAC + Heterocapsa 3 0.72 0.72 0.76 Clay + Heterocapsa 10 0.90 0.90 0.95 Clay + PAC + Heterocapsa 10 0.90 0.90 0.95

Note: Values are single estimates derived from OBS slope analysis. “1st 3 hr” refers to the resuspension event after the first 3-hr settling period, followed by a second 3-hr settling period and resuspension event (“2nd 3 hr”), followed by a 9-hr settling period and final resuspension event (“9 hr”).

54

Table 11. Clay Deposition and Algal Removal Efficiency (RE) in the Racetrack Flume.

*Cell counts may be low due to large amount of clay in sample. Note: Values are reported for the % of clay mass that deposited during the first and second 3-hr settling periods and the 9-hr settling period (as depicted in

Figure 17). RE values are based on the cell counts just prior to clay addition (shown in Figure 18).

% of Clay Mass That Deposited Removal Efficiency (RE) of Heterocapsa

Experiment Description

Flow Speed (cm s-1)

During Settling 1st 3 Hr 2nd 3 Hr 9 Hr 1st 3 Hr 2nd 3 Hr 9 Hr

Clay + Heterocapsa 3 94% 94% 96% 100% 98% 99% Clay + PAC + Heterocapsa 3 94% 95% 96% 78% 25% 19% Clay + Heterocapsa 10 79% 72% 86% 89% 78% 90% Clay + PAC + Heterocapsa 10 39% 31% 40% 28% 15% 69% Clay + Heterocapsa 20 1% 41%*

55

Note: Symbols indicate mean ± range of 2 or 3 counts of each 50-ml sample of flume water. Note log scale. Mean values were used to determine removal efficiency (RE) reported in Table 10. (Source: Beaulieu 2003.)

Figure 16. Cell Counts for Racetrack Flume Runs.

33 PAC1010 PAC20

56

OXYGEN DEMAND IN MESOCOSM EXPERIMENTS

The removal of Karenia brevis in the mesocosm experiments is presented in Figure 17. In the clay treatments, red tide cells were removed in both the surface and deep portion of the mesocosm, leading to an overall RE of 64% in Treatment 1 and 71% in Treatment 2 after 4 hrs. In the untreated mesocosm, K. brevis disappeared from the deeper part of the enclosure, but appeared in the surface, as demonstrated by the large “negative” removal efficiency (Figure 17). The total cell removal for the mesocosm was 2%. Lastly, the cell removal from the ambient site was 67% in the absence of clay.

From the sediment trap analysis, the total settled solids (TSS) was approximately

50% lower in the control mesocosm than the ambient control site (Figure 18). The amount of volatile settled solids (VSS) was also lower in the control mesocosm compared to the ambient site. However, the biological oxygen demand (BOD) in the control (8.8 mg/L) and ambient sites (8.0 mg/L) was approximately the same.

The addition of the phosphatic clay to the mesocosms increased TSS by a factor of 15, VSS by a factor of 6, and the BOD rate by a factor of 1.3. In addition, the input of inorganic clay fraction reduced (or diluted) the relative percentage of VSS from ~26% of the TSS to 10% of the TSS. In effect, the organic material in the surface sediments was thus reduced by 62%.

-60

-40

-20

0

20

40

60

80

100

Ambient control Control Treatment 1 Treatment 2

cell

rem

oval

eff

icie

ncy

(%)

Figure 17. Karenia brevis Removal in Mesocosm Experiments Using IMC-P4 Phosphatic Clay.

surface (0.5 m) depth (2.5 m) surface & depth

57

Figure 18. Biological Oxygen Demand, Total Settled Solids and Volatile Settled Solids from Mesocosm Experiments.

59

CONCLUSIONS AND RECOMMENDATIONS

CLAY EFFECTIVENESS

Our results showed that the removal ability of phosphatic clays from various sources and locations in Florida, were comparable and high (>88% removal efficiency) against Karenia brevis, when clay loading exceeded 0.25 g/L in suspension. Below 0.25 g/L, a wider range of removal efficiencies (RE’s) were observed among the various samples, especially as the clay concentration decreased. The reasons for this variability are unknown. There was no observed correlation between the clay’s removal ability and the percent solids content of the original stock material (Table 1). Differences in the composition and relative proportions of the various minerals among the clay samples may explain these results, however, this analysis was not performed in this study. Likewise, differences in the size frequency distribution among the samples could influence the ability of phosphatic clay to remove K. brevis, as a higher proportion of finer particles – which remain in suspension longer and have a greater opportunity to interact with algal cells – would mean that more of the clay mass will be “active” in flocculating with the organisms. However, particle sizing was not performed in this study. Nevertheless, the important conclusion that can be drawn from this study is that Florida phosphatic clays will remove the Florida red tide organism at a uniformly high level above the threshold concentration of 0.25 g/L, regardless of source location and solids content of the stock material. Therefore, the choice of which material to use in future research or field application will be determined by practical consideration (e.g., handling, processing, dispersing in larger scales), economics, and the environmental impacts. For example, clay samples from Nu Gulf contain unknown amounts of petroleum products which are introduced during the extraction of phosphate rock. Recent efforts at IMC Phosphates to improve the settling of clays involved the addition of certain cationic polymers which can potentially affect this clay’s removal ability (as demonstrated in this study), and its impact in the environment. Impact studies would then be needed to ascertain the consequences of releasing such clays in the environment.

In this study, we found that the removal ability of phosphatic clay (IMC-P2) greatly exceeded that of phosphate rock produced in the conventional manner (see Methodology section). The differences in RE were quite clear, even when both materials where only at 0.025 g/L in suspension. Phosphate rock did not disperse into suspension as readily as phosphatic clay, despite frequent and vigorous agitation. Large pieces of phosphate rock were clearly visible at the bottom of test tubes containing the working suspensions, as well as in the tubes with cell cultures during the experiments. Therefore, the disparity between the effectiveness of phosphate rock and phosphatic clay may be a function of the particle size distributions of these samples (i.e., phosphatic clays have a higher content of finely divided particles which remain suspended longer leading to greater interactions with cells). Additional tests are recommended to determine whether the removal ability of phosphate rock could be improved by crushing the material further, and to ensure that a suspension of finer particles would be produced. If finely crushed phosphate rock proved to be equally or more effective than phosphatic clay, then the

60

choice of using it over phosphatic clay would be governed by the cost of production, and other practical considerations. At this point, however, a significant issue in applying phosphate rock over phosphatic clay in natural waters, is the environmental impact of introducing potentially high amounts of inorganic phosphorus from the ore, which can, for instance, dramatically influence the occurrence of future red tides under the right circumstances. Although phosphatic clays also contain significant amounts of inorganic phosphate (see results of chemical analysis), they have far less than the amount in pure apatite.

In the absence of clay, coagulants like alum and cationic flocculants (e.g., Percol LT-7990 and LT 7991) led to cell loss ranging from 5% to 40%, between 1 and 1000 ppm. By comparison, cell removal with untreated phosphatic clay ranged between 5% and 20% with 0.01 g/L, but quickly increased to > 90% when 0.10 g/L or higher was applied. This result clearly demonstrated the greater efficacy of clay flocculation in removing algal cells compared to some of these conventional additives used in water treatment. Several reasons can be proposed to explain these findings. First, when coagulants and flocculants are used in water treatment, the water column is typically mixed to increase the collision rate among the particles, thus promoting rapid flocculation. In our experiments, alum and the flocculants were added to the surface without further mixing. Hence, any effect that these chemicals might have on the surface charge of the algal cells would be limited near the water column surface. Moreover, any increase in the adhesiveness of algal cells that comes from adding these substances, would not necessarily result in higher flocculation rates, if interparticle collisions do not increase as well. Second, aggregates consisting of algal cells alone may not necessarily sink faster (according to Stokes’ Law), despite their increased size, if the density of these aggregates does not increase as well. Dinoflagellate cells are only slightly denser than seawater, and agglomerates of these particles may not achieve enough of a density to increase their settling rates. On the other hand, clay minerals are two or three times more dense than seawater. The attachment of these particles onto the cell surface, and their incorporation into a growing floc, would clearly increase particle density as the aggregate increases in size. With this argument, it is clear to see how the application of enough clay minerals in suspension with algal cells would yield significant removal compared to chemical enhancers alone.

In this study, the use of alum with a low concentration of phosphatic clay (e.g., 0.01 g/L) did not increase cell removal, relative to alum alone or clay alone, until alum dosage was 1000 ppm. At higher clay concentrations (e.g., 0.10 g/L and 0.50 g/L), the addition of alum did not further improve the (already) high removal ability of clay alone. Thus, we concluded that alum is not a suitable alternative to or enhancer of clay flocculation.

The use of Percol LT-7990 and LT-7991, in combination with phosphatic clay, did result in a four-fold improvement in cell removal at low clay concentrations (i.e., 0.01 g/L). However, the highest observed RE’s were only 37% with 1000 ppm of LT-7990, and 31% with 100 ppm of LT-7991. These were far less than what could be achieved with phosphatic clay alone at higher clay concentrations (e.g., ≥ 0.10 g/L). The use of

61

either flocculant did not improve the removal ability of phosphatic clay at higher clay concentrations (e.g., 0.10 and 0.50 g/L). In fact, RE actually decreased when more flocculant was added (≥ 10 ppm), which is likely the result of very rapid clay flocculation with itself as soon as the flocculant was added to the clay suspension. By the time the clay slurry was added to the culture, most of the clays had already formed aggregates consisting of strong particle contacts that could not be disrupted by vigorous agitation or vortexing. As a result, the large aggregates may have sunk through the culture medium without sufficient opportunity to flocculate with the algal targets, instead removing cells by entrainment during its descent. Although we have only tested two select flocculants, we concluded that these additives, used in this way, would not be suitable for our purposes, as they are too aggressive in flocculating the clay suspension, thus reducing its removal ability. However, alternative methods of combining clay with these substances may be devised in the future that would maintain the clay’s effectiveness.

The most promising flocculant used in this study was polyaluminum chloride (PAC). With 5 ppm of PAC, cell removal was increased significantly in the range of 0.01 and 0.04 g/L of clay. Unlike clays treated with Percol, the clay slurry did not flocculate excessively when PAC was added, and large clay flocs were not observed in the cell cultures during experiments. Based on these results, we concluded that PAC deserved further investigation as a potential enhancer of clay flocculation and cell removal.

Aside from clay concentration, we found that removal efficiency can also vary with the concentration of Karenia brevis in suspension. Previously, we observed that RE was a hyperbolic function with respect to increasing clay concentration, with a maximum value of 70% for most of the phosphatic clays tested (i.e., Figure 5 re-plotted as a series of curves, not shown). These results were found with relatively constant cell concentrations around 10,000 cells/mL. Interestingly, when cell concentration was also varied, ranging between 200 and 13,000 cells/mL, RE was again a hyperbolic function with increasing clay concentration that reached its highest values at ≥ 0.10 g/L (Figure 9). However, the magnitudes of RE (i.e., plateaus) varied widely depending on the concentration of cells being treated, generally increasing with increasing cell concentration. Moreover, 70% RE was not observed until cell concentration reached 1,049 cells/mL in suspension. At lower cell concentrations, cell removal was constant above 0.10 g/L even though more clay was added. These findings suggest that should bloom concentrations fall below a minimum “threshold” level--in this case, 1000 cells/mL--the effectiveness of clay treatment would be limited, despite increasing the dosage. Conversely, clay treatment can attain its greatest efficacy when Karenia concentrations are sufficiently high, above the threshold level. Korean colleagues also reported similar results (H.G. Kim, National Fisheries Research and Development Institute, Pusan, South Korea, pers. comm.). For example, treatment efforts are only mounted when bloom densities (Cochlodinium polykrikoides) surpass 3,000 cells/mL. Furthermore, treatment is initiated around noon when the highest cell concentrations are usually found, as the organisms vertically migrate to and accumulate near the surface. One possible explanation for these results is that the presence and incorporation of the organisms themselves in the flocs influences the removal of other cells farther down in

62

the water column. Aggregates composed of algal cells are much larger than those formed from clay minerals only. This can be seen clearly in experiments conducted in settling columns (see below). Hence, clay-cell flocs, with their larger “projected” surface area, pass through more of the water column during settling, compared to clay-clay flocs, and they have a better chance of entrapping other cells beneath the surface. In situations where the cell concentration is well below 1,000 cells/mL, some of the Karenia cells are removed, but without sufficient formation of clay-cell flocs, the remaining cells were unaffected, even though clay loading increased to inundate the system.

In 2.5-m tall settling columns, the dynamics of clay and cell concentrations over time were clearly evident and provided a better understanding of the flocculation process in space and time. At all three clay loadings (0.04, 0.10 and 0.20 g/L), the added clay near the surface was already detected at 1.75 m within the first 20 min. Most of the clay mass settled through the column within the first hour of dispersal. At each depth, there was a correspondence between the onset of cell removal (i.e., decrease in cellular fluorescence) and the arrival of the clay plume (i.e., increase in spectrophotometric absorbance) (Figure 10). As the clay concentration increased, the surface clay signal (i.e. at 0.25 m) broadened over time. By the time the peak of the clay signal reached 0.50 m and deeper, its magnitude was reduced by about half, and was displaced in time. Cell removal was always highest near the surface (at 0.25 m) where the clay concentration was highest. However, cell removal began to decline over time as the clay sank and diluted, especially at the lower clay loading (i.e., 0.04 g/L). As a result, RE ranged from 75% near the surface to only 25% at 1.75 m. With higher amounts of clay (i.e., 0.10 and 0.20 g/L), greater cell removal was found at deeper depths, all occurring in the first 60 min. This exercise demonstrated that the depth of the water column must be accounted for in calculating the appropriate loading rate to remove cells effectively. In practical terms, loading rates or clay concentrations should be presented as mass per unit volume (e.g., g/m3 or g/L), instead of mass per unit surface area (e.g., g/m2), to account for the length scale within which the clays will be diluted and their efficacy reduced. CELL VIABILITY, RECOVERY AND GROWTH

Clay treatment can kill Karenia brevis. However, the extent of cell mortality depended on the interplay between clay concentration, the frequency of resuspension, and the timing of the first resuspension event. Below 0.10 g/L, virtually all of the flocculated cells eventually escaped and resumed vegetative growth within a few hours after deposition, with or without resuspension. We did observe, however, that recovery became increasingly delayed as more clay was added, in a range between 0.03 and 0.13 g/L, without resuspension. Above 0.50 g/L, high RE was also accompanied by extremely high cell mortality in the sedimented layer, even with frequent resuspension. Cell death was not instantaneous, but increased dramatically between 2.5 and 12 hrs after deposition. At intermediate concentrations (i.e., between 0.10 and 0.50 g/L), cell removal was expectedly high, but the amount of mortality was dependent to when the first resuspension event took place. We found that cells can survive in the floc layer for

63

up to three days, and that these cells can return to vegetative growth. However, resuspension must take place during this time for the cells to escape and recover.

In practical terms, these findings suggest an apparent lower limit to the amount of clay required to treat a Karenia bloom. That is, the loading rate of clay must be sufficient to physically remove the organisms, as well as minimize their escape, and maximize their mortality in the floc layer. Likewise, reducing the amount of clay to limit the potential impacts of treatment, must also be balanced by considering cell escape.

Shirota (1989) proposed that cell death may be the result of Al3+ ions released by the clay which causes disruption of the cell membrane and cell lysis. However, our studies showed that the cause of cell death appeared to be the physical contact between the algae and the minerals, not the release of potentially cytolytic substances into the surrounding media from clay (Sengco and others 2001). A mechanism for cell lysis by direct contact remains unknown. With respect to the impact of cell mortality, an important consideration is the fate and environmental effects of the potent toxins within the cell. Typically, these toxins are only released when the cells are killed and lysed. Our experiments revealed that cell death does not occur immediately, but after several hours of exposure to sufficient amounts of clay. Therefore, we would expect that any impact due to toxin release would take place at or near the bottom. Research is being conducted elsewhere to address toxin dynamics on clay particles. POTENTIAL IMPACTS OF CLAY DISPERSAL

In addition to investigating its efficacy and cost-effectiveness, a vital area of research into clay dispersal deals with the environmental impacts of treatment. The list of potential impacts can be rather long, but they can be categorized as “water column” and “benthic” effects. Moreover, dispersal impacts can be further characterized as physical, chemical, and biological effects. In this section, we discuss results of several laboratory experiments on water column chemistry, and possible effects of settled clays on sediment quality and organic composition. Although our results represent just some of the areas of concern, these studies provide the crucial datasets that will be required to evaluate this control strategy. Seawater Chemistry

Based on our trace metals analysis, the primary element of concern in the phosphatic clay samples we tested appeared to be Cr, which was elevated relative to other phosphatic clay analysis performed by IMC Phosphates and to the report by Altschuler (1980). Our measurements of Cr levels were higher than those for reference sediments (from U.S. EPA, Gulf Breeze, FL), natural sediments (Lewis and others, submitted), and sediment quality parameters proposed by MacDonald and others (1996). Two other important elements in our analysis were Ni and Cd. We found both to be elevated in phosphatic clays versus natural sediments. Based on the criteria set by MacDonald and

64

others (1996), Ni levels were between the Threshold and Probable Effects Levels, while Cd, as with Cr, was above the Probable Effects Level. On the opposite side, the levels for As, Hg, Pb and Zn were consistently below the levels found in reference sediments, natural sediments and the proposed sediment quality parameters for Florida.

It is unknown at this time how these results will influence the decision to use phosphatic clays in further studies, especially regarding Cr and Cd. An investigation into the actual, direct impacts of excess phosphorus, metals and radionuclides on water quality and biota was beyond the scope of this project. In our proposal, our objectives were to quantify bulk composition of several elements in clays and/or the solubility of radionuclides and important algal nutrients (like phosphorus). Our intent here was to provide a baseline quantification of these elements first, but the consequences of such finding would be left for further research (which is now underway elsewhere). Therefore, we recommend that the implications of these findings be further investigated.

Nevertheless, we can begin to make several inferences regarding the potential impact of phosphatic clays on the viability of phytoplankton and benthic (bottom) organisms from other studies in this report, and other studies, which were completed by our colleagues. First, the results from the study of phytoplankton mortality (p 17-19) showed that the red-tide organism, Karenia brevis, can survive clay treatment with clay concentrations < 0.50 g/L and the opportunity to escape the floc layer through resuspension, even after three days of continuous contact with clay particles. Below 0.50 g/L, cells that escaped resumed vegetative growth even with the clay layer present on the bottom of the test tubes, and frequent resuspension events. When cell death was observed, the process was not instantaneous following direct contact with the phytoplankton, even at 0.50 g/L. This was evident from video microscopy of living cells in the presence of clays (Sengco, unpublished data). In the field work described in this report (p 24-26), we also observed that 0.25 g/L of clay was sufficient to remove Karenia brevis enclosed in mesocosms with up to 77 % two hours after treatment. However, the removal efficiency dropped to ca. 60% after the 4-hr experiment due to the apparent escape of cells from the flocs. These observations suggest that phosphatic clays are not inherently deleterious to these algal cells in the small amounts being tested. Other laboratory experiments have also shown that phosphatic clays removed but did not appear to kill Heterocapsa triquetra in the flume experiments. These are only two algal species where such data are available.

Regarding the benthic impacts of phosphatic clay, two studies have now been published. In the first study, phosphatic clay alone and polyaluminum chloride (PAC) alone were shown not to have deleterious effects (i.e., mortality) on four sentinel species when added at 0.25 g/L (Lewis and others 2003). There were no sublethal effects or bioaccumulation of elements determined for this work. When Karenia brevis cells were incorporated into the flocs, significant animal mortalities were observed for three species, relative to clays alone. These results were similar to the effect of sedimented K. brevis alone. Therefore, the authors concluded that the effect of flocculating K. brevis with clays would not be better or worse than the effect of settling the bloom itself onto the benthos.

65

In the second study, the effect of clays on juvenile hard clams, Mercenaria mercenaria, were studies under two flow conditions (Archambault and others 2004). When the clays are added under flow conditions that allowed full sedimentation of clay-algal (i.e., non-toxic, Heterocapsa triquetra or Prorocentrum micans) flocs, there were no significant differences between the survival of juveniles relative to clay-free controls. The clams immediately resumed contact with the overlying water and cleared the area around them from clay. However, then phosphatic clays were added to high flow that kept them suspended (for two weeks continuously), there was a significant effect on the growth rate of the clams, but no mortalities were observed. The bioaccumulation of elements was not determined during this study. The fact that the juveniles did not appear significantly harmed by the clays so long as the clay settled suggest that the clays are not in itself deleterious to these organisms. The effect of incorporating toxic algal cells to the floc was not done in this study, but it is now the focus of on-going research. Both this and the previous benthic study using natural benthic communities above will be investigated again in the field in the new program we are currently engaged in.

Due to their ecological significance, there have been numerous studies on the bioavailability and bioaccumulation of metals on aquatic ecosystems and into benthic organisms. In a recent study, Mountouris and others (2002) developed statistical techniques to examine correlations between heavy metal bioconcentration factors (e.g., for Cd, Cr, Cu, Ni, Pb, Zn, Hg) and sediment characteristics that were expected to affect bioavailability (e.g., concentration of iron, manganese, aluminum oxides, level of organic carbon and acid volatile sulphide). The authors used published data from various locations in their analysis. Indeed, the best models were found only when the sediment characteristic above were factored, instead of simply comparing the concentrations of heavy metals in the biota and the sediment. In other words, there does not appear to be a direct correspondence between the amount of these heavy metals in the sediment and their bioaccumulation into biota.

Using the deposit- (sediment-) feeding peanut worm, Sipunculus nudus, Yan and Wang (2002) studied the uptake rates of Cd, Cr and Zn from the medium and the sediment into the organism. Uptake rates were generally low, especially for Zn, which was disproportionately slower with increasing Zn concentration in the medium. The assimilation efficiencies reported were 6-30% for Cd, 0.5-8% for Cr and 5-15% for Zn. In addition, there was no major difference in the assimilation efficiencies from sediments collected from a contaminated site. The higher assimilation of Cd appeared to be correlated to the relative ease of dissociation of the metal in the animal’s gut fluid, compared with the lower extraction of Cr and Zn in the gut. The bioaccumulation of the three metals appeared to be dominated by sediment ingestion due to the low uptake from the solute phase as well as the high metals concentration in the sediment.

In another study, Geffard and others (2003) studied the bioavailability and toxicity of heavy metals (Cd, Cu, Zn) on the embryos and larvae of the oyster, Crassostrea gigas. This organism was considered as one of the most sensitive of all the classically-used bioassays. In this study, the authors examined the effect of both whole sediment and elutriates. The sediment used contained 2-3 orders of magnitude more of

66

each metal than found in phosphatic clays (Table 5, p. 40). They demonstrated that these metals were accumulated in the tissue of both embryo and larvae, corresponding to the amount exposed in the sediment or elutriates, leading to developmental abnormalities. Cr and Ni were not addressed in these studies.

Although an exhaustive review of the literature can be compiled, the best way of determining the rate and amount of bioaccumulation of metals and other elements into aquatic biota, ultimately, would be to examine it directly. The various characteristics of phosphatic clays under the environmental conditions in the Gulf will greatly influence what elements will be available and accumulated. Factors such as grain size and mineralogy, reactive compounds and complexing agents on the clay surface, the mode of exposure (i.e., direct ingestion or through the liquid phase), and most especially, the biological/ physiological characteristics of the organisms (e.g., gut extraction, assimilation efficiencies) will all be important in gauging the effect of each chemical constituent. In our opinion, such studies will be critical and must be engaged if and when this method approaches the actual implementation.

Turning to radionuclides, the amounts of 266Ra, and 238U released by phosphatic clay are lower than those listed in U.S.EPA effluent standards. Therefore, these two elements may not be a cause for concern with regards to release in natural waters. However, 210Pb levels in phosphatic clay samples from Nu-Gulf were lower than the effluent standards, while those from IMC Phosphates and FIPR were above the standard by a factor of 2 or 3. As with the outcome of the trace metals analysis, it is unknown how these variable results will influence the decision to use phosphatic clays in further work on clay mitigation. We recommend additional impact studies and bioassays to investigate the implications of certain radionuclide release at expected levels of clay addition.

Lastly, we quantified the amount of four inorganic nutrients released or absorbed by clays, and in some cases, clay treated with polyaluminum chloride (PAC). The bentonite, WB-B, significantly added nitrate and silicate to the system. By contrast, the cationic polymer-treated kaolinite, H-DP, did not add any nutrient to the system and was even able to remove phosphate and nitrate when combined with PAC. On the other hand, Florida phosphatic clay did not add nitrate, ammonium and silicate to the surrounding medium, but added, as expected, a significant amount of phosphate. However, the amount of released phosphate was mitigated by combining the clay with PAC. This method of using PAC may be important should phosphatic clay be considered in future research and field applications.

In theory, the concern is that this excess phosphate can eventually counteract the removal of red tide through flocculation, by the stimulation of a new harmful bloom. However, this is not necessarily the only possible outcome. While “new” phosphate can certainly lead to bloom formation, other essential nutrients must also be available at sufficient amounts, including nitrate/nitrite, ammonium, silicate, and a number of trace elements and organic compounds. Ultimately, the important predictor is the relative ratios of these elements in the water column, not just the amount of each nutrient in bulk

67

water. Our main point here is that the triggers for algal blooms are complex and synergistic, encompassing elements of the physical, chemical and biological system. For example, should an algal bloom result it may not be that of a harmful species, but another member of the phytoplankton assemblage. Bloom formation may also be affected by the type and number of animal grazers, the availability of light, and local hydrodynamic and meteorologic processes. We must, therefore, investigate the impact of nutrient release and the likelihood of bloom formation in controlled experiments in the field, such as those in mesocosms and limnocorrals. Only by gaining an appreciation of the local system can we hope to ascertain how nutrient addition, specifically, phosphate addition, can affect the planktonic community. Behavior of Clay Flocs on the Ocean Floor

Results from the racetrack flume experiments documented high (>80%) removal efficiencies of algae in low flow conditions (≤ 10 cm s-1). Results from both sets of flume experiments indicated that consolidation, or dewatering, of a floc layer on the bed over time increased the critical shear stress (τ0crit) for resuspension (i.e., decreased erodibility). The addition of PAC increased the erodibility of the floc layer (i.e., resuspension occurred at lower flow speeds). Overall, we do not advise the addition of PAC in an attempt to reduce the clay loading in future field applications. Though our experiments should be repeated for more confidence, it appeared that PAC decreased the removal efficiency of algal cells in flow. Also, by increasing the likelihood that flocs remain in suspension, the addition of PAC would have greater impact on sensitive, filter-feeding organisms. Oxygen Demand

The addition of phosphatic clay to the limnocorrals increased total settled solids (TSS) by 1426%, volatile settled solids (VSS) by 527%, and the biological oxygen demand (BOD) rate by 31%. The additional amount of inorganic clays reduced the relative percentage of VSS from ~26% of the TSS, to 10% of the TSS. In effect, the organic material in the surface sediments was reduced by 62%. This has implications for benthic organisms that feed on surface sediment with respect to the relative nutritional value of the sediment.

ENGINEERING REQUIREMENTS AND TREATMENT COSTS

In addition to considering its efficacy and environmental impacts, the decision to implement a clay control program must also consider the feasibility and cost of mounting an effort in larger, more complex and dynamic systems. Unfortunately, there is very little information in the literature regarding cost and specific project designs from earlier Japanese and Korean experiments. While some additional technical information has been obtained from personal communications with Korean colleagues, they have not provided

68

sufficient detail to replicate their experiments locally. Therefore, it has been necessary to develop a “conceptual” treatment protocol for the United States guided by previous research and experience, in order to identify the practical and engineering requirements at each step of a large scale treatment. The following sections present the major steps involved in planning and executing a possible field program. Some of these ideas were generated through discussions with engineers at the JAD Enterprises (Macon, GA, USA) and Grain Processing Corporation (Muscatine, IA, USA). Wherever possible, cost estimates are provided. Phosphatic Clay Acquisition and Transport to Treatment Site

Our first task was to calculate the required amount of clay for the operation. Clearly, the amount will be dictated by various factors including the treatment scale (i.e., volume or surface area), loading rate, and the number of repetitions. The treatment scale will depend on the aerial and depth distribution of the bloom which can be determined by aerial and shipboard surveys. The target area can be further defined by the presence of priority sites (e.g., aquaculture sites, marina, public beaches). Once a treatment “volume” is calculated from the surface area and depth distribution of the bloom, this value can then be multiplied by the “optimal” loading rate (in mass per unit volume) as defined by empirical trials.

Our first consideration was the minimum amount of clay needed to remove Karenia brevis effectively (i.e., > 80 % removal efficiency), at algal densities found in nature. The maximum concentration was limited by concerns of environmental and ecological impacts. We chose a concentration range between 0.10 and 0.50 g/L. To maximize cell removal and minimize the escape of cells from the flow, we chose 0.25 g/L as our working concentration. Removal studies in 2-m water columns showed that removal efficiency was evenly maintained throughout the column with this concentration. Furthermore, 0.25 g/L successfully removed up to 77% of the Karenia brevis captured in a mesocosm during a field study.

According to our experimental results, the depth (or volume) of the water column being treated must be considered in determining clay loading to ensure high cell removal. Conversely, considering only the surface dimensions (i.e., m2) does not take into account the dilution of the slurry as it sinks, leading to the steady decline in the effective clay concentration as the plume sinks. As a result of the latter, cell removal occurs only near the surface, leaving the cell below unaffected. In our calculations of clay concentrations for field application, we typically use a depth of 3 m as we expect this to be a suitable estimate of where most of the red tide would be in the water column. In Table 12, we presented our estimates of clay solids needed (in dry amount, metric tons) at various loading rates, treatment scales and frequency of treatment. We expect that multiple treatments would be necessary. In Korea, it is not uncommon to prescribe two or three dispersals over a given area.

69

Table 12. Cost Analysis of Red Tide Treatment Using Phosphatic Clays: Estimated Amount of Clay Required and Freight Charges.

Note: (A) The amount of clay needed (in dry weight in metric tons) was estimated for various treatment scales, loading rates and repetitions. The treatment volume = surface area (m) x 3 m depth, which is an estimate for the depth of occurrence for the bloom. (B) transport cost for each corresponding clay amount using quotes from American freight company. The cost of truck rental to transport material between Bartow and Sarasota or Charlotte Harbor was $799 (25 February 2003). Each truck has a limit of 45,000 lbs (3,500 ft2, 9 ft tall).

To determine the maximum amount of phosphatic clay we might expect to settle

in a given area, the dry weight of clay from Table 12 can be divided by the corresponding surface area of treatment, assuming no lateral transport away from the area. For example, a 100 m2 area (with 3 m depth) treated once with clay at 0.25 g/L would have 8 g/m2 of

A) Amount of dry clay needed

Surface Surface Treatment Surface Surface Loading rate = 0.10 g/L Loading rate = 0.25 g/L Loading rate = 0.50 g/Larea area volume area area 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments(m2) (km2) (L) (acres) (mi2) (m.t. needed) (m.t. needed) (m.t. needed) (m.t. needed) (m.t. needed) (m.t. needed) (m.t. needed) (m.t. needed) (m.t. needed)

100 0.0001 300,000 0.025 0.00004 0.03 0.06 0.09 0.08 0.15 0.23 0.15 0.30 0.452,500 0.0025 7,500,000 0.618 0.0010 0.75 1.50 2.25 1.88 3.75 5.63 3.75 7.50 11.254,047 0.0040 12,140,834 1 0.0016 1 2 4 3 6 9 6 12 18

10,000 0.010 30,000,000 2.5 0.0039 3 6 9 8 15 23 15 30 4540,469 0.040 121,408,337 10 0.016 12 24 36 30 61 91 61 121 182

250,000 0.25 750,000,000 62 0.097 75 150 225 188 375 563 375 750 1,125404,694 0.40 1,214,083,367 100 0.16 121 243 364 304 607 911 607 1,214 1,821

1,000,000 1 3,000,000,000 247 0.39 300 600 900 750 1,500 2,250 1,500 3,000 4,5002,590,000 2.6 7,770,000,000 640 1 777 1,554 2,331 1,943 3,885 5,828 3,885 7,770 11,655

12,950,000 13 38,850,000,000 3200 5 3,885 7,770 11,655 9,713 19,425 29,138 19,425 38,850 58,27525,000,000 25 75,000,000,000 6178 10 7,500 15,000 22,500 18,750 37,500 56,250 37,500 75,000 112,500

(acres) (mi2) 3% clay solid content 0.025 0.00004 1 2 3 3 5 8 5 10 15

0.618 0.0010 25 50 75 63 125 188 125 250 3751 0.0016 40 81 121 101 202 304 202 405 607

2.5 0.0039 100 200 300 250 500 750 500 1,000 1,50010 0.016 405 809 1,214 1,012 2,023 3,035 2,023 4,047 6,07062 0.097 2,500 5,000 7,500 6,250 12,500 18,750 12,500 25,000 37,500

100 0.16 4,047 8,094 12,141 10,117 20,235 30,352 20,235 40,469 60,704247 0.39 10,000 20,000 30,000 25,000 50,000 75,000 50,000 100,000 150,000640 1 25,900 51,800 77,700 64,750 129,500 194,250 129,500 259,000 388,500

3200 5 129,500 259,000 388,500 323,750 647,500 971,250 647,500 1,295,000 1,942,5006178 10 250,000 500,000 750,000 625,000 1,250,000 1,875,000 1,250,000 2,500,000 3,750,000 B)

Cost of Shipping clay from Bartow to Sarasota or Charlotte Harbor

Surface Surface Loading rate = 0.10 g/L Loading rate = 0.25 g/L Loading rate = 0.50 g/Larea area 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments

(acres) (mi2) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $)0.025 0.00004 799 799 799 799 799 799 799 799 7990.618 0.0010 799 799 799 799 799 799 799 799 799

1 0.0016 799 799 799 799 799 799 799 799 7992.5 0.0039 799 799 799 799 799 1,598 799 1,598 2,39710 0.016 799 1,598 1,598 1,598 2,397 3,995 2,397 4,794 7,19162 0.097 3,196 6,392 9,588 7,990 15,181 22,372 15,181 29,563 44,744

100 0.16 4,794 9,588 14,382 11,985 23,970 35,955 23,970 47,940 55,930247 0.39 11,985 23,970 35,955 29,563 59,126 88,689 59,126 115,855 176,579640 1 30,362 61,523 91,885 76,704 152,609 228,514 152,609 304,419 456,229

3,200 5 152,609 304,419 456,229 380,324 760,648 1,140,972 760,648 1,520,497 2,281,1456,178 10 293,233 587,265 880,498 734,281 1,467,763 2,201,245 1,467,763 2,935,526 4,402,490

(acres) (mi2) 3% clay solid content0.025 0.00004 799 799 799 799 799 799 799 799 7990.618 0.0010 1,598 2,397 3,196 2,397 5,593 7,990 5,593 10,387 15,181

1 0.0016 1,598 3,196 4,794 3,995 7,990 11,985 7,990 15,980 23,9702.5 0.0039 3,995 7,990 11,985 10,387 19,975 29,563 19,975 39,151 59,12610 0.016 15,980 31,960 47,940 39,950 79,900 119,051 79,900 159,001 238,10262 0.097 98,277 195,755 294,032 245,293 489,787 734,281 489,787 978,775 1,467,763

100 0.16 159,001 317,203 475,405 396,304 791,809 1,188,113 791,809 1,584,417 2,376,226247 0.39 391,510 783,020 1,174,530 978,775 1,956,751 2,935,526 1,956,751 3,913,502 5,870,253640 1 1,013,931 2,027,063 3,040,994 2,534,428 5,068,057 7,601,686 5,068,057 10,136,114 15,203,372

3200 5 5,068,057 10,136,114 15,203,372 12,669,743 25,339,486 38,008,430 25,339,486 50,678,173 76,016,8606178 10 9,783,755 19,566,711 29,350,466 24,458,988 48,917,177 73,375,366 48,917,177 97,833,555 146,749,933

70

clay. However, we are not likely to find a completely static water column, without currents or motion. Therefore, the clay slurry added to the surface would eventually spread and become diluted as it sinks, which lowers the actual loading rate per surface area. Moreover, the deposition rates and persistence of this fine materials along the sediment will be determined by the bottom boundary conditions and local hydrodynamics in the area. Lastly, clays and other sediments from outside the treated area may also add to the layer over time. At this point, we can only make some estimates of the amount of clay being added to the bottom based on some simple calculations.

IMC Phosphates Company, one of the largest phosphate producers in Florida, has

offered us their clays at no cost (J. Keating, personal communication). However, costs will be associated with labor and transportation to the treatment site, most likely along the Gulf coast of Florida. According to the American Freight Company, which represents a consortium of freight truck companies in Florida, the cost of hiring a truck to move clay between Bartow and Sarasota or Charlotte Harbor is currently set at $799 (25 February 2003). Each truck has a weight capacity of 45,000 lbs with a 3,500-ft2 cargo space and 9-ft clearance. Using this information, the cost of transporting the predicted amount of clays from source to treatment site can be estimated. First, the number of trucks was determined by dividing the mass of clay to be transported in Table 12A, by the capacity of each truck. Then, the total cost of shipping the material by truck was calculated by simply multiplying the number of trucks needed by the cost of each hiring (see Table 12B). As we can see, the cost of treatment increases linearly with the number of repetitions called for at each loading rate. The treatment cost also grows linearly with increasing clay concentrations. Finally, there is a clear cost advantage to securing and using clay samples with higher solids content in terms of shipping cost. For example, the cost of shipping enough dry phosphatic clay to treat a 62-acre area three times with 0.50 g/L each, comes to about $45,000 (Table 12B). In the same table, the cost of shipping enough wet phosphatic clay at 3% solids, for the same treatment scale, loading and frequency, would come to $1.5 million. At higher scales of treatment (e.g., 1 to 5 sq. mi), the cost of shipping alone approaches potentially prohibitive values.

Phosphatic clays, sometimes referred to as phosphate slimes, are materials removed from the phosphate matrix through a process called beneficiation. The solid content of the slurry leaving the beneficiation plant can be low (only up to 3% by weight) (Table 1). The clays are then stored behind earthen dams where they settle and dry. Depending on the age of the pond, the solid content of this material can vary widely. The highest solid content encountered in this study was 70% (IMC-P5). Therefore, the actual quantity of clay to be collected and transported for the treatment reported in Table 12A will increase depending on the solid content or wet weight of the source clay. Since all of the phosphatic clays tested in this study displayed comparable removal abilities regardless of source or solids content (Figure 5), it seems more economical to prioritize sites bearing clays with higher solid content, since more clay mass will be collected and transported per unit effort and cost. Moreover, clays with higher solid content (>30%), which are in a semi-solid state, can be dug, packaged, moved, and stored more readily. By contrast, watery slurries would require water-tight containment from collection to

71

actual dispersal, pumping equipment, and chemical treatment to prevent fouling during long-term storage.

In addition to collecting the clay from its source and its transport to the treatment sites, we considered a number of intermediate steps that could affect the cost of the operation, and expedite steps like clay preparation and dispersal. Since we anticipate that the cost of collecting and transporting the clay can be maximized by selecting samples with the highest solid content, we explored the possibility of further drying the clay at the collection site or at some processing facility before delivery to the treatment site. Apparently, the technology exists to allow “flash drying” of wet clay to a desired moisture content (B. Dahlquist, JAD Enterprises, personal communication). This procedure may allow us to deliver more actual clay solids. However, we do not know at this time whether such a procedure is feasible or cost-effective given the anticipated clay demand (Table 11). In addition, as we explain in the following section, the clay’s dryness can significantly affect its dispersal in water, which can affect its removal ability. Therefore, it will be necessary to investigate this drying procedure further in order to assess its effectiveness, practicality, and cost to the operation.

Based on our experience working with larger volumes (e.g., in flumes or field mesocosms), clay slurry preparation beginning with rock-sized clumps of clay was a tedious, labor-intensive, rate-limiting step which required soaking the clay for several hours, manual and mechanical breakage of softened clay, blending, vigorous mixing and sieving. Commercially available machines for making clay slurries, such as those from Korea or the U.S., call for finely sized clays. Therefore, another crucial step between collection and delivery is the crushing or pulverization of the clays to make it suitable for rapid slurry preparation as described in the following section. Again, the technology for this process is widely available, although it may be necessary to have this task done by a clay company. This step may also be coupled with the flash drying process described previously (B. Dahlquist, JAD Enterprises, personal communication). This is another area that we are currently investigating with regards to practicality and cost. However, this may be one step in the operation that may not be as readily avoided or disregarded as the drying procedure. Preparation and Dispersal of Phosphatic Clay Slurry

The next set of engineering challenges involves the efficient processing of phosphatic clay for shipboard dispersal. We first considered the addition of dry clay powder to the surface of the bloom as this appeared to be the easiest, fastest, least labor-intensive and most cost-effective method. Although the intermediate step of preparing a clay slurry would be omitted, it would still be necessary to grind the clay and find a means of dispersing it over the sea surface. We investigated this in the laboratory by adding dried, crushed and sifted clay powder to the surface of the cell culture (data not shown). Cell removal efficiency was only 20% at 0.25 g/L, and <40% at 0.50 g/L, for both IMC-P1 and IMC-P2, compared to >90% for their respective slurries. We observed clumping of the clay powder at the surface, resulting in large, visible aggregates that fell

72

quickly through the medium. We also found that the dried and crushed powder was difficult to disperse in either freshwater or seawater, even with prolonged soaking and agitation over several days. It appeared that when the mineral crystals are collapsed by dehydration, it becomes difficult to reintroduce water into the lattices. Hence, the removal efficiencies for these “reconstituted” clays were approximately 60% at 0.25 g/L and 85% at 0.50 g/L (data not shown). While these results were higher than using dried powder, these were still lower than diluted clay slurries at the same concentrations. Therefore, both of these findings strongly suggest that using fully dried and crushed phosphatic clay directly or in suspension would not be suitable for this treatment effort. Incidentally, phosphatic clay at 70% solids content (e.g., IMC-P7) softened adequately after several hours of soaking in water, and produced a fine slurry with sufficient mechanical processing (e.g., blending, rotor mixing). More importantly, the removal ability of the clay was comparable to clays with higher moisture levels. Therefore, there may be a dryness limit at which phosphatic clays would remain effective and workable.

To maximize its effectiveness, we concluded that phosphatic clay must be dispersed in a slurry form, similar to the methods used in Japan (Shirota 1989) and South Korea (Bae and others 1998). In our initial experience working in larger volumes (e.g., flumes and mesocosms), slurry preparation and dispersal involved two separate, independent steps. Slurry preparation began by soaking and softening relatively dry clumps of clay (i.e., between 46-72% solids) in seawater. The clumps consisted of large, rock-like pieces taken directly from settling ponds. After several hours, the clumps could be broken by hand, and then processed using a blender. Softened clumps were also processed into slurries placing them in 20-gal buckets with seawater and immersing a rotor attached to a power drill. The final step involved sieving the slurry to remove plant debris and large, unbroken pieces. In all, the preparation took several hours. For dispersing the clay slurry over the sea surface, we used various submersible pumps and power washers attached to a hose and adjustable spray nozzle. Constant mixing by hand was necessary to prevent settling of the clay. Clogging of the pumps and nozzles sometimes occurred. One advantage of this system is the low cost of the various dispersal equipment. However, the labor cost involved in slurry preparation would more than offset the savings from the dispersal equipment.

Presently, we are aware of several machines that can combine slurry preparation and dispersal. Clay suspensions can be prepared rapidly and with greater control over particle size and concentration, as well as the rate of dispersal. In South Korea, a seawater slurry from dry yellow loess is made using a specially-designed machine (Figure 19A). Raw seawater is pumped in and forced through angled channels along the walls of the mixing chamber to generate a vortex (Figure 19B). Dry loess is then added to the mixing chamber. The finished slurry containing particles less than 50 µm flows into a second chamber through an overflow slot at the top of the mixing chamber. It is then pumped directly over the sea surface from boats or barges (Figure 19C and 19D; photos provided by Dr. HG. Kim, C.K. Lee and H.M. Bae from the National Fisheries Research and Development Institute, Busan, South Korea). The greatest advantage of this design is the incorporation of slurry preparation and dispersal.

73

Figure 19. Clay Dispersal Machine from South Korea. (A) Side View, (B) Details, (C) Clay Dispersal in and around Fish Pens, (D) Clay Dispersal from Barge.

A

C D

B

74

Unfortunately, there is no information available on the specifications of this device (e.g., capacity, mixing rate, dispersal rate), or on the exact procedures and quantities such as the clay-to-seawater mixing ratios. The Korean manufacturer offers an iron model at $4,300 and a stainless steel model at $10,800. The shipping cost from Korea to Florida was estimated at $1,000 excluding the tariff. The latest model of this machine includes a chamber where seawater is hydrolyzed by passing a current through it to produce a short-lived sodium oxychloride species (NaOCl) (H.G. Kim, personal communication). Loess is then added to the hydrolyzed water to produce the slurry bearing the cell-killing oxychloride species. The result is higher bloom mortality with less loess. This technology has not been tested for U.S. species. It is unknown whether this newer version can be purchased.

In the U.S., a device is available from Grain Processing Corporation (Muscatine, IA) that can also prepare clay suspensions rapidly and precisely. Dispersal is also easily controlled. The machine can be hired on a per day basis, while the cost of shipping may be more reasonable than the cost of shipping a new sprayer from Korea. Both this and the Korean machine are about the same size. They also require that the clays be of a certain small size that can be wetted relatively quickly. While our contacts at JAD Enterprises have offered to design a machine specifically for our purposes, it may be more advantageous and cost-effective to purchase or rent an existing machine, instead of producing one de novo. Obtaining a pre-existing device also means that the product could be field tested and optimized more rapidly.

Typically, clay has only been added directly and systematically to the surface of

the water column in previous reports and studies. The slurry is diluted as the particles flocculate and sink. We do not have any empirical evidence on how different dispersal methods can affect the overall efficacy of the clay treatment. For example, subsurface addition and/or pulsed addition may enhance removal by preventing excessive clay addition over a given area. Turbulent addition (i.e., injecting the clay to the water surface with a vigorous, more powerful flow) may also be considered in order to maximize the collision frequency between particles near the surface. Whatever method is used, some attention must be given to the effective concentration (or the rate of dispersal) in a given parcel of water and the clay particle size entering the system. These will be the focus of upcoming field trials.

At present, we are unaware of any efforts to apply polyaluminum chloride (PAC)

or other flocculants in the field, in conjunction with clay dispersal. Currently, most reports on the possible benefits of PAC-treatment have been restricted to laboratory trials and mesocosm experiments. At WHOI, some of our experiments in settling columns demonstrated that scaling-up of the PAC loading and treatment did not produce the desired effect (data not shown). When the clays were pretreated with PAC before addition to the column, we observed that the clays immediately formed large, baseball-sized aggregates which immediately sank to the bottom, leaving behind most of the cells in suspension. To avoid this phenomenon, we subsequently added the PAC separately from the clay. Based on our preliminary findings, the best results were found when PAC was added to the seawater first to allow the flocculant to dilute and diffuse through the

75

surface, which was followed by the clay slurry after a 20 min delay. While these latter experiments have shown promising results at these larger scales, more effort must be given to testing the dispersal methods at field treatment scale. Ship and Crew Hiring

A significant portion of this operation, such as slurry preparation and dispersal, will be performed onboard ships and by the ship’s crew. One way of minimizing the cost is to involve volunteers, including those who wish to use their private vessels for the purpose (e.g., local property owners, fishermen, aquaculture). Otherwise, some current ship hiring costs were provided by our colleagues at Mote Marine Laboratory (Sarasota, FL). Their largest research vessel (46 ft) with A frame and full complement of equipment would cost about $1,100 per day. A smaller craft (27 ft) which has the necessary working area for clay dispersal and sample processing would cost $600 per day. Finally, the smallest vessel that would be appropriate for clay dispersal would cost about $300. This craft would have to unload the dispersal apparatus and return with sampling gear. Depending on the size and power, pleasure crafts can be rented in the area for prices ranging from $600 to a few thousand dollars. Other considerations on the choice and number of ships will be the water depth at the treatment site, the surface area to be treated, the size and complexity of the dispersal apparatus or system, and the size of the crew.

Two shipping companies in the Florida region were contacted to inquire about barges and other vessels for clay transport and dispersal. With respect to barges, the Cashman Equipment Corporation provided the following examples (estimates as of 10 April 2003):

(a) 220” x 60” x 14” (L x W x H), with a 3500 ton capacity, at $1000 - $1400 per day (b) 250” x 72” x 16”, 5500 ton capacity, at $1600 - $1800 per day (c) 300” x 100”x 18”, 10,000 ton capacity, at $4500 per day

These costs do not include the cost of renting a tugboat or offshore supply vessel

(OSV) to tow the barges. In addition, this company and others like it keep their fleet based in Louisiana, and would take between 8-10 days to tow the barges to Florida. The barges offer large working areas and cargo room, but may present difficulties in stability and movement from site to site. Hence, this option is less appealing.

Ted Brown Marine Services in Pensacola, FL, proposed that OSV’s would be the most suitable vessels for this purpose. OSV’s are the choice in transporting both dry materials and wet drilling muds to drilling platforms in the Gulf. They often have large storage tanks to keep wet clays. They are also self-powered with plenty of cargo and work space. Their speed and stability offer suitable conditions for working. Most companies that rent OSV’s are also located in LA, but a few are found in Tampa and the Gulf coast of Florida. The cost estimates below include rental and crew who also have

76

familiarity with navigating the shallower, coastal regions where the treatment would likely occur. Therefore, the use of the OSV may be the best option at the moment. One estimate provided by the company is the following: 150”-170” with 500-1,000 ton capacity, at $5,000-$10,000 per day. Using these values, the cost of ship hiring was calculated. Assuming that each ship can transport up to 1,000 tons (or 907.2 metric tons), the number of ships needed to disperse the required amount of clays in Table 12A was calculated. Then, the number of ships to hire was multiplied by the average cost of OSV rental (i.e., averaged at $7,500 per day) to get the total (Table 13). (Values provided by Ted Brown Marine Services in Pensacola, FL: 150”-170” with 500-1,000 ton capacity, at $5000-$10,000 per day. The number of ships needed for each treatment scale was calculated using the clay amounts in Table 12A. Then, the number of ships was multiplied by the cost of hiring each at the average cost of $7,500 per day.)

Table 13. Cost of Hiring Offshore Supply Vessels (OSVs) for Clay Transport and

Dispersal.

As with the cost of transport by truck, the cost of ship hiring and dispersal also increases linearly with both the frequency of treatment at one loading rate, and with increasing loading rates. For example, the cost of treating between 60 and 100 acres using dry clay, with respect to ship rental, is less than $23,000, even with three treatments at 0.50 g/L each. However, using wet clay at 3% solids, we can expect to spend this

Surface Surface Loading rate = 0.10 g/L Loading rate = 0.25 g/L Loading rate = 0.50 g/Larea area 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments

(acres) (mi2) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $)0.025 0.00004 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,5000.618 0.0010 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500

1 0.0016 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,5002.5 0.0039 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,50010 0.016 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,50062 0.097 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 15,000

100 0.16 7,500 7,500 7,500 7,500 7,500 7,500 7,500 15,000 22,500247 0.39 7,500 7,500 7,500 7,500 15,000 22,500 15,000 30,000 37,500640 1 7,500 15,000 22,500 22,500 37,500 52,500 37,500 67,500 97,500

3200 5 37,500 67,500 97,500 82,500 165,000 247,500 165,000 322,500 487,5006178 10 67,500 127,500 187,500 157,500 315,000 465,000 315,000 622,500 937,500

(acres) (mi2) 3% clay solid content0.025 0.00004 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,5000.618 0.0010 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500

1 0.0016 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,500 7,5002.5 0.0039 7,500 7,500 7,500 7,500 7,500 7,500 7,500 15,000 15,00010 0.016 7,500 7,500 15,000 15,000 22,500 30,000 22,500 37,500 52,50062 0.097 22,500 45,000 67,500 52,500 105,000 157,500 105,000 210,000 315,000

100 0.16 37,500 67,500 105,000 90,000 172,500 255,000 172,500 337,500 502,500247 0.39 90,000 172,500 255,000 210,000 420,000 622,500 420,000 832,500 1,245,000640 1 217,500 435,000 645,000 540,000 1,072,500 1,612,500 1,072,500 2,145,000 3,217,500

3200 5 1,072,500 2,145,000 3,217,500 2,677,500 5,355,000 8,032,500 5,355,000 10,710,000 16,065,0006178 10 2,070,000 4,140,000 6,202,500 5,167,500 10,335,000 15,502,500 10,335,000 20,670,000 31,005,000

77

amount on OSV rental when treating only 10 acres twice with 0.25 g/L of phosphatic clay. There may be some time and effort in using wet clays due to lowered demands in processing the clays for dispersal, but other logistical and practical considerations will also be accounted for, such as the need to move large volumes of slurry from truck containers to the OSV’s holding tanks, and storage of wet clays.

Finally, the values in Tables 12B and 13 were combined to provide an estimate for the cost of clay treatment at various scales (Table 14). These values were obtained by adding the estimates from Table 12B (truck transport from central Florida to coast) and Table 13 (OSV rental for clay dispersal). The one time cost of the dispersal machine or the development of a suitable system de novo was not included. In addition, the cost of outside labor was not estimated at this time. As before, using dry clay instead of wet slurry will be more cost effective. Applying wet clay can be anywhere from 7 to 30 times the cost of working with dry clay. Depending on the scale, the combined costs can be as little at $9,000 to tens of thousands of dollars when the treatment scale grows to between 60 and 100 acres (with dry clay). Above 100 acres, the costs climb to a few hundred thousand dollars. Finally, the cost of treating a 10 sq. mile area with at least three doses of dry clay at 0.50 g/L is conservatively estimated at over $5 million. Ultimately, the decision on cost effectiveness will be based on a comparison between clay dispersal and the estimated losses of leaving a bloom untreated. We also expect that considerable cost savings are possible over the rough estimates given here. Table 14. Total Cost of Treatment.

Surface Surface Loading rate = 0.10 g/L Loading rate = 0.25 g/L Loading rate = 0.50 g/Larea area 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments 1 treatment 2 treatments 3 treatments

(acres) (mi2) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $) (cost $)0.025 0.00004 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,2990.618 0.0010 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,299

1 0.0016 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,2992.5 0.0039 8,299 8,299 8,299 8,299 8,299 9,098 8,299 9,098 9,89710 0.016 8,299 9,098 9,098 9,098 9,897 11,495 9,897 12,294 14,69162 0.097 10,696 13,892 17,088 15,490 22,681 29,872 22,681 37,063 59,744

100 0.16 12,294 17,088 21,882 19,485 31,470 43,455 31,470 62,940 78,430247 0.39 19,485 31,470 43,455 37,063 74,126 111,189 74,126 145,855 214,079640 1 37,862 76,523 114,385 99,204 190,109 281,014 190,109 371,919 553,729

3200 5 190,109 371,919 553,729 462,824 925,648 1,388,472 925,648 1,842,997 2,768,6456178 10 360,733 714,765 1,067,998 891,781 1,782,763 2,666,245 1,782,763 3,558,026 5,339,990

(acres) (mi2) 3% clay solid content0.025 0.00004 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,299 8,2990.618 0.0010 9,098 9,897 10,696 9,897 13,093 15,490 13,093 17,887 22,681

1 0.0016 9,098 10,696 12,294 11,495 15,490 19,485 15,490 23,480 31,4702.5 0.0039 11,495 15,490 19,485 17,887 27,475 37,063 27,475 54,151 74,12610 0.016 23,480 39,460 62,940 54,950 102,400 149,051 102,400 196,501 290,60262 0.097 120,777 240,755 361,532 297,793 594,787 891,781 594,787 1,188,775 1,782,763

100 0.16 196,501 384,703 580,405 486,304 964,309 1,443,113 964,309 1,921,917 2,878,726247 0.39 481,510 955,520 1,429,530 1,188,775 2,376,751 3,558,026 2,376,751 4,746,002 7,115,253640 1 1,231,431 2,462,063 3,685,994 3,074,428 6,140,557 9,214,186 6,140,557 12,281,114 18,420,872

3200 5 6,140,557 12,281,114 18,420,872 15,347,243 30,694,486 46,040,930 30,694,486 61,388,173 92,081,8606178 10 11,853,755 23,706,711 35,552,966 29,626,488 59,252,177 88,877,866 59,252,177 118,503,555 177,754,933

78

Using estimates provided by several clay companies on their bentonite products, which we have shown to have comparable removal ability to phosphatic clays, the cost of using phosphatic clays (e.g., shipping) will be much less than the cost of using these other clays (e.g., cost of purchasing the clay and shipping it from the Midwest to Florida with trucks or rail service). Although a number of these companies are based in states closer to Florida, the bentonite clays they sell would have to originate from states like Wyoming where bentonites are mined primarily, adding significantly to the shipping costs. Certainly, the use of these dried clays would seem to have an advantage over the handling of wet phosphatic clays. However, if the proper source of phosphatic clays can be located and tapped (i.e., the concentration of clays needed for dispersal), then time can be saved during the preparation step. Dried clay would also have to be handled and turned to slurry before use. Wet clays can be pumped into the tanker trucks and the holding tanks of the appropriate vessels. At this point, we still recommend phosphatic clays over dry clays with regards to cost.

79

REFERENCES Alldredge AL, Silver MW. 1988. Characteristics, dynamics and significance of marine snow. Progress in Oceanography 20: 41-82. Altschuler ZS. 1980. The geochemistry of trace elements in marine phosphorites. In: Bentor YK, editor. Marine Phosphorites; SEPM Spec. Pub. No. 29, Tulsa, OK. p 19-30. Anderson DM. 1997. Turning back the harmful red tide. Nature 388: 513-4. Anderson DM, Kulis DM, Sullivan JJ, Hall S, Lee C. 1990. Dynamics and physiology of saxitoxin production by the dinoflagellates Alexandrium spp. Marine Biology 104: 511-24. Anderson DM, Kulis DM, Keafer BA, Berdalet B. 1999. Detection of the toxic dinoflagellate Alexandrium fundyense (Dinophyceae) with oligonucleotide and antibody probes: variability in labelling intensity with physiological condition. Journal of Phycology 35: 870-83. Archambault MC, Bricelj VM, Grant J, Anderson DM. 2004. Effects of suspended and sedimented clays on juvenile hard clams, Mercenaria mercenaria, within the context of harmful algal bloom mitigation. Mar. Biol. 144: 553-65. Avnimelech Y, Troeger BW, Reed LW. 1982. Mutual flocculation of algae and clay: evidence and implication. Science 216(4541): 63-5. Bae HM, Choi HG, Lee WC, Yoon SJ. 1998. Control of the red tide by yellow loess dispersion. In: Lee HK, Lee SG and Lee CK, editors. Proceedings of the Korea-China Joint Symposium on Harmful Algal Blooms; National Fisheries Research and Development Institute, Pusan, South Korea, p 53-60. Beaulieu SE, Sengco MR, Anderson DM. Forthcoming. Using clay to control harmful algal blooms: deposition and resuspension of clay/algal flocs. Harmful Algae. Beaulieu SE. 2003. Resuspension of phytodetritus from the sea floor: a laboratory flume study. Limnology and Oceanography 48(3): 1235-44. Boesch DF, Anderson DM, Horner RA, Shumway SE, Tester PA, Whiteledge TE. 1997. Harmful algal blooms in coastal waters: options for prevention, control and mitigation. Silver Spring (MD): U.S. Department of Commerce, National Oceanic and Atmospheric Administration, Coastal Ocean Office. NOAA Coastal Ocean Program Decision Analysis Series nr 10. Special joint report with the National Fish and Wildlife Foundation, Washington, DC. 49 p.

80

Choi HG, Kim PJ, Lee WC, Yun SJ, Kim HG, Lee HJ. 1998. Removal efficiency of Cochlodinium polykrikoides by yellow loess. Journal of the Korean Fisheries Society 31(1): 109-13. Choi HG, Lee PY, Yun SJ, Lee WC, Bae HM. 1999. Control of Cochlodinium polykrikoides blooms and adsorption of nutrients in the seawater by clay and yellow loess. Bulletin of the National Fisheries Research and Development Institute of Korea 57: 105-10. Elimelech M, Gregory J, Jia X, Williams R. 1995. Particle deposition and aggregation: measurement, modelling and simulation. Boston: Butterworth-Heinemann. 441 p. Fries JS, Trowbridge JH. 2003. Flume observations of enhanced fine particle deposition to permeable sediment beds. Limnology and Oceanography 48(2): 802-12. Geffard O, Geffard A, His E, Budzinski H. 2003. Assessment of the bioavailability and toxicity of sediment-associated polycyclic aromatic hydrocarbons and heavy metals applied to Crassotrea gigas embryos and larvae. Mar. Poll. Bull. 46: 481-90. Geissler U. 1958. Das membranpotential einiger diatomeen und seine bedeutung für die lebende kieselalganzelle. Mikroskopie 13: 145-72. Gregory J. 1987. Flocculation by polymers and polyelectrolytes. In: Ardos TF editor. Solid/liquid dispersions. London: Academic Press. p 163-81. Ives KJ. 1956. Electrokinetic phenomena of planktonic algae. In: Proceedings of the Society of Water Treatment Examiners 5: 41-58. Kim HG. 1997. Recent harmful algal blooms and mitigation strategies in Korea. Ocean Research (Seoul) 19: 185-92. Jackson GA, Lochmann SE. 1993. Modeling coagulation of algae in marine ecosystems. In: Buffle J, van Leeuwen HP, editors. Environmental particles: Vol. 2. Boca Raton: Lewis Publishers. p 387-414. Lachat Instruments. 1994. Ammonium in brackish or seawater--QuikChem Method 31-107-06-1-A. Lachat Instruments, 6645 West Mill Road, Milwaukee, WI 53218. Lachat Instruments. 1996. Silicate in brackish or seawater--QuikChem Method 31-114-27-1-B. Lachat Instruments, 6645 West Mill Road, Milwaukee, WI 53218. Lachat Instruments. 1998. Phosphate in brackish or seawater--QuikChem Method 31-115-01-3-A. Lachat Instruments, 6645 West Mill Road, Milwaukee, WI 53218.

81

Lachat Instruments. 1999. Nitrate and/or nitrite in brackish or seawater--QuikChem Method 31-107-04-1-A. Lachat Instruments, 6645 West Mill Road, Milwaukee, WI 53218. Leslie JF, Greasham RL, Hulbert MH. 1984. A cyanobacterial system for consolidation of phosphate slimes. Geomicrobiology Journal 3, 4: 343-58. Lewis MA, Dantin DD, Walker CC, Kurtz JC, Greene RM. 2003. Toxicity of clay flocculation of the toxic dinoflagellate, Karenia brevis, to estuarine invertebrates and fish. Harmful Algae 2(4): 235-46. Maruyama T, Yamada R, Usui K, Suzuki H, Yoshida T. 1987. Removal of marine red tide planktons with acid treated clay. Nissuishi/Bulletin Japanese Society of Science and Fisheries 53(10): 1811-9. MacDonald DD, Carr RS, Calder FD, Long ER, Ingersoll C. 1996. Development and evaluation of sediment quality guidelines for Florida coastal waters. Ecotoxicology 5(4): 253-78. Mountouris A, Voutsas E, Tassios D. 2002. Bioconcentration of heavy metals in aquatic environments: The importance of bioavailability. Mar. Poll. Bull. 44: 1136-41. Na GH, Choi WJ, Chun YY. 1996. A study on red tide control with loess suspension. Journal of Aquaculture 9: 239-45. O'Melia CR, Tiller CL. 1993. Physicochemical aggregation and deposition in aquatic environments. In: Buffle J, van Leeuwen HP, editors. Environmental particles: Vol. 2. Boca Raton: Lewis Publishers. p 353-86. PEDCo Environmental, Inc. 1981. Evaluation of best management practices for mining solid waste storage, disposal and treatment. Presurvey Study. Preliminary Draft. Cincinnati (OH): U.S. Environmental Protection Agency, Resource Extraction and Handling Division, Industrial Environmental Research Laboratory, Office of Research and Development. PEDCo Environmental, Inc. 1983. Evaluation of best management practices for mining solid waste storage, disposal and treatment. Presurvey Study. Revised. Cincinnati (OH): U.S. Environmental Protection Agency, Resource Extraction and Handling Division, Industrial Environmental Research Laboratory, Office of Research and Development. Rounsefell GA, Evans JE. 1958. Large-scale experimental test of copper sulfate as a control for the Florida red tide. U.S. Fish Wildlife Serv. Spec. Sci. Rep. 270. US. Department of the Interior, Fisheries and Wildlife Service, Washington, DC. 57 p.

82

Sengco MR, Li A, Tugend K, Kulis D, Anderson DM. 2001. Removal of red- and brown-tide cells using clay flocculation. I. Laboratory culture experiments with Gymnodinium breve and Aureococcus anophagefferens. Marine Ecology Progress Series 210: 41-53. Sengco MR. 2001. The aggregation of clay minerals and marine microalgal cells: physicochemical theory and implications for controlling harmful algal blooms [Dphil Thesis]. Woods Hole (MA): Massachusetts Institute of Technology/Woods Hole Oceanographic Institution Joint Program. Shirota A. 1989. Red tide problem and countermeasures (2). Int. J. Aq. Fish. Technol. 1: 195-293. Søballe DM, Threlkeld ST 1988. Algal-clay flocculation in turbid waters: variations due to algal and mineral differences. Verh. Internat. Verein. Limnol. 23: 750-4. Steidinger KA. 1983. A re-evaluation of toxic dinoflagellate biology and ecology. Progress in Phycological Research 2: 147-88. Stumm W, Morgan JJ. 1996. Aquatic chemistry: chemical equilibria and rates in natural waters. New York: John Wiley and Sons. Tester PA, Steidinger KA. 1997. Gymnodinium breve red tide blooms: initiation, transport and consequences of surface circulation. Limnology and Oceanography 42: 1039-51. Thomas DN, Judd SJ, Fawcett N. 1999. Flocculation modelling: a review. Water Research 33(7): 1579-92. van Benschoten JE, Edzwald JK. 1990. Chemical aspects of coagulation using aluminum salts. I. Hydrolytic reactions of alum and polyaluminum chloride. Wat. Res. 24(12): 1519-26. Yan Q-L, Wang W-X. 2002. Metal exposure and bioavailability to a marine deposit-feeding pipuncula, Sipunculus nudus. Environ. Sci. Technol. 36(1): 40-7. Yu Z, Zou JZ, Ma X. 1994a. A new method to improve the capability of clays for removing red tide organisms. Oceanologia et Limnologia Sinica 25: 226-32. Yu Z, Zou JZ, Ma X. 1994b. Application of clays to removal of red tide organisms. I. Coagulation of red tide organisms with clays. Chinese Journal of Oceanology and Limnology 12: 193-200. Yu Z, Zou JZ, Ma X. 1994c. Application of clays to removal of red tide organisms. II. Coagulation of different species of red tide organisms with montmorillonite and effect of clay pretreatment. Chinese Journal of Oceanology and Limnology 13: 62-70.

83

Yu Z, Zou JZ, Ma X. 1995. Application of clays to removal of red tide organisms. III. The coagulation of kaolin on red tide organisms. Chinese Journal of Oceanology and Limnology 12: 316-24.

A-1

Appendix A

FLUME EXPERIMENTS

A-2

Figure A-1. Flume Experiments.

Volume (l) DinoflagellateFlow speed above test bed cells per ml PAC concn. (ul/l) Wet mass (g) Dry mass (g) Concentration

Experiment Settling during Flume water Flume water (in fenced box (in fenced box (in fenced box clay including clay excluding Areal loading loading of clayDate Description of experiment run # time (hr) settling (cm/s) temp (C) salinity (ppt) or entire flume) or entire flume) or entire flume) organics organics clay (dry g/m^2) (dry g/l)

17-m flume16-Apr-99 modelling clay in test bed 1 n/a n/a 13 31 n/a n/a n/a n/a n/a n/a n/a19-Apr-99 coarse Nobska beach sand (500 - 1000 um) 1 n/a n/a 11.5 33 n/a n/a n/a n/a n/a n/a n/a21-Apr-99 medium Nobska beach sand (250 - 500 um) 1 n/a n/a 13 31 n/a n/a n/a n/a n/a n/a n/a21-Apr-99 IMC-P3 clay 1 3 0 13 31 3.2 n/a n/a nd nd nd nd24-Apr-99 fine Nobska beach sand (125 - 250 um) 1 n/a n/a 15 31 n/a n/a n/a n/a n/a n/a n/a25-Jul-99 IMC-P3 clay 1 3 0 22.5 32 3.2 n/a n/a 20 9.47 236.65 2.9625-Jul-99 IMC-P3 clay 1 3 0 22.5 32 3.2 n/a n/a 6 2.84 71.00 0.8926-Jul-99 IMC-P3 clay 1 3 0 23 33 3.2 n/a n/a 4 1.89 47.33 0.5928-Jul-99 IMC-P3 clay + PAC 1 3 0 23 32.5 3.2 n/a 14 4 1.89 47.33 0.5928-Jul-99 IMC-P3 clay + PAC 2 3 0 23 32.5 3.2 n/a 14 4 1.89 47.33 0.591-Aug-99 IMC-P3 clay + Gyrodinium 1 3 0 23 32 3.2 4.86E+01 n/a 4 1.89 47.33 0.591-Aug-99 IMC-P3 clay + Gyrodinium 2 3 0 23 32 3.2 4.93E+01 n/a 4 1.89 47.33 0.591-Jul-00 IMC-P4 clay + Heterocapsa 1 3 0 21.5 33 3.2 2.00E+03 n/a 2.8 1.33 33.13 0.411-Jul-00 IMC-P4 clay + Heterocapsa 2 3 0 21.5 33 3.2 2.00E+03 n/a 2.8 1.33 33.13 0.411-Jul-00 IMC-P4 clay + Heterocapsa 3 3 0 21.5 33 3.2 2.00E+03 n/a 2.8 1.33 33.13 0.412-Jul-00 IMC-P4 clay + Heterocapsa 1 9 0 22 33 3.2 1.89E+03 n/a 3.8 1.80 44.96 0.563-Jul-00 IMC-P4 clay + Heterocapsa 2 9 0 22 33 3.2 2.53E+03 n/a 3.8 1.80 44.96 0.563-Jul-00 IMC-P4 clay + Heterocapsa 3 9 0 22 33 3.2 2.43E+03 n/a 3.23 1.53 38.22 0.484-Jul-00 IMC-P4 clay + Heterocapsa 1 24 0 23 33 3.2 2.35E+03 n/a 4.8 2.27 56.80 0.715-Jul-00 IMC-P4 clay + Heterocapsa 2 24 0 23 33 3.2 3.28E+03 n/a 4 1.89 47.33 0.596-Jul-00 IMC-P4 clay + Heterocapsa 3 24 0 23 33 3.2 3.23E+03 n/a 4 1.89 47.33 0.597-Jul-00 IMC-P4 clay + PAC + Heterocapsa 1 3 0 23.5 33 4.8 2.68E+03 5 2.8 1.33 33.13 0.287-Jul-00 IMC-P4 clay + PAC + Heterocapsa 2 3 0 23.5 33 4.8 1.64E+03 5 2.8 1.33 33.13 0.287-Jul-00 IMC-P4 clay + PAC + Heterocapsa 3 3 0 23.5 33 4.8 1.64E+03 5 2.8 1.33 33.13 0.288-Jul-00 IMC-P4 clay + PAC + Heterocapsa 1 9 0 23.5 33 4.8 1.80E+03 5 3.2 1.51 37.86 0.329-Jul-00 IMC-P4 clay + PAC + Heterocapsa 2 9 0 23.5 33.5 4.8 2.46E+03 5 3.2 1.51 37.86 0.329-Jul-00 IMC-P4 clay + PAC + Heterocapsa 3 9 0 23.5 33.5 4.8 2.43E+03 5 3.2 1.51 37.86 0.32

10-Jul-00 IMC-P4 clay + PAC + Heterocapsa 1 24 0 24 33.5 4.8 2.39E+03 5 4 1.89 47.33 0.3911-Jul-00 IMC-P4 clay + PAC + Heterocapsa 2 24 0 24 33.5 4.8 2.77E+03 5 4 1.89 47.33 0.3912-Jul-00 IMC-P4 clay + PAC + Heterocapsa 3 24 0 24 33.5 4.8 3.36E+03 5 4 1.89 47.33 0.39

Racetrack flume5-Apr-01 IMC-P4 clay + Heterocapsa 1 5.5 3 21 nd 2537 3.50E+02 n/a 1200 567.96 33.21 0.226-Apr-01 IMC-P4 clay + PAC + Heterocapsa 1 5.5 3 21 nd 2503 6.23E+02 4 1200 567.96 33.21 0.232-Jul-01 IMC-P4 clay + Heterocapsa 1 3 10 21 30 2052 5.11E+02 n/a 1200 567.96 33.21 0.284-Jul-01 IMC-P4 clay + PAC + Heterocapsa 1 3 10 22 30 2052 7.02E+02 5 1200 567.96 33.21 0.286-Jul-01 IMC-P4 clay + Heterocapsa 1 3 3 23 30 2052 7.54E+02 n/a 1200 567.96 33.21 0.288-Jul-01 IMC-P4 clay + PAC + Heterocapsa 1 3 3 23 30 2052 3.32E+02 5 1200 567.96 33.21 0.28

10-Jul-01 IMC-P4 clay + Heterocapsa 1 3 20 21 30 2052 1.88E+02 n/a 1200 567.96 33.21 0.2812-Jul-01 IMC-P4 clay + Heterocapsa 2 3 20 20.5 30 2052 1.52E+02 n/a 1200 567.96 33.21 0.28

A-3

Figure A-1 (Cont.). Flume Experiments.

Volume dispensedinto fenced box (ml) Flushness Total Assumed (based on Kinematic LDV (or OBS) files for estimating

or into entire flume (l) of layer thickness of Estimated settled tube results) or Cells per ml viscosity Fluid density critical shear velocities General Wave Mean flow MassDescription of experiment (in SW unless noted) on test bed layer (mm) floc volume (ml) calculated RE in settled floc (cm^2/s) (g/cm^3) Initial motion bedload resuspension resuspension erosion

17-m flumemodelling clay in test bed n/a n/a n/a n/a n/a n/a 1.24E-02 1.023 n/a n/a n/a n/a n/acoarse Nobska beach sand (500 - 1000 um) n/a n/a n/a n/a n/a n/a 1.30E-02 1.025 19A6 nd nd nd 19A7medium Nobska beach sand (250 - 500 um) n/a n/a n/a n/a n/a n/a 1.24E-02 1.023 21A4 nd nd nd 21A5IMC-P3 clay nd nd > 10 400 n/a n/a 1.24E-02 1.023 nd nd 21A6 nd ndfine Nobska beach sand (125 - 250 um) n/a n/a n/a n/a n/a n/a 1.00E-02 1.020 24A4 nd nd nd 24A5IMC-P3 clay 400 (in DI H2O) 1 cm above flush 12 480 n/a n/a 9.79E-03 1.022 nd nd nd nd ndIMC-P3 clay 300 (in DI H2O) 2 mm above flush 5 200 n/a n/a 9.79E-03 1.022 25J1 25J2 25J3 25J4 ndIMC-P3 clay 200 (in DI H2O) flush 3 120 n/a n/a 9.70E-03 1.022 nd nd 26J1 nd ndIMC-P3 clay + PAC 230 = 200 (clay in DI H2O) + 30 (PAC) flush 5 200 n/a n/a 9.69E-03 1.022 28J1 28J2 28J3 nd ndIMC-P3 clay + PAC 260 = 200 (clay in DI H2O) + 60 (PAC in DI H2O) ~2 mm above flush nd nd n/a n/a 9.69E-03 1.022 28J4 28J5 28J6 nd ndIMC-P3 clay + Gyrodinium 300 = 200 (clay in DI) + 100 (dinos) ~1 mm above flush 4 160 nd nd 9.69E-03 1.022 01A1 01A2 01A3 nd ndIMC-P3 clay + Gyrodinium 300 = 200 (clay in DI) + 100 (dinos) ~1 mm above flush 5 200 nd nd 9.69E-03 1.022 01A4 01A5 01A6 nd ndIMC-P4 clay + Heterocapsa 300 = 200 (clay) + 100 (dinos) flush 2 80 80 6.4E+03 1.00E-02 1.023 01J1_080 01J2_080 01J3_080 01J4_080 ndIMC-P4 clay + Heterocapsa 300 = 200 (clay) + 100 (dinos) flush 2 80 80 6.4E+03 1.00E-02 1.023 01J5_080 01J6_080 01J7_080 01J8_080 ndIMC-P4 clay + Heterocapsa 300 = 200 (clay) + 100 (dinos) flush 2 80 80 6.4E+03 1.00E-02 1.023 01J9_080 1J10_080 1J11_080 1J12_080 ndIMC-P4 clay + Heterocapsa 300 = 200 (clay) + 100 (dinos) < 1 mm above flush 2.5 100 80 4.8E+03 9.91E-03 1.023 02J1_080 02J2_080 02J3_080 02J4_080 ndIMC-P4 clay + Heterocapsa 300 = 200 (clay) + 100 (dinos) 1 mm above flush 3 120 80 5.4E+03 9.91E-03 1.023 03J1_080 03J2_080 03J3_080 03J4_080 ndIMC-P4 clay + Heterocapsa 270 = 170 (clay) + 100 (dinos) flush 2 80 80 7.8E+03 9.91E-03 1.023 03J5_080 03J6_080 03J7_080 03J8_080 ndIMC-P4 clay + Heterocapsa 250 = 150 (clay) + 100 (dinos) 1 mm above flush 3 120 80 5.0E+03 9.70E-03 1.022 04J1_080 04J2_080 04J3_080 04J4_080 ndIMC-P4 clay + Heterocapsa 225 = 125 (clay) + 100 (dinos) 1 mm above flush 3 120 80 7.0E+03 9.70E-03 1.022 05J1_080 05J2_080 05J3_080 05J4_080 ndIMC-P4 clay + Heterocapsa 225 = 125 (clay) + 100 (dinos) < 1 mm above flush 2.5 100 80 8.3E+03 9.70E-03 1.022 06J1_080 06J2_080 06J3_080 06J4 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) flush 2 80 nd nd 9.60E-03 1.022 07J1_080 07J2_080 07J3_080 07J4_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) flush 2 80 nd nd 9.60E-03 1.022 07J5_080 07J6_080 07J7_080 07J8_080 ndIMC-P4 clay + PAC + Heterocapsa 375 = 175 (clay) + 100 (dinos) + 100 (PAC in DI) flush 2 80 nd nd 9.60E-03 1.022 07J9_080 1J10_080 1J11_080 1J12_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) flush 2 80 nd nd 9.60E-03 1.022 08J1_080 08J2_080 08J3_080 08J4_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) flush 2 80 nd nd 9.61E-03 1.023 09J1_080 09J2_080 09J3_080 09J4_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) < 1 mm above flush 2.5 100 nd nd 9.61E-03 1.023 09J5_080 09J6_080 09J7_080 09J8_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) flush 2 80 nd nd 9.51E-03 1.023 10J1_080 10J2_080 10J3_080 10J4_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) < 1 mm above flush 2.5 100 nd nd 9.51E-03 1.023 11J1_080 11J2_080 11J3_080 11J4_080 ndIMC-P4 clay + PAC + Heterocapsa 400 = 200 (clay) + 100 (dinos) + 100 (PAC in DI) < 1 mm above flush 2.5 100 nd nd 9.51E-03 1.023 12J1_080 12J2_080 12J3_080 12J4_080 nd

Racetrack flumeIMC-P4 clay + Heterocapsa nd (~2.5 carboys dinos, 18*3 l clay) n/a ~1.5 mm after 1.5 hr 25650 89.71 3.1E+03 nd nd n/a n/a n/a ndIMC-P4 clay + PAC + Heterocapsa nd (~2.5 carboys dinos + 18*3 l clay + 18 l PAC in tap water) n/a ~3.5 mm after 1.5 hr 59850 95.98 2.5E+03 nd nd n/a n/a n/a ndIMC-P4 clay + Heterocapsa 99 = 54 (clay) + 45 (dinos) n/a ~2 mm after 3 hr 34200 89.10 2.7E+03 1.01E-02 1.021 n/a n/a n/a 2JUL2001.capIMC-P4 clay + PAC + Heterocapsa 122 = 54(clay) + 50 (dinos) + 18 (PAC in DI) n/a on the order of 1 mm 17100 28.02 2.4E+03 9.88E-03 1.020 n/a n/a n/a 4JUL2001.capIMC-P4 clay + Heterocapsa 104 = 54 (clay) + 50 (dinos) n/a ~2 mm after 3 hr 34200 99.90 4.5E+03 9.67E-03 1.020 n/a n/a n/a 6JUL2001.capIMC-P4 clay + PAC + Heterocapsa 122 = 54(clay) + 50 (dinos) + 18 (PAC in DI) n/a ~3 mm after 3 hr 51300 78.02 1.0E+03 9.67E-03 1.020 n/a n/a n/a 8JUL2001.capIMC-P4 clay + Heterocapsa 104 = 54 (clay) + 50 (dinos) n/a n/a n/a n/a n/a 1.01E-02 1.021 n/a n/a n/a 10JUL2001.capIMC-P4 clay + Heterocapsa 104 = 54 (clay) + 50 (dinos) n/a n/a n/a n/a n/a 1.02E-02 1.021 n/a n/a n/a 12JUL2001.cap

A-4

Figure A-1 (Cont.). Flume Experiments.

Critical shear velocities (ustartotal; cm/s)calculated for LDV measured flow General Wave Mean flow Mass WSEQ video Lab notebook no.

Description of experiment Initial motion bedload resuspension resuspension erosion VHS video images captured and pages therein Comments

17-m flumemodelling clay in test bed n/a n/a n/a n/a n/a n/a n/a 1; 115-118 goal of these runs was to make sure that u* at 10.9 m downstream was similar to u* over test bedcoarse Nobska beach sand (500 - 1000 um) 1.61 nd nd nd 1.74 Tape 1 yes, from above 1; 133-136 goal to check against Shields curve for initial motionmedium Nobska beach sand (250 - 500 um) 1.35 nd nd nd 1.60 Tape 1 yes, from above 1; 137-140 goal to check against Shields curve for initial motionIMC-P3 clay nd nd 0.58 0.70 nd Tape 1 no 1; 141-142 layer too thick and acts like fluid hitting sides of test bedfine Nobska beach sand (125 - 250 um) 1.12 nd nd nd 1.55 Tape 2 yes, from above 1; 126-128 goal to check against Shields curve for initial motionIMC-P3 clay nd nd nd nd nd Tape 5 no 2; 56 did not run flume test because layer extended too high above test bedIMC-P3 clay 0.41 0.55 0.72 0.89 1.06 Tape 5 yes 2; 57 - 59 can see surge due to wave-induced stress; confused b/c layer had to flattenIMC-P3 clay nd nd 0.78 nd 0.98 Tape 5 no 2; 59 - 60 sand not perfectly flat; resuspension is intermediate b/w initial and mean flow resuspnIMC-P3 clay + PAC 0.40 0.54 0.75 nd nd Anderson tape no 2; 65 - 66 PAC mixed with flume water so does NOT flocculate as when in DI H2O; perfect layerIMC-P3 clay + PAC 0.40 0.48 0.72 nd nd Anderson tape yes 2; 67 - 68 PAC solution in distilled water; flocs are more cottony than 28-Jul Run 1; can see surgeIMC-P3 clay + Gyrodinium 0.44 0.57 0.76 nd nd Anderson tape no 2; 75 - 76 incredibly smooth, perfect layer; intermediate b/w initial and mean flow resuspensionIMC-P3 clay + Gyrodinium 0.45 0.54 0.74 nd nd Anderson tape yes 2; 76 - 78 layer not perfectly level due to several clay chunksIMC-P4 clay + Heterocapsa 0.40 0.49 0.61 0.78 nd nd no 3; 98-99 clay seems sticky to itself but not to sandIMC-P4 clay + Heterocapsa 0.42 0.54 0.70 0.82 nd Anderson tape no 3; 99-100 loose, fluffy aggregates on top rolled first and were easy to seeIMC-P4 clay + Heterocapsa 0.42 0.53 0.67 0.86 nd Anderson tape no 3; 101IMC-P4 clay + Heterocapsa 0.41 0.53 0.70 0.88 nd Anderson tape no 3; 102-103 loose, less compacted aggregates eroded first, suggesting erosion is a function of compactionIMC-P4 clay + Heterocapsa 0.44 0.54 0.65 0.90 1.14 Anderson tape no 3; 103-104 no video of bedloadIMC-P4 clay + Heterocapsa 0.43 0.54 0.68 0.89 1.14 Anderson tape no 3; 105-106 D. Anderson watched this flume runIMC-P4 clay + Heterocapsa 0.45 0.58 0.73 0.95 1.18 Anderson tape no 3; 107-108 clay floc seems stickier than the previous 3 and 9hr runsIMC-P4 clay + Heterocapsa 0.43 0.55 0.72 0.93 nd Anderson tape no 3; 109-110IMC-P4 clay + Heterocapsa 0.43 0.55 0.75 0.90 1.16 Anderson tape no 3; 110-111IMC-P4 clay + PAC + Heterocapsa 0.42 0.50 0.59 0.75 nd Anderson tape no 3; 111-113 A. Li watched flocculation and resuspensionIMC-P4 clay + PAC + Heterocapsa 0.42 0.51 0.63 0.77 1.00 Anderson tape no 3; 113-114 A. Li took digital photos of flocculation; definitely appears easier to erode than longer duration tests withIMC-P4 clay + PAC + Heterocapsa 0.41 0.49 0.61 0.76 1.01 Anderson tape no 3; 114-115 some turbidity in fenced box water may be PAC precipitateIMC-P4 clay + PAC + Heterocapsa 0.42 0.53 0.68 0.83 1.08 Anderson tape no 3; 115-116 test bed surface seems very fluffyIMC-P4 clay + PAC + Heterocapsa 0.41 0.52 0.63 0.83 1.06 Anderson tape no 3; 116-117 test bed surface seems very fluffy until during/after mean flow resuspensionIMC-P4 clay + PAC + Heterocapsa 0.41 0.50 0.65 0.79 nd Anderson tape no 3; 117-118 gave subsample of test bed floc to Mario (preserved in formalin)IMC-P4 clay + PAC + Heterocapsa 0.44 0.56 0.66 0.82 1.16 Anderson tape no 3; 118-119 test bed surface fluffy but not so much as 9hr PAC treatmentIMC-P4 clay + PAC + Heterocapsa 0.44 0.54 0.66 0.85 nd Anderson tape no 3; 119-120 gave subsample of test bed floc to Mario (preserved in formalin)IMC-P4 clay + PAC + Heterocapsa 0.42 0.54 0.66 0.84 nd Anderson tape no 3; 120-121

Racetrack flumeIMC-P4 clay + Heterocapsa nd n/a n/a 4; 105-109 see summaryIMC-P4 clay + PAC + Heterocapsa nd Quicktime movie n/a 4; 111-113 see summaryIMC-P4 clay + Heterocapsa 0.90, 0.90, and 0.95 n/a n/a 5; 9-14 see notebook data sheetsIMC-P4 clay + PAC + Heterocapsa 0.90, 0.90, and 0.95 n/a n/a 5; 15-18 see notebook data sheetsIMC-P4 clay + Heterocapsa 0.76, 0.76, and 0.76 n/a n/a 5; 19-22 see notebook data sheetsIMC-P4 clay + PAC + Heterocapsa 0.72, 0.72, and 0.76 n/a n/a 5; 23-25 see notebook data sheetsIMC-P4 clay + Heterocapsa nd n/a n/a 5; 26-27 LOST DUE TO POWER OUTAGE; see notebook data sheetsIMC-P4 clay + Heterocapsa 0.58 largest decrease in slope n/a n/a 5; 28-29 see notebook data sheets