|UFDC Home||myUFDC Home | Help|
1 CHARACTERIZATION OF SELECTED PESTICIDES IN THE CALOOSAHATCHEE RIVER AND THE POTENTIAL EFFECTS OF THE MOST COMMON MIXTURE ON THE AQUATIC MACROPHYTE LEMNA MINOR By RAMONA D. SMITHBURRELL A DISSERTATION PRESENTED TO THE GRADU ATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Ramona D Smith Burrell
3 To my family. Thank you all for your patience and support. I love you.
4 ACKNOWLEDGMENTS I thank my advisors, Samira Daroub and Chris Wilson, for their expert guidance and thoughtful help throughout my doctoral work ; I have learned so much from you and I greatly appreciate your efforts I thank Dr. Ramesh Reddy for his support and encouragement, as well as the generous department assistantship. I thank Cathleen Hapeman for introducing me to a new field and showing me how productive a scientific team can be. I thank my parents, Rollie and Donna Smith, for encouraging me throughout my life and reminding me to pay attention to what is important. I thank my sister for making me laugh when things didn't seem funny. I thank my children for being so fun and wonderful that this project sometimes didn't seem worth finishing. Lastly, I thank my incredible husband. When I started this project, we were dating. Now we are married with two children! Your support, patience, and love bless my life; thank you for recognizing how much this degree means to me, and thank you for being so amazing. I love you.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 9 LIST OF ABBREVIATIONS ........................................................................................... 13 ABSTRACT ................................................................................................................... 15 1 INTRODUCTION .................................................................................................... 17 Caloosahatchee Basin ............................................................................................ 17 Agriculture in the Caloosahatchee Area: Glades, Hendry, Charlotte, and Lee Counties .............................................................................................................. 19 Pesticide Use in Florida and the Caloosahatchee Area .......................................... 22 Pesticides and Aquatic Ecotoxicology Studies ........................................................ 25 Choice of a Study Subj ect ....................................................................................... 28 Study Goals ............................................................................................................ 28 2 SPATIAL AND TEMPORAL VARIABILITY OF PESTICIDES AND PESTICIDE DEGRADATION PRODUCTS OF THE CALOOSAHATCHEE RIVER, FLORIDA .. 47 Background ............................................................................................................. 47 Materials and Methods ............................................................................................ 49 Results .................................................................................................................... 54 Summary of Pesticide Concentrations ............................................................. 55 Spatial and Temporal Comparisons ................................................................. 56 Site Comparisons ............................................................................................. 58 Temporal Comparisons .................................................................................... 60 Target Analyte Comparisons and CoOccurrence ............................................ 62 Discussion .............................................................................................................. 64 3 EFFECT OF HERBICIDE MIXTURES (ATRAZINE AND METOLACHLOR) ON THE AQUATIC MACROPHYTE, LEMNA MINOR .................................................. 89 Background ............................................................................................................. 89 Materials and Methods ............................................................................................ 93 Plant Culturing .................................................................................................. 93 Test Solution Preparation ................................................................................. 93 Bioassay Procedures ....................................................................................... 95 Data Analysis ................................................................................................... 97 Results .................................................................................................................. 100 Validation of Test Solution Concentrations ..................................................... 100
6 Individual Concentration Response Bioassa ys ............................................... 101 Tests for Synergy ........................................................................................... 103 Frond production ...................................................................................... 103 Fresh Weights .......................................................................................... 104 Root Length ............................................................................................. 105 Photosystem II Electron Transport (Fv/Fm) ............................................... 105 Mixture Comparisons to Individual Test Chemicals ........................................ 106 Discussion ............................................................................................................ 107 4 PROJECT CONCLUSIONS .................................................................................. 126 LIST OF REFERENCES ............................................................................................. 131 BIOGRAPHICAL SKETCH .......................................................................................... 141
7 LIST OF TABLES Table page 1 1 Top ten natural communities by acreage found within the Caloosahatchee basin. Adapted from FDEP, 2003. ...................................................................... 30 1 2 Listed animal species found in the Caloosahatchee basin (SFWMD, 2000). Status is indicated as E (Endangered), T (Threatened), SSC (Species of special concern). ................................................................................................ 31 1 3 Population estimates in 2000 and 2007 by County in the Caloosahatchee Region. (Adapted from US Census Bureau, 2008). ............................................ 33 1 4 Land Use Percentages in the Caloosahatchee Basin (SFWMD GIS data, 1998) .................................................................................................................. 33 1 5 Top ten pesticides in Florida by total pounds of active ingredient (AI) reportedly applied per crop per year, adapted from Shahane, 2008. .................. 34 1 6 Pesticides reported on key crops of the Caloosahatche e area. Adapted from Shahane, 2008. .................................................................................................. 34 1 7 Herbicide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003 ..................................................... 36 1 8 I nsecticide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003. .................................................... 37 1 9 Fungicide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003. .................................................... 38 1 10 Other pesticide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003. ....................................... 39 1 11 Pesticide application in Florida citrus: percent of area with pesticide applied, average number of applications per season, and average appl ication rate in kilograms of active ingredient per hectare per application. Adapted from Shahane, 2003. .................................................................................................. 40 1 12 Land use by hectares and percentage of the Lake Okeechobee Watershed. Adapted from James & Zhang, 2008. ................................................................. 41 2 1 Pesticides and Pesticide Degradation Products Measured with associated average spike recoveries, percent relative standard deviation (%RSD) and method detecti on limits (MDL). ........................................................................... 70
8 2 2 Site coordinates, pH, salinity, conductivity, and temperature on monthly sampling trips between December 2004 and April 2006 (n= 16 for sites A, B & C and n=9 for sites D & E). ............................................................................. 71 2 3 Caloosahatchee River Flow data from three of our pesticide sampling sites, December 2004 through April 2006 (n=517). Flow measured in cubic feet per second (adapted from DBHYDR O, SFWMD). .................................................... 71 2 4 Detection frequencies of pesticides and their concentration ranges in water samples collected along the Caloosahatchee River, Florida. ............................. 72 2 5 Mean, minimum, median, maximum concentrations, and percent relative standard deviation (%RSD) of the ten most commonly detected pesticides (organized by site) in the Caloosahatchee River. ............................................... 74 3 1 Summary of chemical and physical properties of atrazine and metolachlor. .... 112 3 2 Target concentrations, measured concentrations, and percent from expected with the associated standard deviation (Deviation SD) in the individual atrazine and metolachlor tests. The calculated toxic units for each measured concentration are included as well. ................................................................... 113 3 3 Computed quantities for testing assumption that the separate regression slopes of metolachlor and atrazine individual doseresponse curves do not differ. F0.001[1,12] = 18.6. ............................................................................... 113 3 4 Target c oncentrations, measured concentrations, and percent from expected with the associated standard deviation in the mixture (atrazine+metolachlor) tests. Resulting toxic unit adjustments for the atrazine (Atr) and metolachlor (Met) mixture study are also show n. TU= toxic units. SD= Standard Deviation .......................................................................................................... 114
9 LIST OF FIGURES Figure page 1 1 Geopolitical map of the Caloosahatchee Basin (FDEP, 2003) ........................... 42 1 2 Caloosahatchee Area Counties .......................................................................... 42 1 3 Citrus harvest season (FASS, 2007). ................................................................. 43 1 4 Planting and harvesting seasons of selected Florida field crops (FDACS, 2007). ................................................................................................................. 44 1 5 Planting and harvesting seasons of selected Florida vegetables (FDACS, 2007). ................................................................................................................. 45 1 6 Lemna minor in culture during a laboratory experiment, control culture. ............ 46 2 1 Five pesticide sampling sites used along the Caloosahatchee River, Florida. From: www.evergladesplan.org. ......................................................................... 76 2 2 Pesticides detected (ng/L) in at least one sample in the Caloosahatchee River from December 2004April 2006 (n= 75 for compounds ethropr op through methoxychlor and n= 90 for diazinon through mirex). The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whis kers (error bars) above and below the box indicate the 90th and 10th percentiles. Outliers are plotted with black circles. ................ 77 2 3 SIMPROF results of a hierarchical cluster analysis of all target analytes detected at least once from all sample sites and sampling months (permutations=999, = 0.05). Samples connected by red lines cannot be differentiated at the 0.05 significance level. ........................................................ 78 2 4 Two Dimensional MDS plot (Stress = 0.11) showing similarities between samples by sampling month and site. Similarity between samples are superimposed from a hierarchical cluster analysis. ............................................ 79 2 5 Concentrations of the ten most frequently detected pesticides from all sampling sites and dates along the Caloosahatchee River, Florida (n= 75 for atrazine through ametryn and n=90 for malathion through chlorothalonil above). The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles. Outliers are plotted with black circles. ....................................................................................................... 80
10 2 6 Average number of analytes detected at each sampling site between January 2005 and April 2006 with the associated standard errors (n=15 for sit es A, B, analyses using the Holm Sidak method; different letters indicate significant difference. ........................................................................................................... 81 2 7 Average concentr ation of atrazine (ng/L) in samples collected between January 2005 and January 2006 and the associated standard errors shown analyses with Tukey's test shown. ...................................................................... 82 2 8 Average concentration of metolachlor (ng/L) in samples collected between January 2005 and January 2006 and the associated standard errors. Results Tukey's test shown. ......................................................................................................... 83 2 9 Average number of pesticides detected per sample over the sampling months of December 2004 to January 2006 and the associated standard errors. While the overall oneway ANOVA indicated significant difference between the number of pesticides detected in these months, no pairwise comparisons indicated significant differences. Bars with no error bars indicate an error so small that they do not show up on the figure. ...................... 84 2 10 SIMPROF results of a hierarchical cluster analysis based on concentration of all target analytes detected at least once from all sample sites and sampling months (Resemblance= Bray Curtis Samples connected by black lines are significantly different. ............................. 85 2 11 Two Dimensional MDS plot (Stress = 0.12) showing similarities between samples by month and site after transformation of data to presence/absence. Similarity between samples are superimposed from a hierarchical cluster analysis. The first plot showed one extreme outlier so the subsequent plot is a subset of the smaller top image without this point. .......................................... 86 2 12 SIMPROF results of a hierarchical cluster analysis of all target analytes detected at least once from all sample sites and sampling months transformed to presence/absence data. Black lines indicate significant differences at nodes shown (Resemblance= Simple Matching, Permutations= ...................................................................................................... 87 2 13 The cooccurrence of the top ten most commonly detected com pounds (atrazine, metolachlor, CIAT, CEAT, simazine, ametryn, malathion, chlorpyrifos oxon, dieldrin, chlorothalonil) over all sampling months and sites along the Caloosahatchee River, Florida (42 analytes and 75 sampling events). ............................................................................................................... 88
11 3 1 Lemna minor frond production plotted as percentages of the corresponding controls against A) logs of measured atrazine concentrations and B) logs of measured metolachlor concentrations. ............................................................. 115 3 2 Transformed logit model against A) log atrazine concentration on day 6 and regression results used to determine EC50 of atrazine and B) log metolachlor concentration on day 6 and regression results used to determine EC50of metolachlor ....................................................................................................... 116 3 3 Frond production of Lemna minor as a percentage of controls in a mixture of atrazine and metolachlor. One toxic unit is expected to produce a 50% mortality e ffect (shown in "Theoretical TU Expected Response" line), so this relationship indicates a synergistic effect based on the 50% mortality effect occurring at lower than expected toxic units. The green line illustrates the expected frond production based on values directly from the individual response curves. .............................................................................................. 117 3 4 Mean final Lemna minor frond count (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with different letters ......................................................... 118 3 5 Mean final Lemna minor fresh weight (D ay 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with ................................. 119 3 6 Mean final Lemna minor root length (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with different letters ......................................................... 120 3 7 Mean final Lemna minor Fv/Fm values (Day 6, n=8) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with ................................. 121 3 8 Mean final Lemna minor chlorophyll a and b values (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. ............................................................................................................... 122 3 9 Mean final Lemna minor total chlorophyll and total carotenoid values (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with different letters indicating significant differenc e ........................................................................................................... 123
12 3 10 Day 6 relative Lemna minor frond counts expressed as a percentage of the controls in individual atrazine and metolachlor tests as well as mixture tests, equalized against toxic units (which were calculated from frond count). .......... 124 3 11 Day 6 relative Lemna minor fresh weights expressed as a percentage of the controls in individual atrazine and metolachlor tests as well as mixture tests, normalized against toxic units (which were calculated from frond counts). ...... 125 3 12 Morphology changes in fronds produced in toxicity tests. In (a), the control culture is pictured with normal frond production. In (b), leaves exposed to mixture test concentrations exhibit different sized frond and different pigmentation. .................................................................................................... 125
13 LIST OF ABBREVIATIONS ANCOVA Analysis of covariance, univariat e statistics ANOSIM Analysis of similarity, multivariate statistics ANOVA Analysis of variance, univariate statistics BCF Bioconcentration factor CEAT 6 amino2 chloro4 isopropylaminostriazine atrazine degradation product CIAT 6 amino2 chloro4 ethylamino s triazine atrazine degradation product DBHYDRO South Florida Water Management District's corporate environmental database which stores hydrologic, meteorologic, hydrogeologic and water quality data. DDD 1,1dichloro2,2bis[p chlorophenyl]ethane, DDT degradation product DDE 1,1dichloro2,2bis[4 chlorophenyl]ethylene, DDT degradation product DDT 1 trichloro 2,2bis [p chlorophenyl]ethane EC50 Concentration which causes 50% of a given effect EPA United States Environmental Protection Agency FDACS Fl orida Department of Agriculture and Consumer Services FDEP Florida Department of Environmental Protection HCH ,2,3,4,5,6 hexachlorocyclohexane KOC Adsorption coefficient KOW Partition coefficient for a compound between noctanol and water phases LC50 Concentration which causes 50% mortality in a given test population LD50 Dose which causes 50% mortality in a given test population LOEC Lowest observed effect concentration
14 LOEL Lowest observed effect level MDS Multidimensional Scaling NASS National Agricultur al Statistics Service NOEC No observed effect concentration NOEL No observed effect level RED Re registration eligibility decision RSD Relative standard deviation SD Standard deviation SE Standard error SFWMD South Water Management District SIMPER Similarity percentages, multivariate analysis SIMPROF Similarity profile, multivariate analysis TU Toxic Unit USGS United States Geological Survey WHO World Health Organization
15 Abstract of Dissertation Presented to the Graduate School of the University of Fl orida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHARACTERIZATION OF SELECTED PESTICIDES IN THE CALOOSAHATCHEE RIVER AND THE POTENTIAL EFFECTS OF THE MOST COMMON MIXTURE ON THE AQUATIC MACROPHYTE LEMNA MINOR By Ramona SmithBurrell December 2012 Chair: P. Christopher Wilson Cochair: Samira H. Daroub Major: Soil and Water Science Land use immediately adjacent to waterways is often used to predict the types of pesticides found in those surface waters. Monthly samples collected from the Caloosahatchee River between December 2004 and April 2006 were analyzed for 42 pesticides and pesticide degradation products. This study highlighted a lack of spatial patterns in pesticide concentration, despite sampling sites located in primarily agricultural or urban areas. This is likely due to the magnitude and scale of the drainage system (far removed from the areas adjacent to the river), the persistence of the chemicals, the patchiness of usepatterns, and the unpr edictable but dramatic change in flow to meet water management needs. Temporal differences depended on the compound. Atrazine and metolachlor were most commonly detected, with detection frequencies of 98.7% and 94.7%, respectively. Of the ten compounds det ected most frequently, six were herbicides, three were insecticides and one was a fungicide. When all 42 analytes were evaluated, between two and twelve compounds were always detected (n=75) per sample. These data suggest that organisms living in these regions
16 of the Caloosahatchee River are frequently exposed to several pesticides at low concentrations. Many risk assessments identify suites of chemicals in the environment and then use toxicity data for the chemicals based on doseresponse curves generated for single toxicant exposures. Because of the potential interactions of pesticides it is more desirable to compare field concentrations to toxicity data based on the relevant combinations of contaminants to which species may be exposed. To do this effectiv ely, integrated multi level studies related to field detections are needed. Atrazine and metolachlor, identified as most commonly occurring in our field data from the Caloosahatchee River, were used in laboratory toxicity assays alone and in mixtures to id entify effects on growth, reproduction and health of Lemna minor (duckweed). Greater than additive (synergistic) effects on the frond production were observed, as were sign ificantly greater fresh weights and root length when individual compounds were compared to mixtures. This indicates that the predicted exposure responses for the individual compounds actually underestimate the effects of mixtures of atrazine and metolachlor on the aquatic macrophyte Lemna minor
17 CHAPTER 1 INTRODUCTION Caloosahatchee B asin The Caloosahatchee River and Basin stretches 113 km westward from Lake Okeechobee to San Carlos Bay within four counties of Southwest Florida (Figure 1 1). The entire watershed contains five drainage basins (East Caloosahatchee, West Caloosahatchee, Estuarine Caloosahatchee, Telegraph Swamp, and Orange River), as well as one national wildlife refuge, two state aquatic preserves, and one wildlife management area (FDEP, 2003). Originally, the Caloosahatchee River was a shallow, meandering river with headwaters near Lake Hicpochee, slightly southwest of Lake Okeechobee (FDEP, 2003). This natural watershed sustained a variety of plant communities such as pine flatwoods, saw palmetto prairies, sand pines, xerophytic oaks, hardwood swamp forests, prairie gr asslands, mangrove swamps, and coastal marshes (Kimes & Crocker, 1998). S ome major natural ecosystems and unmanaged areas still exist in the region, and these are summarized in Table 11 While grasslands and agriculture dominate these undeveloped areas, marshes, swamps, pinelands and hardwood hammocks are found as well. Today, this area serves as habitat to 22 avian, 10 mammalian and 14 reptilian, amphibian and fish species that are classified as endangered, threatened, or species of special concern (Table 1 2) In 1881, the Caloosahatchee was connected to Lake Okeechobee to lower the water table (FDEP, 2005). To allow for better navigation, flood control, and land reclamation, the freshwater portion of the river was canalized into Canal C 43 beginning i n the 1930s (FDEP, 2003). Currently, discharge structures and locks control the flow of
18 Lake Okeechobee through the Caloosahatchee River to the Gulf of Mexico. Water releases from Lake Okeechobee occur through a series of locks when lake levels exceed the US Army C orps of Engineers criteria for flood protection. Water is also released to flush algae blooms and salt water out of the river to protect potable water supplies from contamination (Flaig & Capece, 1998). T he Moore Haven Lock, S 77, separates Lak e Okeechobee from the beginning of the canal on the eastern side and the Franklin Lock, S 79, separates the canal from the salinity of the estuary on the western side. The freshwater portion of the river is about 60 km long and the tidal Caloosahatchee (w est of the Franklin Lock) extends downstream for about 45 km (FDEP, 2005). T he Caloosahatchee River is part of the Okeechobee Waterway, which links the Atlantic Ocean to the Gulf of Mexico through the St. Lucie River and Canal, Lake Okeechobee, and finally the Caloosahatchee Canal and River. Much of the land around the Caloosahatchee has been converted to intensive agricultural use, and several side canals were constructed along the banks of C 43 to support the growing agricultural needs of the area as a result. The river is the major source of surface water supply for much of the southwest coastal region; it is used for agricultural irrigation needs, to recharge shallow wellfields, and to provide a potable water supply for Lee County and the city of Fort M yers (FDEP, 2003; SFWMD, 2000). Complicating the prioritization of water use in the region, the population of southwest Florida is growing rapidly There are large population centers in the basin including Fort Myers, Cape Coral, North Fort Myers, Lehigh A cres, LaBelle, and Moore Haven. Table 13 summarizes population growth in the area over the last several years. IN the 7 years between 2000 and 2007, Charlotte County population increased
19 by 8%, Glades by 5%, Hendry by 9% and Lee by 34%. This growth woul d be expected to influence the water needs in the area. While the tidal basin is primarily urbanized, the most common land uses in the freshwater portions of the basin are agriculture and rangeland (FDEP, 2005). Table 14 summarizes the major land use in the Caloosahatchee basin (SFWMD, 1998) illustrating that over 50% of the basin is urbanized or used in agriculture, with water demands In 1995, the tidal Caloosahatchee River, as part of the Charlotte Harbor system, was recognized as an estuary of national significance and was accepted into the National Estuary Program (FDEP, 2003). Nevertheless, freshwater discharge from the river to the estuary at S 79 is great. Doering & Chamberlain (1999) report that enough fresh water enters the estuary at the Franklin Lock to fill the entire volume over eight times per year; the median discharge at S 79 is 870 x 106 m3 (Flaig & Capece, 1998). Agriculture in the Caloosahatchee Area: Glades, Hendry, Charlotte, and Lee Counties Data in this section describes the agri culture in the area of study during the time period of the study (20042006). While there is more current data available, the information pertinent to the period of study is included for its relevance. Florida was among the top ten agricultural states in t he United States with annual farm cash receipts totaling nearly $7 billion (Shahane, 2003). Florida growers produced more than 70% of the countrys citrus, nearly 25% of its vegetables, and 48% of its sugarcane, besides its contribution in row crops and ornamental plants (Shahane, 2003). Pasture, citrus, and sugarcane we re the top three crops by acreage in the state of Florida, totaling 2,454,153 hectares 251,460 hectares, and 163898 hectares respectively
20 (Shahane, 2008). Of the 67 counties in Florida, four counties bank the Caloosahatchee River; Glades and Hendry Counties border the river to the north, and Charlotte and Lee Counties border the river to the south (Figure 1 2). Although oranges are grown in all four counties, each has a unique agricultural profile. In Glades County ( 255, 400 hectares total) the top five crops by acreage included sugarcane for sugar, oranges, sugarcane for seed, forage land used for all hay and haylage, grass silage, and greenchop, and sod (NASS, 2002). Glades County ranked third in the state for sugarcane for sugar and for seed, and rank ed eighth in the state for the number of cattle and calves with over 66,000 beef and milk cows (NASS, 2002). Approximately 165, 091 hectares of the county was in farms in 2002, with most specific use not permitted to be disclosed (NASS, 2002). The estimated market value of agricultural production in Glades County wa s $72 million, with about 70% of that attributed to crops, including nursery and greenhouse, and 27% coming from livestock, p oultry, and their products (NASS, 2002). By a rea, the five main crops in Hendry County ( 308, 200 hectares total) include d oranges, sugarcane for sugar, combined vegetables, grapefruit, and tomatoes (NASS, 2002). A close sixth in terms of acreage, sugarcane for seed wa s a major crop in this county, and Hendry rank ed second in the state for this crop (NASS, 2002). Hendry also rank ed second in the state for both orange and sugarcane for sugar production, fourth in the state for both grapefruit and tomato production, and fifth in the state for vegetables (NASS, 2002). Considering all citrus combined produced, Hendry County rank ed second in the state (NASS, 2002). It also rank ed sixth in the state for the number of cattle and calves, with over 73,000 beef and milk cows (NASS, 2002).
21 Among the 223529 hectares of the county in farms in 2002, 54% were in cropland, 30% in pasture, 9% in woodland, and 7% listed as other uses (NASS, 2002). T he market value of agricultural production in Hendry County was over $375 m illion with approximately 95% of that coming from the value of crops, including nursery and greenhouse crops, and 5% attributed to livestock, poultry, and their products (NASS, 2002). The top five crops by area in Charlotte County ( 222, 500 hectares total) we re oranges, grapefruit, combined vegetables, watermelons, and forage land used for all hay and haylage, grass silage, and greenchop (NASS, 2002). Charlotte County rank ed 26th in the state for cattle and calves with approximately 21,000 beef and milk cows (NASS, 2002). As of the 2002 Census of Agriculture, there was approximately 77, 509 hectares of Charlotte County in farms, 50% of which was in woodland condition, 22% in crop lands, 18% in pasture, and 9% listed as other uses (NASS, 2002). The estimated market value of agricultural production in Charlotte County wa s $48 million, with about 88% from crops, nurseries, and greenhouses, and 12% from livestock, poultry, and their products (NASS, 2002). Lee County ( 313, 900 hectares total) list ed oranges, com bined vegetables, cucumbers and pickles, potatoes, and nursery stock as the top five crops grown by area (NASS, 2002). It rank ed third in the state for cucumbers and pickles, fifth in the state for potatoes, and produce d the states second highest acreage of mangoes and fifth most colonies of bees (NASS, 2002). It rank ed 31st in the state for all cattle and calves, with over 16,000 beef and milk cows (NASS, 2002). Of the estimated 50, 990 hectares in farms in Lee County in 2002, 35% was pastureland, 28% w as cropland, 23% was
22 woodland, and 12% was listed in other uses (NASS, 2002). The market value of agricultural production in Lee County was approximately $113 million, with 96% attributed to crops, including nursery and greenhouse, and 4% attributed to li vestock, poultry, and their products. Citrus and sugarcane we re among Floridas key crops and are heavily represented within the agricultural area of the four Caloosahatchee counties. Figures 13 and 14 illustrate the expected harvest seasons for these c rops. Figure 15 shows the expected seasons for several commonly grown Florida vegetable crops. These are included because they were expected to illustrate and explain the temporal patterns in detection of pesticides and pesticide degradation products in the region, since it might be expected to find certain compounds at particular harvest or planting times. Pesticide Use in Florida and the Caloosahatchee Area P esticides are used in agriculture, structural pest control, aquatic weed control, right of way maintenance, lawn pest control, landscaping, and golf courses Considering only agricultural uses, there are approximately 191 active ingredients registered for use on crops in the state of Florida: 43 fungicides, 64 herbicides, and 19 other pesticides (Sh ahane, 2008). Reporting on pesticide use in Florida between 20032006, Shahane (2008) indicated that 79 million pounds of pesticide active ingredients per year were reportedly applied to Florida crops But these numbers might not be perfectly accurate bec ause of the lack of laws requiring regular, specific reporting to the state. Of the fungicides used in Florida, the top three reported, in terms of total pounds of active ingredient, are copper hydroxide (1,751,900), chlorothalonil (1,435,833), and m ancozeb (1,027,107) (Shahane, 2008). The top three herbicides in terms of total
23 pounds of active ingredient reportedly applied in the state are 2,4D (2,702,917), g lyphosate (1,690,282), and a trazine (1,478,834) (Shahane, 2008). Of insecticides, the top three reported in terms of total pounds of active ingredient are petroleum distillate (46,006,600), sulfur (1,336,700), and a ldicarb (250,171) (Shahane, 2008). In the class considered other pesticides, m ethyl bromide tops the list with 9,242,742 pounds of activ e ingredient reported, chloropicrin has 3,501,581 total pounds of active ingredient reported, and 1,3d ichloropropene has 761,325 pounds reported (Shahane, 2008). The ten pesticides active ingredients applied in the highest amounts in Florida are summariz ed in Table 15. The pesticides applied to the commodities of interest in the Caloosahatchee River basin (sugar, pasture and citrus like oranges and grapefruit) are summarized in Table 16 Only 7 pesticides are reported for use on sugarcane, four on past ure, 53 on oranges and 48 on grapefruit (Shahane, 2008). But while Florida requires the maintenance of records regarding what pesticides are applied to crops and the availability of such data upon request, the statute does not require voluntarily reporting such information to the state. So this data is not always complete (modified from Shahane, 2008). Sugarcane reported 1 364, 019 kg of herbicide per year applied in Florida, and no other pesticide use on 186,296 hectares (Shahane, 2008). Pasture reported 1 208, 400 kg per year insecticide application and no other pesticide use in the same time period on 2,454,153 hectares. On the reported 28,733 hectares of grapefruit, total pesticide application was estimated at 2,705,316 kg per year, with 214,639 kg fungici de/year, 134,808 kg herbicide/year, and 2,355,868 kg insecticide/year. On the 219,259 reported
24 hectares of oranges, a total of 20,320,621 kg of pesticide per year was reported; this included 317,152 kg/year fungicide, 1,476,579 kg/year herbicide, and 18,526,890 kg/year insecticide (Shahane, 2008). In tables 17 through 110, the statewide citrus pesticide use is broken down by class to grapefruit, oranges, tangelos, tangerines, and temples. These are compounded into the following tables by pesticide type, showing the pesticides for each citrus crop ranked in order of AI applied per crop year. Table 1 11 describes general characteristics of pesticide application in citrus, grouped by the percent of area on which the pesticide type was applied, the average number of applications per season, and the average application rate (lbs. of active ingredient per acre per application) (adapted from Shahane, 2003). Most of the statistics presented in this portion of the report focus on statewide averages based on the com modities of interest in the Caloosahatchee area. P esticides will enter the Caloosahatchee River from areas far from Glades, Hendry, Lee, and Charlotte Counties however Because Lake Okeechobee drains primarily through the Caloosahatchee River, with som e outflow through the St. Lucie River, contaminants in the Lake Okeechobee watershed can potentially affect the Caloosahatchee River as well. L and uses within the Lake Okeechobee watershed are summarized in Table 1 12. In this table, the term "natural ar eas" includes areas that are undeveloped for any agricultural or urban uses. From this table, the largest land use by crops includes citrus, sugar, and pasture, representing 39% of the total land use. Pesticides applied to these crops can affect not only t he Lake Okeechobee watershed, but the Caloosahatchee watershed as well But it should be noted that there are patchy land use patter n s along this river, with areas of high agricultural activity interspersed
25 with areas of high urban density, natural areas, and rural regions. U rban uses of pesticides, such as for mosquito control, termite treatment, roadway management, golf courses, turf grass, and landscape maintenance are also expected to affe ct this system notably Pesticides and Aquatic Ecotoxicology Stu dies Since most contaminants released into the environment are eventually deposited into aquatic ecosystems by direct discharge, terrestrial runoff, or atmospheric deposition, aquatic toxicology is deemed especially important (Pritchard, 1993). I n this Ca loosahatchee River watershed, a variety of chemical toxicants might be expected in the water based on the heavy agricultural and residential uses within the watershed and from further upstream in the Lake Okeechobee watershed. Predicting aquatic toxicolog ical responses are often based on laboratory testing exposing organisms to a particular contaminant via water, sediment, or food for a specified time. From this, the effects are estimated for either lethal or sublethal impacts to the individual organisms (Kendall et al., 2001; Newman & Unger, 2003). Very often, regression models are then applied to acute exposure data to predict measures like the LC50, the concentration killing 50% of exposed individuals by the end of exposure, or the EC50, the concentrat ion having an effect of some defined type on 50% of exposed individuals by the end of exposure (Zhao & Newman, 2005). Statistical tests like the ANOVA may be applied for chronic exposures or more subtle effects. Examples of these include the LOEC, lowest observed effect concentration, or lowest exposure with a statistically significant effect, and the NOEC, no observed effect concentration, or the highest
26 exposure concentration with no statistically significant effect (Kendall et al., 2001; Newman & Unger 2003). Because of the sensitive nature of ecosystems in South Florida as well as the economic reliance of the state on agriculture in the region, much research has been conducted to quant ify and estimate the effects of agricultural chemicals, including p esticides, in these environments (SFWMD, 2000; Shahane, 2007; FDEP, 2005; HarmanFetcho et al., 2005; Downing et al., 2004; Doering & Chamberlain, 1998). Often, field studies which quantify pesticides found in targeted ecosystems are then compared to toxi city indices (Downing et al., 2004; HarmanFetcho et al, 2005). Although organisms in the ecosystem would not realistically be exposed to a single toxicant at a time, many studies examine suites of chemicals identified and quantified from field studies and then base the toxicity of these chemicals on the doseresponse curves of organisms exposed to a single toxicant at a time (Munn & Gilliom, 2001; Banks et al., 2003). Because there exists the possibility of potentiation, additivity, synergism, or antagonism when organisms are exposed to mixtures of xenobiotics like pesticides (Newman & Unger, 2003), it may be desirable to compare field concentrations to toxicity data based on combinations of pesticides to which species may be exposed. To do this effecti vely, integrated multilevel studies are considered necessary to interpret these effects because chemical contamination may act at such a variety of levels (Porter et al., 1993). A common method for determining effects of mixtures in toxicity testing is the toxic unit (TU) concept (Sprague, 1970). Toxic units express amounts or concentrations of different toxicants in units of lethality like LC50 or LD50 (Newman & Unger, 2003).
27 Based on the idea of concentration additivity, concentrations of toxicants, adj usted for relative potency, can be added together to predict the resulting effects. But because many toxicants show a sigmoid, rather than linear, relationship between concentration (or dose) and lethality, this summation of potency adjusted concentrations may not be valid (Newman & Unger, 2003; Berenbaum, 1985). Considered a more valid methodology to explore additivity, toxicant concentrations should not be added together. Instead, effect levels predicted with models like the probit or logistic model for each toxicant in a mixture should be added together (Newman & Unger, 2003). In addition, toxicants in a mixture that have different modes of action must be considered differently, as well. This would be expected to be applicable to pesticide mixtures containing different general types of pesticides, such as herbicides and insecticides. Finney (1947) described this independent joint action of toxicants as existing if each toxicant produces an effect independent of the other and by a different mode of act ion. In this case, the mortalities, not the concentration (doses) would be additive. Using pesticides with different modes of action can complicate this toxic unit approach significantly. Van der Geest et al. (2000), working with the insecticide diazinon and the metal copper, explain that many studies published actually violate the toxic unit concept by ignoring the requirement for similarly shaped doseresponse relationships among mixture components. To test if the modes of action are similar, the slo pes of the probit models can be compared; parallel slopes indicate similar modes of action (Newman & Unger, 2003; Finney, 1947). Even with different modes of action when dose response curves of individual toxicants and the resulting mixture show significa nt difference, the toxic unit approach can still be meaningful using effect
28 concentrations other than the LC50 (Van der Geest et al., 2000; Van Gestel & Hensbergen; Banks et al., 2003). Choice of a Study Subject By using a U.S. EPA recommended test species Lemna minor (Figure 16), that is also a native species of aquatic macrophyte, a more effective consideration of the pesticide mixtures occurring in the river on the delicate food webs and ecosystems of the Caloosahatchee River is possible. Lemna minor ( common duckweed) is a perenn ial, aquatic macrophyte of the family Lemnaceae (USDA NRCS, 2012) It has a rapid growth rate, making it a good test subject to use in the laboratory. It lives on the surface of the water and does not root into the sediment ma king it a good choice for relating toxicity of pesticide residues found in surface waters without considering the effects of sediment (USDA NRCS, 2012) It is also an established test subject as described in the EPA Ecological Effects Test Guidelines (1996) for determining aquatic plant toxicity in Tier I and Tier II toxicity testing, indicating it is a reasonable surrogate for aquatic plants in toxicity tests. Because primary producers are the base of all food webs, studying the effects on nontarget aquati c plants allows us to predict potential effects on the rest of the ecosystem within the Caloosahatchee River. Study Goals T he goals of this project we re two fold. The f irst portion of t his project aimed to illustrate the types of pesticide mixtures that can be found in the Caloosahatchee River and the estuary ecosystems to which it drains The main goal of this portion of the study was discovery. There have not been enough exploratory studies in this region to characterize the pesticides found in the Caloos ahatchee River. Within this discovery plan, I questioned whether there would be spatial patterns in the types of pesticides
29 detected along the river, and I expec ted that more pesticides and degradation products would be present and at higher concentrations at the sampling sites located within more agricultural areas and less in the more residential areas. I also questioned whether there would be seasonal patterns in the concentrations and types of pesticides detected, and I expected that local agriculture w ould drive the presence of detected pesticides and pesticide degradation products based on harvest or planting seasons of local land use. In the second portion of this project toxicity tests with commonly detected pesticides individually and as mixtures were conducted in an effort to provide a more reasonable ecological risk assessment than those based on individual pesticides alone. I hypothesized that the tests with mixtures of atrazine and metolachlor would reflect the results of individual pesticide toxicity assays, and I predicted that such mixtures have an additive effect on the frond count of Lemna minor Given the inclusion of the Caloosahatchee system in the northern Comprehensive Everglades Restoration Plan and its listing by Florida Department of Environmental Protection as an impaired waterway this study should be a relevant contribution to the understanding of water quality within this important ecosystem.
30 Table 11 Top ten natural communities by acreage found within the Caloosahatchee basin. Adapted from FDEP, 2003. Community Hectare s 30% Grassland and Agriculture 10423 9 16.6% Barren land 57703 16.1% Dry Prairie 56038 14.5% Pinelands 50437 6.6% Shrub and Brush 22890 5.6% Freshwater Marsh 1934 9 4.2% Hard wood Hammock 14509 3.0% Water 1039 6 2.2% Cypress Swamp 7565 0.3% Shrub Swamp 114 6
31 Table 12 Listed animal species found in the Caloosahatchee basin (SFWMD, 2000) Status is indicated as E (Endangered), T (Threatened), SSC (Species of special concern). Scientific n ame Common n ame Federal s tatus State s tatus Amphibians Rana capito Gopher frog SSC Reptiles Alligator mississippiensis American alligator T(S/A) SSC Caretta caretta Loggerhead sea turtle T T Chelonia caretta Green sea turtle E E Dermochelys coriacea Leatherback sea turtle E E Drymarchon corais couperi Eastern indigo snake E T Eretmochelys imbricata Hawksbill sea turtle E E Gopherus polyphemus Gopher tortoise SSC Lepidochelys kempii Kemp's ridley sea turtle E E Crocodylus acutus American crocodile E E Pituophis melanoleucus mugitus Florida pine snake SSC Birds Ajaia ajaja Roseate spoonbill SSC Aphelocoma coerulescens Florida scrub jay T T Aramus guarauna Limpkin SSC Caracara plancus Audubon's crested caracara T T Charadrius alexandrinus tenuirostris Southeastern snowy plover T Charadrius melodus Piping plover T T Egretta caerulea Little blue heron SSC Egretta thula Snowy e gret SSC Egretta tricolor Tricolored heron SSC Eudocimus albus White ibis SSC Falco peregrinus tundrius Arctic peregrine falcon E
32 Table 12. Continued Scientific n ame Common n ame Federal s tatus State s tatus Falco sparverius paulus South eastern American kestrel T Grus canadensis pratensis Florida sandhill crane T Haematopus palliatus American oystercatcher SSC Haliaeetus leucocephalus Bald eagle T T Mycteria americana Wood stork E E Pelecanus occidentalis Brown pelican SSC Picoides borealis Red cockaded woodpecker E T Phyncops niger Black skimmer SSC Rostrhamus sociabilis plumbeus Everglades snail kite E E Speotyto cunicularia floridia Florida burrowing owl SSC Sterna antillarum Least tern T Mammals B larina brevicauda shermanii Sherman's short tailed shrew SSC Felis concolor coryi Florida panther E E Felis concolor Mountain lion T E Mustela vison evergladensis Everglades mink T Oryzomys palustris sanibelli Sanibel Island rice rat E SSC Podomys floridanus Florida mouse SSC Sciurus niger avicennia Big Cypress fox squirrel T Trichechus manatus latirostris Florida manatee (subspecies of the West Indian manatee) E E Sciurns niger shermani Sherman's fox squirrel SSC Ursus americanus floridanus Florida black bear T Fish Acipenser oxyrhynchus Atlantic sturgeon SSC T Centropomus undecimalis Common snook SSC Cyprinodon variegatus hubbsi Lake Eustis pupfish SSC
33 Table 13 Population estimates in 2000 and 2007 by County in the Caloosahatchee Region. (Adapted from US Census Bureau, 2008). County Census 2000 p opulation Estimated 2007 p opulation Charlotte 141,627 152,814 Glades 10,576 11,109 Hendry 36,210 39,611 Lee 440,888 590,564 Table 14 Land Use Percentages in the Caloos ahatchee Basin (SFWMD GIS data, 1998) Type Percent of b asin Urban and Built Up 15.8% Agriculture (including improved pasture) 38.7% Rangeland 5.9% Upland Forest 18.5% Water (including open bay) 2.6% Wetlands 16.2% Barren Lands 0.9% Tr ansportation, Communication, and Utilities 1.4%
34 Table 15 Top ten pesticides in Florida by total pounds of active ingredient (AI) reportedly applied per crop per year, adapted from Shahane, 2008. Pesticide a ctive i ngredient Total p ounds AI per c rop y ear Pesticide t ype Petroleum distillate 46,006,600 Insecticide Methyl bromide 9,242,742 Other pesticide Chloropicrin 3,501,581 Other pesticide 2,4 D 2,702,917 Herbicide Copper hydroxide 1,751,900 Fungicide Glyphosate 1,690,282 Herbicide Atrazine 1,478,834 Herbicide Chlorothalonil 1,435,833 Fungicide Sulfur 1,336,700 Insecticide Mancozeb 1,027,107 Fungicide Table 16 Pesticides reported on key crops of the Caloosahatchee area. Adapted from Shahane, 2008. Sugarcane Pasture Citrus (oranges ) Citrus (grapefruit) 2,4 D 2,4 D 2,4 D 2,4 D Ametryn Dicamba BEE Abamectin Asulam Hexazinone isopropylamine salt Aldicarb Atrazine Triclopyr Abamectin Azoxystrobin Halosulfuron Acephate Bacillus subtilus Metribuzin Aldicarb Basic copper sulfa te Pendimethalin Azoxystrobin Benomyl Bacillus subtilus Bromacil Bacillus thuringiensis (Bt) Carbaryl Basic copper sulfate Carbofuran Benomyl Chlorpyrifos Bromacil Copper chloride hydroxide Butoxyethyl triclopyr Copper hydroxide Capt an Copper oxychloride Carbaryl Copper sulfate Carfentrazone ethyl Dicofol Chlorothalonil Diflubenzuron Chlorpyrifos Diuron Clomazone Ethion Copper ammonium complex Fenbuconazole Copper hydroxide Fenbutatin oxide
35 Table 16 Continued Citrus (oranges) Citrus (grapefruit) Copper oxychloride Fenpropathrin Copper resinate Ferbam Copper sulfate Fosetyl al Dicofol Glyphosate ammonium salt Diflubenzuron Glyphosate isopropylamine salt Diuron Harpin protein Ethion H ydrogen peroxide Fenamiphos Imidacloprid Fenbuconazole Malathion Fenbutatin oxide Mefenoxam Fenpropathrin Norflurazon Ferbam Paraquat Fosetyl al Petroleum distillate Glyphosate Petroleum oil Glyphosate ammonium salt Phosmet Imida cloprid Phosphorous acid Mefenoxam Propargite Norflurazon Pyraclostrobin Oxamyl Pyridaben Paraquat Pyriproxyfen Petroleum distillate Sethoxydim Petroleum oil Simazine Pyridaben S Methoprene Pyriproxyfen Sulfosate Sethoxydim Sulfu r Simazine Thiophanate methyl S Methoprene Trifloxystrobin Spinosad Sulfosate Sulfur Triclopyr Trifluralin
36 Table 17. Herbicide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003 Grapefruit Oranges Tangelos Tangerines Temples Glyphosate Glyphosate Glyphosate Glyphosate Glyphosate Diuron Diuron Diuron Diuron Diuron Simazine Simazine Norflu razon Norflurazon Simazine Norflurazon Bromacil Simazine Bromacil Norfl urazon Bromacil Norflurazon Bromacil Simazine Bromacil Sulfosate Sulfosate Paraquat Paraquat Sulfosate 2,4 d Paraquat 2,4 d Sulfosate Clopyralid Diquat Sethoxydim Oryzalin 2,4 d Mcpa Paraquat 2,4 d Sulfosate Diquat Oryzalin Sethoxydim Diquat Thiazo pyr Oryzalin Paraquat Thiazopyr Oryzalin Thiazopyr Thiazopyr Triclopyr Pendimethalin Thiazopyr Triclopyr
37 Table 18. Insecticide use in Florida citrus ranked in order by active ingredient applied per crop year. A dapted from Shahane, 2003. Grapefruit Oranges Tangelos Tangerines Temples Petroleum distillate Petroleum distillate Petroleum distillate Petroleum distillate Petroleum distill ate Sulfur Sulfur Sulfur Sulfur Sulfur Ethion Aldicarb Ethion Ethion Eth ion Chlorpyrifos Carbaryl Fenbutatin oxide Fenbutatin oxide Chlorpyrifos Aldicarb Ethion Chlorpyrifos Pyridaben Fenbutatin oxide Fenbutatin oxide Dicofol Dicofol Diflubenzuron Dif lubenzuron Dicofol Chlorpyrifos Pyridaben Chlorpyrifos Pyridaben Pyridab en Diflubenzuron Carbaryl Albamectin Aldicarb Diflubenzuron Pyridaben Diflubenzuron Aldicarb Carbaryl Abamectin Albamectin Abamectin Carbaryl Carbofuran Acephate Fenpropathrin Ald icarb Carbofuran Dicofol Bt S methoprene Fenpropathrin Dicofol Fenpropath rin Carbaryl Acephate Imidacloprid Fenpropathrin Imidacloprid Carbofuran Bifenthrin Neem oil Imidacloprid Propargite Fenoxycarb Bt Propargite Lindane S methoprene Fenpropathrin F enbutatin oxide S methoprene Methidathion Imidicloprid Fenoxycarb Neem oil Methidathion Imidicloprid Propargite Neem oil Methidathion S methoprene Propargite Permethrin S methoprene Neem oil Spinosad Oxamyl Pro pargite Spinosad
38 Table 19. Fungicide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003. Grapefruit Oranges Tangelos Copper hydroxide Copper hydroxide Copper hydroxide Cop per chloride hydroxide Basic copper sulfate Copper sulfate Basic copper su lfate Mefenoxam Basic copper sulfate Ferbam Fenbuconazole Azoxystrobin Copper sulfate Benomyl Benomyl Fosetyl al Azoxystrobin Copper ammonium complex Fenbuconazole Basic copper zinc sulfate Copper chloride hydroxide Mefenoxam Copper ammonium complex C opper oxide Benomyl Copper chloride hydroxide Copper oxychloirde sulfate Azoxystrobin Copper oxychloride sulfate Fenbuconazole Copper ammonium complex Copper oxychloride Ferbam Copper oxide Copper sulfate Fosetyl al Copper oxychloride sulfate Ferbam M ancozeb Flutolanil Fosetyl al Maneb Maneb Iprodione Mefenoxam Metalaxyl Maneb Phosphorous acid Phosphorous acid Metalaxyl Streptomycin Phosphorous acid Tangerines Temple Copper hydroxide Copper hydroxi Copper sulfate Ferbam Ferbam Benomyl Fosetyl al Azoxystrobin Azoxystrobin Basic copper sulfate Fenbuconazole Copper chloride hydroxide Basic copper sulfate Copper oxychloride sulfate Benomyl Co pper sulfate Copper ammonium complex Fenbuconazole Copper chloride hy droxide Maneb Copper oxychloride sulfate Mefenoxam Maneb Mefenoxam Phosphorous acid
39 Table 110. Other pesticide use in Florida citrus ranked in order by active ingredient applied per crop year. Adapted from Shahane, 2003. Grapefruit Oranges Tangelos Tangerines Temples Gibberellic acid Ammonium soap Giberellic acid Giberellic acid Giberellic acid Harpin protein Giberellic acid Harpin protein Harpin protein Zinc phosphide Harpin protein NAA PT807 hcl
40 Table 111. Pesticide application in Florida citrus: percent of area with pesticide applied, average number of applications per season, and average application rate in kilogram s of active ingredient per hectare per application. Adapted from Shahane, 2003. Grapefruit Herbicides Insecticides Fungicides Percent area applied 94 97 N/A Number of Applications per season 1.1 2.8 1.0 1.9 1.0 3.9 Application Rate ( kg/hectare ) 0.99 1.96 0.01 31.25 0.26 8.45 Oranges Herbicides Insecticides Fungicides Percent area applied 95 91 61 Number of Applications per season 1.0 2.6 1.0 2.2 1.0 1.8 Application Rate (kg/hectare) 0.22 2.09 0.007 41.15 0.15 2.67 Tangelo Herbicides Insecticides Fungicides Percent area applied 79 86 71 Number of Applications per season 1.0 2.6 1.0 2.3 1.0 2.7 Application Rate (kg/hectare) 0.37 2.44 0.012 35.83 0.25 2.04 Tangerine Herbicides Insecticides Fungicides Percent area appli ed 95 94 77 Number of Applications per season 1.0 2.3 1.0 2.1 1.0 2.3 Application Rate (kg/hectare) 0.18 2.66 0.01 36.13 0.12 7.65 Temple Herbicides Insecticides Fungicides Percent area applied 93 98 66 Number of Applicatio ns per season 1.3 3.0 1.0 2.3 1.0 2.9 Application Rate (kg/hectare) 0.89 2.61 0.01 32.68 0.28 8.64
41 Table 112. Land use by hectares and percentage of the Lake Okeechobee Watershed. Adapted from James & Zhang, 2008. Entire w atershed Land u se Hecta re s Percentage Citrus 94 95 1 7% Dairy 9 07 8 1% Improved pasture 272 902 20% Natural areas 518 915 37% Ornamentals 1 89 7 <1% Other areas 11 15 6 1% Row crops 9 371 1% Sod farms 15 81 6 1% Sugarcane 161 757 12% Tree plantations 20 10 8 1% Unimproved pa sture 56,75 7 4% Woodland/rangeland 74 616 5% Urban 149 282 11%
42 Figure 11 Geopolitical map of the Caloosahatchee Basin (FDEP, 2003) Figure 12 Caloosahatc hee Area Counties
43 Figure 13. Citrus harvest season ( FASS, 2007 ).
44 Figure 14 Planting and harvesting seasons of selected Florida field crops (FDACS, 2007).
45 CROP Usual Planting Dates 1/ Usual Harvesting Dates Begin Mo st Active End JUL AUG SEP OCT NOV DEC JAN FEB MAR AP MAY JUN JUL 2/ Snap Beans Blueberries Cabbage Carrots Cantaloupes Celery Sweet Corn C ucumbers Eggplant Escarole/Endiv e Lettuce/Romain e Peppers Potatoes Radishes 3/ Squash Strawberries Tomatoes Watermelon JUL AU SEP OCT NO DEC JAN FEB MA APR MA JUN JUL 1/ Usual date direct seeded or transplanted. 2/ Includes pole beans. 3/ A small acreage of summer squash is marketed locally during July and August. Figure 15 Planting and harvesting seasons of selected Florida vegetables (FDACS, 2007).
46 Figure 16. Lemna minor in culture during a laboratory experiment, control culture.
47 CHAPTER 2 SPATIAL AND TEMPORAL VARIABILITY OF PESTICIDES AND PESTICIDE DEGRADAT ION PRODUCTS OF THE CALOOSAHATCHE E RIVER, FLORIDA Background Much research has demonstrated that pesticides and pesticide degradation product s are present in worldwide surface waters and across the United States (Gao et al.2008, Gilliom et al., 2007, Hively et al., 2011, Konstantinou et al., 2006; Laabs et al., 2002, McConnell et al., 2007, NinoDeGuzman et al., 2012, Whitall et al., 2010, Wilson and Boman, 2011, Wilson et al., 2012, Wilson et al., 2007). There are concerns regarding the potential effects of such compounds on human health and the vitality of nontarget ecosystems ( EPA 2004, WHO 2010, 2011). In the United States, the overall intensity of pesticide application for agricultural production was determined to be greatest in the croplands of the Corn Belt, the Mississippi River Va lley, Florida, the coastal plain of the Southeast and MidAtlantic states, and irrigated areas of the West (Gilliom et al., 2007). In addition, pesticides are also used for nonagricultural purposes such as for management of turf grasses and landscaping, golf courses, public parks, and residential areas for mosquito and termite control and for roadway management. Gilliom et al. (2007), reported agricultural and urban areas have similar numbers of detectible pesticides in surface waters. The multiple land u ses draining into surface water bodies often complicates determination of sources of pesticides in nontarget aquatic systems. The multiple uses for pesticides described above are particularly relevant in south and southwest Florida, where the subtropical climate, long growing season, and application frequency of pesticides is complicated by the mixeduse nature of the state: intense agriculture in close proximity to sprawling urban development and sensitive
48 ecosystems like the Everglades (Carriger & Rand, 2008). This is also true of the present study area, the Caloosahatchee River. The Caloosahatchee River stretches westward from Lake Okeechobee to San Carlos Bay and the Gulf of Mexico in Southwest Florida. The Caloosahatchee watershed drains an area of about 3700 km2 (Flaig & Capece, 1998) within four counties of Southwest Florida. The entire watershed contains five drainage basins (East Caloosahatchee, West Caloosahatchee, Estuarine Caloosahatchee, Telegraph Swamp, and Orange River), as well as large population centers (including Fort Myers, Cape Coral, North Fort Myers, and others), one national wildlife refuge, two state aquatic preserves, and one wildlife management area (FDEP, 2003). Water releases from Lake Okeechobee occur through a series of locks on the Caloosahatchee River (C43) when lake levels exceed the US Army corps of Engineers criteria for flood protection or to flush algae blooms and salt water out of the river to protect potable water supplies from contamination (Flaig & Capece, 1998). T his management results in much variation in flow rates and volumes along the river throughout the year. The Moore Haven Lock (S 77) separates Lake Okeechobee from the beginning of C43 on the eastern side, and the Franklin Lock (S 79) separates C43 from the salinity of the estuary on the western side (Figure 21). The Caloosahatchee River is the major source of surface water supply for much of the southwest coastal region; it is used for agricultural irrigation, to recharge shallow wellfields, and to provide a potable water supply for Lee County and the city of Fort Myers (FDEP, 2003; SFWMD, 2000).
49 Numerous studies have shown the presence of pesticides in south Florida surface waters, some linking agricultural usage to the predominant pesticides present (Car riger & Rand, 2008; Miles & Pfeuffer, 1998; HarmanFetcho et al., 2005; Wilson and Bowman, 2011; Wilson et al., 2012; Wilson et al., 2007). Frequently, herbicides have been the most common pesticides found in Florida surface waters regardless of the most c ommon adjacent crop (Pfeuffer &. Rand, 2004; HarmanFetcho et al., 2005; Wilson and Bowman, 2011). Pesticides have also been detected in the tissues of organisms, such as oysters, that grow in estuaries fed by these surface waters (Volety, 2008), leading t o speculation about the potential toxicity of pesticides in surface waters to aquatic organisms (Carriger & Rand, 2008). In order to assess the potential risks of pesticides to aquatic organisms within a given waterbody, information is needed to identify w hich pesticides aquatic organisms are likely to encounter. Unfortunately this information is often lacking. This study was developed to identify the presence or absence of 39 pesticidal active ingredients and/or degradation products in the Caloosahatchee R iver (Florida) over an 18 month period. Results are useful for characterizing the temporal and spatial variation of pesticides in surface water along the length of the river. This information is needed to identify pesticides that may be of concern to aquat ic resources in the river. Materials and Methods Five water sampling sites were selected along the Caloosahatchee River, with three on the eastern Caloosahatchee reach and two within the estuary (Figure 21). Sites A, B and C were located downstream of exi sting water control structures maintained by the South Florida Water Management District and/or the US Army Corp of Engineers (Franklin, Ortona, and Moore Haven locks, respectively; Figure 21). Sites
50 D and E were located in the estuary using GPS; Figure 21. Water samples were collected during 18 monthly sampling trips between December 2004 and April 2006. Water was collected at a depth of 1 m using a stainless steel bucket and brought to the lab in a pre cleaned, 20L stainless steel metal canisters (Pepsi, Hyatsville, MD) with airtight lids. Samples were stored on ice until processing (<5 hours). Water pH, salinity, conductivity, dissolved oxygen, and temperature were recorded at each site using a precalibrated, multifunctional, portable probe (model 556 MPS, YSI Environmental, Yellow Springs, OH). For quality control, duplicate samples were collected in the same manner described during each sampling trip to evaluate method and matrix spike recoveries. Blank samples consisted of distilled water placed in precleaned stainless steel canisters identical to those used for sample collection. Water from the stainless steel collection canisters was filtered using an 1100 gallon per hour (GPH) submersible marine pump (Rule Industries, Gloucester, MA) connected directly to two in line, highpressure, stainless steel filter holders (Millipore, Bedford, MA) housing a 2.7 m pore size GF/D glass fiber prefilter (Whatman, Middlesex, UK) and a 0.7 m pore size GF/F filter (Whatman), respectively, with Teflon and stai nless steel tubing. The filtered water was pumped into a precleaned 2 L glass graduated cylinder. Using these graduated cylinders, exactly 4 or 8 L from each of the original water samples was measured into new precleaned stainless steel canisters for e xtraction (<12 hours). In the upper portions of the river (sites A, B, and C; Figure 21 ), 4L were filtered per sampling site. In the estuarine portion of the river (sites D and E; Figure 21 ), 8L were filtered per sampling site. These volumes were chosen based on five previous runs of the sampling protocol prior to beginning this study. Filters were
51 changed and all filtration equipment was cleaned between each sample by pumping several liters of a 1:1 organic free water/methanol mixture through the entire filtration system to avoid crosscontamination. Field blanks included pumping 4 or 8 liters of organic free water through the sampling and filtration system into a clean stainless steel canister to evaluate the equipment for contamination. Replicat e samples were collected at alternated sites (randomly selected) on each sampling trip. These were handled as all of the other samples to evaluate precision of the methods. Prior to extraction, all samples were fortified with an extraction efficiency sur rogate standard, diazinon [diethyl d10] (Cambridge Isotope Laboratories, Andover, MA). Each sample was then drawn by vacuum through an ENV+ extraction resin solidphase extraction (SPE) cartridge containing 200 mg of hyper crosslinked styrenedivinylben zene copolymer (Argonaut Inc., Redwood City, CA). After extraction, the cartridges were dried with high purity nitrogen gas and eluted with 6 mL of dichloromethane followed by 9 mL of 3:1 acetone/acetonitrile (Fisher Scientific, Fair Lawn, NJ). All solve nts were chromatographic grade. The 15 mL extract was concentrated to a final volume of 0.5 mL under highpurity nitrogen. Internal standards, atrazine [ethylamined5] (Cambridge Isotope Laboratories) and PCB 204 (2,2,3,4,4,5,6,6 octachlorobiphenyl) [A ccuStandard, New Haven, CT), were added to the final extracts and standards. Sample extracts were analyzed by gas chromatography mass spectrophotometry (GC MS), using the methods described by HarmanFetcho et al. (2005). Full scan spectra were acquired using a Varian 3800 GC coupled to a Saturn 2000 Ion Trap MS equipped with a DB 17MS (Agilent Technologies, Inc., Palo Alto, CA)
52 capillary column (30 m, 0.25 mm i.d., 0.25 m film thickness). The GC inlet was operated in splitless mode. The carrier gas was ultrahighpurity helium at a constant flow rate of 1.0 mL/min. Operational temperatures were as follows: injection port, 260C; oven, 130C (1 min), increasing to 280C at 5C/min; transfer line 280C; and ion trap, 220C. The ion trap MS was operated i n electron impact (EI) selective ion storage (SIS) mode, scanning for ions with masses of 70450. To allow for better sensitivity to halogenated analytes, a Hewlett Packard (HP) 5890 GC coupled to a HP 5989A quadrapole MS in the negative chemical ionizati on (NCI) mode was used (HarmanFetcho at al. 2005). The GC inlet was operated in splitless mode with a column identical to that used on the Varian GC MS described above. The carrier gas was ultrahigh purity helium at a constant flow rate of 1.12 mL/min. Operational temperatures were as follows: injection port, 280C; oven, 130C (1 min), increasing 280C at 6C/min; transfer line 280C; source, 150C; and quadropole, 100C. The NCI reagent gas was methane and the source pressure was 1.6 Torr. The ins truments were calibrated using a mixture of analytes with at least five different concentrations across the expected sample range. Calibration curves were repeated after every 20 sample injections. Sample results were quantified using the internal standard method. In laboratory experiments using fortified organic carbon free water, the extraction technique was reported to be effective at isolating the target compounds included in this study (Lehotay, et al., 1998). Using the same method, greater than 80% of spiked pesticides were recovered from water samples (n=11) (Liu et al., 2002). Even in the case where lower average recovery was observed using this
53 technique (such as for aldrin, fipronil, and p,p DDE), most data points were still within the acceptable range suggested by the EPA standard methods (80120%) (HarmanFetcho et al., 2005; EPA, 2004). Field blank samples using this technique were shown to be devoid of compounds at greater than the minimal detection limits and recovery of the surrogate com pound diazinon [diethyl d10] has been measured reliably in all sample extracts, blanks, spikes, and replicates (HarmanFetcho et al., 2005). A summary of the pesticides and pesticide degradation products analyzed in this study, average recoveries, method detection, practical quantitation limits, mass ions monitored and instrument mode are shown in Table 21. The pesticides analyzed include fungicides, herbicides and insecticides. In the previously published studies, DI was used for the field spikes (sur rogate and target analytes) (HarmanFetcho et al., 2005). In the current study, water from the river was used for spikes instead of dI water. As a result, the variability in matrix spike recoveries were much greater, possibly due to interactions of salini ty and dissolved/suspended materials, but most likely associated with variations in the river conditions (i.e. flow) at the time of sampling. Microsoft Excel (including the XL Toolbox Version 4.01), Sigmaplot 12.0, and SigmaStat 4.1 were used for statisti cal calculations and analyses such as mean, median, standard deviation, variance, and univariate statistics like analysis of variance (ANOVA) and associated post hoc analyses. A twoway ANOVA was used to determine patterns in the spatial and temporal detec tions of the ten most commonly detected compounds by site and month. To identify pesticide detection similarities at each site or sampling month, oneway ANOVAs were employed. Oneway ANOVAs were used to
54 explore the temporal characteristics of the two mos t commonly detected compounds (aytrazine and metolachlor) more thoroughly by month. The Holm Sidak or Tukey's method was used for pair wise comparisons in all post hoc analyses. Due to the large number of samples and target analytes, comparisons between al l of the analytes (detection frequency and concentration) in the entire data set across sites and months required use of nonparametric analyses. These analyses were performed using Primer software (version 6.1.12), using interpretation from Clarke and Warw ick (2001) and Clarke and Gorley (2006). A twoway crossed analysis of similarity (ANOSIM no replicates test ) test based on the Spearman rank correlation method was performed to compare concentrations of all target analytes (transformed by square root) ac ross all sampling si tes and times A similarity profile test (SIMPROF) was conducted within an hierarchical clustering of all of the target analytes that were detected at least once to determine if they differed in multivariate structure. Pesticide occurrence data was transformed to presence/absence data for some analyses by ANOSIM and SIMPROF to evaluate the similarity in detections without potential bias from the actual concentrations detected. Results The physical locations of the sampling sites as well as the environmental parameters measured during the sampling period are summarized in Table 22 The pH and temperature were relatively stable at each site, but patterns were observed for salinity and conductivity relative to sites. Salinity was highest at Site E, closest to the tidal influx at the mouth of the estuary, with site D being intermediate and the river sites (A,B,C) having lower salinities.
55 The flow of the Caloosahatchee River varies dramatically based on water management needs (Table 23). This highly variable flow of water and dissolved constituents from the surrounding watershed is likely responsible for the broad range of salinity and conductivity at each site. The range of temperatures likely reflects seasonality, since the sampling period spanned December 2004April 2006. Summary of Pesticide Concentrations Target analytes included some of the most frequently used pesticides in this region as well as some persistent legacy pesticides that are no longer used. Results for the 7590 water samples analyzed during the sampling period from December 2004 to April 2006 are summarized in Table 2 4. The top ten most commonly detected pesticides were atrazine ( 99%), metolachlor (95%), CIAT (81 %), CEAT (75 %), simazine (60%), ametryn (47%), malathion ( 40%), chlorpyrifos oxon (26%), dieldrin (17 %), and chlorothalonil (11%). The top six detected by frequency were herbicides, followed by three insecticides and one fungicide. Other compounds were detected at lower frequencies. Twenty pesticides were never detected. Concentrations for pesticides analyzed are summarized using the box and whisker plots in Figure 2 2. Only five compounds were detected in more than 50% of the samples: atrazine, metolachlor, CIAT, CEAT, and simazine. These were all herbicides. No insecticide was found in more than 40% of the samples, and only five targeted insecticides were detected more than once (heptachlor, diazinon, dieldrin, malathion, and chlorpyrifos oxon). The most prevalent insecticide detected was malathion, with a detect ion frequency of 40%, a maximum concentration of 31 ng/L, a minimum concentration of 1.6 ng/L and a median concentration of 3.5 ng/L (%RSD: 120). The only fungicide
56 detected was chlorothalonil, which was present in 11% of the samples, (maximum 11 ng/L, minimum 2.4 ng/L, and median 4.0 ng/L (%RSD: 55)). According to the WHO (2010), the use and production of aldrin, DDT, dieldrin, and heptachlor has been prohibited or severely restricted by the Stockholm convention on persistent organic pollutants since 2004. However, the DDT degradation products DDE & DDD were each detected once, aldrin was detected once, heptachlor was detected twice, and dieldrin was detected in 15 separate samples (17% detection frequency, 9th most commonly detected pesticide in this study ). Ethoprop, considered an extremely hazardous active ingredient by the WHO, was found in one sample. Of the top ten most frequently detected compounds, two are considered to be unlikely to cause acute hazards (simazine and chlorothalonil), five are considered slightly hazardous (atrazine, CEAT, CIAT, metolachlor, and malathion), two are considered to be moderately hazardous (ametryn and chlorpyrifos oxon) and one (dieldrin) is regulated by the Rotterdam Convention on Prior Informed Consent and restricted by the Stockholm Convention on Persistent Organic Pollutants (WHO, 2010). The herbicides with the three highest median concentrations in this study were atrazine (72 ng/L), CIAT (21 ng/L), and metolachlor (18 ng/L). The insecticides with the three highest median concentrations in this study were 4,4'DDD (26 ng/L), diazinon (15 ng/L), and methoxyclor (15 ng/L). The compounds detected with the highest reported toxicities included four prohibited or severely restricted pesticides (WHO 2010): dieldrin, heptachlor, aldrin, and mirex. Spatial and Temporal Comparisons Significant differences between sampling sites were observed based on the ANOSIM analysis of pesticide concentrations (Global rho = 0.17; p=0.03). However,
57 there were no significant differences between sample sites based on pair wise comparisons due to the very low rho values even though the global tests allow rejection of the null hypothesis that all sites were the same. This is more likely due to the high variability at each site, not the differences in pesticide concentrations detected and number of compounds detected per site. For example, the samples from Site A contained the highest total combined concentration of all pesticides over all months (6400 ng/L),with an average of 6.1 detections per month. The samples from Site D contained a total combined concentration of all pesticides over all months of 1700 ng/L (about one third of the Site A value) but with an average of 7.2 detections per month (higher than Site A). Significant differences between sampling months were observed based on the ANOSIM analysis (Global rho = 0.29; p<0.001). However, no pattern was observed relating the sampling months, as shown in the SIMPROF results (Figure 23). Some obvious outliers included the sample at Site E in J une 2005, May 2005, and January 2006 (p<0.001), which yielded fewer than the normal number of detections and concentrations, as well as unusual detections, such as the only detection of ethion in June 2005. The overall relationship between sampling site and sampling month is shown using a 2dimensional ordination of nonmetric multi dimensional scaling (MDS) ( Figure 2 4 ) Within this figure, points that are close together represent samples that are similar in target analyte composition and concentration, and points far apart correspond to very different samples based on value and variance in the samples The samples are shown with the percentage of similarity (20%, 40%, 60% or 80%) superimposed from a
58 hierarchical cluster analysis of the data to illustrate similarities between samples. Data from site E f o r June 2005 were removed to allow visualization of the other individual points Results from the outlier Site E masked the other results to the point where all other points appeared as one in the overall MD S. The similarities illustrated in the MDS and cluster analyses confirm some patterns seen in the data. For example, March 2005 exhibited the highest combined pesticide concentrations (6200) but close to the median frequency of compound detections (26) Sa mples from all of the sites in this case were at least 60% similar Total pesticide concentrations of samples from December 2004 and February 2005 were 80% similar During both of these months a relatively low number of compounds were detected (18 in bot h cases) but combined pesticide concentrations (5788 and 4083 ng/L, respectively) were high relative to the other months T he high degree of similarity between July 2005, December 2005, and January 2006 can be partially attributed to the lower number of c ompounds detect ed (23,20, and 17, respectively), but more likely to their low combined pesticide concentrations (380, 3 60, and 400 ng/L, respectively). There wa s a high degree of similarity between the Site E samples from August 2005, October 2005, November 2005 and December 2005 (80%). However, this grouping is also part of the larger grouping with 60% similarity containing all sample sites and at least one sample from every month except March 2005. Site Comparisons The ten compounds detected most frequent ly (atrazine, metolachlor, CIAT, CEAT, simazine, ametryn, malathion, chlorpyrifos oxon, dieldrin, and chlorothalonil) were compared across all sampling sites using a twoway ANOVA. Concentrations of these ten compounds at all sampl ing sites over all sampli ng months are summarized in
59 Figure 25 Differences between concentrations and detection frequencies by site are highlighted in Table 25 N o significant differences were observed between sampling However, significant differences were observed between the c p=<0.001; power=0.999). A trazine concentrations were significantly greater than all other compounds (p<0.05) except for chlorothalonil (p=0.336) based on the post hoc analyses using the Holm Sidak method. This lack of difference is likely due to their high variances more than similarity in their concentrations sinc e the median atrazine concentration was almost 20x the median concentration of chlorothalonil N o other significant differences between any other compounds were observed. The lack of differences is probably due to the high variability in detection frequencies of most compounds. Only atrazine and metolachlor were detected in more than 90% of samples. A significant difference in the number of pesticides and analytes detected at each site was observed as shown in Figure 2 6. The samples from Site D contained the most cumul ative pesticide detections (79) over all sampling months, with an average of 7.2 detections per month. The samples from site E had the least number of detections (47) over all sampl ing months, with an average of 4.3 detections per month. Sites A, B, and C exhibited an intermediate frequency of pesticide detections (75,65, 79, respectively). S ignificantly more analytes were detected at site D than sites C or E, but no significant differences were observed between any of the other sites. Interestingly, site D was
60 located downstream from the sites with lower detection frequencies, indicat ing that not all compounds were entering the system upstream in the watershed. Temporal Comparisons Concentrations of the ten most frequently detected compounds (atrazine, met olachlor, CIAT, CEAT, simazine, ametryn, malathion, chlorpyrifos oxon, dieldrin, and chlorothalonil) were compared across all sampling months in a twoway ANOVA. A summary of the concentrations of these compounds at all sampling sites over all sampling months is shown in Figure 22. Significant differences were observed between concentrations were significantly higher than any of the other compounds based on the Holm Sidak method. Additionally, atrazine concentrations were significantly higher than metolachlor. Atrazine was detected in 99% of samples at a maximum concentration of 2900 ng/L. The closest compound in terms of detection frequency and maximum concentration detected was metolachlor (95% and 270 ng/L, respectively). Not surprisingly, mean concentrations of atrazine and metolachlor were significantly different since the median concentration of atrazine was more than three times that of metolachlor (atrazine median concentration was 72 ng/L and metolachlor median concentration was 18 ng/L). In contrast, dieldrin, while being one the ten most frequently detected pesticides, was present at significantly lower concentrations than the other compounds (17% detection frequency, and 1.5 ng/L was the maximum detected) (Table 2 5). Significant differences in pesticide concentrations were also observed between the significantly higher in March and Apr il of 2005 than in most other months based on the
61 Holm Sidak post hoc analysis. Samples from July 2005, December 2005, and January 2006 contained the lowest concentrations of pesticide. Oneway ANOVAs were conducted to compare the concentrations of the m ost commonly detected individual compounds across sampling months. While malathion (p=0.149), chlorothalonil (p=0.134), and dieldrin (p=0.550 concentrations were not significantly different between sampling months, chlorpyrifos oxon, ametryn, simazine, CEAT, and CIAT were all different (p<0.001). Concentrations of most compounds were higher in the spring and lower in December and July. The highest mean concentration of chlorpyrifos oxon (n=18) was detected in March 2006 (6.7 ng/L) and at the lowest mean concentration in January 2006 (1.3 ng/L). However, it was not present for 10 months (56%, December 2004 through July 2005, and November and December 2005). The highest mean concentration (6.2 ng/L) ametryn (n=15) was detected in March 2005, while the lowest m ean concentrations were detected in January 2006 (0.28 ng/L) and July 2005 (0.29 ng/L). No samples contained ametryn in December 2005. The highest mean concentration of simazine (n=15) was detected in March 2005 (9.0 ng/L) and at the lowest mean concentrat ion in July 2005 (0.67 ng/L). Simazine was detected at least once during every monthly sampling event. Similarly, CEAT and CIAT were detected in every monthly sampling event (n=15) at most sites. Both compounds were found in their highest concentrations in March 2005 (8.0 ng/L and 11 ng/L, respectively) and lowest concentrations in December 2005 (0.41 ng/L and 0.39 ng/L, respectively). Considering the two most commonly detected compounds, significant differences were observed between the concentrations of atrazine and metolachlor detected in
62 different sampling months (Figures 27 & 2 8, respectively), with higher concentrations of both occurring during the spring months (March through May). The number of pesticides detected in the samples also varied signi ficantly by p =0.0046). Samples from March 2005 had the highest numbers of analytes detected (average 8.7 detections per sample). Detections during m ost other months did not differ significantly from each other based on the Holm Sidak post hoc analysis method. Target Analyte Comparisons and CoOccurrence Atrazine and metolachlor, both herbicides, were detected the most frequently in the samples. Of 75 samples, there was only one (1.3%) that did not contain either atrazine or metolachlor. These two compounds were found together in 71 of 75 samples (95%). 73% of the 75 samples analyzed for all 42 analytes contained CIAT, CEAT and atrazine together. 72% contained CIAT, CEAT, atrazine and metolachlor together. Six compounds (CIAT, CEAT, atrazine, simazine, ametryn, and metolachlor) were detected together in the the same sample in 36% of the samples collected. A similarity profile test (SIMPROF) was conducted within an hierarchical clustering to compare all of the target analytes that wer e detected at least once to determine if they differed in multivariate structure. The results of the SIMPROF test are summarized in Figure 210, with solid black lines of the dendogram representing statistically significant differences between pesticides. Note that the ten most frequently detected compounds (atrazine, metolachlor, CIAT, CEAT, simazine, ametryn, malathion, chlorpyrifos oxon, dieldrin, and chlorothalonil) were separated by the first similarity comparison, being significantly different from th e other pesticides detected. The only frequently detected fungicide,
63 chlorothalonil, was separated at the second node of that group. This test would group the pesticides, based on concentrations, similarly to the way species are compared in taxonomic groupings. In this way, this figure can be used to describe how the pesticides co occur and behave chemically in the environment, or how often they are found together. These results confirm the observation that atrazine, metolachlor, simazine, CIAT, CEAT, and ametryn were frequently found together, as they remained a cohort in the SIMPROF results through three nodes and greater than 40% similarity. Considering that atrazine was found at higher concentrations than the other compounds (Figure 24), there was concern that even with transformation to smooth the data for such outliers, a bias may exist based on concentration. Although concentrations were all measured in ng/L, the toxicity of each compound is not equivalent (e.g. toxicity aldrin). To evaluate the similarity of detections without potential bias from concentrations, the data were transformed to a presence/absence matrix. Significant differences were again observed in the ANOSIM between sampling sites (Global Rho= 0.21, p=0.01, 999 permutations) and sampling months (Global Rho= 0.241, p=0.006, 999 permutations). However, pair wise comparisons by site revealed no significant differences. There also was no discernible pattern in the differences tested by month other than one sample from June 2005 that was an obvious outlier (Figure 211). Comparisons were also made including target compounds detected at least once (Figure 2 12). Compounds never detected were removed to make the figure more readable. While exclusion of the nondet ected pesticides did change the groupings of the top ten most commonly detected compounds from the original square root
64 transformation, it confirmed the relationship between the most commonly detected compounds represented in the raw data. For example, atr azine and metolachlor were most often found together, just as the raw data counts (without compound concentrations) indicate (Figure 212). Relating this back to the raw frequencies of detection, the ten most commonly detected pesticides are shown in the f irst column in Table 212. The percentage of times that compound is found with one of the other top ten compounds is broken down in 10% increments in subsequent columns. Discussion Results reported in this study are similar to those reported by SFWMD They summarized detections of atrazine in surface water (DBHYDRO, 2012) during the period from 19922007 at some of the same sites along the Caloosahatchee River (sites A, B, and C). They reported maximum concentrations of 8.5 g/l at site A, 2.2 g/l at site B, and 2.3 g/l at site C. While they did not analyze for metolachlor at those sites during the same time period, they did monitor for it at their citrus agricultural sampling locations (including our sites B and C) and reported a maximum concentration of 1.2 g/l (1200 ng/L) and median of 0.11 g/l (11.1 ng/L) Also during the same time period, they reported a maximum concentration of 1.4 g/l (1400 ng/L) and a median of 0.14 g/l (140 ng/L) at their Everglades Agricultural Area (EAA) sampling locations ( includes our Site A of present study). Differences between the sites along the river (Sites A, B, and C) and those in the estuarine portion of the river (D & E) were expected. Sites A C were located within more agricultural areas, while sites D E were in m ore urbanized areas. No significant differences were observed for pesticide detections between the sites. Intuitively, one
65 might expect that the landus es immediately adjacent to the sampling sites would most influence they types of pesticides possibly present in surface water samples. However, this does not appear to be the case in the present study. The homogeneity of pesticide presence at the sites may partially be explained by the drainage systems within the watershed. The majority of this watershed is drained using a complex network of ditches and drainage canals since there is very little elevation difference to facilitate drainage. These drainage systems deliver surface water to the Caloosahatchee River from a variety of mixed use land areas. The connection of the river to Lake Okeechobee further extends the drainage watershed to include the entire Lake Okeechobee watershed. As a result, there is a much higher probability for pesticides used in areas not adjacent to the river to ultimately enter into it. Given this complexity, it is nearly impossible to pinpoint exact sources of the presence of pesticides in this river system without more intensive monitoring of inflows from specific drainage basins Unexpectedly, no clear differences were observ ed in the types of pesticides detected by season, which might be expected as agricultural activity changes throughout the year (Figure 26). While pesticide detection frequency was highest in March, detections during most other months were not significantl y different. This may also be expected based on the lack of differences between the sampling sites. If agricultural or urban usage drives seasonal fluctuations of pesticide concentrations in surface waters, a lack of significant differences between agricul tural and urban sampling sites supports the finding that there is not strong seasonality in the numbers of analytes detected. However, the magnitude and scale of the drainage system (far removed from the areas adjacent to the river), the persistence of the chemicals, the patchiness of use-
66 patterns, and the unpredictable but dramatic change in flow to meet water management needs are likely more responsible for the lack of seasonality here. All of these diverse factors combine to create a well mixed system. Pesticide residues in such a system would not likely follow the seasonal patterns observed elsewhere (HarmanFetcho et al., 2005; Hampson et al., 2000; Cater et al., 1995; Thurman et al., 1991). Atrazine and metolachlor were the most frequently detected compounds. This is in agreement with national data compiled by the USGS, which found that atrazine and metolachlor were the top two compounds detected most frequently in water (Gilliom et al., 2007). The lack of significant differences between sampling sites for both atrazine (p=0.46) and metolachlor (p=0.113) reinforces the finding that pesticide usage in the immediate area is not a clear predictor of the types of pesticides found in the surface waters nearby. While both atrazine and metolachlor have been previously reported in higher concentrations closer to areas with more agricultural land use (McConnell et al., 2007), in this study, they were both equally likely to be present in the river from Lake Okeechobee to the estuary. The inputs of some of the pes ticides may have occurred at the farthest upstream location, resulting in detections at downstream locations. However, this is not the case for all of the pesticides since the total number of different pesticides detected was greatest at Site D, located downstream of sites with lower numbers of compounds detected. During six of the eleven sampling months that included the estuarine, urban sites, Site D had the highest number of compounds detected of all five sites supporting the supposition that some of the compounds originated locally.
67 Interestingly, the fungicide chlorothalonil was only detected in samples from sites D and E, never being detected along the river near the more agricultural sites A, B, and C. However, this finding is not statistically signif icant because of the low detection frequency. No significant differences were observed between any of the sites (n=18, chlorothalonil nor malathion were detected frequently enough to allow comparison of concentrations by month. Significant differences in monthly concentrations were observed for atrazine and metolachlor. Atrazine concentrations tended to increase from January to May. Metolachlor concentrations did not consistently increase or decrease throughout the monitoring period, but instead spiked during the spring, August and November. Both atrazine and metolachlor are labeled for agricultural and nonagricultural uses (e.g. turfgrass and roadside management).The lack of seasonality may also be partially due to the use of these herbicides throughout the year for these nonagricultural uses. In addition to the widespread use, the half life of these compounds may have also contributed to their presence throughout the monitor ed period. The half life for metolachlor in aquatic systems in natural light is estimated to be 70 days (U.S. EPA, 1995) and atrazine has been shown to remain in aquatic systems for at least as long (EPA, 2003). As a result, detections in this study may have also been herbicide residues from a previous application. None of the other frequently occurring pesticide concentrations were significantly different between sampling sites (CIAT, p=0.21; CEAT, p= 0.16; simazine, p=0.63; ametryn, p=0.35; chlorpyrifos oxon, p=0.41; chlorothalonil, p=0.88; malathion, p=0.39;
68 dieldrin had too few samples to run the test; no transformation would allow it to pass the system is well mixed; th e areas immediately adjacent to a sampling site do not indicate the types of pesticides found there well. No significant differences between sampling month (p=0.55, p=0.13, p=0.15, hion. observed for CIAT, CEAT, simazine, ametryn, and chlorpyrifos oxon. However, the patterns were not always easily seen. CIAT had highest concentrations in the winter and spring, while CEAT and ametryn peaked in spring without the higher concentrations in winter. Simazine concentrations were highest during both sampling events of April 2005. Chlorpyrifos oxon concentrations were higher in fall, late winter, and early spring, wit h lower concentrations in mid winter, late spring and summer. This lack of a clear pattern corresponding to a particular land use or crop also supports the finding that this mixed system might be influenced by pesticide residues from elsewhere rather than only adjacent land use. An interesting aspect of results presented here is that when all 42 analytes were evaluated, between two and twelve compounds were always detected (n=75). The average number of detectable pesticides in field samples was 5.6, and the median number of analyte detections was 6, with a percent relative standard deviation of 38.5%. Considering the two compounds detected most frequently, 71 of 75 samples had both atrazine and metolachlor present (94.7%). While this study only included mont hly
69 sampling, repeated detections of these pesticides suggest that the compounds are there frequently. These results showed that six herbicides (atrazine, metolachlor, CIAT, CEAT, simazine, and ametryn) occurred together much of the time. It should be not ed that CIAT and CEAT are degradation products of atrazine, so they are related in that if increased atrazine concentrations are found, one might reasonably expect that CEAT & CIAT concentrations would increase as well. However, the data suggests that organisms living in these regions of the Caloosahatchee River are frequently exposed to several pesticides at one time, even if they are found at low concentrations. Current toxicological studies generally analyze compounds individually, but our data suggests that this may not likely be relevant to a system where every sample contained more than one compound.
70 Table 2 1. Pesticides and Pesticide Degradation Products Measured with associated average spike recoveries, percent relative standard deviation (%RSD) and method detection limits (MDL). Compound Average % Recovery %RSD MDL (ng/L) Instrument mode Mass ions monitored (m/z)h Acetochlor 1 20 40 1.5 EI 162, 174, 223 225 Alachlor 12 0 43 1.4 EI 160 188 Aldrin 68 66 0.2 NCI 237 330, 332 Ametryn 13 0 42 1.4 EI 212, 227, 229 Atrazine 10 0 64 0.9 EI 173, 200 215 CEAT a 1 30 37 0.9 EI 173 158, 145 Chlordane 52 63 0.2 NCI 408, 410 412 Chlordane 36 65 0.2 NCI 408, 410 412 Chlorothalonil 91 60 0.2 NCI 264, 266, 268 Chlorpyrifos 97 55 0.2 NCI 214, 313 315 Chlorpyrifos oxon 1 20 57 0.2 NCI 298 297, 299 CIAT b 1 30 29 0.8 EI 172 187, 174, 145 Cyanazine 1 40 27 1.8 EI 198, 212 225 4,4' DDD c 80 59 0.2 NCI 248 250, 320 4,4' DDE d 41 75 0.2 NCI 281 316, 318, 320 4,4' DDT e 60 49 0.2 NCI 71, 248 318 Diazinon f 1 10 56 1.0 NCI 169 303 p,p' Dicofol 90 39 1.5 EI 111, 139 Dieldrin 1 20 49 0.2 NCI 2 37, 346 380 Endosulfan 72 52 0.2 NCI 406 408, 410 Endosulfan 72 66 0.2 NCI 404, 406 408 Endosulfan s ulfate 85 70 0.2 NCI 384, 386 388 Ethion 1 20 44 1.5 EI 203, 231, 338 Ethoprop 1 20 31 1.5 EI 203, 231, 338 Fipronil 34 53 0.2 NCI 331, 384 400 HCH g 74 70 0.2 NCI 71 255, 257 HCH g 83 59 0.2 NCI 71 255, 257 Heptachlor 47 61 0.2 NCI 266 300, 232 Malathion 1 20 57 1.6 NCI 157, 172 Methoxychlor 12 0 37 1.8 EI 114, 152, 228, 344 Metolachlor 12 0 39 1.0 EI 162 238 Metribuzin 1 20 44 1.5 EI 198 Mirex 16 46 0.2 NCI 334, 370 404, 439 cis Nonachlor 55 45 0.2 NCI 442, 444 446 trans Nonachlor 76 60 0.2 NCI 442, 444 446 Pendamethalin 11 0 80 1.6 EI 252 Phorate 79 36 1.5 EI 121, 231 Simazine 1 20 64 1.6 EI 138, 186, 201 Trifluralin 65 60 0.2 NCI 305, 335 3 36
71 a CEAT= 6 amino2 chloro4 isopropylaminostriazine b CIAT= 6 amino2 chloro4 ethylaminostriazine c 4,4'DDD = 1,1 dichloro2,2bis[p chlorophenyl]ethane d 4,4'DDE = 1,1 dichloro2,2bis[4 chlorophenyl]ethylene e 4,4'DDT = 1 trichloro 2,2bis [ p chlorophenyl]ethane f Diazinon diethyl d10 was used as an extraction efficiency surrogate & diazinon was a target analyte g hexachlorocyclohexane h quantifying ions in italics Table 2 2 Site coordinates, pH, salinity conductivity, and temperature on monthly sampling trips between December 2004 and April 2006 (n= 16 for sites A, B & C and n=9 for sites D & E). Site Coordinates pH Salinity Conductivity Temperature Average SD (Range) (ppt) Average SD (Range) (mS) Average SD (Range) (C) Average SD (Range) A N 26.5021 W 81.0507 8.3 0.4 (7.7 9.2) 0.3 0.2 (0.2 0.9) 0.49 0.36 (0.25 0.18) 24.6 4.2 (18.1 31.4) B N 26.7885 W 81.3014 8.0 0.3 (7.6 8.5) 0.2 0.05 (0.1 0.3) 0.40 0.083 (0.30 0.60 ) 24.6 4.2 (17.9 31.0) C N 26.8392 W 81.0850 8.0 0.4 (7.1 8.5) 0.2 0.07 (0.1 0.4) 0.40 0.14 (0.23 0.76) 24.4 4.6 (16.8 31.3) D N 26.6844 W 81.8309 7.4 0.5 (6.8 8.0) 0.62 0.66 (0.15 2.0) 1.20 1.2 (0.31 3.7) 26.0 4.5 (20.3 31.7) E N 26.4895 W 82.0149 7.9 0.2 (7.6 8.3) 25.2 5.3 (13 30) 40 7.9 (21 48) 24.7 4.7 (18.6 31.1) Table 2 3 Caloosahatchee River Flow data from three of our pesticide sampling sites, December 2004 through April 2006 (n=517). Flow measured in cubic feet per second (adapted from DBHYDRO, SFWMD). Flow (CFS) S77 (Site A) S78 (Site B) S79 (Site C) Maximum 8188 9372 16900 Minimum 0 6 0 Mean 2497 3110 4130 St an d ard d ev iation 2317 2673 3916 Median 1720 2098 2570
72 Table 2 4 Detection f requencies of pesticides and their concentration ranges in water samples collected along the Caloosahatchee River, Florida. Pesticide name Use Chemical type n Number of detections Frequency (%) d etection Highest conc. detected (ng/L) Lowest conc. detected (ng/L) Median conc. (ng/L) Atrazine Herbicide Triazine 75 74 99 280 0 13 72 Metolachlor Herbicide Chloroacetanilide 75 71 95 2 70 2.5 1 8 CIAT Herbicide Triazine 75 61 81 18 0 3.9 21 CEAT Herbicide Triazine 75 56 7 5 9 4 3.8 17 Simazine Herbicide Triazine 75 45 60 12 0 2.7 9.4 Ametryn Herbicide Triazine 75 35 47 87 2.0 8.4 Malathion Insecticide Organophosphate 90 36 40 31 1.6 3.5 Chlorpyrifos oxon Insecticide Organophosphate 90 23 26 8.8 3.7 7.1 Dieldrin Insecticide Organochlorine 90 15 1 7 1.5 0.2 0.6 Chlorothalonil Fungicide Chloronitrile 90 10 11 11 2.5 4.0 Diazinon Insecticide Organophosphate 90 9 10 17 1 3 15 Metribuzin Herbicide Triazine 75 5 6.7 4 7 9.0 12 Pendamethalin Herbicide Dinitroaniline 75 4 5.3 18 7.1 10 Heptachlor Insecticide Organochl orine cyclodiene 90 2 2.2 6.2 5.2 5.7 Ethion Insecticide Organophosphate 75 1 1.3 3.3 3.3 3.3 Ethoprop Insecticide Organophosphate 75 1 1.3 8.3 8.3 8.3 Methoxychlor Insecticide Organochlorine 75 1 1.3 1 5 1 5 1 5 Aldrin Insecticide Organochlorine 90 1 1.1 3.1 3.1 3.1 4,4' DDD Insecticide Organochlorine 90 1 1.1 26 26 26 4,4' DDE Insecticide Organochlorine 90 1 1.1 2.8 2.8 2.8 Endosulfan sulfate Insecticide Chlorinated h ydrocarbon 90 1 1.1 3.2 3.2 3.2
73 Table 24. Continued. Pesticide name Use Chemical t ype n Number of detections Frequency (%) d etection Highest conc. detected (ng/L) Lowest conc. detected (ng/L) Median Conc. (ng/L) Mirex Insecticide Organochlorine 90 1 1.1 0.9 0.9 0.9 Acetochlor Herbicide Chloroacetanilide 75 0 0.0 nd nd nd Alachlor He rbicide Chloroacetanilide 75 0 0.0 nd nd n d Cyanazine Herbicide Triazine 75 0 0.0 nd nd nd Trifluralin Herbicide Dinitroaniline 90 0 0.0 nd nd nd Chlordane Insecticide Organochlorine 90 0 0.0 nd nd nd Chlordane Insecticide Organochlorine 90 0 0.0 n d nd nd Chlorpyrifos Insecticide Organophosphate 90 0 0.0 nd nd nd 4,4' DDT Insecticide Organochlorine 90 0 0.0 nd nd nd p,p' Dicofol Insecticide Organochlorine 75 0 0.0 nd nd nd Endosulfan Insecticide Chlorinated hydrocarbon 90 0 0.0 nd nd nd End osulfan Insecticide Chlorinated hydrocarbon 90 0 0.0 nd nd nd Fenamiphos Insecticide Organophosphate 75 0 0.0 nd nd nd Fipronil Insecticide Phenyl pyrazole 90 0 0.0 nd nd nd HCH Insecticide Organochlorine 90 0 0.0 nd nd nd HCH Insecticide Organochl orine 90 0 0.0 nd nd nd cis Nonachlor Insecticide Organochlorine 90 0 0.0 nd nd nd trans Nonachlor Insecticide Organochlorine 90 0 0.0 nd nd nd cis Permethrin Insecticide Pyrethroid 75 0 0.0 nd nd nd trans Permethrin Insecticide Pyrethroid 75 0 0.0 nd nd nd Phorate Insecticide Organophosphate 75 0 0.0 nd nd nd
74 Table 2 5 Mean, minimum, median, maximum concentrations and percent relative standard deviation (%RSD) of the ten most commonly detected pesticides ( organized by site) in the Caloosahatchee River Site Pesticide n Detection Frequency (%) Max detected (ng/L) Min detected (ng/L) Median detected (ng/L) % RSD Site A Atrazine 15 93 2 800 39 76 240 Metolachlor 15 93 18 0 11 2 7 10 0 CIAT 15 73 18 0 6.8 22 1 40 CEAT 15 73 9 4 8.4 1 8 100 Simazine 15 73 12 0 5.0 9.8 1 60 Ametryn 15 40 1 6 5.2 7.8 46 Malathion 18 39 1 6 2.7 4.3 64 Chlorpyrifos oxon 18 11 7.8 7.8 7.8 6 Dieldrin 18 28 1.5 0.6 0.8 41 Chlorothalonil 18 0 0.0 0.0 0.0 0 Site B Atrazine 15 100 6 70 28 76 102 Metolachlor 15 100 1 10 5.5 27 86 CIAT 15 80 89 8.2 25 82 CEAT 15 73 4 4 6.6 19 48 Simazine 15 67 9 3 6.1 10 1 30 Ametryn 15 47 27 5.6 8.3 66 Malathion 18 22 5.8 3.2 4.0 25 Chlorpyrifos oxon 18 11 8.4 7.8 8.1 9 Dieldrin 18 0 0.0 0.0 0.0 0 Chlorothalonil 18 0 0.0 0.0 0.0 0 Site C Atrazine 15 100 1 30 0 3 3 79 1 50 Metolachlor 15 100 1 80 4.2 10 1 70 CIAT 15 80 86 6.9 2 8 76 CEAT 15 67 58 10 21 60 Simazine 15 27 1 20 1 2 28 1 00 Ametryn 15 40 87 5.4 9.2 1 20 Malathion 18 6 28 28 28 91 Chlorpyrifos o xon 18 17 8.8 7.8 8.8 8 Dieldrin 18 6 0.8 0.8 0.8 0 Chlorothalonil 18 0 0.0 0.0 0.0 0
75 Table 25. Continued. Site Pesticide n Detection Frequency (%) Max detected (ng/L) Min detected (ng/L) Median detected (ng/L) % RSD Site D Atrazine 9 100 2 40 2 6 7 2 73 Metolachlor 9 100 92 8.7 20 84 CIAT 9 89 26 6.8 1 3 44 CEAT 9 89 20 7.2 1 4 33 Simazine 9 89 24 3.0 1 0 64 Ametryn 9 89 24 2.0 3.1 12 0 Malathion 12 83 31 1.6 2.6 1 60 Chlorpyrifos oxon 12 58 6.5 3.8 4.3 23 Dieldrin 12 58 0.7 0.3 0.4 34 C hlorothalonil 12 33 7.5 2.5 4.7 43 Site E Atrazine 9 89 9 8 1 3 2 7 77 Metolachlor 9 67 1 9 2.5 8.8 70 CIAT 9 67 6.9 3.9 4.5 28 CEAT 9 67 7.4 3.8 5.2 26 Simazine 9 56 9.6 2.7 4.8 48 Ametryn 9 0 0 0 0 0 Malathion 12 75 2.8 1.6 2.3 17 Chl orpyrifos oxon 12 17 5.2 3.7 4.5 24 Dieldrin 12 8 0.2 0.2 0.2 0 Chlorothalonil 12 50 11.3 2.6 3.6 65
76 Figure 21. Five pesticide sampling sites used along the Caloosahatchee River, Florida. From: www.evergladesplan.org.
77 ethoprop CIAT CEAT atrazine simazine ametryn metolachlor metribuzin pendamethalin ethion methoxychlor diazinon heptachlor chlorothalonil aldrin malathion chlorpyrifos-oxon 4,4'-DDE dieldrin 4,4'-DDD endo-sulfate mirex Measured Concentrations (ng/L) 0.1 1 10 100 1000 10000 F igure 22. Pesticides detected (ng/L) in at least one sample in the Caloosahatchee River from December 2004April 2006 (n= 75 for compounds ethroprop through methoxychlor and n= 90 for diazinon through mirex). The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles. Outliers are plotted with black c ircles
78 Figure 23. SIMPROF results of a hierarchical cluster analysis of all target analytes detected at least once from all sample sites and sampling months (permutations=999, = 0.05) Samples connected by red lines cannot be differentiated at the 0.05 significance level.
79 Figure 24. Two Dimensional MDS plot (Stress = 0.11) showing similarities between samples by sampling month and site. Similarity between samples are superimposed from a hierarchical cluster analysis.
80 Compounds Atrazine Metolachlor CIAT CEAT Simazine Ametryn Malathion Chlorpyrifos-oxon Dieldrin Chlorothalonil Concentration (ng/L) 0.1110100100010000 Figure 25. Concent rations of the ten most frequently detected pesticides from all sampling sites and dates along the Caloosahatchee River, Florida (n= 75 for atrazine through ametryn and n=90 for malathion through chlorothalonil above). The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 90th and 10th percentiles. Outliers are plott ed with black circles.
81 Figure 26. Average number of analytes detected at each sampling site between January 2005 and April 2006 with the associated standard errors (n=15 for sites A, B, and C; n= 12 for sites D hoc analyses using the Holm Sidak method; different letters indicate significant difference.
82 Figure 27. Average concentration of atrazine (ng/L) in samples collected between January 2005 and January 2006 and the associated standard errors shown analyses with Tukey's test shown.
83 Figure 28. Average concentration of metolachlor (ng/L) in samples collected between January 2005 and January 2006 and the associated standard errors. Results shown.
84 Figure 29. Average number of pesticides detected per sample over the sampling months of December 2004 to January 2006 and the associated standard errors. While the overall oneway ANOVA indicated significant difference between the number of pesticides detected in these months, no pairwise comparisons indicated significant differences. Bars with no error bars indicate an error so small that they do not show up on the figure.
85 Figure 210. SIMPROF results of a hierarchical cluster analysis based on concentration of all target analytes detected at least once from all sample sites and sampling months (Resemblance= Bray Curtis similarity, Permutations= 999, black lines are significantly different
86 Figure 211. Two Dimensional MDS plot (Stress = 0.12) showing similarities between sam ples by month and site after transformation of data to presence/absence. Similarity between samples are superimposed from a hierarchical cluster analysis. The first plot showed one extreme outlier so the subsequent plot is a subset of the smaller top image without this point.
87 Figure 212. SIMPROF results of a hierarchical cluster analysis of all target analytes detected at least once from all sample sites and sampling months transformed to presence/absence data. Black lines indicate significant differe nces at nodes shown (Resemblance= Simple Matching, Permutations=
88 Figure 213. The cooccurrence of the top ten most commonly detected compounds (atrazine, metolachlor, CIAT, CEAT, simazine, ametryn, malathion, chlorpyrifos oxon, dieldri n, chlorothalonil) over all sampling months and sites along the Caloosahatchee River, Florida (42 analytes and 75 sampling events).
89 CHAPTER 3 EFFECT OF HERBICIDE MIXTURES (ATRAZINE AND METOLACHLOR) ON THE AQUATIC MACROPHYTE, LEMNA MINOR Background Her bicides are used for weed control to reduce competition and enhance yields in agricultural crops, and for visual impact in landscapes. Of the 684 million pounds of pesticidal active ingredients applied to agricultural lands in the United States in 2007, 6 5% were herbicides (Fishel, 2007). Many herbicides typically have longer half lives in the environment (relative to insecticides) to provide extended weed control action. As a result of their persistence and chemical physical properties, herbicides frequently move from their sites of application with surface runoff or drainage water. Several studies have reported the presence of herbicides in surface water worldwide. Many of those studies have focused on surface water within the state of Florida due to it s abundant surface water resources, the sensitive nature of ecosystems in South Florida, and the prevalent use of herbicides due to the economic reliance of the state on agriculture in the region(SFWMD, 2000; Shahane, 2007; FDEP, 2005; HarmanFetcho et al. 2005; Downing et al., 2004; Doering& Chamberlain, 1998; Wilson and Boman 2011; Wilson et al. 2012; Wilson et al. 2007). Often, herbicides are detected most frequently in surface waters nationally and in Florida (EPA, 1995; USGS, 2007; HarmanFetcho et al ., 2005; Scott et al., 2002; Wilson and Boman 2011). Because of the diversity of land management practices within a given watershed, aquatic resources may be exposed to multiple pesticides or stressors at any given time. In a ten year USGS study of mixeduse streams across the country by Gilliom et al. (2007), 51 major river basins and aquifer systems were monitored for 75 pesticides and 8 pesticide degradation product s. They reported that pesticides usually occur in
90 mixtures of multiple compounds. A gricu ltural, urban, and mixeduse streams had at least 5 pesticides per sample at least 50% of the time. They found that streams with developed watersheds (those dominated by agricultural, urban, or mixed use land use) contained at least 2 pesticidesmore than 9 0% of the time and more than 9 pesticides about 20% of the time (Gilliom et al., 2007). They predicted that toxicity is likely underestimated when assessments are based on individual compounds since most compounds occur in mixtures (Gilliom et al., 2007). Many studies in Florida have reported the presence of more than one pesticide in a given sample (Carriger & Rand, 2008; HarmanFetcho, 2005; USGS 2007; Wilson and Boman 2011). HarmanFetcho reported 91% and 89% of their surface water samples contained atrazine and metolachlor, respectively (n=88). They showed seasonal patterns of mixtures including atrazine, chlorothalonil, chlorpyrifos, endosulfan species, and metolachlor, with increases in the spring compared to fall This study also found mixtures of endosulfan species, chlorothalonil, and chlorpyrifos occurring together at most sample sites in March 2003. A combined analysis of pesticide detections in surface waters of South Florida canals was compiled by Carriger and Rand (2008). They reported 92% of 185 freshwater samples in 1999 and 100% of 106 samples taken in 2000 had detectable atrazine concentrations while 28% and 26% contained metolachlor in 1999 and 2000, respectively. Frequently, atrazine was detected in the same samples as endosulfan, metolac hlor, chlorpyrifos, and malathion (Carriger & Rand, 2008). Wilson and Boman (2011) surveyed canals for pesticides in southern Florida weekly for three years. They reported binary and tertiary combinations of the
91 herbicides atrazine, bromacil, metolachlor, and norflurazon in up to 91 of the 211 samples collected (Wilson & Boman, 2011). Aquatic resources in the environment may be exposed to multiple stressors (including pesticides) at any given time. Realization of this is especially important in ecologic al risk assessments. Potential stressor interactions may enhance the toxicity of one another (potentiation), reduce toxicity (antagonism), or have no effects (additive) on the toxicity of one another (Sprague, 1970). Unfortunately, little information is available on many of the potentially cooccurring contaminants in aquatic environments. As a result, many risk assessments identify suites of chemicals in the environment and use toxicity data for the chemicals based on doseresponse curves generated for single toxicant exposures to assess risks (Munn & Gilliom, 2001; Banks et al., 2003). Because of the potential interactions of contaminants (Newman & Unger, 2003), it is more desirable to compare field concentrations to toxicity data based on the relevant combinations of contaminants to which species may be exposed. To do this effectively, integrated multilevel studies are needed (Porter et al., 1993). Two herbicides that have been detected frequently in surface water include atrazine and metolachlor. These herbicides are used for production of many agricultural crops, as well as for maintenance of weedfree lawns and landscapes in urban areas. Nationally, metolachlor and atrazine are in the top five herbicides detected in surface water samples (EPA 2006; EPA 1995). Drinking water criteria were established for atrazine because of its prevalence in ground water, especially in the midwestern United States (EPA, 1995; EPA 2006). Both of these herbicides are soluble in water, are not
92 volatile, and are relatively long lived in the environment as evidenced by their half lives (Table 31). In an unpublished study, Smith et al. surveyed the Caloosahatchee River in Southwest Florida monthly over a period of 17 months (Smith et al., unpublished). The most frequently detected pesticides in that study were atrazine and metolachlor, detected in 98.7% and 94.7% of 75 samples, respectively. Atrazine was detected in the range of 12.92854.0 ng/L with a median concentration of 72.2 ng/L. Metolachlor was detected in the range of 2.5268.3 ng/L with a median concentration of 17.7 ng/L. Only one of 75 samples did not contain either atrazine or metolachlor, and the two compounds were detected together in 71 of the 75 samples. Atrazine and metolachlor were also frequently detected together in surface and ground water in the National Water Quality Assessment Program of the U.S. Geological Survey between 1992 and 1999 (Squillace et al. 2002). Given the widespread use of atrazine and metolachlor, and their frequent detection in surface water, research is needed to assess whether they may interact together to possibly enhance or reduce the toxicity to nontarget aquatic macrophytes. The purpose of this study was to determine the effects of atrazine and metolachlor on aquatic macrophyte ( Lemna minor ) growth, reproduction and health. This study evaluated the effects of these compounds separately and in a mixture with the goal of identifying risks to an aquatic ecosystem that are not adequately explained by toxicity assessments bas ed on individual compounds.
93 Materials and Methods Plant Culturing Lemna minor is a wellestablished test or ganism in toxicological studies It is a small floating plant (leaves 24 mm wide), each with a single root suspended in the water column (Wang 1990, Correll & Correll 1972). This species is especially applicable to natural systems in Florida because it is a native species commonly found in waterways Mats of L. minor are habitat for small invertebrates, provide food and shelter for fish, and are a k ey diet component for many waterfowl and marsh birds (Hillman, 1961; Newmaster et al., 1997). Fairchild (1998) reported in a study of several herbicides that L. minor is of intermediate sensitivity to herbicides and that algae and macrophytes are of similar sensitivity to herbicides. Lemna minor is ideal for toxicity testing due to its small size, ease of culturing, fast growth rate, and small space requirements. Original cultures were purchased from Carolina Biological Supply. Algal free cultures were maintained at the UF/IFAS Indian River Research and Education Center (Fort Pierce, FL). Plants were grown in 20% Hoagland's nutrient media made with reconstituted, moderately hard water made using reagent grade nanopure water (APHA et al. 1998). Cultures and tests were maintained at 25 C under a 12 hour photoperiod. Toxicity tests were conducted in a separate room from where cultures were maintained. Test Solution Preparation Commercial formulations of simazine and metolachlor were used for all toxicity t ests. Formulations included Atrazine 4L (Tenkoz, Inc., Alpharetta, GA) and Dual Magnum (Syngenta Crop Protection, Greensboro, NC). These formulations contained 42.6%and 83.7% atrazine and metolachlor, respectively. Stock solutions (479.3 mg/L
94 atrazine an d 91.55 mg/L metolachlor) were prepared by diluting the appropriate amount of either metolachlor or atrazine into reconstituted moderately hard water. Stock solutions were made immediately before use. Initially, the toxicity of atrazine and metolachlor to L. minor were evaluated separately in order to derive individual EC50 values under these test conditions. For these assays, test concentrations for atrazine and metolachlor were prepared by pipetting the required amount of stock into known volumes of tes t media to produce nominal test concentrations of 0, 0.008, 0.015, 0.030, 0.060, 0.125, 0.250, 0.5, and 0.750 g/ml. The highest atrazine concentration used in the dilution series was approximately equal to four to eight times the published 96 h median eff ective concentration values (EC50), based on frond production (EPA Ecotox, 2000 & Fairchild, et. al, 1998). The highest metolachlor concentration used in the dilution series was approximately equal to two times the published 96 h EC50, based on frond production (Fairchild, et.al, 1998). Lower concentrations were also included, and all subsequent concentrations were prepared by serial dilution using a 50% dilution factor. Results from the singlechemical toxicity assays were used to estimate the EC50 of each herbicide under test conditions. Using the estimated EC50 values, mixture concentrations for evaluation were then selected and mixed as previously described. Test concentrations were determined using the toxic unit (TU) approach (Sprague, 1970). With this approach, the EC50 for each herbicide was set equal to 1 TU. Equal fractions of TUs for both herbicides were then assayed along with nontreated controls. The TUs used included (expressed as TUs atrazine + TUs metolachlor) were 0+0, 0.25+0.25, 0.5+0.5, 0.75+0.75, and 1+1. This corresponds to the following nominal
95 concentrations (expressed as atrazine/metolachlor in g/ml): 0/0, 0.058/0.033, 0.116/0.066, 0.174/0.099, and 0.232/0.132 g/ml. The concentration of atrazine and metolachlor within the prepared test solutions was measured at the beginning of the experiments for both individual compounds and the mixture. For confirmation, two 100 ml or 1000 ml sample aliquots (based on concentration) from excess treatment solution for each treatment concentr ation were extracted using liquidliquid extractions based on the method of Wu et al. (2010). A Hewlett Packard 5890 Series II gas chromatograph (Agilent Technologies, Wilmington, DE) equipped with dual electron capture detectors and Rxi 5ms (Restek, Bell efonte, PA) and SGEPX350 25 (SGE Incorporated, Austin, TX) columns was used to measure actual test concentrations within test solutions. Analysis quality control included instrument and method blanks, matrix spikes, and calibration check standards. Measur ed concentrations differed the most (up to 17.6%) from target concentrations at the lower test concentrations (Table 32). Results are presented using the measured, rather than nominal, concentrations of compounds. Bioassay Procedures Exposures were conducted in 200 ml glass crystallization dishes filled with 150 ml of test solution. Four replicates containing 12 fronds each were used for every treatment. All tests were conducted under static, nonrenewal conditions. Previous assays indicated that atrazine and metolachlor are stable under the test conditions used in this study (unpublished). All tests were conducted with a 12:12h light (90+ E):dark photoperiod. Containers were randomized every two days to avoid light difference bias. Individual atrazine and metolachlor tests were conducted for 14 days to assess longer
96 term impacts of exposure. The mixture study was conducted for six days based on the results of individual exposure tests. Frond counts and root length measurements were made every 2 days. Digital images were also taken every two days to monitor changes in color. Root length measurements were taken from three sets of randomly selected replicates per treatment. The longest roots were measured from the outside of the glass vessel to avoid brea kage, handling stress, and contamination. These measurements were averaged for each replicate of each test concentration. At the end of the study, fresh weights of all fronds were measured using a Denver Instruments TL 104 balance after gently blotting dry In addition, measurements of SPAD (leaf greenness), photosystem II electron transport efficiency, and chlorophyll/carotenoid content were measured. SPAD measurements were made using a portable Minolta Chlorophyll meter (SPAD 502). Enough plant tissue was randomly selected from each vessel to cover the measurement area on the meter. The highest of three measurements was used for each experimental replicate. Tissue was discarded after measurement since this was somewhat destructive and any damaged plant material may have influenced the other photosynthesis related measurements. Electron transport efficiency of PSII was measured using an Opti Sciences OS5p portable fluorometer. Approximately 1520 fronds from each experimental replicate were loaded into the mouth of15 mL glass vials filled with nanopure water. The diameter of the mouth of the vial was the same as that of the fluorometer probe, which facilitated measurements. Vials with accompanying plants were placed in a rack and allowed to acclimat e to the dark for at least 30 minutes. Cardboard shields were used to
97 shade surrounding vials when fluorometer readings were being made. Two replicate measurements for each sample vial were made (with a 30 minute rest period between measurements) and the Fv/Fm measurements were averaged. For measurement of chlorophyll and carotenoid content, approximately 30 mg of fresh tissue from each test vessel was frozen with liquid nitrogen. These samples were then analyzed for chlorophyll a, chlorophyll b, and carotenoid contents using a Beckman DU 64 spectrophotometer following the procedures of Arnon (1949) and calculated with correction values described in Lichtenthaler (1987). Data Analysis There are a variety of ways to explore the effects of multiple toxicants, and we used the toxic unit approach for this study (Sprague, 1970).Toxic units are used to express the concentrations of different toxicants in units of effect (EC50). The EC50 for each compound is then set as 1 TU for that compound; in this way, effect l evels predicted with the logit model for each toxicant can be added together. To explore additivity of two compounds, if we add 0.5 TU of each compound in the mixture, a 1TU combined effect, or the EC50 of the mixture, is expected. Care must be taken when using this method to confirm that the slopes of the individual compound logit models are parallel, indicating similar modes of action (Van der Geest et al., 2000). If this condition is met, the joint toxic effect of two toxicants can be determined as addit ive, more than additive (synergistic) or less than additive (antagonistic) depending on the comparison of the EC50 to the toxic units of the mixture. When adding 0.5 TU of each atrazine and metolachlor: if the EC50 of the mixture of atrazine and metolachlor is equal to 1 TU, the effect is additive, if the EC50 of the mixture is less than 1 TU, the mixture is more than additive, and if the EC50 is more than 1 TU, the mixture is less than additive. Using the
98 additive index of Marking & Dawson (1975)allows qua ntification of this mixture additivity approach (Equation 1). Equation 1.Additive Index of Marking & Dawson (1975). S= Am/Ai + Bm/Bi Where Am & Bm are the incipient EC50 of toxicants A & B when present in mixture, and Ai & Bi the toxicity of A & B when tested separately. Using pesticides with different modes of action can complicate the toxic unit approach significantly. Van der Geest et al. (2000), working with the insecticide diazinon and the metal copper, explained that many studies published actually violate the toxic unit concept by ignoring the requirement for similarly shaped doseresponse relationships among mixture components. To test if the modes of action are similar, the slopes of the probit models can be compared; parallel slopes indicate si milar modes of action (Newman & Unger, 2003; Finney, 1947). Even with different modes of action where doseresponse curves of individual toxicants and the resulting mixture have significantly different slopes, the toxic unit approach can still be meaningf ul using effect concentrations other than the LC50 (Van der Geest et al., 2000; Van Gestel & Hensbergen; Banks et al., 2003). Care was taken in the present study when considering the joint effect of mixtures of atrazine and metolachlor since the mode of a ction of metolachlor is not well understood, but expected to inhibit seedling development, while atrazine inhibits photosystem II (EPA, 1995; EPA 2006).
99 In the single compound experiments, concentrationresponse data were analyzed using the transformed logit model (Equation 2). Equation 2.Transformed logit model. Transformed Logit= [Logit(P)/2] + 5 where P is the proportion mortality by the end of the exposure and Logit(P) = ln[P/(1 P)]. The resulting linear relationship between the log concentration and the transformed logit was used to generate a best fit model by regression. Back calculation based on the regression model allowed estimation of an EC50 for each individual pesticide based on frond count. In order to use the mixture additivity index, slopes of the doseresponse curves of the candidate compounds must be similar if the compounds do not share the same modeof action. Slopes of the transformed logit versus log concentration plots for metolachlor and atrazine were compared prior to determinati on of the mixture toxic unit additivity. This was accomplished using the homogeneity of regression slopes assumption test from an ANCOVA test to check that the slopes of the regression lines are parallel. No differences in slopes were observed (Table 33). Since the slopes were similar, the toxic units (TU) for each compound were set to equal 1 at the measured EC50 concentration for each pesticide (0.232 g/ml atrazine and 0.132 g/ml metolachlor). The joint toxic effect of atrazine and metolachlor was defi ned as being concentration additive if 0.5 TUatrazine + 0.5 TUmetolachlor = 1 TU, more
100 than additive if 0.5 TUatrazine + 0.5 TUmetolachlor <1, or less than additive if 0.5 TUatrazine + 0.5 TUmetolachlor >1.This was quantified using the additiv ity index (Eq uation 1). For all endpoints measured (frond count relative to controls, fresh weight, root length, and Fv/Fm photosystem II electron transport), a oneway ANOVA was completed after tests for normality and homogeneity of variance were passed. The P values for the tests for normality (listed first) and homogeneity of variance (listed second) for those tests were: frond count (P=0.646, P=0.091), fresh weight (P=0.0936, P=0.268), root length (P=0.972, P=0.118), Fv/Fm (P=0.824, P=0.798). For the Fv/Fm measurem ents, a t test was conducted to assure that the two replicates of measurements could be combined into a mean value. The difference in the mean values of the two replicates was not great enough to reject the possibility that the difference was due to random sampling variability. There were no statistically significant differences between the two replicates (P = 0.449). As a result, the combined means were used for Fv/Fm comparisons. Results Validation of Test Solution Concentrations Data for confirmation of target concentrations of atrazine and metolachlor in the individual compound toxicity assays are summarized in Table 3 2. Once toxic units were determined for each individual pesticide (1 TU = EC50), the test concentrations used were converted to toxic units, as shown in Table 3 2. Table 3 3 summarizes the results of the homogeneity of regression slopes test typically used as a prerequisite to an ANCOVA. Since the results were not significant, parallel slopes for the individual atrazine and metolachlor c oncentrationresponse curves were assumed.
101 Validation data for the mixture (atrazine+metolachlor) assays are shown in Table 3 4. Since the actual measured concentrations differed from the targeted concentrations, measured values were used to estimate the t oxic units (Table 34). Individual Concentration Response Bioassays Results for the individual concentrationresponse bioassays for all 14 days are shown in Figure 3 1 Growth re sults are presented as a percentage of controls versus the log of the compound concentration. Considering atrazine, no concentration at or below the 0.064 g/ml concentration caused a reduction in growth of 50% compared to controls. On day 2, the growth rate ranged from 66% at the highest concentration to 109% of controls at 0.16 g/ml, with five concentrations producing growth rates greater than the controls (0.0068 g/ml, 0.016 g/ml, 0.033 g/ml, 0.065 g/ml, and 0.12 g/ml). Certain concentrations at day 4, and day 10 also produced growth rates greater than the controls (Day 4: 0.016, 0.033, 0.065; Day 10: 0.0068 g/ml, 0.016 g/ml). This prevented the calculation of logit and transformed logit calculations for those days. Day 6 growth rates ranged from 20% to 99% of controls at 0.78 g/ml and 0.0068 g/ml, respectively (the highest and lowest concentrations to which they were exposed). The EC50 was determined to be 0.23 g/ml at Day 6. Day 8 growth rates followed a similar pattern, ranging from 12% to 98% of controls at 0.78 g/ml and 0.0068 g/ml, respectively (the highest and l owest concentrations to which they were exposed). The EC50 was 0.16 g/ml at Day 8. Day 12 growth rates also followed the described pattern, ranging from 3% to 97% at the highest and lowest concentrations, respectively. The EC50 was 0.07 g/ml on Day 12.Af ter 14 d exposure, growth (frond production) ranged from 2% (at 0.78 g/ml) to 89% (at 0.016 g/ml), and the EC50 was 0.05 g/ml. Results
102 from days 12 and 6 appeared to be the best candidates for the mixture study based on atrazine concentrations. For metolachlor, all concentrations above 0.0071 g/ml reduced frond production by 50% (relative to controls) at some point during the 14 day study. On days 2 and 4, no EC50 could be determined because no test concentration reduced growth by half compared with controls. On day 2, the growth rate ranged from 98% at 0.0071 g/ml to 85% of controls at 0.74 g/ml. On day 4, a similar pattern was observed, as the growth rate decreased with increasing concentration (range= 100% of controls at 0.0071 g/ml and 86% at 0.74 g/ml. While no test concentration caused a 50% reduction in growth compared to controls on days 2 and 4, these growth reductions were observed on all days after these. On day 6, frond production ranged from 88% of controls at the lowest test concentration to 31% at the highest. On day 8, frond production ranged from 68% at the lowest test concentration and 19% at the highest test concentration. From day 10 through day 14, the 50% reduction in frond production was reached between the lowest and next lowest test concentrations (0.0071 g/ml and 0.0123 g/ml). Six (EC50 = 0.13 g/ml) and 8 (EC50 = 0.017 g/ml) days exposure were considered the best candidates for the duration of the mixture study since mean metolachlor responses varied more across the concent ration range tested (i.e. some treatments with no differences and some with great differences relative to controls). Based on the results, a 6 d exposure period was chosen for conducting the mixture toxicity assays. The 6 d exposure period resulted in the most balanced distribution of responses around the EC50, with three (atrazine) and five (metol achlor) concentrations below it.
103 Before comparisons of the mixtures can be made using toxic units, the slopes of the individual concentrationresponse curves mus t be shown to be parallel (Van der Geest et al.,2000, Banks et al.2003). T he homogeneity of regression slopes assumption for an ANCOVA indicated that the slopes for both metolachlor and atrazine concentrationresponse curves were not significantly different, even though they do not look parallel upon visual inspection (p>0.05, F0.005[1,12] = 4.60, Fs = 3.76). Tests for Synergy Frond production Based on the responses observed (Figure 3 2 and Table 32 ), toxic units were calculated from the transformed logit model Solving for the EC50, the regression model fit to the atrazine concentrationresponse data (y=0.7539ln(x) + 6.1023, r2=0.9735) indicated an EC50 of 0.232 g/ m L. Likewise, the model fit to the metolachlor exposure data (y=0.2662ln(x) + 5.5399, r2=0. 9478) produced an EC50 value of 0.132 g/ m L. Based on this relationship, theoretically the test concentrations of 0.5 TU + 0.5 TU in the mixture study should result in 1 TU of effect, or 50% reduction in frond production. Since the measured and target conc entration values differed, the actual TU in the present study was 1.07 TU. 1 TU of effect was observed at a lower concentration, closer to the 0.25 + 0.25 test mixture, indicating a synergistic effect of the mixture (Figure 3 3).This conclusion was confir med using the mixture additivity approach and the additive index of Marking & Dawson (1975) by calculating S as described in Equation 1. In this case with atrazine and metolachlor, S= 1.05, indicating a synergistic relationship between the toxicants. Comparing the mixture study test concentrations with the frond count endpoint, t he differences in the mean values among the treatment groups wer e greater than
104 Multiple Comparison Procedures (Holm Sidak method) showed differences between the means as shown in Figure 3 4, with same letters indicating no significant diff erence between means and different letters indicating significant difference between the mean. While not all treatment mixtures were significantly less than the immediately lower test concentration, the mean frond count appears to decrease as test concentr ation increases. A 50% reduction in frond production was observed at the 0.25 + 0.27 and greater TUs compared to controls. A 72% reduction in total fronds was observed at the highest tested mixture TU compared to the control. Between the lowest tested mix ture (0.25 + 0.27) and the highest tested mixture (1.04 + 1.18) there was another reduction of almost half (56% of control). Fresh Weights Significant reductions in final fresh weights after 6 d exposure were observed between treatment groups <0.001, power=1.000) The Holm Sidak method for the All Pairwise Multiple Comparison Procedures indicated differences between the means as shown in Figure 3 5. All treatments had significantly lower fresh weights compared to controls Fresh weights were reduced by 48%, 69%, 89% and 87% (relative to controls) at the 0.25 + 0.27 TU, 0.53 + 0.54 TU, 0.79 + 0.83 TU and 1.04 + 1.18 TU concentrations, respectively. The 0.53 +0.54 TU mixture seemed to be the threshold for effects with no further reductions in f resh weight occurring at greater concentrations (Figure 35). Generally, as test concentration increases, the mean final fresh weight decreased. A 50% reduction in fresh weight was observed at all TUs 0.53 + 0.54 TUs and greater compared to controls. The highest mixture concentrations (0.79 + 0.83 TUs and 1.04 + 1.18 TUs) resulted in a reduction in weight greater than 85%
105 compared to controls. At least a 75% reduction was detected between the highest mixture concentrations (0.79 + 0.83 TUs and 1.04 + 1.18 TUs) and the lowest (0.25 + 0.27 TU). Root Length Significant reductions in root length were observed at all mixture combinations (Holm Sidak method) indicated differences between the means as shown in Figure 3 6, with same letters indicating no significant difference between means and different letters indicating significant difference between the means. In this case root length was reduced at all mixture concentrations. There was at least a 50% reduction in root length in all test mixtures 0.53 + 0.54 TUs and above when compared to controls. Photosystem II Electron Transport (Fv/Fm) No significant reductions in Fv/Fm were observed except for at the 0.75+0.75 TU mixture, indi cating no effects on PSII electron transport. This result is surprising because atrazine is a PSII electron transport inhibitor. The Fv/Fm measurements made on Day 6 of the mixture study showed significant differences between treatment groups 0.007, power=0.781). The Holm Sidak method for the All Pairwise Multiple Comparison Procedures identified differences between the means as shown in Figure 3 7. Only the 0.79 + 0.83 TU mixture was different than the others, with a slightly lower mean Fv/Fm. But all averages were within 4% of each other. Interestingly, two mixtures, 0.25 + 0.27 TU and 0.53 + 0.54 TU had final Fv/Fm values that were higher than controls. No significant differences in chlorophyll a, chlorophyll b, and total chlorophyll content were observed for any of the mixture concentrations (P=0.086, P=0.211,
106 P=0.140, respectively) of atrazine and metolachlor. Patterns are shown in Figure 3 8. However, total carotenoids content was greater than in the power=0.5 95). The Holm Sidak method for the All Pairwise Multiple Comparison Procedures identified differences between the means (Figure 3 9). Total carotenoids, total chlorophyll, and both chlorophyll a and chlorophyll b increased in all test mixtures compared wit h the control. All showed a reduction at the 0.79 + 0.83 TU mixture, although it was not statistically significant and still contained higher concentrations than the controls. Mixture Comparisons to Individual Test Chemicals The relative frond count after 6 d exposure (expressed as a percentage of the control) for the individual atrazine and metolachlor tests are shown with those from the mixture assay (Figure 3 10). At 1 TU, the atrazine + metolachlor mixture resulted in relative frond counts slightly high er than either atrazine or metolachlor alone. Atrazine reduced the number of fronds more than metolachlor at the 1 TU point when compared to controls. At 2 TU, atrazine seems to reduce the relative frond count more than either the mixture or metolachlor al one when compared to controls. However, at the lower test to have reduced the relative frond count more than the mixture or atrazine alone. Similarly, the final fresh we ight (as percentage of the controls ) was plotted for the individual atrazine and metolachlor tests on the same graph as the mixture results. (Figure 3 11). It should be noted that the toxic units plotted in this figure were based on frond production, not f resh weight. Fresh weight appeared to be more influenced by metolachlor at all concentrations than the mixture or atrazine alone. Relative fresh weight was more greatly reduced by atrazine than the mixture at concentrations of
107 0.259 g/ml and below. At concentrations above 1 TU (combined) in the mixture study, the mixture reduced fresh weight more than atrazine did alone, relative to controls (0.79 + 0.83 TU and 1.04 + 1.18 TU). As noted in previous sections, including comparisons of the different pigments Fv/Fm, mean fresh weight, and mean root length, there seems to be a reduction in response between the 0.79 + 0.83 TU and 1.04 + 1.18 TU mixtures, causing less of an effect at this highest mixture concentration. Discussion Differences of opinion regarding the effects of mixtures containing atrazine and metolachlor have been reported. Lin (1999) reported that photodegradation of atrazine and metolachlor in natural systems decreases their toxicities. Kotrikla (1999) showed antagonism in toxic concentrations lower than the EC50 with mixtures of atrazine and metolachlor in studies with Chlorella However, in a study that looked at root development in Typhalatifolia, the common cattail, Moore & Locke (2012) reported that a combination of atrazine and metolachlor affected root development unlike the singular compounds. Nontarget species may be affected by mixtures that may not be expected to cause effects individually (Leboulanger et al., 2011). Given the frequency of detection of these two herbicides in surface water samples, the effects of these mixtures on plants in natural ecosystems is important since plants may be susceptible to the herbicidal mode of action and they occupy critical niches within ecosystems. Based on toxic unit theory the lower concentrati ons in this study appear to cause a greater than additive (synergistic) response in the frond production of Lemna minor exposed to mixtures of atrazine and metolachlor, because the frond production was reduced beyond what would be predicted by the toxic units. This is confirmed with the additivity index of Marking and Dawson (1975) This is an important finding because
108 both of these compounds are frequently detected in surface waters, and often cooccur (Carriger and Rand, 2008; HarmanFetcho, 2005; USGS 2007; Wilson and Boman 2011) These findings demonstrate that mixtures of herbicides may affect macrophytes more severely than the herbicides alone. But a t the higher concentrations tested, the mixtures appear to be less than additive since the response was not as great as predicted by the TUs. This may result from detoxification mechanisms in the plants being triggered between the 0.5 + 0.5 and 0.75 + 0.75 mixture concentrations. This is supported in the findings from the pigments and Fv/Fm data as well. However, based on the individual response curves, these mixtures may not t ruly be greater than additive at any concentration. For example, based on the 0.25 + 0.25 TU mixture in figure 31, an atrazine concentration of 0.058 mg/L would be expected to produce frond production closer to 88% of the controls (a 12% reduction) according to the individual atrazine response curve. At the same time, 0.036 mg/L of metolachlor would be expected to produce frond production closer to 61% of the controls (a 39% reduction). So while the toxic unit approach predicts that a mixture of about 0.5 TU would reduce the frond production by 25%, the individual response curves suggest that the response to that mixture should actually reduce the frond production by 51% (or 49% of the c ontrols). Comparing these data to figure 33, the mixture experiments resulted in frond count s closer to the values predicted by the individual response curves. So while these results appear to be greater than additive, this may not truly be the case based on the individual response curves. A cavaet to this conclusion should be made based on how the EC50s were estimated. The EC50s, and therefore, the toxic units, were calculated based on the
109 transformed logit model, not from the response curve data directly So comparing the responses from individual response curves to the toxic units in the mixtures may not be completely accurate. Using the response curves alone, for example, rather than the transformed logit to determine EC50s, the EC50 for metolachlor would be estimated closer to 0.10 and the EC50 for atrazine would be estimated closer to 0.17. This would change the toxic units of the mixtures, shifting them equal to one toxic unit at lower concentrations. The 0.75 + 0.75 TU concentrations would be more si milar to our estimated 0.5 + 0.5 TU mixture, making the predicted theoretical TU expectations closer to what was observed. Significant effects on frond count, root length, fresh weight, and carotenoid content of Lemna minor were found between atrazine and metolachlor mixture treatments. As the mixture concentrations increased, the frond count decreased with significant differences between most treatments. Root length was affected at all exposure concentrations. Fresh weight was also reduced at all exposur e concentrations, with the higher concentration mixtures (12 total TU) being more damaging than those at lower mixture concentrations. Interestingly, carotenoid concentrations appeared to increase as mixture concentrations increased. While the only statis tically significant difference is between the 0TU and 2TU averages, this is an interesting effect, perhaps related to a response on the part of the plant to environmental stress. No effects were observed for chlorophyll a and b content or Fv/Fm measurement s in the treatment mixtures. This is important, especially since the chemical mode of action for metolachlor is largely unknown (other than that is affects seedling
110 development). The chemical mode of action of atrazine involves the disruption of Photosystem II electron transport, so an effect on pigments and fluorescence might be expected. These data could be illustrating the ability of Lemna minor to metabolize atrazine into nontoxic metabolites. Solomon et al. (1996) and Huckins et al. (1986) reported that atrazine concentrations at or below 20 g/L result in little to no difference in Lemna minor biomass after 6 weeks exposure. But they also noted that dissolved oxygen concentrations decreased early in the exposure, implying a decrease in primary producti on. Huckins et al. (1986) reported that this primary productivity decreased by 23% with 2 d exposure to 10 g/L atrazine and 32% with 100 g/L exposure s over 2 days. In both cases, there was full recovery by day 7. In addition, Simard et al. (1990) report ed structural changes in chloroplasts in Lemna minor plants exposed to sublethal concentrations of atrazine. Grenier et al. (1989) reported that sublethal concentrations of atrazine stimulate the lipid metabolism of L. minor to form more thylakoid m embranes. Further studies need to examine this possibility, since some evidence here (Fv/Fm, total chlorophyll, total carotenoid content, chlorphyll a content and chlorophyll b content) implied that at the highest mixture concentration of atrazine and metolachlor (1.04 + 1.18 TU), the effect on Lemna minor is not as severe as at a slightly lower concentration (0.79 + 0.83 TU). It should also be noted that while some measures of physiology showed no significant effects, this does not mean a lack of symptoms. Sever al changes in morphology were observed in the higher mixture concentration treatments, including interveinal chlorosis in mature leaves, smaller new leaves, misshapen young leaves, and darker pigmentation in young leaves (Figure 312) While the frond count might
111 indicate growth is occurring, this metric does not describe the health of the plants. This is further supported in Figures 310 and 311, which show that while metolachlor appeared to have the least toxicity in terms of relative frond count when compared to the mixtures and atrazine alone, it was responsible for the most toxicity in terms of relative fresh weight when compared to the mixtures and atrazine alone. So although frond counts might not be reduced, the health of the new fronds is questiona ble since they were smaller and weighed less. This might be related to the Fv/ Fm data as well. While it might be expected that this metric would show decreases after treatment with pesticide mixtures including atrazine, these differences did not occur in t his study. But it might be due to the plant response to the mixture; if the plants increased their pigment content in response to the effects of the low doses of pesticide, it would compensate for the reduction of frond size, frond malformation and decreas ed efficiency of photosystem II. More research is required to identify low dose chronic effects of pesticides in aquatic systems, including morphological changes of aquatic species and the possibility of their recovery after exposure to such mixtures. Furt her, because mixtures found in natural systems likely contain low concentrations of multiple pesticides, and this study suggests that the mixtures may be more toxic to the aquatic macrophyte Lemna minor at low concentrations than either compound alone, mor e studies should focus on the effects of the presence of many pesticides of different types found in low concentrations on nontarget aquatic organisms. Lastly, further studies including wider ranges of concentrations would be useful to make more accurate conclusions regarding the additivity of atrazine and metolachlor in mixtures.
112 Table 31. Summary of chemical and physical properties of atrazine and metolachlor. Atrazine Metolachlor Chemical Structure Chemical name 6 chloro N2 ethyl N4 iso propyl 1,3,5 triazine 2,4diamine 2 chloro N (2 ethyl 6 methylphenyl) N 2 methoxy 1 methylethyl) acetamide Chemical Family Triazines Chloroacetanilide Empirical Formula C8H14ClN5 C 15H22ClNO2 Molecular Weight 215.7 283.8 Vapor Pressure 40 Pa at 20C 1. 7 x 10 3 Pa at 25 C Kow 2.76 794 (Weed 1994) Water solubility 33 g/ml at 25C 530 g/ml at 20C Koc indications Medium to High mobility in soil 22 310 Moderate to very high mobility in soil Henry's Law Constant 2.6 x 10 9 atm cu m/mole 9.0 x 10 9 atm cu m/mole BCF indications Low to moderate Low to moderate Aqueous photolysis natural sunlight Stable 70 d Soil photolysis (natural sunlight) 87 d 8 d Aerobic soil half life 90 120 d 67 d Aerobic aquatic half life n/a 47 d Anaerobic soil half life 33 0 d 81 d Anaerobic aquatic half life 608 d 78 d Adapted from EPA (1995 & 2006); HSDB (2012a & b)
113 Table 32 Target concentrations, measured concentrations, and percent from expected with the associated standard deviation (Deviation SD) in the individual atrazine and metolachlor tests. The calculated toxic units for each measured concentration are included as well. Target (g/ ml ) Measured a trazine (g/ ml ) Deviation SD (%) Atrazine t oxic u nits Measured m etolachlor (g/ ml ) Deviation SD (%) Metol achlor t oxic u nits 0.0 0 0 0 .0 0 .0 0 0 .0 0.008 0.007 14.75 0.001 0.03 0.007 11.25 0.001 0.05 0.015 0.017 10.13 0.001 0.07 0.012 17.68 0.001 0.09 0.03 0.033 9.35 0.003 0.14 0.026 14.31 0.001 0.2 0.06 0.065 7.74 0.002 0.28 0.065 8.23 0.004 0.49 0.125 0.121 3 0.005 0.52 0.116 6.84 0.016 0.87 0.25 0.259 3.72 0.002 1.12 0.264 5.76 0.002 2 .0 0.5 0.464 7.27 0.009 2 .0 0.459 8.13 0.014 3.47 0.75 0.782 4.33 0.009 3.37 0.735 2.02 0.06 5.57 Table 33 Computed quantities for testing as sumption that the separate regression slopes of metolachlor and atrazine individual doseresponse curves do not differ. F0.001[1,12] = 18.6. Source of v ariation df SS Y SP XY SS X b YX SS df SS YX MS YX Atrazine 7 11.21 1.98 0.53 3.7 2 7.37 6 3.84 0.64 Metolachlor 7 1.51 0.67 0.48 1.3 8 0.92 6 0.59 0.099 Sum of g roups 12 4.43 0.3 7 Among b i s 1 1.39 1.39 Fs = 3.76 ns
114 Table 34 Target concentrations, measured concentrations, and percent from expected with the associated standard deviation in the mixture (atrazine+metolachlor) tests. Resulting toxic unit adjustments for the atrazine (Atr) and metolachlor (Met) mixture study are also shown. TU= toxic units SD= Standard Deviation Target m ixture Target (g/ m L) Measured (g/ m L) Deviation from expected SD (%) Measured t otal c ombination (TU) Atr Met A tr M et A tr M et TU mix 0 + 0 TU (0TU) 0 0 0 0 0 0 0 0 0 TU (0TU) (0TU) (0TU) (0TU) 0.25 + 0.25 TU (0 5TU) 0.058 0.033 0.058 0.036 0.37 0.002 9.94 0.002 0.52 TU (0.25TU) (0.25TU) (0.25TU) (0.27TU) 0.5 + 0.5 TU (1TU) 0.116 0.066 0.124 0.072 6.51 0.022 9.55 0.000 1.07 TU (0.50TU) (0.50TU) (0.53TU) (0.54TU) 0.75 + 0.75 TU (1.5TU) 0.174 0.099 0.183 0.11 5.06 0.007 11.01 0. 012 1.62 TU (0.75TU) (0.75TU) (0.79TU) (0.83TU) 1 + 1 TU (2TU) 0.232 0.132 0.241 0.156 3.75 0.006 18.37 0.005 2.22 TU (1TU) (1TU) (1.04TU) (1.18TU)
115 A B Figure 31 Lemna minor frond production plotted as percentages of the correspond ing controls against A) logs of measured atrazine concentrations and B) logs of measured metolachlor concentrations.
116 A B Figure 32 Transformed logit model against A) log atrazine concentration on day 6 and regression results used to determine EC50 of atrazine and B) log metolachlor concentration on day 6 and regression results used to determine EC50of metolachlor
117 Figure 33 Frond production of Lemna minor as a percentage of controls in a mixture of atrazine and metolachlor. One toxic unit is expected to produce a 50% mortality effect (shown in Theoretical TU E xpected R esponse" line), so this relationship indicates a synergistic effect based on the 50% mortality effect occurring at lower than expected toxic units. The green line illustrates the expected frond production based on values directly from the individual response curves.
118 Figure 34 Mean final Lemna minor frond count (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with
119 Figure 35 Mean final Lemna minor fresh weight (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Sig nificant differences between means are indicated on the figure with
120 Figure 36 Mean final Lemna minor root length (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with
121 Figure 37 Mean final Lemna minor Fv/Fm values (Day 6, n=8) of atrazine and metolac hlor mixture study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with
122 Figure 38 Mean final Lemna minor chlorophyll a and b values (Day 6, n=4) of atrazine and metolachlor mixture study by toxic units and their associated standard errors.
123 Figure 39 Mean final Lemna minor t otal chlorophyll and total carotenoid values (Day 6, n=4) of atrazine and metolachlor mixt ure study by toxic units and their associated standard errors. Significant differences between means are indicated on the figure with different letters indicating signific ant difference
124 Figure 310 Day 6 relative Lemna minor frond counts expressed as a percentage of the controls in individual atrazine and metolachlor tests as well as mixture tests, equalized against toxic units (which were calculated from frond count)
125 Figure 311 Day 6 relative Lemna minor fresh weights expressed as a percentage of the controls in individual atrazine and metolachlor tests as well as mixture tests, normalized against toxic units (which were calculated from frond counts) Figure 312 Morphology changes in fronds produced in toxicity tests. In (a), the control culture is pictured with normal frond production. In (b), leaves exposed to mixture test concentrations exhibit different sized frond and different pigmentation. a b
126 CHAPTER 4 PROJECT CONCLUSIONS This study aimed to characterize some of the pesticides present in the Caloosahatchee River of Southwest Florida over an 18 month period and then identify the effects of the most prevalent mixture of pesticides on the nont arget aquatic macrophyte, Lemna minor In the first portion of the study that characterized the pesticides along the Caloosahatchee River, the prediction that the pesticides detected would be governed by spatial and temporal patterns were examined. I hypot hesized that there would be spatial patterns in the types of pesticides detected along the river, and I predicted that more pesticides and degradation products would be present and at higher concentrations at the sampling sites located within more agricult ural areas and less in the more residential areas. This prediction was not supported by the data collected. The ten most commonly detected pesticides in the surface water samples included atrazine, metolachlor, CIAT, CEAT, simazine, ametryn, malathion, chl orpyrifos oxon, dieldrin and chlorothalonil. There were no significant differences between sampling sites in terms of pesticide concentrations detected. However, there were significant differences in the number of pesticides detected at the different sites with samples from site D averaging 7.2 detections per month, higher than the other sites. Given that site D was in a res idential area ( not agricultural ), this result was not expected. Also, since these samples repeatedly contained higher numbers of detec tions than upstream sites, it is unlikely these pesticides were simply traveling downstream. I also hypothesized that seasonal patterns in the concentrations and types of pesticides detected would emerge, and I predicted that local agriculture would drive the
127 presence of detected pesticides and pesticide degradation products based on the harvest or planting seasons for the agricultural product closest to the sampling site. T here were some patterns based on seasonality and land use. Overall, there were higher concentrations of pesticides in March and April than in July, December, and January. Some compounds, like malathion, chlorothalonil and dieldrin were found with similar concentrations year round. Others, like chlorpyrifos oxon, ametryn, simazine, CEAT a nd CIAT were detected with higher concentrations in spring than in winter and summer months. The highest number of pesticides was found in March 2005, when an average of 8.5 pesticides were detected per sample from all sampling sites While there were some pesticides that varied by month, the data cannot support the prediction that pesticide concentrations were driven by the agricultural usage adjacent to sampling sites. Instead, t he lack of strong seasonality and spatial patterns suggests that the presence of pesticides cannot be predicted in this system based on adjacent land use alone. This is likely due to the extensive drainage basin and the patchiness of land use throughout the basin. Urban development is mixed with agricultural usage, and the vast net work of streams and canals links the different land uses. In addition, many pesticides are us ed for both agricultural and residential applications. A critical finding of this study is that all samples analyzed detect ed more than one pesticide, and some up to 12 pesticides per sample. Most (99%) of the samples contained atrazine and 95% of the samples contained metolachlor. Of 75 samples, there was only one (1.3%) that did not contain either atrazine or metolachlor. These two compounds were found together in 71 of 75 samples (95%). O f the 75 samples
128 analyzed for all 42 analytes 73% contained CIAT, CEAT and atrazine together and 72% contained CIAT, CEAT, atrazine and metolachlor together. Six compounds (CIAT, CEAT, atrazine, simazine, ametryn, and metolachlor ) were detected together in the same sample in 36% of the samples collected. These results indicate that organisms of the Caloosahatchee River are frequently exposed to low dose concentrations of multiple pesticides at a time. This highlights a concern that ecological assessments based on toxicity tests of individual pesticides likely misrepresent the effects of these pesticides in the environment since they often occur in mixtures. These data also do not support the hypotheses that adjacent land use alone determines the occurrence of pesticides along the Caloosahatchee River. Since the most common mixture encountered in the field samples was the mixture of atrazine and metolachlor (present in 71 of 75 samples), a laboratory mixture study was conducted to identify the response of the aquatic macrophyte, Lemna minor to these pesticides. I hypothesized that the tests with mixtures of atrazine and metolachlor would reflect the results of individual pesticide toxicity assays, and I predicted that such mixtures would have an additive effect on the frond count of Lemna minor Results indicated that the mixture of atrazine and metolachlor might be greater than additive compared to the effects of the individual pesticides. Lower than expected concentrations caused a 50% decrease in frond count and a change in morphological characteristics when the pesticides were present together in mixture. This synergistic effect on frond count indicates that aquatic macrophytes, of importance as primary producers in surface waters and habitat to other organisms, could be impacted more
129 severely than expected based on current ecotoxicity estimates. Since these compounds were so frequently found together in low doses in the related field work, it is possible that the ecosystem of the Caloosahatchee River is more greatly impacted by atrazine and metolachlor than the individual toxicity tests, which indicate rather low ecotoxicity, would suggest. However, looking at the individual response curves for atrazine and metolachlor and comparing them to the mixture responses, it might be that the responses are not truly greater than additive and are skewed by the toxic unit approach. Because the EC50 values were estimated from the transformed logit model s rather than from the response curves alone, there is some difference between the toxic unit concentrations and those predicted directly from the individual response curves (transformed logit model atrazine r2=0.97, metolachlor r2=0.95) This difference may account for the uncertainty in the determination of whether the mixture has an additive or greater than additive effect on Lemna minor in mixtures of atrazine and metolachlor. The concentrations detected in the 18month field sampling surface water project connected this work to the laborat ory mixture experiments with Lemna minor In the field, the highest concentration of atrazine detected (2900 ng/L= 0.0029 mg/L) was almost eighty times low er than the calculated laboratory atrazine EC50 for frond production in Lemna minor ( 0. 23 m g/L). The highest concentration of metolachlor detected (270 ng/L= 0.00027 mg/L) was approximately five hundred times lower than the calculated laboratory metolachlor EC50 for frond production in L minor ( 0. 130 m g/L). That means that during our sampling period, th e re were no samples in which the river surface water contained concentrations of atrazine and metolachlor greater than our
130 calculated EC50 for the individual compounds But given the synergistic response of Lemna minor to mixtures of atrazine and metolachlor, it is not clear whether these comparisons to individual EC50 values are meaningful. Further research is needed to quantify the pesticides of the Caloosahatchee River on a finer scale. With a monthly sampling regime, a good picture of the types of pestic ides present in the river was obtained; but to get a clearer picture of the more subtle patterns, a more frequent sampling regime for a longer duration would be beneficial. In addition, the new technique produced slightly lower spike recoveries than the established technique; further studies could be conducted to optimize this new method under the unique conditions of the Caloosahatchee River There is a lack of available data on mixtures in surface waters that deserves attention. A large scale study with the goal of identifying the effects of multiple compounds (greater than 2) would be greatly beneficial to resource managers and those trying to understand the true effects of pesticides on the organisms in the Caloosahatchee River. Other aspects that coul d be explored in such experiments besides increasing the number of compounds considered, might include a focus on the lower concentrations that are more common in the surface waters, the effects of turbidity due to the mixing at water management structures along the river and the possibility of organism recovery in between pulses of pesticides detected in the river.
131 LIST OF REFERENCES An YJ Lee WM. Decreased toxicity to terrestrial plants associated with a mixture of methy; tert butyl ether and its metabolite tert butyl alcohol. Env iron Tox Chem 2007; 26(8) : 171116. Aliberti MA, Allan E, Allard S, Bauer DJ, Beagen W, Brad t SR, Carlson B, Carlson SC, Doan UM, Godkin WT, Greene S, Haney JF, Kaplan A, Maroni EE, Melillo S, Murby AL, Smith (Nowak) JL, Ortman B, Quist JE, Reed S, Rowin T, Schmuck M, Stemberger RS An ImageBased Key To The Zooplankton Of The Northeast (USA). Version 3.0. UNH Center for Freshwater Biology, Department of Biological Sciences : University of New Hampshire; 2 009. Arnon D Copper enzymes in isolated chloroplasts. Polyphenoloxidase in Beta vulgaris. Plant Physiol 1949; 24:1 15. Banks KE, Wood SH, Matthews C, Thuesen KA. Joint acute toxicity of diazinon and copper to Ceriodaphnia dubia. Environ Toxicol Chem 2003; 22(7):15627. Berenbaum MC. The expected effect of a combination of agents: the general solution. J Theo Biol 1985; 114: 41331. Carriger JF, Rand GM. Aquatic risk assessment of pesticides in surface waters in and adjacent to the Everglades and Biscayne National Parks: I. Hazard assessment and problem formulation. Ecotox 2008; 17(7) :66079. Carter DS Ly dy MJ, Crawford C G Water quality assessment of the White River Basin, Indianaanalysis of available information on pesticides, 1972 92: U.S. Geological Survey Water R esources Investigations Report 944024; 1995. Clarke KR, Gorley RN. Primer v6: User Manual/Tutorial. Plymouth, UK: PRIMER; 2006. Clarke KR, Warwick RM. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation. 2nd edition. Plym outh, UK : PRIMERE ; 2001. Day KE, Hodge V The toxicity of the herbicide metolachlor, some transformation products and a commercial safener to an alga (Selenastrum capricornutum), a cyanophyte (Anabaena cylindrica) and a macrophyte (Lemna gibba). Water Q ual Res Jo Canada 1996; 31(1) : 197214. Doering PH, Chamberlain RH. Water Quality in the Caloosahatchee Estuary, San Carlos Bay, and Pine Island Sound, Florida. In: Proceedings of the Charlotte Harbor Public Conference and Technical Symposium; March 151 6, 1997, Punta Gorda, Florida, pp. 7380. Charlotte Harbor National Estuary Program Technical Report No. 9802, West Palm Beach, Florida, South Florida Water Management District ; 1998.
132 Doering PH, Chamberlain RH. Water quality and source of freshwater discharge to the Caloosahatchee Estuary, Florida. Jo Amer Water Res Assoc 1999; 35(4) : 793806. Downing HF, Delorenzo ME, Fulton MH, Scott GI, Madden CJ, Kucklick JR Effects of the a gricultural p esticides a trazine, chlorothalonil, and e ndosulfan on South Florida m icrobial a ssemblages: South Florida Ecosystems (Guest Editors: Gary M. Rand and Piero R. Gardinali) Ecotox 2004; 13(3) :24560. Environment Canada. Biological Test Method: Test for Measuring the Inhibition of Growth Using the Freshwater Macrophyte Lemna minor. Method Development and Application Section. Report EPS1/RM/37 Ottawa, ON : Environmental Technology Centre Environment Canada; 1999. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Document: Heptachlor. List A. C ase 0175. Office of Pesticide Programs, Special Review and Reregistration Division ; 1992. EPA (U.S. Environmental Protection Agency) Methods for measuring the acute toxicity of effluents and receiving waters to freshwater and marine organisms. EPA 600/ 4 90/027F. Cincinnati, OH ; 1993. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED): Metolachlor. Prevention, pesticides and Toxic Substances (7508W). EPA 738R 95006; 1995. EPA (U.S. Environmental Protection Agency) Ecological Effects Test Guidelines. OPPTS 850.4400 Aquatic Plant Toxicity Test Using Lemna spp., Tiers I and II. Prevention, Pesticides and Toxic Substances (7101). EPA 712C 96156; 1996. EPA (U.S. Environmental Protection Agency) Reregistration Eligib ility Decision (RED): Pendimethalin. Prevention, pesticides and Toxic Substances (7508W). EPA 738R 97007; 1997. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED): Metribuzin. Prevention, pesticides and Toxic Substances (7508W). EPA 738R 97006; 1998. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED): Chlorothalonil. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 99004; 1999. EPA (U.S. Environmental Protection A gency) Reregistration Eligibility Decision (RED) for Ethion. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 01002; 2001.
133 EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED) for Endosulfan. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 02013; 2002. EPA (U.S. Environmental Protection Agency) Interim Reregistration Eligibility Decision (IRED): Atrazine. Case 0062; 2003. EPA (U.S. Environmental Protection Agency) National Primary Drink ing water Regulations ; 2004a http://www.epa.gov/safewater/mcl.html#mcls Accessed September 30, 2012. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED): M ethoxychlor. Prevention, pesticides and Toxic Substances (7508W). EPA 738R 04010; 2004b. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED): Ametryn. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 05006; 2005. EPA (U.S. Environmental Protection Agency) List of Impaired Waters (Florida), Charlotte Harbor, Caloosahatchee 3090205; 2006a. http://www.epa.gov/surfgulf/florida/imp caloosahatchee.html Accessed September 30, 2012. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED) for Chlorpyrifos. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 01007. February 2002 (interim). A ccepted July 2006; 2006b. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED) f or Diazinon. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 04006. May 2004 (interim). Accepted July 2006; 2006c. EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED) for Ethoprop. Prevention, pesticides and Toxic Substances (7508C). EPA 738R 06018. September 2001 (interim). Accepted February 2006; 2006d EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED) for Simazine. Prevention, pesticides and Toxic Substances ( 7508P). EPA 738R 06008; 2006e EPA (U.S. Environmental Protection Agency) Reregistration Eligibility Decision (RED) for Malathion. Prevention, pesticides and Toxic Substances (7508P). EPA 738R 06030; 2009. Fairchild JF, Ruessler DS, Heverland P S, Carlson AR. Comparative Sensitivity of Selenastrum capricornutum and Lemna minor to Sixteen Herbicides. Arch Environ ContamToxicol 1997; 32:3537
134 Fairchild JF, Ruessler DS, Ron A Comparative sensitivity of five species of macrophytes and six species of algae to atrazine, metribuz in, alachlor, and metolachlor. Env Tox icol Chem 1998; 17( 9 ) :1830 34. FASS ( Florida Agricultural Statistics Service ) F lorida Agriculture: Citrus Chemical Usage. August 2000. USDA, NASS; FDACS, DMD; UF, IFAS; 2000. FASS ( Florida Agricultural Statistics Service ) Citrus Summary 2005 2006; 2007 FDACS ( Florida Department of Agriculture and Consumer Services ) Florida Agricu ltural Statistical Directory ; 2007. FDEP ( Florida Department of Environmental Protection) Basin Status Report: Caloosahatchee. Division of Water Resource Management, South District, Group 3 Basin ; 2003. FDEP ( Florida Department of Environmental Protecti on) Phase I Draft, South Florida Water Quality Protection Program. V. Caloosahatchee Study Area 2005. http://www.dep.state.fl.us/southeast/WRMEP/wqpp/wqpp.htm. Accessed September 30, 2012. Finney DJ. Probit analysis: A statistical treatment of the sigmoid response curve. Cambridge, U K : Cambridge Uni versity Press; 1947. Fishel FM. Pesticide Use Trends in the U.S.: Global Comparison. University of Florida IFAS Extension Publicatio n PI 143.Pesticide Information Office, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida; 2007. Flaig EG Capece J. Water Use and Runoff in the Caloosahatchee Watershed. In: Proceedings of the Char lotte Harbor Public Conference and Technical Symposium; March 15 16, 1997, Punta Gorda, Florida, pp. 7380. Charlotte Harbor National Estuary Program Technical Report No. 9802, West Palm Beach, FL: South Florida Water Management District ; 1998. Gao J Liu L Liu X, Lu J, Zhou H, Huang S, Wang Z., S pear PA. Occurrence and distribution of organochlorine pesticides DDT,and heptachlor epoxide in surface water of China. Environ Internat 2008; 34(8):10971103. Gianessi L P, Marcelli M B. Pesticid e use in U.S. Crop Production. National Summary Report ; 1997. Gilli om RJ, Barbash JE, Crawford CG, Hamilton PA, Martin JD, Nakagaki N, Nowell LH, Scott JC, Stackelberg PE, Thelin GP Wolock DM The Quality of Our Nations Waters Pesticides in the Nations Streams and Ground Water, 1992 2001 (Revised). U.S. Geological Survey Circular 1291; 2007.
135 Grenier G, Proteau L, Beaumont G. Lipid synthesis by isolated duckweed (Lemna minor) chloroplasts in the presence of a sublethal concentration of atrazine. Can Jo Bot 1989; 67(8) : 22615. Haanstra L, Doelman P, Oude Voshaar JH. The use of sigmoidal doseresponse curves in soil ecotoxicological research. Plant Soil 1985; 84: 293 7. Hampson PS, Treece Jr. MW, Johnson GC, Ahlstedt SA, Connell JF. Water quality in the Upper Tennessee River Basin, Tennessee, North Carolina, Virginia, and Georgia 1994 98: U.S. Geological Survey Circular 1205 ; 2 000. HarmanFetcho JA, Hapeman CJ, McConnell LL, Potter TL, Rice CP, Sadeghi AM, Smith RD, Bialek K, Sefton KA, Schaffer BA, Curry R. Pesticide occurrence in selected south Florida canals and Biscayne Bay during high agricultural activity. Jo Ag Food Chem 2005; 53: 60408. Havens KE, Bull LA Warren GL, Crisman TL, Phlips EJ, Smith JP. Food Web Structure in a Subtropical Lake Ecosystem. Oikos 1996; 75(1) : 2032. HSDB ( Hazardous Substances Data B ank ) A trazine. CASRN: 1912 249 ; 2012. http://toxnet.nlm.nih.gov/cgi bin/sis/search/f?./temp/~sKEicH:1 Accessed September 30, 2012. HSDB ( Hazardous Substances Data Bank ) M etolachlor CASRN: 51218 452 ; 2012. http://toxnet.nlm.nih.gov/cgi bin/sis/search/a?dbs+hsdb:@term+@DOCNO+6706. Accessed September 30, 2012. Hendley P, Giddings J. Draft report of the Aquatic Workgroups of ECOFRAM (Ecological committee on FIFRA risk assessment) ; 1999. Hillman WS. The Lemnaceae or Duckweeds: A review of the description and experimental literature. Bot Rev 1961; 27: 221 87. Hively WD, Hapeman CJ, F isher TR, Rice C, Mccarty GW, Mcconnell LL, Downey PM, Nino De Guzman G, Bialek Kalinski KM, Lang MW. Relating nutrient and herbicide fate with landscape features and characteristics of 15 subwatersheds in the Choptank River Watershed. Jo Ag Food Chem 2011; 409: 3866 78. Huckins JN, Petty JD, England DC Distribution and impact of trifluralin, atrazine, and fonofos residues in microcosms simulating a northern prairie wetland. Chemosphere 1986; 15: 563 88. James RT, Zhang J. Chapter 10: Lake Okeechobee Pr otection Program State of the Lake and Watershed. In: 2008 South Florida Environmental Report Volume I West Palm Beach, FL: South Florida Water Management District ; 2008.
136 Kendall RJ, Anderson TA, Baker RJ, Bens CM, Carr JA, Chiodo LA, Cobb GP, Dicker son RL, Dixon KR, Fram LT, Hooper MJ, Martin CF, McMurray ST, Patino R, Smith EE, Theodoakis CW. Ecotoxicology. In CD Klaassen (ed): Cassarett and Doulls Toxicology: The basic science of poisons, pp. 673 709. New York : McGraw Hill; 2001 Kimes CA, Croc ker L C The Caloosahatchee River and its Watershed. A Historical Overview Submitted to Florida Gulf Coast University Library Services Under Subcontract From the Florida Center for Environmental Studies. Fort Myers, Florida; 1998. Konstantinou IK, Hela D G, Albanis TA The status of pesticide pollution in surface waters (rivers and lakes) of Greece. Part I. Review on occurrence and levels. Environ. Pollut 2006; 141(3) : 55570. Kotrikla A, Gatidou G, Lekkas TD. Toxic effects of atrazine, deethyl atrazine, deisopropyl atrazine and metolachlor on Chlorella fuscavar fusca. Global Nest 1999; 1 ( 1 ) : 3945. Kotrikla A, Lekkas T, Bletsa G. Toxicity of the herbicide atrazine, two of its degradation products and the herbicide metolachlor on photosynthetic microorganisms. Fresenius Environ Bull 1997; 6 : 5027. Laabs V, Amelunga W, Pintob AA, Wantzenc M, da Silvab CJ, Zecha W. Pesticides in Surface Water, Sediment, and Rainfall of the Northeastern Pantanal Basin, Brazil. Jo of Environ Qual 2002; 31(5) :163648. Leboulanger C, Bouvy M, Carr C, Cecchi P, Amalric L, Bouchez A, Pagano M Sarazin G. Comparison of the Effects of Two Herbicides and an Insecticide on Tropical Freshwater Plankton in Microcosms. Arch of Environ Contam and Tox icol 2011; 61(4) : 599613. Lehotay S J, Harman Fetcho JA, McConnell LL. Agricultural pesticide residues in oysters and water from two Chesapeake bay tributaries. Mar Pollut Bull 1998; 37(1) : 3244. Lichtenthaler HK Wellburn AR. Determinations of total carotenoids and chlorophylls a and b of leaf extracts in different solvents. Biochem Soc Trans 1983; 11:5912. Lin YJ, Karuppiah M, Shaw A, Gupta G. Effect of simulated sunlight on atrazine and metolachlor toxicity of surface waters. Ecotoxicol Environ Saf 1999; 43(1) : 357. Litchfield JT. A method for rapid graphic solution of time per sent effects curves. Jo of Pharmacol Exp Theo r 1947; 97: 399408. Liu B, McConnell L, Torrents A. Herbicide and Insecticide Loadings from the Susquehanna River to the Northern Chesapeake Bay. J o Agric Food Chem 2002; 50: 435892.
137 Marking LL, Dawson VK. Method for assessment for toxicity or efficacy of mixtures of chemicals. U.S. Fish Wild Serv Fish Control 1975; 67:18 McConnell LL, Rice CP, Hapeman CJ, Drakeford L, Harman Fetcho JA, Bialek K, Fulton MH, Leight AK, Allen G Agricultural Pesticides and selected degradation products in five tidal regions and the main stem of Chesapeake Bay, USA. Environ Tox icol Chem 2007; 26( 12) : 256778. Miles CJ Pfeuffer RJ Pesticides in Canals of South Florida. Arch. Environ Cont am Tox icol 1997; 32(4) : 33745. Merritt RW, Cummins KW, Berg MB, Novak JA, Higgins MJ, Wessell KJ, Lessard JL. Development and application of a macroinvertebrate functional group approach in the bioassessment of remnant river oxbows in southwest Florida. Jo North Amer Benth Soc 2002; 21(2) : 290310. Moore MT, Locke M A Phytotoxicity of atrazine, s metolachlor and permethrin to Typhalatifolia (Linneaus) germination and seedling growth. Bull. Environ Contam Tox 2012; 89 : 29295. Munn MD, Gilliom RJ. Pesticid e Toxicity Index for Freshwater Aquatic Organisms. Water Resources Investigations Report 014077. Sacramento, California: US Geological Survey. National Water Quality Assessment Program ; 2001. NASS ( National A gricultural Statistics Service ) Census of Agriculture, State and County Profile ; 2002. Newman MC Aplin M Enhancing toxicity data interpretation and prediction of ecological risk with survival time modeling: An illustration using sodium chloride toxicity to mosquitofish (Gambusia holbrooki). Aquatic Toxicol 1992; 23: 8596. Ne wman MC McCloskey JT Time to event analysis of ecotoxicity data. Ecotox 1996; 5 : 187 96. Newman, M C and Unger, M A. ( 2003) Fundamentals of Ecotoxicology. 2nd ed. Lewis Publishers : Boca Raton, FL, USA. 458 pp. Newmaster SG Harris AG, Kershaw LJ. Wetland Plants of Ontario. Edmonton, Alberta: Long Pine Publishing; 1997. Nino De Guzman GT, Hapeman C J, Prabhakara K, Codling EE, Shelton DR, Rice C, Hively WD, Mccarty G W, Torrents A. Potential pollutant sources in a Choptank River subwatershed: Influence of agricultural and residential land use and aqueous and atmospheric sources. Sci Tot Env 2012; 430: 2709. Ott WR. Environmental Indices: Theory and Practice. Ann Arbor, MI : Ann Arbor Science Publishers ; 1978.
138 Overmeyer JP, A rmbrust K L, Noblet R Susceptibility of black fly larvae (Diptera: Simuliidae) to lawn care insecticides individually and as mixtures. Environ Tox icol Chem 2003; 22(7) : 158288. Pait AS, DeSouza DR, Faroow D. Agricultural Pesticides in Coastal Areas: A Nat ional Summary Strategic Environmental Assessments Division. Rockville, MD: ORCA/NOS/NOAA; 1992 Pape PA, Lydt MJ Synergistic toxicity of atrazine and organophosphate insecticides contravenes the response addition mixture model. Environ Tox icol Chem 1 997; 16(11) : 241520. Pfeuffer RJ Rand GM. South Florida Ambient Pesticide Monitoring Program. Ecotox 2004; 13(3) : 195205. Porter WP, Green SM Debbink NL, Carlson I. Gro u ndwater pesticides: Interactive effects of low concentrations of carbamates aldicarb and methomyl and the triazine metribuzin on thyroxine and somatotropin levels in white rats. J Toxicol Environ Health 1993; 40 : 1534. Pritchard JB. Aquatic Toxicology Past, present, and prospects. Environ. Health Perspect 1993; 100:24957. Roy R, Campbel P. Survival time modeling of aluminum and zinc in soft water at low pH. Aquatic Tox 1995; 33: 155176. SFWMD. Geographic Information Systems Data Catalog ; 1998. http://www.sfwmd.gov/po rtal/page/portal/levelthree/GIS Accessed September 30, 2012. SFWMD ( South Florida Water Management District ) Caloosahatchee Water Management Plan; 2008. http://www.sfwmd.gov/portal/page/portal/xrepository/sfwmd_repository_pdf/caloo s_mngmt_plan.pdf Accessed September 30, 2012. SFWMD ( South Florida Water Management District ) DBYHDRO ; 2012. ht tp://www.sfwmd.gov/dbhydro. Accessed September 30, 2012. Shahane AN Summary of Agricultural Pesticide Use in Florida: 19992002. Florida Department of Agriculture and Consumer Services ; 2003. Shahane, AN. Summary of Agricultural Pesticide Use in Florida: 20032006. Florida Department of Agriculture and Consumer Services ; 2008. Simard S, Grenier G; Beaumont G Morphometric analysis of ultrastructural changes induced by a sublethal concentration of atrazine on young Lemna minor chloroplasts. Jo Plant Phys and Biochem (Paris) 1990; 28(1) : 4955.
139 Solomon KR, Bake r DB, Richards RP, Dixon KR, Klaine SJ, La Point T W Kendall RJ, Weisskopf CP, Giddings JM, Giesy JP, Hall Jr LW, Williams WM. Ecological risk assessment of atrazine in North American surface water s. Env Tox icol and Chem 1996; 15(1) : 3176. Sprague JB. Measurement of pollutant toxicity to fish. II. Utilizing and applying bioassay results. Water Res 1970; 4 : 3 32. Squillace PJ, Scott JC, Moran MJ, Nolan BT, Kolpin DW VOCs, p esticides, n itrate, and t h eir m ixtures in g roundwater u sed for d rinking w ater in the United States. Environ Sci Tech 2002; 36(9) : 192330. Stephan CE, Mount DI, Hansen DJ, Gentile JH, Chapman GA, Brungs WA. Guidelines for deriving numerical national water quality criteria for the protection of aquatic organisms and their uses: U.S. Environmental Protection Agency, PB 85227049; 1985. http://www.epa.gov/waterscience/criteria/85guidelines.pdf Accessed Septem ber 30, 2012. Stokes ME, Davis CS, Koch GC. Categorical Data Analysis Using the SAS system. Cary, NC: SAS; 1995. Taraldsen JE NorbergKing TJ. New method for determining effluent toxicity using duckweed (Lemna minor). Environ Tox icol Chem 1990; 9 : 7617. Thurman EM, Goolsby DA, Meyer MT, Kolpin DW. Herbicides in surface waters of the midwestern United States the effect of spring flush. Environ Sci Tech 1991; 25( 10 ) : 1794 6. US Census Bureau. United States Census Bureau State & County QuickFacts ; 2012. http://quickfacts.census.gov/qfd/states/12/12071.html Accessed September 30, 2012. USDA (US Department of Agriculture) Natural Resources Conservation Service. Plants Database. Conser vation Plant Characteristics for Lemna minor ; 2012. http://plants.usda.gov/java/charProfile?symbol=LEMI3. Accessed September 30, 2012. USDA (US Department of Agriculture) Quick Stats ; 2012. http://quickstats.nass.usda.gov/data/maps/ Accessed September 30, 2012. USGS (US Geological Survey) The Quality of Our Nations Waters: Pesticides in the Nation s Streams and Ground Water, 19922001. Circular 1291; 2007. http://pubs.water.usgs.gov/cir1291. Accessed September 30, 2012. Van der Geest HG, Greve GD, Boivin M, Kraak MS, Van Gestel CM. Mixture toxicity of copper and diazinon to larvae of the mayfly (Ephoron virgo) judging additivity at different effect levels. Environ Tox icol Chem 2000; 19(12) : 29002905.
140 Van gestel CM, Hensbergen PJ. Interaction of Cd and Zn toxicity for Folsima candida willem (Colembola:Isotomidae) in rel ation to bioavailability in soil. Environ Tox icol Chem 1997; 16: 117786. Vencill WK. Herbicide handbook. 8th ed. Lawrence, KS: Weed Science Society of America; 2002. Volety AK. Effects of salinity, heavy metals and pesticides on health and physiology of oysters in the Caloosahatchee Estuary, Florida. Ecotoxicology 2008; 17(7) : 57990. Whitall D, Hively WD, Leight AK, Hapeman CJ, Mcconnell LL, Fisher T, Codling EE, Rice C, Mccarty GW, Sadeghi AM. Pollutant fate and spatiotemporal variability in the C hop tank R iver estuary: factors influencing water quality. Sci Tot Env 2010; 408:20962108. Wilson PC, Boman B J Albano JP. Copper losses in surface runoff from flatwoods citrus production areas. Bull Environ Contam Toxicol 2012; Accepted 6/17/2012. Wilson PC, Boman BJ. Characterization of selected organonitrogen herbicides in South Florida canals: exposure and risk assessments. Sci Tot Environ 2010; 412(13) : 119126. Wilson PC, Boman BJ FergusonFoos J. Norflurazon and simazine losses in surface runoff wate r from flatwoods citrus production areas. Bull Environ Contam Toxicol 2007; 78(5) : 341344. WHO (World Health Organization). The WHO recommended classification of pesticides by hazard and guidelines to classification: 2009. Geneva, Switzerland: WHO Press ; 2010. WHO (World Health Organization). Guidelines for drinkingwater quality 4th ed. Geneva, Switzerland: WHO Press ; 2011. Zhao Y Newman MC. Ecotoxicology entry for environmental encyclopedia edited by L. Shugart, Elsevier, Inc ; 2005. http://www.vims.edu/people/newman_mc/pubs/ZhaoNewman2005.pdf Accessed October 1, 2012.
141 BIOGRAPHICAL SKETCH Ramona SmithBurrell grew up in Pittsburgh, Pennsylvania. She attended Florida I nstitute of Technology for her undergraduate work, earning a B.S. in m arine b iology and a B.S. in e cology in 1997. She earned an M.S. in e cology and conservation b iology from Florida Institute of Technology in 2002, studying the demographics of the rare, e ndemic scrub mint, Dicerandra thinicola. She completed her doctoral work at the U niversity of Florida in the Soil & Water Science Department with an advisory committee consisting of Dr. Chris Wilson (UF), Dr. Samira Daroub (UF), Dr. Zhenli He (UF), Dr. Ste ve Roberts (UF), and Dr. Cathleen Hapeman (USDA). She is an Associate Professor of b iology at Brevard Community College and lives in Malabar, Florida with her husband, Jeff Burrell their two children, and two dogs