Attenuation of atrazine and its major degradation products in a restored riparian buffer system

MISSING IMAGE

Material Information

Title:
Attenuation of atrazine and its major degradation products in a restored riparian buffer system
Physical Description:
Book
Language:
English
Creator:
Gay, Paige Adams
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 28077359
oclc - 847496255
System ID:
AA00022650:00001

Table of Contents
    Title Page
        Page i
        Page i-a
    Dedication
        Page ii
    Acknowledgement
        Page iii
    Table of Contents
        Page iv
        Page v
        Page vi
    List of Tables
        Page vii
        Page viii
    List of Figures
        Page ix
        Page x
        Page xi
    Abstract
        Page xii
        Page xiii
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
    Chapter 2. Literature review of atrazine in the environment
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
    Chapter 3. Study site in Tifton, Georgia
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
    Chapter 4. Water study
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
    Chapter 5. Soil study
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
    Chapter 6. Summary, conclusions, and recommendations
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
    Appendix A. Ground water well dataset
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
    Appendix B. Surface runoff water dataset
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
    Appendix C. Surface runoff sediment dataset
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
    Appendix D. Dairy wetland soil properties
        Page 152
        Page 153
    Appendix E. Soil dataset
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
    Appendix F. Contour maps of atrazine concentrations in wells
        Page 167
        Page 168
        Page 169
        Page 170
    Appendix G. Contour maps of deethylatrazine concentrations in wells
        Page 171
        Page 172
        Page 173
        Page 174
    Appendix H. Contour maps of bromide concentrations in wells
        Page 175
        Page 176
        Page 177
        Page 178
    List of references
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
        Page 184
        Page 185
        Page 186
    Biographical sketch
        Page 187
        Page 188
        Page 189
        Page 190
Full Text











ATTENUATION OF ATRAZINE AND ITS MAJOR DEGRADATION PRODUCTS IN
A RESTORED RIPARIAN BUFFER SYSTEM














By

PAIGE ADAMS GAY















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005






LD 1780
12L































This document is dedicated to Russell Alexander Gay.














ACKNOWLEDGMENTS

I would like to thank Dr. Joseph Delfino, Dr. Paul Chadik, and Dr. Craig Kvien for serving on my graduate committee. Special thanks go to Dr. Daniel Spangler for continuing to serve on my committee after retirement. Dr. George Vellidis was instrumental in funding and design of this project. His support was unyielding. I would like to thank Ms. Wynn Page, Ms. Sandy Allen, Ms. Debbie Coker, Ms. Linsey Boone, and Ms. Megan Ivey of the Water Quality Laboratory of the University of Georgia's Department of Biological and Agricultural Engineering in Tifton, Georgia, for the support and dedication in completing this project. Mr. Andy Knowlton maintained the field site and collected samples for the duration of the study. Statistical support was obtained from Mr. Benjamin Mullinix. Ms. Vickie Garrick provided much needed computer support. Many thanks go to my parents and brother for the belief, and to my daughter and husband for allowing me the time.















TABLE OF CONTENTS

page

A CKN OW LED GM EN TS ................................................................................................. iii

LIST OF TABLES ............................................................................................................ vii

LIST OF FIGU RES ........................................................................................................... ix

ABSTRA CT ...................................................................................................................... xii

CHAPTER

I IN TRODU CTION ........................................................................................................ 1

2 LITERATURE REVIEW OF ATR-AZINE IN THE ENVIRONMENT ...................... 8

Chem ical D escription of Atrazine ................................................................................ 9
Problem s A ssociated with Atrazine U sage ................................................................. 10
Fate of Atrazine in the Environm ent .......................................................................... 13

3 STUD Y SITE IN TIFTON GEORGIA ..................................................................... 17

Field Location ............................................................................................................. 17
Dairy W etland Restoration ......................................................................................... 19
Instrum entation ........................................................................................................... 20
M onitoring W ell N etwork ................................................................................... 21
Surface Runoff Collector N etw ork ...................................................................... 22
Flow M easurem ent at D airy W etland Outlet ....................................................... 23
Precipitation M easurem ent .................................................................................. 25

4 W ATER STUD Y ........................................................................................................ 26

M aterials and M ethods ............................................................................................... 26
Atrazine Application ........................................................................................... 26
Field Sam pling Strategy ...................................................................................... 28
Sam ple Preparation .............................................................................................. 30
Instrum ent Analysis ......... ................................................................................... 32
Chem icals and Standards ................... ................................................................. 33
Quality Control/Quality A ssurance Plan .................................................................... 34
D ata Analysis .............................................................................................................. 41



iv








M ass Loads .......................................................................................................... 41
Dilution Param eters ............................................................................................. 45
Statistical Analysis .............................................................................................. 46
Groundwater Concentration M aps ...................................................................... 47
Results and Discussion ............................................................................................... 47
Ground W ater Concentrations ............................................................................. 47
Groundwater M ass Loads .................................................................................... 55
Surface Runoff W ater Concentrations ................................................................. 60
Surface Runoff W ater Statistical Analyses ......................................................... 63
Surface Runoff Sedim ent Concentrations ........................................................... 70
Dilution Results ................................................................................................... 73
Surface Runoff M ass Loads ................................................................................ 78

5 SOIL STUDY ............................................................................................................. 84

M aterials and M ethods ............................................................................................... 84
Field Configuration and Sampling Strategy ........................................................ 84
Sample Preparation .............................................................................................. 86
Instrum ent Analysis ............................................................................................. 88
Quality Control/Quality Assurance Plan ............................................................. 89
M oisture Determination ....................................................................................... 91
Soil Characterization ........................................................................................... 91
Results and Discussion ............................................................................................... 92

6 SUMMARY, CONCLUSIONS, and RECOMMENDATIONS ................................ 96

Summ ary ..................................................................................................................... 96
Conclusions ................................................................................................................. 99
Recomm endations for Future W ork ......................................................................... 100

APPENDIX

A GROUND W ATER W ELL DATASET ................................................................... 103

B SURFACE RUNOFF W ATER DATASET ............................................................. 142

C SURFACE RUNOFF SEDIM ENT DATA SET ....................................................... 147

D DAIRY W ETLAND SOIL PROPERTIES .............................................................. 152

E SOIL DATASET ...................................................................................................... 154

F CONTOUR MAPS OF ATRAZINE CONCENTRATIONS IN WELLS ............... 167

G CONTOUR MAPS OF DEETHYLATRAZINE CONCENTRATIONS IN
W ELLS ..................................................................................................................... 171

H CONTOUR MAPS OF BROMIDE CONCENTRATIONS IN WELLS ................. 175


v








LIST O F REFER EN CES ................................................................................................. 179

B IO G RA PH ICA L SK ETCH ........................................................................................... 187

















































vi














LIST OF TABLES

Table page

1. Properties of A trazine at 20-25'C .................................................................................. 10

2. Ground water well samples QA/QC Summary .............................................................. 36

3. Surface runoff water samples QA/QC Summary ........................................................... 39

4. Surface runoff sediment samples QA/QC Summary ..................................................... 40

5. Groundwater weighted well concentrations averaged over the entire study period
following application for each landscape position .................................................. 48

6. Groundwater weighted well concentrations averaged over the immediate after
application period (4/24/03-7/31/03) for each landscape position ........................... 48

7. Groundwater weighted well concentrations averaged over the final post application
period (8/l/03-4/5/04) for each landscape position .................................................. 49

8. Groundwater weighted well mass loads summed over entire study period following
application for each landscape position .................................................................... 56

9. Groundwater weighted well mass loads for the immediate after application period
(4/24/03-7/31/03) summed for each landscape position .......................................... 58

10. Surface runoff water concentrations averaged over the immediate after application
period (4/24/03-7/31/03) for each landscape position ........................................... 61

11. Surface runoff water concentrations averaged over the final post application period
(8/l/03-4/5/04) for each landscape position ............................................................. 62

12. Surface nmoff sediment concentrations averaged over the immediate after
application period (4/24/03-7/31/03) for each landscape position ........................... 70

13. Surface runoff sediment concentrations averaged over the final post application
period (8/l/03-4/5/04) for each landscape position .................................................. 71

14. Atrazine concentration changes, dilution factors, and estimated concentration
changes due to dilution and other factors for six runoff events ............................... 74




vii








15. Deethylatrazine concentration changes, dilution factors, and estimated
concentration changes due to dilution and other factors for six runoff events...75 16. Deisopropylatrazine concentration changes, dilution factors, and estimated
concentration changes due to dilution and other factors for six runoff events....76 17. Hydroxyatrazine concentration changes, dilution factors, and estimated
concentration changes due to dilution and other factors for six runoff events....77 18. Surface runoff water mass loads summed over entire after application study period
for each landscape position.............................................................. 79

19. Surface runoff sediment loads summed over entire after application study period
foreach landscape position............................................................... 79

20. Mass loads of each analyte and the total water volume exiting the forested buffer
system through the H-flume into the farm pond........................................ 81

21. Removal efficiency from the groundwater, surface runoff water, and surface runoff
sediment matrices complied from mass loads summed for the entire year
following application ..................................................................... 82

22. Soil samples QAIQC Summary ............................................................ 90






























viii















LIST OF FIGURES

Figure 01. Chemical structures for atrazine, deethylatrazine, deisopropylatrazine, and
hydroxyatrazine ............................................................................. 6

2. Three interactions of atrazine within a plant cell ............................................ 9

3. Location of study site in the coastal plain area of Georgia................................ 17

4. Diagram showing typical patterns of subsurface flow in the Tifton Upland............ 18

5. Dairy Wetland restoration plan using the 3-zone buffer system......................... 20

6. Network of paired groundwater monitoring wells used to quantify subsurface flow
through the Dairy Wetland ............................................................... 21

7. Location of the runoff collectors............................................................. 23

8. Photograph depicting the runoff collector configuration .................................. 24

9. Final outlet flume at the Dairy Wetland, 450mm. H-flume................................ 25

10. Network of paired groundwater monitoring wells showing the contributing areas
assigned to each well in order to calculate groundwater loads........4

11. Network of LIFE samplers showing the contributing areas assigned to each in
order to calculate surface runoff loads................................................... 44

12. Atrazine least square means of natural log transformed well concentration data
grouped by study phase................................................................... 50

13. Atrazine least square means of natural log transformed well concentration data
grouped by landscape position for the two study periods following application..51 14. Deethyltrazine least square means of natural log transformed well concentration
data grouped by study phase............................................................. 52

15. Deethylatrazine least square means of natural log transformed well concentration
data grouped by landscape position for the two study periods following
application.................................................................................5



ix








16. Bromide least square means of natural log transformed well concentration data
grouped by study phase ................................................................... 54

17. Bromide least square means of natural log transformed well concentration data
grouped by landscape position for the two study periods following application..55

18. Precipitation at the Dairy Wetland ......................................................... 59

19. Atrazine least square means of natural log transformed surface runoff water data
grouped by study phase................................................................... 63

20. Atrazine least square means of natural log transformed surface runoff water data
grouped by landscape position for the two study periods following application..64

2 1. Deethylatrazine least square means of natural log transformed surface runoff water
data grouped by study phase ............................................................. 65

22. Deethylatrazine least square means of natural log transformed surface runoff water
data grouped by landscape position for the two study periods following
application.................................................................................. 65

23. Hydroxyatrazine least square means of natural log transformed surface runoff
water data grouped by study phase...................................................... 66

24. Hydroxyatrazine least square means of natural log transformed surface runoff
water data grouped by landscape position for the two study periods following
application.................................................................................. 67

25. Deisopropylatrazine least square means of natural log transformed surface runoff
water data grouped by study phase...................................................... 67

26. Deisopropylatrazine least square means of natural log transformed surface runoff
water data grouped by landscape position for the two study periods following
application.................................................................................. 68

27. Bromide least square means of natural log transformed surface runoff water data
grouped by study phase................................................................... 69

28. Bromide least square means of natural log transformed surface runoff water data
grouped by landscape position for the two study periods following application..69

29. Hydroxyatrazine least square means of natural log transformed runoff sediment
data grouped by study phase ............................................................. 72

30. Hydroxyatrazine least square means of natural log transformed runoff sediment
data grouped by landscape position for the two study periods following
application.................................................................................. 72




x








3 1. Deethylatrazine least square means of natural log transformed runoff sediment
data grouped by landscape position for the two study periods following
application.................................................................................. 73

32. Average soil concentrations of each analyte grouped by landscape position ......... 93 33. Location and field number of all monitoring wells and runoff collectors ............ 142















































xi














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

ATTENUATION OF ATRAZINE AND ITS MAJOR DEGRADATION PRODUCTS IN A RESTORED RIPARIAN BUFFER SYSTEM By

Paige Adams Gay

August, 2005

Chair: Joseph J. Delfino
Major Department: Environmental Engineering Sciences

The fate and transport of atrazine and three of its major degradation products were studied in a restored forested riparian buffer system to examine the effectiveness of the system in attenuating the chemicals. Concentrations of atrazine, deethylatrazine, deisopropylatrazine, and hydroxyatrazine were measured in groundwater, surface runoff water, surface runoff sediments, and soils throughout the buffer system following application of the parent compound on a 10 x 100 m strip immediately upslope from the buffer system. The restored buffer consisted of an 8 m grass zone adjacent to the application strip, a 20 m pine forested zone, and a final streamside hardwood zone. The grass and upper 10 m of pines were most instrumental in removal of the target chemicals based on net concentrations per meter of zone width. Deethylatrazine was the predominant degradation product in both water matrices, but removal efficiency (84.2% in ground water and 92.0% in surface runoff water) was also lowest for deethylatrazine. Atrazine, deisopropylatrazine, and hydroxyatrazine each exhibited between 94.7% and



xii








100% removal between mass loads entering the upsiope landscape position beginning at the application strip and the final stream side landscape position in both ground water and surface runoff water. Loads of each analyte associated with surface runoff sediments were least efficiently removed (67.4%-100%) within the system. Adsorption of the analytes to subsurface soils was minimal in this system. It is suggested that further degradation of the analytes and plant uptake are the primary processes involved in the removal of the compounds within the system.








































xiii













CHAPTER 1
INTRODUCTION

Atrazine use remains a source of great controversy. While the Environmental

Protection Agency (EPA) released an Interim Reregistration Decision (IRED) in October, 2003, the human health and environmental impacts continue to be debated. In the IRED (EPA, 2003), the EPA maintained the status of atrazine as not likely to be a human carcinogen. The EPA further concluded that the effects on aquatic plants and organisms are likely to be greatest in areas with concentrations which exceed 10-20 gtg L- on a recurrent or prolonged basis. It also concluded reports of endocrine disruption in frogs, largemouth bass, and salmon as inconclusive and in need of further validation. It has been estimated the national economic impact of banning atrazine would be two billion dollars a year on growers of corn, sorghum, and sugar cane (Farm Monitor, 2003).

Atrazine is the most frequently used pesticide in the United States in agricultural production systems (Aspelin, 1994). The Environmental Protection Agency estimates 36.4 X 106 kg are applied annually in the United States (EPA, 2003). The widespread usage has led to its detection in surface waters and ground waters across the nation, especially in the midwestern cornbelt where its use in corn production is extensive. Vegetative buffer strips and riparian buffer systems are two agricultural Best Management Practices (BMP's) which have been successful in decreasing atrazine concentrations reaching surface and ground waters (Paterson and Schnoor 1992, Lowrance et al., 1997, Vellidis et al., 2002)





2


Vegetative buffer strips trap sediment and increase water infiltration processes

which act to reduce herbicide runoff. In a review of removal of herbicides by vegetated systems, Baker et al. (1995) calculated an average removal rate of 48% of the herbicides in runoff. The range of removal in the studies was 9% to 91 %, and varied based on soil type and moisture content, runoff volume, and concentration of the herbicides, buffer width, and vegetation type. Mickelson and Baker (1993) reported that a 4.6 mn grass buffer system reduced atrazine in runoff by 35% while a 9.1 mn strip reduced atrazine by 59.5%. Atrazine loss in runoff was reduced by 30% using a series of 3 Bermuda grass buffers (9.1 mn wide) and 57% using a series of 3 wheat buffers (9.1 mn in width) in a study by Hoffman (1995).

Baker et al. (1995) further report that Misra et al. (1994) found the amount of

atrazine (in %) removed by grass filter strips increased as the concentration of atrazine in simulated runoff increased. This study attributed the reduction of atrazine to infiltration of the runoff water and herbicide adsorption in the filter strips. Arora et al. (1993) found only 9.3% to 12.5% of atrazine in runoff from a maize field was removed by 20 mn grass filter strips. The diminished removal compared to other studies was attributed to wet antecedent soil conditions which reduced water infiltration into the buffer strips. Hatfield (1995) examined the soils in the grass filter strips used by Arora et al. (1993) and confirmed atrazine from the runoff had been adsorbed by the soil. Following the first runoff event, 20% of the atrazine applied in simulated runoff was recovered in the soil and surface organic matter. The soil concentration decreased over time presumably due to degradation (Hatfield, 1995).





3


In addition to intercepting surface runoff of herbicides, forested riparian systems

have been shown to reduce pesticide concentrations in subsurface flow. Forested riparian buffer systems have been endorsed by the USDA Forest Service as a best management practice (BMP) for the reduction of some pesticides leaving agricultural fields in surface runoff and subsurface flow (Welsch, 1991). Three studies have shown that atrazine concentrations can be effectively diminished in these buffer systems and these are discussed below.

The fate of atrazine in a riparian zone field site in Iowa was addressed by Paterson and Schnoor (1992). The field site was composed of a Nodaway soil series which was stratified and consisted of a 15 cm topsoil layer, 45 cm till layer, a clay layer to 1 m, and a clayey soil extended to the water table. The riparian zone was subdivided into three plots: a barren plot, a plot planted to corn, and a plot planted with deep rooted poplar trees (Populus spp). Atrazine was applied directly to the plots. A series of mass balance equations were solved to quantify the fate of the herbicide in each pathway. Soil biotransformation reactions were found to be the dominant fate pathway for atrazine in the poplar plots, with 77.8% of the atrazine applied being biotransformed. Mechanisms and products of the biotransformation reactions were not addressed in the study. Rate of transformation was estimated based on the disappearance rate in soil or water. Runoff contained 10.2% of the atrazine applied. Soil-water biotransformations accounted for

4.8% of the atrazine, while plant uptake and soil retention each accounted for 3.6% of the atrazine applied. No atrazine was percolated through the soil out of the study boundaries. The authors note the uptake by poplar trees acted to decrease atrazine available for accumulation or transformation in the pore water by competing for the water containing






4


the pesticide. The barren plot accumulated 5.8% of the atrazine applied compared to

3.6% in the poplar plot. Only 65% of the poplar root mass was in the study zone; thus plants with a larger near-surface root zone could further reduce the accumulation of atrazine in a riparian zone.

The runoff pathway was the second largest fate pathway, by percentage, for the mass applied, superseded only by soil biotransformations. Recall, runoff losses were estimated to account for 10.2% of the mass of atrazine applied. However, there was only one runoff collector and it was situated at a site that collected runoff from the corn plot only. The authors assumed runoff concentrations in the barren and poplar plots were proportional to those found at the corn plot, which may not be the case. The authors also noted that the lack of measured residue in the soil and plants represented a large source of uncertainty. Actual measurement of concentrations in runoff, soil, and plant tissue, instead of mathematical calculation, would further validate the major processes involved in the attenuation of atrazine in a riparian system.

Lowrance et al. (1997) found atrazine concentrations were reduced significantly during movement through a mature riparian forest buffer system in both the surface runoff and shallow ground water in a study conducted in Tifton, Georgia. The buffer system consisted of 3 zones emanating from the stream bed. Zone I included the streambed and 10 mn of hardwood trees. Pine trees composed zone 2 which was 40 mn to 55 mn wide. Upsiope of zone 2 and adjacent to the upland field was zone 3, an 8 mn wide grass filter strip. Atrazine transport in surface runoff occurred primarily in rainfall events within a short period after application. During this time, edge of field concentrations






5


averaging 34.1 /ptg L-1 were reduced to 1 ,tg L-1 or less as they entered the zone nearest the stream.

In shallow groundwater, atrazine concentrations were below analytical detection limits (0.16 yig L-) for the first two years of the study. During the third year (1994), atrazine concentrations were above detection limits in the entire buffer system following application with the exception of a small portion of the zone closest to the stream. Pooled data from the third year of the study showed average atrazine concentrations were reduced from 1.29 Mtg L-' at the field edge to less than the analytical detection limit of

0. 16 /,tg L-' in the streamside zone wells. Though not addressed in this study, the authors suggested that infiltration and adsorption were the primary factors causing net concentration decreases.

Another study (Vellidis et al., 2002) was conducted in Tifton, Georgia, at the same site as the current research. The effects of a recently planted three zone forested riparian buffer system on the transport of atrazine in surface and subsurface water were examined. The three zones were, again, hardwood species (zone 1), pine trees (zone 2), and a grass filter strip (zone 3). Zones 1 and 2 were interspersed with various shrubs and grasses. Atrazine was applied onto a 10 m wide by 100 m long application strip immediately upsiope and adjacent to zone 3. Most of the atrazine movement in surface runoff occurred shortly after application. The restored buffer system was effective in diminishing the atrazine concentrations in surface runoff from 12.7 ug U-1 at the field edge to 0.66 ptg L-1 near the stream. Concentration reduction was greatest in the grass buffer zone. There was very little transport of atrazine in the shallow groundwater. This was presumably due to a low application concentration. The authors attempted to apply





6

concentrations comparable to the edge of field concentrations measured by Lowrance et al. (1997), but underestimated the quantity needed to produce equivalent concentrations.

These studies showed that filter strips and forested riparian buffer systems are capable of reducing concentrations of atrazine in both surface and subsurface flow. However, the processes involved in the attenuation have not been clearly defined and measured. It is imperative to understand the processes that affect attenuation to maximize the effectiveness of these systems. In addition, none of the studies examined the fate and transport of the major degradation products of atrazine (Figure 1).

C1

j---N ---N. -.
H H

Atrazine
OH
Cl Cl t.

H N N N N NH2 H H
H H

Deethylatrazine Deisopropylatrazine Hydroxyatrazine


Figure 1. Chemical structures for atrazine, deethylatrazine, deisopropylatrazine, and
hydroxyatrazine.

These degradation products have been found in surface and ground waters nationwide. Deethylatrazine and deisopropylatrazine are both phytotoxic and speculated to retain the same toxicity as the parent compound (Liu et al., 1996).

The overall goal of this research was to study the attenuation of atrazine and its major degradates (deethylatrazine, deisopropylatrazine, and hydroxyatrazine) in a riparian forest buffer system. The specific objective of the study was to determine and





7

quantify the fate and transport of atrazine and its major degradates between the edge of the riparian zone and the stream channel in

,-shallow groundwater

,-surface runoff water and sediments

,-forest floor litter and soils

This included assessing the concentration and mass load changes of atrazine and its degradation products as they were formed and transported through the wetland, as well as determining which degradation processes were most prevalent in the system. The hypothesis tested was that the concentration of the parent compound plus each of the degradation compounds would diminish as the distance of the landscape position from the application site toward the stream was increased.












CHAPTER 2
LITERATURE REVIEW OF ATRAZINE IN THE ENVIRONMENT

In 1994, the Environmental Protection Agency estimated pesticides were used on 900,000 farms and in and around 69 million homes nationwide (Aspelin, 1994). Atrazine and metolachlor were the most widely used pesticides by volume. Atrazine was also the most widely used pesticide in agricultural crop production. Atrazine is used for preemergence and post emergence weed control of broadleaf and grassy weeds in agricultural and industrial applications. It is frequently used in agricultural production of sorghum, orchard crops, sugarcane, and most extensively in corn in the midwest. Atrazine is also used for weed control in conifer restoration and Christmas tree plantations, and vegetative control on fallow land. Application is on the soil surface or by weed foliar spray at a nominal rate of 0.9-1.8 kg season-1. The EPA estimates approximately 31.7-34.0 x 106 kg of active ingredient atrazine are applied annually in the United States (EPA, 2003). In Georgia for the year 1997 (National Center for Food and Agricultural Policy, 1997), the predominant uses for atrazine and approximate kilograms of active ingredient applied statewide were corn production (1.6 x 105 kg), sorghum production (1.3 x 104 kg), and sod production (3.6 x 103 kg).

Atrazine is taken into the plant through the roots or foliage and moves in the xylem to the plant leaves where it kills the plant by inhibiting photosynthesis (Gunsolus and Curran, 1998). If applied foliarly, atrazine is less mobile and does not move out of the





9







cHtoRoPLAs Ho00c NUCLEUS




NON-PHYTOTOXIC
METABOLISM WATER-SOLUBLE
ATRAZiNE METABOLITES




ATRAZINE
Figure 2. Three interactions of atrazine within a plant cell (1) atrazine metabolism (2)
inhibition of photosynthesis (3) reduction of atrazine in the chloroplast and
recovery of photoactivity. Source: Shimabukuro and Swanson (1969)

leaf tissue. Atrazine is absorbed into the cells (Shimabukuro and Swanson, 1969) and enters the chloroplasts (Figure 2) where it interrupts the oxidation of water in the Hill reaction. Plants which are able to tolerate atrazine do so by metabolizing atrazine to less toxic forms. The Hill reaction is shown below where A is an electron acceptor:

2 H20 + 2 A > (light, chloroplasts) > 2 AH2 + 02.

Chemical Description of Atrazine

Atrazine [6-chloro-N-ethyl-N'-(1-methylethyl)-1,3,5-triazine-2,4-diamine] is a chlorinated triazine herbicide. Triazines are weakly basic and their physicochemical properties vary greatly based on the substitutions occurring at the 2, 4, and 6 positions on the ring structure. The structure for atrazine is given in Figure 1. The molecular formula for atrazine is CsH14C1N5 and the compound has a molecular weight of 215.69 g. In their review of the literature for atrazine properties, Hornsby et al. (1996) selected the following values at 20-250 C (Table 1).






10


Table 1. Properties of Atrazine at 20-251C
Water Solubility 33 mg L-'
Field Half Life 60 days
Sorption Coefficient (Koc) 100 ml g-1 organic carbon Vapor Pressure 2.89 X 10-7m.H
pKa 1.68
Log Kow* 2.68
(From Hornsby et al., 1996, and *Paterson and Schnoor, 1992)

Its water solubility is 33 mg L-1; thus atrazine is only slightly soluble in water. This low solubility is particularly important in terms of adsorption as will be discussed later in the paper. The accepted sorption coefficient (Koc) is 100 ml g-1 organic carbon (Paterson and Schnoor, 1992) and is based on the assumption that pesticide sorption by soils is entirely caused by soil organic matter. Vapor pressure is a measure of the tendency of the pesticide to volatilize in its normal pure state and is given as 2.89 x 10-' mm Hg. The authors of this pesticide database (Hornsby et al., 1996) list the pKb value as 12.32, by subtraction from pH 14 gives a pKa of 1.68. This constant represents the tendency of the atrazine to ionize. In normal soils with a pH range 4-8, atrazine molecules should exist as neutral species and behave as nonionic species. Under extremely acidic conditions atrazine can ionize by accepting a proton. However, this only occurs at pH levels less than the pKa, of 1.68. The half life of a pesticide is the time required for the pesticide to dissipate or degrade to half of its original concentration. The value varies based on location, soil type, and climatic conditions; thus the authors recognized the uncertainty of the selected 60 day half life for atrazine in the field.

Problems Associated with Atrazine Usage

The Maximum Contaminant Level (MCL) for atrazine in drinking water has been set by the EPA at 3 gag L-'(EPA, 2005 a). While the EPA has determined it is not likely to be a human carcinogen, it does acknowledge the occurrence of mammary gland tumors in





11

rats (EPA, 2005b). In addition, the EPA (2005b) lists potential health effects from the ingestion of water containing atrazine as cardiovascular system problems and reproductive difficulties. The EPA (2005b) gives a range of bioconcentration factors for atrazine in six fish species (0.3 to 2.0), and suggests that atrazine is not expected to bioconcentrate.

The metabolites of atrazine do not have MCLs or Health Advisory Levels set by the EPA. Deisopropylatrazine and deethylatrazine are both phytotoxic (Liu et al., 1996). Since deisopropylatrazine and deethylatrazine have similar molecular structures as atrazine, including the chlorine atom, it is speculated that they should exhibit human toxicity similar to atrazine (Liu et al., 1996). Lerch et al. (1995) give a brief review of the toxicity of hydroxylated degradation products of atrazine. Generally, the hydroxylated forms are considered nontoxic as they exhibit no phytotoxity in plants or to aquatic photosynthetic microorganisms. However, hydroxyatrazine has an acute LD50 of >3000 mg kg1 in rats resulting from kidney damage. Hydroxyatrazine has been found to be non mutagenic and non teratogenic.

While atrazine has been extremely beneficial in terms of improving crop yields, its use has also elicited environmental concerns. Atrazine has been found in surface and ground waters in numerous nationwide studies. The EPA (1990) conducted a National Pesticide Survey encompassing 1300 community water systems and rural domestic wells to determine the frequency and levels of pesticides, pesticide degradates, and nitrate nationwide. Atrazine was the second most commonly detected pesticide, superseded only by metabolites of dimethyl tetrachloroterephthalate (DCPA). The minimum reporting limit for the survey was 0.12 g L- .





12

The EPA estimated that atrazine was present at or above the reporting limit in 1.7% of community wells and 0.7% of rural domestic wells nationwide. Atrazine has been detected year round in groundwater and streams in the midwestern states which have high usage rates of atrazine (Thurman et al., 1991, Burkart and Kolpin, 1993) often at levels greater than the MCL of 3 gg L-1. Atrazine plus two of its major metabolites, deethylatrazine and deisopropylatrazine, were frequently detected in a study of near surface aquifers in the midcontinental United States (Burkart and Kolpin, 1993). Maximum deethylatrazine concentration measured was 2.32 /tg L-', while the maximum concentration measured for atrazine was 2.09 4g L1. A study of the groundwater in Iowa spanning years 1982-1995 (Kolpin et al., 1997) resulted in 19.5% of 1485 samples having atrazine concentrations exceeding 0.10 /tg L-, with the maximum concentration of atrazine recorded as 21.0 4g L-1. However, only 0.8% of the samples exceeded the MCL.

A nationwide survey of pesticides in 58 rivers and streams in the United States was conducted by the U.S. Geological Survey (Larson et al., 1999). The streams were located in agricultural basins, urban basins, and basins with complex mixed land uses. Atrazine had the highest annual mean detection frequency (-80%) of the 46 pesticides tested for in the study, and deethylatrazine had an annual mean detection frequency of about 50%, third in frequency following atrazine. Less than 5% of monthly median concentrations of atrazine were greater than 0.1 ug L1 and less than 0.1% was greater than 1 gg L1. Atrazine was found to exceed the maximum contaminant level (3 4g L-) at 16 of the 50 sites where samples were collected in the study.





13

In Missouri, a study of surface water quality in Goodwater Creek Watershed was made in 1993-1994 (Donald et al., 1998). Atrazine and deethylatrazine were detected year round in the watershed at concentrations greater than 0.1 4tg L-, with maximum concentrations occurring soon after application. Maximum atrazine concentrations at the watershed outlet were 62.4 /tg L- in 1993 and 22.9 4g L-' in 1994. Maximum concentrations of deethylatrazine (5.6 and 6.2 gg L1 in 1993 and 1994, respectively) and deisopropylatrazine (values not given) lagged slightly behind the peak concentrations for atrazine. In a study of Goodwater Creek, Missouri (Lerch et al., 1995), maximum concentrations of 112 /tg L-1 for atrazine, 23 /g L- for deethylatrazine, and 5.7 4g L-1 for hydroxyatrazine were measured in the creek soon after application. This occurrence of maximum concentrations immediately following application of atrazine has been termed "spring flush" in a study by Thurman. et al. (1991). A 1992-1996 study conducted in the coastal plain area of western Georgia and western Florida (Berndt et al., 1998) found atrazine in three streams draining agricultural basins at concentrations up to 0.86 gg L-. Atrazine was also detected in ground water wells though concentrations were not given.

Fate of Atrazine in the Environment

Once applied, atrazine can be transported to surface water bodies in aqueous form or attached to soil particles by runoff, leach to groundwater, volatilize, or be degraded by photolytic processes. In the soil environment, atrazine can be degraded by microbes, chemically transformed, or adsorbed to solid surfaces.

Runoff of atrazine is most likely to occur only if there is a substantial rainfall event directly after application. Studies have shown that loss of 2% of the total mass of applied atrazine is generally lost via runoff, although losses up to 5% may occur (Hall et





14

al., 1972, Wauchope, 1978). The majority of atrazine lost in runoff is in the dissolved phase. Basta et al. (1997) found that up to 99.8% of atrazine occurred in the water phase. Thus, erosional losses of atrazine are presumed to be quite small or insignificant.

The leaching potential of atrazine varies based on the property of the soils (soil texture, amount and composition of organic material, hydraulic conductivity, pH) amount of degradation occurring in the soil, and rainfall patterns (Nicholls, 1988). Adsorption of atrazine occurs mainly by hydrophobic partitioning from the water phase to the organic fraction of the soil and to a lesser extent to the clay fraction. Soils which adsorb more atrazine will have less atrazine percolating through them. Degradation will also diminish leaching. Leaching will be maximized when heavy rainfall occurs shortly after application. Flury (1996) found 0.1% 5% of atrazine applied was lost by leaching during worst case rain events in loamy and clayey soils.

Volatilization of atrazine should be low since the vapor pressure is 2 x 107mm Hg. Studies have reported losses of
The major abiotic degradation product in soils is hydroxyatrazine (Lerch et al., 1995). Degradation to hydroxyat'azine occurs via photolytic hydrolysis at the soil surface or by chemical hydrolysis of infiltrated atrazine. Biological degradation of atrazine in soil produces deethylatrazine and deisopropylatrazine through dealkylation reactions catalyzed by enzymes (Erickson and Lee, 1989).

Photolysis of atrazine occurs both by direct and indirect processes. Direct photolysis occurs primarily by dechlorination and hydroxylation yielding hydroxyatrazine (Pape and Zabik, 1970, Torrents et al., 1997). Indirect photolysis occurs





15


via the production of *OH radicals from photosensitizers such as dissolved organic matter and nitrate (Torrents et al., 1997). Deethylatrazine and deisopropylatrazine are the major products of indirect photolysis.

Chemical hydrolysis of atrazine to hydroxyatrazine is very slow under neutral pH conditions (Armstrong et al., 1967, Comber, 1999). However, the reaction is catalyzed by highly acidic or alkaline conditions. Under acidic conditions, protonation of a ring or chain nitrogen increases the electron deficiency of the carbon bonded to the chlorine atom. This increases the nucleophilic displacement of the chlorine by the weak nucleophile water. In alkaline hydrolysis the electron deficiency is enough to allow displacement of the chlorine atom by the hydroxyl ion which acts as a strong nucleophile. At neutral pH values, water is not able to displace the chlorine atom. Adsorption of atrazine to the organic fraction in soils can also promote hydrolysis by increasing the electron deficiency of the carbon bonded to the chlorine atom; thus allowing water to displace the chlorine.

Adsorption of atrazine by soil particles occurs by several mechanisms (Koskinen and Harper, 1990, Bohn et al., 1985). Hydrogen bonding, van der Waals forces, and protonation of the atrazine molecule followed by cation exchange with soil particles are minor processes involved in the adsorption of atrazine. Since atrazine mainly exists as a nonionic and nonpolar species at the pH value of most soils, hydrophobic partitioning is the primary mechanism of adsorption. Hornsby et al. (1996) give a sorption coefficient (Koc) of 100 mL g-' of organic carbon at 25' C. This value varies greatly in the literature based primarily on soil composition and pH. For a variety of sand, till, and alluvial soils, Roy and Krapac (1994) calculated Koc values which ranged from 112 1680 mL g-1 of





16

organic carbon. Studies have shown that hydroxyatrazine is more strongly adsorbed than atrazine while deethylatrazine and deisopropylatrazine are less strongly adsorbed than the parent compound (Adams and Thurman, 1991, Lerch et al., 1998).

Numerous bacteria and fungi which are capable of degrading atrazine have been isolated and identified (Kaufman and Kearney, 1970, Adams and Thurman, 1991). Atrazine biodegradation occurs by dealkylation, hydroxylation, or ring cleavage. The dominant mechanism is dealkylation to deethylatrazine and deisopropylatrazine. Enzyme catalysis is suggested to be the method of dealkylation by soil microorganisms (Erickson and Lee, 1989). The dealkylated forms can be further degraded to a variety of products including deaminated species, cyanuric acid, and CO2 and NH3.

Because all of the previously discussed processes are active in a riparian forest

buffer system, it was hypothesized that like the parent compound, atrazine's metabolites can also be effectively attenuated by a riparian forest buffer system.













CHAPTER 3
STUDY SITE IN TIFTON, GEORGIA

Field Location

The field site used in this project is located on the University of Georgia's Animal & Dairy Science Research Farm and is called the Dairy Wetland. The Research Farm is part of the University of Georgia Tifton Campus in Tifton, Georgia (Figure 3). Tifton lies in the coastal plain geologic province of Georgia. The Tifton Upland district of the coastal plain is characterized by undifferentiated Neogene sediments (Watson, 1976) and by intensive crop production on moderately to well-drained upland soils and riparian forests and wetlands on poorly and very poorly drained soils along streams (Lowrance and Leonard, 1988). Upland soils range from flat to 12% slopes. Dense dendritic stream networks result in the typical fann being drained by two or three small streams.






Piedmont





6 a s t a i ri






Figure 3. Location of study site in the coastal plain area of Georgia. (Source: Carl
Vinson Institute of Government, 1999.)


17






18


South Georgia has a long growing season with a 30 year median season length of 259 days (Southeast Regional Climate Center, 1997). The 30 year mean (1961-1990) daily temperature is 18.80 C while the extreme average monthly temperatures are 8.70 C in January and 27.2' C in July. Average annual precipitation in Tifton is 121.7 cm, with the majority occurring during winter, spring, and summer. Precipitation follows a definite seasonal pattern with generally low rainfall from September through November and an increase in precipitation in December through early May. Rainfall typically decreases again in May and early June. Summer thunderstorms and tropical depressions cause July and August to be wetter months on average (Southeast Regional Climate Center, 1997).

Most of the excess precipitation in the Tifton Upland moves either in surface runoff during storm events or infiltrates and creates layers of perched water tables which flow laterally in shallow saturated and unsaturated flow (Figure 4). This is attributed to both a plinthic soil horizon (irreversibly hardened mixture of iron sesquioxides and quartz) beginning at a depth of 1 to 1.5 m and the presence of the Hawthorn Formation, a geologic formation which limits deep recharge to the regional aquifer system,













Figure 4. Diagram showing typical patterns of subsurface flow in the Tifton Upland. The
plinthite layer which acts as an aquiclude, forces flow through the biologically
active root zone of riparian buffers. Aquiclude indicates location of the
plinthite layer.





19


The general hydrology of the Tifton Upland is reflected at the Dairy Wetland and makes this region and the particular site ideal for the study of surface -runoff and shallow subsurface transport of agricultural pollutants into riparian ecosystems. Because top soils generally become shallower near stream channels, subsurface flow is forced through the biologically active root zone of riparian buffers (Figure 4) thus enhancing the potential for degradation of pollutants transported in subsurface flow.

The soil at the Dairy Wetland is an Alapaha loamy sand (fine-loamy, siliceous, acidic, thermic Typic Fluvaquents). The soil of the adjacent upland area is a Tifton loamy sand (fine-loamy, siliceous, thermic, Plinthic Kandiudult) (Calhoun, 1983).

During summer and autumn, overland flow is generally limited to intense rainfall events and the perched water tables frequently recede to a depth below 3 m. In contrast, during winter and spring, the soil profile is often saturated and runoff events are frequent as is ponding in the lowest portion of the landscapes. During this time, stream flow through the wetland is evident.

Dairy Wetland Restoration

The mature riparian forest of the Dairy Wetland was logged in 1985 and replaced with a wet pasture. Two smaller ditches in the headland which met to form a larger ditch were cut after logging to facilitate drainage into a farm pond constructed downslope from the wetland. Over the next 5 years, the ditches filled in with eroded sediment,

Restoration of approximately I ha of the Dairy Wetland began in February, 199 1, when a three zone riparian buffer system as prescribed by the USDA-Forest Service specifications (Welsch, 1991) was established by planting hardwoods in Zone I





20





herbicide
.. ...... application strip
















Figure 5. Dairy Wetland restoration plan using the 3-zone buffer system recommended
by Welsch (1991).

and slash pine in Zone 2 (Figure 5) (Vellidis et al., 1993). Zone 1 is a 10 m wide band of trees with mixed hardwoods in rows including swamp black gum (Nyssa sylvatica var biflora Marsh.), green ash (Fraxinus pennsylvanica L.), and yellow poplar (Liriodendron tulipifera L.). Zone 2 is a 20 m wide band of slash pine (Pinus elliottii Engelm.) in rows. Rows were about 2 m apart. Zone 3 is an 8 m wide strip of Tifton 44 bermudagrass (Cynondon dactylon L.). The entire three zone buffer averages 38 m in width. The site and restoration efforts are described in detail by Vellidis et al. (1993, 2003).

Instrumentation

The Dairy Wetland research site was originally developed to examine the fate and transport of nutrients moving downslope in surface runoff or shallow ground water flow from an upland forage production field receiving liquid dairy manure (Vellidis et al.,





21


1994). Instrumentation for a precursor herbicide study was installed beginning in January, 1993 (Vellidis et al., 2002). Monitoring Well Network

A network of 75 pesticide-quality monitoring wells was installed on the east slope of the Dairy Wetland (Table 6). The network consists of 26 shallow and 29 deep wells. Shallow wells are screened from 0. 1-0.6 m below the soil surface while deep wells are screened from 0.6-2.0 m below the soil surface. Wells were assembled from threaded PVC pipe and screened with 0.25 mm well screen.














Figre6. etorkofpaiedgrondate mnitrig wll uerbdetuniysbufc flow though thaDairycetland











aniaprriateengtseio of elsredrudaen tonaionp eorated setoqunoftpipe. Wusrael





were installed by augering a hole approximately 0.08 ma in diameter to





22


0. 15 mn below the 0.6 ma or 2.0 m screened depth for a well pair. The well was placed in the hole and washed sand poured around the well to just above the screened section. The remainder was filled with a mixture of bentonite and soil extracted from the hole to ensure water would not be able to run down the sides of the wells.

Six well transects were installed with each transect containing 12 wells a shallow and deep well pair at each of 6 positions in the transect. The transect positions, in meters from the edge of the application strip, were: 0 mn (the upslope edge of the Zone 3 grass buffer), 8 m (the upslope edge of the Zone 2 managed forest buffer pines), 13 m, 18 m, 28 m (upslope edge of the Zone 1 forest buffer hardwoods) and 38 mn (near the stream channel (Figure 6). Three wells were also installed at the perimeter of the field to act as controls

Surface Runoff Collector Network

Three transects of four 0.3 m wide Low-Impact Flow Event (LIFE) samplers (Sheridan et al., 1996) were installed to sample surface runoff. In each transect, the collectors were installed at the upslope edge of Zones 1, 2, 3, and at the mid-point of Zone 2 -0, 8, 18, and 28 m, respectively, from the upslope edge of Zone 3 (Figure 7).

Two transects had instruments that retained 10% of the collected sample. The third transect had instruments that retained 1% of the collected sample. This design ensures that a measurable volume can be collected over a wide range of runoff events. The 10% collection is made by splitting the flow into 10 pathways at the back of the collector and collecting flow from one pathway. The 1 % sample is collected by connecting two 10% samplers in series. The sample receptacle is large enough to contain runoff from approximately a 10 year return interval event in the I% samplers. This allowed for large and small runoff events to be sampled without overfilling the sample receptacle.





23
















Figure7. Loationof therunoflcollctors











The water portion being collected passed from the LIFE sampler into a 0. 1 m diameter PVC pipe (Figure 8). The 1.0 m long pipe entered into the upper section of a 57 L plastic barrel which had been buried except for the upper 0.25 m. Approximately, 0.05 m of pebble ballast was placed in the bottom of the barrel both to weight the barrel and prevent it from floating during high water table conditions and to allow for height adjustment of the sample collection vessel a 9.5 L glass barrel jar. Inside the barrel, the PVC inlet pipe took a 900 turn downward into the jar. A teflon bag (0.35 m x 0.69 m) was placed inside a polyethylene bag and the two placed into the jar to receive the incoming sample. Flow Measurement at Dairy Wetland Outlet A 450 mm H-flume (Figure 9) was installed at the wetland outlet in the previous study by Vellidis et al. (1993), and was used in this study to measure atrazine and metabolite concentrations and loads discharged into the farm pond. This provided data with which





24






"If










g 4\ 4 It





Figure 8.Pogrp depicting the LIFE runoff collector configuration. Insets show the
flume divisions, sample containment system, and sample collection vessel. to make calculations on the overall effectiveness of the wetland in the attenuation of atrazine and its major metabolites in the wetland. Two tapered earthen berms were constructed on each side of the flume to route surface flow exiting the wetland through the flume.

Flow proportional composite samples from the H-flume were collected with an

automated system used in Georgia, since 1997 (Vellidis et al., 1997, 1999). The system consists of an ISCO composite sampler, a Campbell Scientific data logger, and a Druck pressure transducer. The pressure transducer installed in the flume's stilling well provided the data logger with continuous stage measurements. Stage data were used by the data logger to determine flow rate from the flume rating curve. The ISCO sampler was instructed to pull a sample for every 2.5 mn3 of flow volume passing through the flume. Consequently, samples were collected more frequently during periods of high





25

























it
Figure 9. Final outlet flume at the Dairy Wetland, 450mm H-flume. flow and less frequently during periods of low flow. Again, a 9.5 L sample jar with both a teflon and polyethylene bag was used to collect the composite sample which, depending on flow, varied in volume from less than 0. 1 L to 10 L. Precipitation Measurement

The Dairy Wetland was equipped with a Model TR-5251 tipping bucket style rain gauge (Texas Electronics, Inc., Dallas, TX) to measure precipitation. In addition, a HOBO Event Data Logger was connected to the rain bucket to record rates, time, and duration of rainfall.

















CHAPTER 4
WATER STUDY

Materials and Methods

Atrazine Application

Sampling began February 17, 2003, and extended to April 5, 2004. Initial

samplings (February through April, 2003) served to establish background conditions and document the presence or absence of residual atrazine and degradation products. A 0. 1 ha (10 m x 100 m) application strip, immediately upslope from the forested buffer system was harrowed twice in the fall of 2002, and again April 4, 2003, just before application. Atrazine was applied in the form of Atrazine 4L (Agrisolutions) to the application strip at on April 24, 2003. Application was originally scheduled for March to emulate application in conventional corn production but was delayed due to extremely wet conditions which prevented used of farm equipment at the site.

The atrazine was applied at a rate of 26.8 kg ha- of active ingredient atrazine to simulate edge of field loadings measured by Lowrance et al. (1997) in a nearby riparian buffer study using recommended rates of application for a corn field. Potassium bromide was applied at a rate of 270 kg ha'1 (172.9 kg ha- Br). The bromide was intended for use as a conservative tracer. The atrazine and bromide were dissolved in water to give a total 151 L tank mixture. The solution was applied by broadcast spray from a tractor-mounted sprayer.




26





27


Nineteen Whatman 7 cm glass fiber filter papers were placed in a zigzag pattern

across the application strip and were used to estimate the actual concentrations of atrazine reaching the soil. A 120 ml specimen cup was placed near each filter paper to allow a measurement of the application rate of the potassium bromide. Following application, each filter paper was placed into a glass jar with Teflon lined cap and frozen until extraction and analysis was performed. The specimen cups were capped and returned to the laboratory where 100 ml of deionized water was added and samples were refrigerated at 4' C overnight. The following day the cups were thoroughly mixed and bromide concentration determined using an ion selective probe.

Glass fiber filters were placed on the application strip were extracted and prepared using the soil extraction protocol (Chapter 5) which consisted of vortexing with solvent followed by a solid phase extraction step. Following evaporation and reconstitution steps in the soil method, the application filter samples were diluted 500 times to provide a concentration that would fall within the range of the calibration standards.

The mass of atrazine and bromide recovered from the filters and cups placed on the application strip varied greatly. The average atrazine recovery from the filters was 16.1 kg ha1 (60.0% of the atrazine applied), however, the range of recovery of atrazine from the nineteen filters was 7.6 kg ha'I to 23.7 kg ha-1. Glass fiber filters (n=10) were spiked with atrazine in the lab to yield 6 mg per filter simulating field application quantities. The average recovery from the laboratory spiked filters was 90.2% using the same laboratory methods as the actual application filters. Bromide recovery from the application strip exhibited variance across the application strip. The average bromide recovery was 14.4 kg ha'1 (80.4% of the 17.9 kg Br ha'1 applied) with a range of 6.7 to





28

27.4 kg ha- In addition to normal variance associated with spray application, two other factors were also instrumental in contributing to the application variance. The sprayer was calibrated to distribute the solution over the application strip in four round trip passes, i.e., sprayer width was 2.5 m so up one half of the strip and back on the other half composed a round trip. After two and a half round trips the solution had been applied. In addition, there was a noticeable wind toward the forested buffer from the application strip on the day of application, though no site measurements were made Field Sampling Strategy

Ground water. Ground water samples were collected from the ground water

monitoring wells between February 17, 2003 and April 5, 2004, whenever a measurable volume could be extracted. Samples were collected biweekly from February November, 2003, then monthly from December, 2003 to the end of the study in April, 2004. Water table depth was measured at each well prior to sampling with an electronic water level indicator. A peristaltic pump with silicone tubing was used to collect samples. Samples were collected in 500 ml glass Qorpak bottles which had been acid and solvent washed and baked in a 300' C oven. Lids were teflon lined. Samples were transported to the lab within two hours of collection, stored at 4' C, and extracted within 7 days. A total of 1703 groundwater samples were collected during the 14 month study.

Aquifer conductivity. The hydraulic conductivity was determined for each of the pesticide wells using the bail test technique (Freeze and Cherry, 1979) and Horslev (195 1) method during three different seasons of the year by Vellidis et al. (2003). Drought conditions and consequentially water tables below the screened portion of the wells prohibited any hydraulic calculations for summer and sporadically during the other seasons; thus all measured conductivities were averaged for use in the Darcian flow





29


equations. For this study, the conductivities developed by Vellidis et al. (2003) were used to calculate an average aquifer hydraulic conductivity of 45.09 min hr-1.

Hydraulic gradients were calculated using the water table elevations measured each time a well was sampled. Elevations were measured from a local reference point. Gradients for each well were then calculated as the difference in hydraulic head between landscape positions above and below the well in question divided by the distance between the two adjacent wells. Wells in the 0 in and 38 in positions had change in hydraulic head calculated based on the change in the water table between itself and the well in nearest landscape position (0 m paired with 8 in and 38 in paired with 28 m).

Surface runoff. Surface runoff was collected in 12 Low-Impact Flow Event

(LIFE) samplers (Sheridan et al., 1996) and an H-flume (Brakensiek et al., 1979) during this study. There were a total of 202 runoff samples generated during 46 measurable runoff events during the course of this study. The LIFE sampler collection vessel was weighed before going to the field and reweighed with the sample in it after being returned to the laboratory to determine weight for conversion to volume using the density of water. Samples in LIFE collectors were retrieved by simply swapping out the glass jar for a clean one. If a rain event occurred during working hours, samples were retrieved immediately following the rain event. If an event occurred outside normal working hours, the samples were collected at the beginning of the next work day. The collected samples arrived at the lab within 2 hours of collection, were vigorously shaken to ensure a representative sample, and I L poured into a glass jar for storage at 4' C. Samples were processed for chemical analysis within 7 days. As with the LIFE samplers, samples from the H-flume were collected immediately after an event during working days. If an event





30

occurred after working hours, the sample was collected on the morning of the next working day. If the event was so small as not to activate the automated sampler and the flume was still flowing when visited, a grab sample was collected. Samples were transported to the lab within two hours of collection, stored at 40 C, and extracted within

7 days.

Sample Preparation

Ground water. Each sample was vigorously shaken and 200 ml vacuum filtered through a Buchner funnel and glass fiber filter for pesticide residue analysis. Another 100 ml was placed into a plastic specimen cup, capped, and stored at 4' C for bromide analysis.

The method for extraction of pesticide residues was adapted from a method

provided by the USDA-Agricultural Research Service-Southeast Watershed Research Laboratory-Pesticide Residue Laboratory (USDA-ARS-SEWRL-PRL) located in Tifton, Georgia. All ground water samples were vacuum filtered (-70 kPa) through a 7 cm Whatman GF/B glass fiber filter into a 500 ml glass jar. The filter and solids were discarded. Samples were spiked with 100 pl of surrogate (10 mg L- terbuthylazine) or surrogate plus analyte (10 mg U1 deisopropylatrazine, deethylatrazine, hydroxyatrazine, and atrazine) for matrix recovery checks. Each sample was then extracted through 12 cm3 500 mg Oasis HLB solid phase extraction cartridges (Waters Chromatography Corporation). Solid phase columns were conditioned with 3 ml methylene chloride, 3 ml methanol, 5 ml deionized water, and a final 5 ml deionized water. Samples were vacuum extracted at less than -10 kPa through Teflon lines leading from sample jar to solid phase cartridge. Cartridges were attached to a Supelco 24 port extraction manifold. Following sample input, cartridges were washed with 5 ml deionized water and dried for 15 minutes





31

at -50 kPa. Samples were then eluted into glass test tubes with 3 ml of methanol followed by 3 ml of methylene chloride at less than -10 kPa. The samples were concentrated to dryness using a nitrogen gas evaporator (Zymark TurboVap LV). Nitrogen pressure was approximately -0.25 kPa. Samples were reconstituted in 1 ml of 40% methanol and 10 gl of 2-chlorolepidine (10 mg L-1) was added to each tube to serve as an internal standard. The samples were placed in an ultrasonic bath for 5 minutes and then vortexed for 15 seconds to ensure adequate mixing. Samples were then syringe filtered (13 mm, 0.45 im nylon Whatman filter) into an autosampler vial and stored at 40 C until analyzed by HPLC, usually within two weeks.

Surface runoff water. A 200 ml portion of each runoff sample was filtered through a

7 cm GF/B Whatman glass fiber filter and the filtrate processed as described above for groundwater samples. Each runoff sample filter was spiked with 100 gl of surrogate (10 mg U terbuthylazine) or surrogate plus analyte (10 mg L-' deisopropylatrazine, deethylatrazine, hydroxyatrazine, and atrazine) for matrix recovery checks. The filter plus adhering solid material was extracted by vortexing with a mixed mode extractant developed by Lerch and Yong-Xi (2001). The mixed mode extraction solution was designed to maximize the recovery of hydroxylated atrazine degradation products from soils by disrupting cation exchange mechanisms in addition hydrophobic interactions and hydrogen bonding mechanisms which are most important with the chlorinated atrazine species. The extractant consisted of a 3:1 mixture of 0.5 M potassium phosphate buffer (pH adjusted to approximately 7.5) and acetonitrile and will henceforth be referred to as the mixed mode extractant. The mixed mode extractant (50 ml) was added to each runoff filter sample and the samples were placed on a rotary bed shaker (Labline Orbital Shaker)





32

at 100 rpm for 2 hours. The extractant was vacuum filtered (-70 kPa) through another 7 cm glass fiber filter (Whatman, GF/B) and the extraction process repeated two more times. After the third extraction the entire contents of the sample vessel were emptied onto the glass fiber filter and the sample jar rinsed twice with 10 ml acetonitrile onto the filter. The filter was vacuum dried for 30 minutes at -70 kPa vacuum. The filter was discarded and the filtrate placed in a nitrogen evaporator (Zymark TurboVap LV) to remove the acetonitrile. Nitrogen pressure was approximately -25 kPa, and the samples were allowed to evaporate for 3 hours to remove the bulk of the acetonitrile. Deionized water (150 ml) was added to each sample and the samples were solid phase extracted in the same manner as the groundwater samples and water portion of the runoff samples. Instrument Analysis

Pesticide residue analysis. All extracted water samples were stored at 4' C until analysis was performed. The samples were analyzed using a Waters Chromatography Corporation High Pressure Liquid Chromatograph (HPLC) equipped with an ultraviolet detector operated at a 220 nm wavelength. A LC8-DB column (25 cm x 4 mm id, 5 gm particle size; Supelco) was used to separate the analytes, employing a gradient mobile phase. In addition, a 2 cm LC8-DB guard column was used to extend the life of the analytical column during analysis of well and runoff samples

The HPLC mobile phase consisted of 0.005 M phosphate buffer (pH-7.5) and methanol, and the flow rate was 1.25 ml min' during the 22 minute run time. The gradient utilized was 70% phosphate buffer and 30% methanol for 12 minutes with a linear change to 40% buffer and 60% methanol within 16 minutes, then slow equilibration back to the original conditions at the final 22 minute mark. Retention times for the analytes were approximately 6.8 minutes for deisopropylatrazine, 8.9 minutes for





33

deethylatrazine, 9.8 minutes for hydroxyatrazine, 14.8 minutes for atrazine, 16.9 minutes for the internal standard 2-chlorolepidine, and 19.3 minutes for the surrogate compound terbuthylazine.

Bromide analysis. Bromide measurements were performed on 100 ml aliquots of the ground water and runoff water samples which were transferred to plastic specimen cups upon arrival from the field site. The analyses were conducted using a Thermo Orion bromide ion selective electrode and single junction reference electrode (Fisher Scientific) in conjunction with a model 50 Accumet pH/Ion/Conductivity meter (Fisher Scientific). A 1000 mg L-1 certified bromide stock standard solution (Fisher Scientific) was diluted to produce calibration standards ranging from 0.1 to 30 mg U1. Following daily calibration, a standard check was performed every 10 samples and if the results were outside 10% of the true value, the meter was recalibrated and the previous 10 samples reanalyzed. Bromide Ionic Strength Adjuster (Fisher Scientific) was added to each sample (2 ml) to minimize differences in ionic strength of the samples due to ions other than bromide.

During application of atrazine and the bromide tracer, empty specimen cups were placed on the application strip along with the filter paper. Deionized water (100 ml) was added to the cups immediately upon return to the lab following chemical application and the mixture vigorously shaken. The bromide concentration was then measured following calibration with bromide standards ranging from 1.0 to 750 mg L-1. Chemicals and Standards

The following calibration standards were obtained from Chem Service (West Chester, PA): atrazine (98% pure), deisopropylatrazine (98% pure), deethylatrazine (99.5% pure), and hydroxyatrazine (98% pure); and from Crescent Chemical (Islandia, NY): terbuthylazine (99% pure). A stock standard was prepared in methanol at 10 mg L-1





34

for each of the analytes and the surrogate compound terbuthylazine. Working calibration standards for the HPLC were prepared from the stock solution in 40% methanol at concentrations of 0.1 to 5.0 mg L-. 2-chlorolepidine (99% pure, Sigma-Aldrich, St. Louis, MO) served as the internal standard and was added to the autosampler vials with the calibration standard and to the samples immediately prior to reconstitution in 1 ml of 40% methanol and placement into autosampler vials. All solvents and KH2PO4 were HPLC grade.

Samples were analyzed on the HPLC in sets composed of 38 samples. Following calibration standards and each set of 10 samples (8 in the last set) within a run was a set of check standards including a 0.3 mg L-1 calibration standard check and a 1.0 mg U NIST traceable standard obtained from Crescent Chemical (Islandia, NY). The NIST traceable standard was prepared at 10 mg L-1 for each of the analytes and surrogate and diluted to 1 mg L1 for injection on the HPLC. Another 10 mg L-1 solution of terbuthylazine was prepared to serve as the surrogate spiking solution for samples. Each ground water sample, runoff water sample, and runoff sediment sample received 100 tiL of the terbuthylazine solution and soil samples each received 200 jiL of the surrogate solution. In addition, another 10 mg U1 solution of atrazine, deisopropylatrazine, deethylatrazine, and hydroxyatrazine was prepared for matrix spike recovery analyses.

Quality Control/Quality Assurance Plan Ground water. On most well sampling dates, samples from six wells were prepared and analyzed in duplicate to determine the precision of the procedures. In addition, percent recovery was determined for the same samples by spiking two more aliquots with a known concentration of atrazine plus each of the degradate compounds and surrogate. The samples were spiked with 100 gL of the 10 mg L-1 mixed standard





35


solution (atrazine, deisopropylatrazine, deethylatrazine, and terbuthylazine) prior to solid phase extraction. A field blank was collected on each well sampling day by taking a 2 L aliquot of laboratory grade water to the field and running it through the same pumping procedures used for the samples. The field blank was processed as a sample in the laboratory. The field blank was utilized to determine if sample contamination had occurred during sampling, preparation, or analysis procedures. A duplicate set of laboratory grade water samples was processed with each well set, and another set of laboratory grade water samples was spiked for recovery determination. Instrument blanks were composed of reagent grade methanol being pipetted directly into autosampler vials and analyzed every ten samples to determine if contamination or carryover was occurring during the instrument analysis.

A summary of the results of the QA/QC plan for groundwater is given in Table 2. Abbreviations for the analytes in table 2 and henceforth are AT atrazine, DEA deethylatrazine, DIA deisopropylatrazine, HA hydroxyatrazine, and BR bromide. The limits of detection (LOD) and limits of quantitation (LOQ) were calculated by analyzing seven replicate samples spiked at a level near the estimated detection limit. The LOD was determined as three times the standard deviation and the limit of quantitation as 10 times the standard deviation, following the recommendations of Keith et al. (1983). The LOD is therein defined as the lowest concentration of analyte that can be shown to be statistically different from a blank at the 99% confidence level. All measured values which were determined to be below the limit of detection were entered into the dataset as zero. The LOQ is defined as the concentration above which quantitative results can be obtained with a certain degree of confidence. Keith et al. (1983) indicate two concerns








36





IZ



Ln Ln cq











cq cq

C

-H

It 00



00 cq
M











rq 11C 00 C4-0
kn

-H
-H






omo





00
rn



> >
u v

bi) bi) 4- 4En z U)
tn 7,

4,
0 C13
(4-4 4-4 0 0





CY CIS C) o





37

with data measurements near the detection limit, (1) the uncertainty associated with a measurement can approach and may even equal the reported value; and (2) confirmation can become difficult or impossible depending on the presence of interferents and the sensitivity of the method. In Table 2, the LOD's are given in jig L1 (to agree with unconcentrated field units) for atrazine (0.29), hydroxyatrazine (0.15), deethylatrazine (0.20), deisopropylatrazine (0.17) and mg L- for bromide (0.02). Matrix spike recoveries were averaged over the course of the study and ranged from 93.4% recovery for atrazine to 107% and 108% for deethylatrazine and deisopropylatrazine, respectively. Bromide matrix spike recovery was 93.4%. The range of the averaged recoveries of spiked laboratory blanks for the triazines was 84.1% for atrazine to 102% for deisopropylatrazine, while blank recovery was 106% for bromide. All groundwater field blanks were averaged together for the entire study period, and the values were below the LOD for each triazine analyte while bromide was 0.05 mg L-1, which is very near the

0.02 mg L-1 LOD.

Three control wells located at the perimeter of the field were sampled throughout the study. The average concentration of each analyte for all control wells collected (n=61) during the study was below the LOD.

Surface runoff. Surface runoff was filtered prior to extraction of atrazine and its degradation products thus yielding a water portion and a sediment portion of the sample. Bromide was measured on an unfiltered sample. The QA/QC results for the water portion of the surface runoff for atrazine, its degradation products, and bromide are summarized in Table 3. Analytical limits were calculated in conjunction with the groundwater limits as the extraction and instrument methods were identical. Average matrix spike





38

recoveries ranged from 102% for atrazine to 112% for hydroxyatrazine. Similar to the groundwater, blank spike recoveries for the triazine analytes in surface runoff water range from 84.9% for atrazine to 99.3% for deethylatrazine, with deisopropylatrazine at 99.3% and hydroxyatrazine at 91.1%. Bromide blank spike recoveries were 119%, a 13% higher recovery than the groundwater. Again, the field blank recoveries were below detection limits for all analytes except bromide which was close to the LOD at 0.06 mg L-.

Matrix spike recoveries for the sediment portion of surface runoff samples averaged over the course of the study were 90.2% for atrazine, 94.6% for deisopropylatrazine, 96.4% for hydroxyatrazine, and 98.9% for deethylatrazine. Average matrix spike recoveries from the particles and blank spike recoveries were both lower than for the groundwater and surface runoff water, most likely due to the adsorption of the molecules to the solid particles (Table 4).

Mean field blank concentrations were below detection limits for all analytes except deethylatrazine which had a mean value of 0.24 gg L-1 just above the 0.23 gg Ldetection limit for deethylatrazine.







39











N
C 0 -H






C Ln







00







CN
-H





cq 11C 140




vs







Cd



w 01


> 7 I=

cn z En cn z
aj -H -H -H

N
C) C) cd 4 -4

.1;L4 C-1 uc o CIO

cn
z 9: Ov








40











N;











m 00.)6
C 00




U; en 0 00

00 0 CI00
-H -H006


m~ 00

00




-~ Cl00 0 00





q) a. UN l C



cn zH -H


00n







0.)






0) 0
C) C Z.






41


Data Analysis

Mass Loads

Ground water. Groundwater well measurements were combined for each paired set of monitoring wells into a weighted well concentration. This provided a concentration for shallow groundwater found between the soil surface and 2 m below the surface. Weighting was calculated using the following equation:

Weighted Concentration = (C,*WT,) + (Cd*WTd) /(WT,WTd)

CS= Concentration in the shallow monitoring well (g~g L-1 triazines and mg L-1

for BR)

WTs= Water table depth in the screened portion (0. 1-0.6 mn) of the shallow

monitoring well (in)

Cd= Concentration in the deep monitoring well (gg L-1 and mng L-' for BR)

WTd= Water table depth in the screened portion (0.6-2.0 mn) of the deep

monitoring well (in)

Well concentrations were also averaged together along landscape positions or tiers to obtain average landscape position concentrations. Mass loads of pollutants and tracer entering each landscape position were calculated by assigning contributing areas to each well (Figure 10). For the subsequent discussions, well transects are defined to run up and down the slope while well tiers or landscape positions are parallel to elevation contour lines. The distance between well transects was divided in half (green lines in Figure 10). The weighted concentration data for each well location were assumed to represent shallow groundwater conditions along the corresponding tier within the contributing area.





42























Figure 10. Network of paired groundwater monitoring wells showing the contributing
areas assigned to each well in order to calculate groundwater loads.

The well concentration data were converted to mass loads using the following equation:

Mass Load (g)=- C (gm"J) *K(m hr-l) *dh/dl *T (hr) *D (m) *L (m)
C Weighted concentration over time period T


K = Hydraulic Conductivity
dh/dl = Hydraulic gradient: change in head over change in distance

T = Time period (rs) for which a load is calculated


(2 or 4 weeks)
D = Depth of water in the well over time period T


L = Length of tier associated with the well





43


The load of atrazine and its degradates associated with each well location in a tier were summed to determine the load moving past each landscape position for the entire study following application of the atrazine on April 24, 2003. In keeping with two previous studies of atrazine in riparian buffers (Lowrance et al., 1997, Vellidis et al., 2002), the study was also broken into three smaller phases; the pre-application phase which encompasses all data acquired before application; the three months following application were termed the after application phase; and the final nine months of the study were termed the post application phase. The studies by Lowrance et al. (1997) and Vellidis et al. (2002) found that a substantial portion of the applied atrazine was transported through the system during the initial three months following application. Preliminary data analysis in this study also showed that the majority of transport occurred during the after application period. To be consistent with data analysis in these previous studies and to ensure high concentrations during this initial phase of the study were not muted by grouping with the low concentrations in the final phase of the study, the current study was grouped into three phases. The exact dates for each phase of the study are:

Pre-application = all samples collected prior to application on 4/24/03

After application Samples collected 4/25/03 through 7/31/03

Post application = Samples collected 7/31/03 through 4/5/04

In addition to summing loads for the entire year of study as a whole, loads for each landscape position were summed for the two periods of the study following application, i.e., the after and post application phases of the study.

Surface runoff. Concentration data from the LIFE samplers were used to calculate loads entering the upslope edge of each vegetation zone 1, 2, and 3, and a midpoint in





44


vegetative zone 2 during each runoff event. As with the wells, runoff collectors were assigned contributing areas (Figure 11) which were used to calculate mass loads of chemicals. The load entering the runoff collector was assumed to be constant over the linear distance assigned to the collector within the contributing area. The equation used to calculate the runoff loads was:

Mass Load (g) (C (g M-3) V (M) / CW (m)) L (m)

C = Concentration in each collector

V = Volume of water measured in each collector

CW Runoff collector width

L Length of zone associated each collector


i T-A




contributing area
attributed to a runoff
collector fbr load
calculations













Figure 11. Network of LIFE samplers showing the contributing areas assigned to each to
calculate surface runoff loads.





45


Again, mass loads of analyses in surface runoff water were summed for each

landscape position for time periods after application (4/24/03 through 7/31/03) and post application (8/l/03 through 4/5/04). Mass loads per landscape position were also calculated for both periods following application.

Stage height measured in the stilling well of the final H-flume was used to calculate total water volume being discharged to the farm pond from the buffer system. A linear regression was performed on discharge data and stage height from a flume rating table for the 450 mm flume (Brakensiek et al., 1979). The slope equation was then used to determine discharge for stage heights recorded by the data logger. Concentrations for composite samples were then multiplied by the discharge calculated for the sample composition time to provide total mass loads exiting the system. Loads were summed for each analyze following application to determine total mass exiting the system. Dilution Parameters

For six runoff events which produced two or three samples per landscape position (three is maximum possible), dilution factors due to the rainfall input were calculated to examine the effects of rainfall dilution on concentration changes between the upper two vegetative zones. Dilution factors were calculated utilizing the total rainfall volume entering the buffer zone of interest and the volume of surface runoff entering the buffer zone, where:

Dilution Factor = Rainfall Input / (Rainfall Input + Surface Runoff Entering) Changes in concentration due to dilution were calculated by multiplying the total concentration change across a buffer zone times the dilution factor. The portion of the concentration change across a buffer zone not due to dilution is termed the net concentration change and is attributed to infiltration and/or degradation processes.






46


Dilution factors and net concentration changes were calculated for Zone 3 (the 8 mn grass strip), Zone 2a (upper portion of pine trees see Figures 7 and 11), and Zone 2b (lower portion of pine trees). With these calculations, a net concentration change or actual chemical attenuation was calculated for the grass and pine portions of the three-zone buffer system.

In addition, the net percent concentration change was calculated and divided by the buffer zone width (8 mn for Zone 3, and 10 m each for Zones 2a and 2b, respectively) to produce the percent concentration change per meter of flow length. These calculations allowed for comparisons to be made between the buffer zones to determine the efficiency of each vegetative zone in diminishing analyte concentrations. Statistical Analysis

Stastistical analyses were performed on concentration data for the three datasets, weighted well (Appendix A), surface runoff water (Appendix B), and surface runoff sediments (Appendix C) using a mixed model procedure (Littel et al., 1996). Previous studies of atrazine in riparian systems (Vellidis et al., 2002, Lowrance et al., 1997) used nonparamnetric procedures to analyze the data, and these procedures prohibited means separation. Data were natural log transformed prior to analysis due to the large number of values below detection limits which were entered as zero for statistical purposes and to minimize variance between the maximum and minimum values. The mixed model procedure allows fixed and random effects to be specified (Littel et al., 1996) and use of least square means of the transformed data allows assuming that the data are normally distributed and appropriate for use in an analysis of variance. Each dataset was grouped by landscape position. Comparisons of landscape positions were made for each season of the study; pre-application, the immediate after application period (April 24, 2003 through






47


July 31, 2003), and for the final post application period (July 31, 2003 through April 5, 2004). Each landscape position was also compared to itself for each of the phases (0 mn pre-application compared to 0 mn after and post application phases) Groundwater Concentration Maps

Surface contour maps were developed for the weighted groundwater concentrations for each sampling date using the Surfer (Golden Software, Inc.) mapping package. The first step in this process was to interpolate among the discrete data to create a surface. This was done within Surfer using the natural neighbor technique which is based on Thiessen polygons and uses weighted values of neighboring observations (Sibson, 198 1). The natural neighbor technique does not extrapolate contours beyond the boundaries of the data. Distances (in) in the north and east directions of wells and collectors from a local benchmark are designated on the contour map axes.

Results and Discussion

Ground Water Concentrations

During the pre-application period of the groundwater study, no significant

differences were detected among the landscape position average concentrations for any of the analytes. There was an increase in concentration for all the analytes following application, though minimally so for deisopropylatrazine and hydroxyatrazine. Deethylatrazine was the predominant degradation product in groundwater following application. Other studies have shown deethylatrazine to be more abundant degradation product in groundwater than deisopropylatrazine or hydroxyatrazine (Adams and Thurman, 199 1, Panshin et. al., 2000) and to be more mobile than the parent or other degradation products being studied. Kruger et al. (1996) found the mobility order in





48


groundwater to be deethylatrazine>atrazine>deisopropylatrazine>hydroxyatrazine in a

variety of 10 soil types found in Iowa.

The average concentration for each analyte in all six landscape positions is given in

Table 5. Averages are for the entire year following application period. In a similar

riparian system in Tifton, Georgia, average atrazine concentrations were reduced from

1.29 gg L-1 at the field edge to 0.14 gg L"1 at a streamside zone approximately 50 m down

slope (Lowrance et al., 1997). Averaging concentrations of atrazine found in

groundwater for the entire year following application produced similar results with the

upslope 0 m landscape average being 3.59 jig L-1 and the final 38 m landscape position

average being 0.17 gg L-1. The capacity of the current buffer system in removing

Table 5. Groundwater weighted well concentrations averaged over the combined two
periods of the study following application (4/24/03-4/5/04) for each landscape
position. Mean followed by standard deviation (n ranges between 120 and
124 for each parameter)
Landscape BR AT DEA DIA HA
position mg L-1 gg L-1 gg L-1 jg L-I gg L1
0 m 2.943.12 3.5917.92 0.55+2.07 0.02+0.10 0.050.13
8 m 1.241.26 1.238.69 0.280.49 0.020.07 0.03-0.09
13 m 0.880.82 0.461.26 0.250.34 0.020.15 0.020.06
18 im 0.720.86 0.230.47 0.170.19 0.000.02 0.010.05
28 m 0.560.45 0.110.31 0.110.15 0.000.01 0.000.02
38 m 0.610.88 0.170.64 0.080.13 0.000.01 0.000.02

Table 6. Groundwater weighted well concentrations averaged over the immediate after
application period (4/24/03-7/31/03) for each landscape position. Mean
followed by standard deviation (n ranges between 38 and 42 for each
parameter)
Landscape BR AT DEA DIA HA
position mg L-1 gg L-' jg L-' gg L' gg L'
0 m 2.533.10 10.3130.03 1.074-3.49 0.050.15 0.060.14
8 m 0.751.07 3.3814.74 0.370.79 0.040.12 0.050.11
13 m 0.751.03 1.201.94 0.280.47 0.020.07 0.020.07
18 m 0.520.85 0.570.67 0.080.17 0.000.03 0.020.05
28 m 0.450.55 0.260.49 0.020.08 0.000.01 0.000.01
38 m 0.871.47 0.411.04 0.010.05 0.000.00 0.000.02





49


atrazine from groundwater is more apparent when looking at the concentration reduction in the initial 3 month period following application. The average concentration in the 0 m landscape position was 10.3 gg L-land was reduced to 0.41 [tg LIby the 38 m position (Table 6).

Average concentrations entering the 0 m landscape position were lower in the final

9 months of the study than the initial 3 months following application. Deethylatrazine exhibited the largest average concentration (0.29 gg L-1) for the final 9 months (Table 7). Concentrations of deisopropylatrazine and hydroxyatrazine were consistently below detection limits as were the average concentrations for the entire study following application (Table 5).

Statistical analysis of each analyte with landscape position data pooled into the 3

phases of the study, (pre-application, immediately after application (4/24/03-7/31/03) and post application (8/1/03-4/5/04) periods, showed significant differences (p<0.05) among the three phases for atrazine, deethylatrazine, and bromide. The concentration of atrazine for the immediately after application period was significantly different from the pre and Table 7. Groundwater weighted well concentrations averaged over the final post
application period (8/1/03-4/5/03) for each landscape position. Mean followed by standard deviation (n ranges between 79 and 84 for each
parameter)
Landscape BR AT DEA DIA HA
position mg Ll gg U jg gg L-1 gg U
0 m 3.12+3.13 0.190.44 0.290.41 0.000.04 0.040.12
8 m 1.471.29 0.110.22 0.240.22 0.000.03 0.020.08
13 m 0.930.71 0.080.22 0.230.24 0.030.17 0.010.05
18 m 0.820.85 0.050.13 0.210.19 0.000.00 0.010.04
28 m 0.614-0.39 0.030.08 0.150.15 0.000.01 0.000.01
38 m 0.580.33 0.050.14 0.110.14 0.000.01 0.000.00









50






ATRAZINE ATRAZINE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
OM FROM AFFLICTION STRIP ISM FROM APPLICATION STRIP
2 2

_________a b a,






PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE



ATRAZINE ATRAZINE

WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
8M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2

t5 a h A I1.5 A


05


00.
PRE APP AFTER APP POSTAPP0
PRE APP AFTER APP POST APP
PHASE PHASE




ATRAZINE ATRAZINE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
213 M FROM APPLICATION STRIP 38M FROM APPLICATION STRIP

L1.5 aw-a
1 5
1 l





PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE



Figure 12. Atrazine least square means of natural log transformed well concentration

data grouped by study phase (pre-application, after application, and post

application) for each of the landscape positions. Differing alphabet

designations indicate study periods were significantly different (p<0.05).



post application periods for every landscape position (Figure 12). The final post



application atrazine concentrations were not found to differ from the pre-application



concentrations.



Bars on the graph with different alphabet letters indicate the data were significantly



different according to the mixed model procedure (SAS Institute). For example, a bar



with a statistical designation "a" was not found to be significantly different from any



other bar designated with "a"9, "ab", "ac", or any combination of letters which contained



the "a" designation. However the bar designated "a" was found to differ from "b", "c"






51


"bc". Significant differences were not found between a bar marked "b" and any other bars with any "b" or combination of letters containing "b". A bar with the "b" designation was found to differ from "a", "c", or any other designation not containing the designation "b".

The higher concentrations found for atrazine in the immediate after application period diminished as distance from the application site increased, though only the landscape position 0 mn from the application strip was found to be significantly different (Figure 13). During the post application period of the study, atrazine concentrations were much lower and significant differences were only found between the landscape positions

0 m from application strip and those at 18, 28, and 38 mn from the application strip. For both phases of the study following application, atrazine concentrations in the 0 mn position were found to be significantly larger than the final 38 mn position based at p<0.05.


ATRAZINE ATRAZJNE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
AFTER APPLICATION POST APPLICATION
2 12
1.5 '2 b~ h~b h A. ab ab b

0.5 0.5 _______________0 0
OM aM 13 M 18M 28 M 38 M OM M8 13 M lam 28 M 38 M
LANDSCAPE POSITION LANDSCAPE POSITION


Figure 13. Atrazine least square means of natural log transformed well concentration
data grouped by landscape position for the two study periods following
application. Differing alphabet designations indicate landscape positions were
significantly different (p<0.05).

Deethylatrazine concentrations at the 0 m position for the study phase immediately after application and the post application period were significantly different from the pre-







52




application period (Figure 14) thus indicating the continued influx of deethylatrazine into



the system from the application strip during the final 9 months of the study.



DEETHYLATRAZINE DEETHYLATRAZINE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
OM FROM APPLICTION STRIP ISM FROM APPLICATION STRIP

2 2
'7', 1.5 a e _a b
cc- 1 1


0.5 0.5


PREAPP AFTER APP POSTAPP PREAPP AFTER APP POSTAPP
PHASE PHASE




DEETHLYATRAZINE DEETHYLATRAZINE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
8M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2


1 .s c ----0.5 0.5__ _ _ __ _

0 -0
PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE




DEETHYLATRAZINE DEETHYLATRAZINE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
13 M FROM APPLICATION STRIP 38M FROM APPLICATION STRIP
2 2

__1.5 a c -------a 1.5




0 -0
PREAPP AFTER APP POSTAPP PREAPP AFTER APP POSTAPP
PHASE PHASE


Figure 14. Deethylatrazine least square means of natural log transformed well

concentration data grouped by study phase (pre-application, after application, and post application) for each of the landscape positions. Differing alphabet

designations indicate study periods were significantly different (p<0.05).


In analysis of landscape positions within each period following application


(Figure 15), the 0 m landscape position exhibited significantly higher concentrations than


the final 38 m position, as did the parent compound, further attesting to the effectiveness


of removal in the current buffer system. However, unlike atrazine, deethylatrazine





53


concentrations for the 0, 8, and 13 mn positions were not found to differ at p<0.05 during the immediately after application period.

DEETHYLATRAZINE DEETHYLATRAZINE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
AFTER APPLICATION POST APPLICATION
S2 z 2
1.5 a a s bfr h- 2 1.5 he ~b~h Ca

C,
0.5 0 m mm5
0
O M 8 M 13 M 18 M 28 M 38 M OM S M 13 M 18 M 28 M 38 M
LANDSCAPE POSITION LANDSCAPE POSITION

Figure 15. Deethylatrazine least square means of natural log transformed well
concentration data grouped by landscape position for the two study periods
following application. Differing alphabet designations indicate landscape
positions were significantly different (p<20.05).

Potassium bromide was applied as a tracer in a previous study at this site by

Vellidis et al. (2003). Even though the earlier application was made in 1994, there were still measurable concentrations of bromide in the system during the current study's preapplication period. However, as Figure 16 demonstrates, there was a discernible increase in concentration following application which became more apparent during the final post application period. Comparisons of bromide concentrations for the three periods of the study show significant differences between the pre-application and both periods after application in the 0 mn landscape position (Figure 16). The next 4 positions (8, 13, 18, 28 mn) had significant differences only between the pre-application and post application periods, while the immediately after application concentration fell between the two. There were no significant differences detected among the three study periods at the 38 mn position.







54




BROMIDE BROMIDE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
OM FROM APPLICTION STRIP 18M FROM APPLICATION STRIP
2 2

1.5 ah 1.5 a .. h



Z I .-
W 0.50.




PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE




BROMIDE BROMIDE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
8M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2

1.5ab b- .5 a.2 aab




W 0.5 - -0.5_ __ __

0 0
PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE




BROMIDE BROMIDE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
13 M FROM APPLICATION STRIP 38M FROM APPLICATION STRIP
2 2

< i .2 b L1.5 a _____I2 _ 02 __ __1_
I-ZI

'J -Lm-]
1. -a h 1.
CY0




PREAPP AFTER APP POSTAPP PRE APP AFTER APP POSTAPP
PHASE PHASE


Figure 16. Bromide least square means of natural log transformed well concentration

data grouped by study phase (pre-application, after application, and post

application) for each of the landscape positions. Differing alphabet

designations indicate study periods were significantly different (p<0.05).


Within each of the study phases following application, the larger concentrations


entering the 0 m landscape position were found to be significantly different (p<0.05) than


the streamside 38 m landscape positions for the parent compound and the preeminent


degradation product, deethylatrazine. Bromide was also found to exhibit significantly





55


BROMIDE BROMIDE
WEIGHTED WELL LEAST SQUARE MEANS WEIGHTED WELL LEAST SQUARE MEANS
AFTER APPLICATION POST APPLICATION
2 2
b- ---b-- b-----b---b-- b b b -b0.5 0.5
0 00 M 8 M 13 M 18 M 28 M 38 M 0 M 8 M 13 M 18 M 28 M 38 M
LANDSCAPE POSITION LANDSCAPE POSITION

Figure 17. Bromide least square means of natural log transformed well concentration
data grouped by landscape position for the two study periods following
application. Differing alphabet designations indicate landscape positions were
significantly different (p<0.05).

larger concentrations in the 0 m position than the final 38 rn position for the two phases of the study following application (Figure 17).

Bromide was used as a conservative tracer in this study because it is not sorbed by most soils, not subject to biological or chemical transformations, and is normally present at low concentrations (Bowman, 1984). In the current study, there was clearly removal of bromide within the system. Average concentrations of bromide entering the 0 m landscape position for the entire year following application was 2735.8 g and was reduced to and average concentration of 625.7 g in the 38 rn position. This results in an average 77. 1 % removal efficiency within the system demonstrating the ineffectiveness of this ion as a conservative tracer for this buffer system. It is suggested the removal might be attributed to plant uptake within the system. Removal of bromide has been shown to be substantial (53% of that applied) in studies of uptake by potatoes (Kung, 1990) and 30% of that applied in a field study removed by grass (Owens, 1985). Groundwater Mass Loads

The mass loads summed for each analyze for the entire period following application delineates the importance of deethylatrazine in the atrazine degradation process in





56


groundwater (Table 8). The mass of deethylatrazine (DEA) entering the upper landscape position adjacent to the application strip was about 10% as much as the parent compound

(AT).

Table 8. Groundwater weighted well mass loads summed over the combined two periods
of the study following application (4/24/03-4/5/04) for each landscape
position (n ranges between 120 and 124 for each parameter.
Landscape BR AT DEA DIA HA
position g mg mg mg mg
0 m 2735.85 3810.74 378.88 16.84 43.88
8 m 1276.55 2058.81 370.46 29.30 39.91
13 m 1783.37 1055.59 470.47 36.43 34.73
18 m 1762.25 375.57 263.82 7.40 18.44
28 m 762.72 142.29 121.33 1.19 3.94
38 m 625.77 174.14 59.94 0.08 2.32
%Removal
Efficiency 77.1 95.4 84.2 99.5 94.7
Removal Efficiency =
(Mass load entering at 0 m Mass load Leaving 38 m / Mass load entering at 0 m)*100


In contrast, only about 1.0% as much hydroxyatrazine (HA) entered and 0.4% as much deisopropylatrazine (DIA) entered the upslope position. Predominance of deethylatrazine in comparison to deisopropylatrazine has been shown by Mills and Thurman (1994) to be due to the faster removal (2-3 times) by microorganisms of the ethyl side chain than the isopropyl side chain of the atrazine molecule. Unfortunately, the system was less efficient at removing the deethylatrazine (84.2%) mass entering or being formed than it was the deisopropylatrazine (99.5%) and hydroxyatrazine (94.7%) possibly due to the greater mobility of deethylatrazine. Krutz et al. (2003) compared adsorption, desorption, and mobility of atrazine and the major degradation products in a vegetated filter strip soil and a cultivated Houston Black clay in Temple, Texas. This study found atrazine and hydroxyatrazine sorption were significantly higher in the vegetative filter strip, and attributed the difference to the increase in organic carbon





57


content of the soil in the filter strip. In contrast, the study found no significant differences in adsorption of deethylatrazine and deisopropylatrazine between the cultivated soil and the vegetated filter strip and suggests that these analytes may not be substantially diminished in vegetated filter strips.

Except for the streamside 38 m position, the mass load for atrazine decreased as it passed each landscape position. There was ponding at the 38 m position following some events. This ponded area may have retained atrazine and its metabolites which moved downslope in surface or subsurface flow and thus bypassed the attenuating capacity of the riparian buffer and may have contributed to the larger mass of atrazine, bromide, and hydroxyatrazine measured at the 38 m position. During the bail test conducted by Vellidis et al. (2003), the soils around the wells at this landscape position were found to have lower conductivities than the other landscape positions. It was also possible that the use of the averaged hydraulic conductivity artificially increased the calculated load at the 38 m position.

There were increases in summed mass loads of deethylatrazine,

deisopropylatrazine, and hydroxyatrazine primarily in the 8 and 13 m landscape positions, presumably due to conversion of the atrazine in the system, or again perhaps influenced by the use of the average conductivity.

The majority of the parent compound moved through the system in the immediate after application period which encompassed approximately three months after application (Table 9). Of the total 3810 mg of atrazine calculated entering the upslope edge of the grass buffer (Zone 3), 96% entered during the first 3 month after application period.





58


Concentration contour maps (Appendix F) for atrazine show the immediate

movement of the applied atrazine into the dairy wetland in conjunction with the first two Table 9. Groundwater weighted well mass loads for the immediate after application
period (4/24/03-7/31/03) summed for each landscape position (n ranges
between 38 and 42)

Landscape BR AT DEA DIA HA
position 9mg mg mg mg
Om 870.83 3663.54 211.86 10.88 15.87
8m 327.85 1929.44 162.66 23.88 19.31
13mi 522.17 878.20 183.46 15.86 11.80
18 mn 428.42 296.48 45.99 7.40 8.56
28 m 222.72 114.26 9.08 0.82 0.66
38 m 328.86 142.09 1.47 0.00 2.32


rainfall events which occurred the evening of, and two days following, application (Figure 18). A total 2.71 cm of rainfall was measured for the combined rain events. By April 28, the first well sampling date after application, atrazine was found in each landscape position, though only one well in the 38 mn position was above the LOD. Following another 1.55 cm of rain, the largest plume of atrazine seen on all the contour maps occurred on May 12, 2003. There continued to be higher concentrations (>0.50 Ig L1'), particularly in the upper two landscape positions through the end of July, 2003. Between July 17 and September 6, 2003, another 29.9 cm of rain had fallen, which appears to have promoted the final flush of atrazine into the buffer system. Following the September 15, 2003, well sampling date, little atrazine was measured in the wetland as clearly shown on the concentration contour maps.

In contrast, the deethylatrazine movement into or formnation of deethylatrazine within the buffer system first appears extensively in the wetland on the concentration contour map from the June 9, 2003 sampling (Appendix G). The concentration plume on





59

this date exhibited the same shape characteristics as the atrazine map for the same date. The contour denoting concentrations >1.00 [ig L-1 extends further into the wetland (13 and 18 m landscape positions) on the northern side of the wetland and includes the upper Dairy Wetland Rainfall

6
C
5
4 3


0 j ,
CO CO CO CO CO CO CO (0 MO I tt t
D 0 0 C CD CD 0 0 C 0 0 0 CD 0 0
C D 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Q\ Q1 QN Zj IN QN QN CN C' N Q' CN QN CN N CN


Date

Figure 18. Precipitation at the Dairy Wetland. Atrazine was applied April 24, 2003.



portion of the 4 transects closest to the farm pond. The well with the largest hydraulic conductivity (240 mm hr1) as measured previously by Vellidis et al. (2003), is located in the 0 m landscape position in the second transect south from the farm pond.

It is likely that soils which permitted higher transport rates are responsible for the larger flux of atrazine and deethylatrazine on the north end of the wetland. Areas of higher concentration on the atrazine and deethylatrazine contour maps continue to show similarities through the September 15, 2003, sampling date. Unlike atrazine, deethylatrazine continued to be in the system throughout the rest of the study, though at diminished concentrations presumably due to conversion of residual atrazine to deethylatrazine in the application strip. Very little atrazine was found in the wetland





60


during the final 9 month phase of the study. Panshin et al. (2000) found that deethylatrazine and deisopropylatrazine were more persistent in pore water than atrazine which exhibited a rapid rise and fall over the first 3-4 months following application to a Crosby silt loam soil and a Brookstone silty clay loam in Indiana.

Though bromide was detected in the wetland prior to the application date, concentrations were very low as shown in the contour map for April 14, 2003, (Appendix H). The average bromide concentration and standard deviation for the 180 pre-application samples was 0.240.11 mg L-1. While bromide concentrations increased in the wetland, as expected, with the initial influx of atrazine (April 28, 2003 map, Appendix H), areas of highest bromide concentration appeared to reach a maximum on the September 2, 2003, sampling date. Bromide continued to exhibit elevated concentrations throughout the remainder of the study, as depicted on the contour maps suggesting a much slower transport rate than the triazine compounds. Surface Runoff Water Concentrations

The atrazine dosing rate on the application strip was calculated to more closely

approximate edge of field loadings of atrazine in surface runoff water when applied at the manufacturer recommended corn application rate. The concentration of atrazine exiting a conventional corn field application and entering an adjacent grass section of a riparian buffer was measured in an earlier study by Lowrance et al. (1997). The nearby riparian system addressed in that study was successful in attenuating average edge of field atrazine concentrations from 34.1 pgg L-1 to 1.0 tg L-. Atrazine transport occurred primarily during a 3-month period immediately after application.

A prior and similar study at the current site (Vellidis et al., 2003) attempted to replicate the edge of field loadings measured by Lowrance et al. (1997) during





61


application of atrazine to an adjacent 10 m-wide strip. They applied at a rate of 17.1 kg ha' of active ingredient atrazine. The average concentration of atrazine entering the upslope edge of the riparian buffer in surface runoff was 12.7 [Ig L-1 which was reduced to 0.66 [tg L-l near the downslope edge of the buffer. Again, the majority of the atrazine movement in the system occurred within a two and a half month period immediately following application.

The present study increased the application rate originally used by Vellidis et al.

(2002) from 17.1 to 26.8 kg ha1 of active ingredient atrazine to more closely simulate the findings by Lowrance et al. (1997). As a result, average atrazine concentrations in surface runoff at the upslope edge of the riparian buffer for the 3-month period immediately after application were more than twice the concentrations measured by Lowrance et al. (1997) in a nearby site (Table 10). The average atrazine concentration Table 10. Surface runoff water concentrations averaged over the immediate after
application period (4/24/03-7/31/03) for each landscape position. Mean
followed by standard deviation (n)
Landscape BR AT DEA DIA HA
position mg Ll kg L-' kg Ll tg L-1 ktg U
0 m 14.5542.68 78.79255.39 1.82+2.27 0.740.69 1.041.11
(10) (11) (11) (11) 011)
8 m 8.9929.76 70.02268.77 1.091.71 0.200.49 0.560.63
(15) (15) (15) (15) (15)
18 m 1.232.96 9.4929.49 0.250.28 0.060.20 0.050.17
(10) (10) (10) (10) (10)
28 m 1.092.52 2.187.70 0.090.22 0.020.07 0.000.00
(14) (14) (14) (14) (14)


entering the 0 m landscape position in surface runoff during the initial three-month period after application was 78.79 tg L which was reduced to 2.18 [g L-at the final LIFE sampler landscape position. It should be noted that this was 28 m from the application strip while the buffer extended another 10 m (38 m total) to the stream bank.





62


As in the ground water, deethylatrazine was again the degradation product with the largest average concentration (1.82 [tg L-) at the 0 m landscape position. However, the deisopropylatrazine (0.74 tg L) and hydroxyatrazine (1.04 pg L-) were both present at substantially higher concentrations in the upper landscape position than found in the groundwater in the immediate after application period (0.05 [tg L1 and 0.06 tg L1, respectively). Through infiltration, degradation, and/or uptake, the buffer system was effective in reducing the average concentration of atrazine in the surface runoff water from 79.8 tg L- to 2.18 [tg L- near the stream, while each degradation product had an average concentration of<0.09 [tg L1 at the 28 m position. Bromide in surface runoff water was reduced 14 fold in average concentration as it moved through the buffer, with the average concentration found to be 1.09 mg L1 at the 28 m position (Table 8).

Though some transport of triazines was measured during in the final 9 months of the study (post application period), average concentrations were substantially lower (Table 11). The system was able to reduce atrazine and hydroxyatrazine to below detection limits. The highest average concentration of all atrazine degradates at 28 m was measured for deethylatrazine 0.08 kg L-. Table 11. Surface runoff water concentrations averaged over the final post application
period (7/31/03-4/5/04) for each landscape position. Mean followed by
standard deviation (n)
Landscape BR AT DEA DIA HA
position mg L-1 kg L-' gg L-' gg L' g L0 m 3.854.49 0.280.47 0.380.41 0.11+0.21 0.310.37
(23) (23) (23) (23) (23)
8 m 0.950.98 0.05+0.19 0.16-0.26 0.010.04 0.16+0.29
(25) (25) (25) (25) (40)
18 m 0.420.24 0.05+0.17 0.000.00 0.00-0.00 0.000.00
(14) (14) (14) (14) (14)
28 m 0.370.21 0.000.00 0.080.20 0.030.14 0.000.00
(23) (23) (23) (23) (23)






63


Surface Runoff Water Statistical Analyses

Atrazine. As with the atrazine concentrations in groundwater, the immediate after

application concentrations were found to be significantly higher than the pre and post

application phases in the 0 m position adjacent to the application strip (Figure 19).

Although concentrations were visibly higher during the period immediately after

application for each of the other landscape positions (8, 18, and 28 m), there were not

statistically significant differences probably due to the high variability in the data set.

In contrast, groundwater concentrations of atrazine were significantly higher in the three

month period following application than the preapplication or final 9 month phase in

every landscape position (Figure 12).


ATRAZINE ATRAZINE
RUNOFF WATER RUNOFF WATER
OM FROM APPLICATION STRIP 8M FROM APPLICATION STRIP
2 2
., 1.5 a 1.5

0.5.
0 0
PRE APP AFTER APP POSTAPP PRE APP AFTER APP POST APP
PHASE PHASE


ATRAZINE ATRAZINE
RUNOFF WATER RUNOFF WATER
18M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2
1.5 A 15 a a a
U) 1 n 1
0.5 ____0.50 0
PRE APP AFTERAPP POSTAPP PRE APP AFTER APP POSTAPP
PHASE PHASE
Figure 19. Atrazine least square means of natural log transformed surface runoff water
data grouped by study phase (pre-application, after application, and post
application) for each of the landscape positions. Differing alphabet
designations indicate study periods were significantly different (p<0.05).

When comparing landscape positions during the 3 month period immediately after

application, atrazine concentrations were significantly higher in the 0 m position than the





64


28 m position (Figure 20). However, concentrations were not significantly different between the 0 m, 8 m, and 18 m positions, suggesting that the pine vegetative zone was also instrumental in decreasing concentrations of atrazine in runoff water. During the post application period, there were no significant differences between landscape positions.


ATRAZINE ATRAZINE
RUNOFF WATER RUNOFF WATER
AFTER APPLICATION POST APPLICATION
2 21.5 a ab ab b F .,1.5 a a a a

U0.5 M 0
0 I t N 0DSCAP
OM 8M 18M 28M OM 8M 18M 28M
LANDSCAPE POSITION LANDSCAPE POSITION

Figure 20. Atrazine least square means of natural log transformed surface runoff water
data grouped by landscape position for the two study periods following
application. Differing alphabet designations indicate landscape positions were
significantly different (p<0.05).

Deethylatrazine. Differences in deethylatrazine concentrations between the study periods were found to be significant only at the 8 m landscape position (Figure 21). In the 8 m position, both study periods following application were significantly higher than the pre application period in the 8 m landscape position. While there was clearly an increase of deethylatrazine in the 0 m landscape position following application, no significant differences were found between the three phases of the study probably due to the high variability in the data set.

Comparison between landscape position concentrations within the after

application and post application periods produced significant differences (Figure 22). Concentrations at both the 0 m and 8 m positions were significantly higher than the final







65




DEETHYLATRAZINE DEETHYLATRAZINE
RUNOFF WATER RUNOFF WATER
OM FROM APPLICATION STRIP 8M FROM APPLICATION STRIP
2 2
Ww
1.5 1. _1



0.5 0.5

0 0
PREAPP AFTERAPP POSTAPP PREAPP AFTERAPP POSTAPP
PHASE PHASE




DEETHYLATRAZINE DEETHYLATRAZINE
RUNOFF WATER RUNOFF WATER
18M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2
wW
F- 1.5 a a a I 1.5



0 j -- O 1 I
,,, 0 .5 .( 0.5



PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE


Figure 21. Deethylatrazine least square means of natural log transformed surface runoff
water data grouped by study phase (pre-application, after application, and post
application) for each of the landscape positions. Differing alphabet
designations indicate study periods were significantly different (p<0.05).



DEETHYLATRAZINE DEETHYLATRAZINE
RUNOFF WATER RUNOFF WATER
AFTER APPLICATION POST APPLICATION
2 2 -- -
Ww
1.5 a b- 1.5 b c



0.5 <- 0.5 "
0 0j 00M 8M 18M 28M OM 8M 18M 28M

LANDSCAPE POSITION LANDSCAPE POSITION


Figure 22. Deethylatrazine least square means of natural log transformed surface runoff
water data grouped by landscape position for the two study periods following application. Differing alphabet designations indicate landscape positions were
significantly different (p<0.05).







66



28 m position in the immediate after application period. The 0 m position continued to


be significantly higher than all other downslope positions during the post application


period of the study.


Hydroxyatrazine. As with the parent compound, hydroxyatrazine concentrations


were only significantly different (p<0.05) at the 0 m landscape position between the preapplication period and the immediate after application period (Figure 23). As with the


other analytes, comparison between landscape position concentrations within the after


application and post application period showed significant differences between the


upslope and downslope landscape positions (Figure 24).



HYDROXYATRAZINE HYDROXYATRAZINE
RUNOFF WATER RUNOFF WATER
OM FROM APPLICATION STRIP 8M FROM APPLICATION STRIP
22
LU LU
1.5 f 1.h a a a



4 L- 0.5 __ 0.5
0 0_ 0

PREAPP AFTERAPP POSTAPP PREAPP AFTERAPP POSTAPP
PHASE PHASE




HYDROXYATRAZINE
RUNOFF WATER
18M FROM APPLICATION STRIP

LU



U4
2.


0.5


PRE APP AFTER APP POST APP
PHASE


Figure 23. Hydroxyatrazine least square means of natural log transformed surface runoff

water data grouped by study phase (pre-application, after application, and post
application) for each of the landscape positions. Differing alphabet
designations indicate study periods were significantly different (p<0.05).

There was insufficient data to statistically analyze data in the 28 m position.







67




Ln HYDROXYATRAZINE Ln HYDROXYATRAZINE
RUNOFF WATER AVERAGES RUNOFF WATER AVERAGES
AFTER APPLICATION POST APPLICATION
2 -2 - - -- --
2

1.5 a ab ab b 1.5 a ab b b
-_ z-J

o 0.5 o 0.5

0 -0
OM 8M 18M 28M 0M 8M 18M 28M
LANDSCAPE POSITION LANDSCAPE POSITION


Figure 24. Hydroxyatrazine least square means of natural log transformed surface runoff
water data grouped by landscape position for the two study periods following application. Differing alphabet designations indicate landscape positions were
significantly different (p<0.05).






DEISOPROPYLATRAZINE DEISOPROPYLATRAZINE
RUNOFF WATER RUNOFF WATER
OM FROM APPLICATION STRIP 8M FROM APPLICATION STRIP
2 2
ILU LUI
1.5 a b a 1. a a a

Cn 1 c 1
CY CY
j 0.5 0.5

0 0 .0m...
PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE




DEISOPROPYLATRAZINE DEISOPROPYLATRAZINE
RUNOFF WATER RUNOFF WATER
18M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2
uJ uJ
T1.5 a a a 41.5 a a a

1a C 1
0.5 < 0.5

0- 0
PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE


Figure 25. Deisopropylatrazine least square means of natural log transformed surface
runoff water data grouped by study phase (pre-application, after application, and post application) for each of the landscape positions. Differing alphabet
designations indicate study periods were significantly different (p<0.05).





68


Deisopropylatrazine. Significant differences in deisopropylatrazine

concentrations between the three study periods were again only detected in the 0 m landscape position at p<0.05. Deisopropylatrazine concentrations were significantly higher in the 3-month period immediately following application than the pre and post application phases (Figure 25) at the upslope 0 m landscape position. Position concentration differences for deisopropylatrazine mimicked the parent compound as well. The only significant difference between landscape positions was in the immediate after application phase where the 0 m position was significantly higher than the 28 m position (Figure 26).


DEISOPROPYLATRAZINE DEISOPROPYLATRAZNE
RUNOFF WATER RUNOFF WATER
AFTER APPLICATION POST APPLICATION
2 _ _2
S. a ab ab b 1.a .
15 1.5 a a
I
0, __ 1 0.50 E 1 0
0M 8M 18M 28M 0M 8M 18M 28M
LANDSCAPE POSITION LANDSCAPE POSITION

Figure 26. Deisopropylatrazine least square means of natural log transformed surface
runoff water data grouped by landscape position for the two study periods following application. Differing alphabet designations indicate landscape
positions were significantly different (p<0.05).

Bromide. No significant differences in bromide concentrations were detected between the three periods of the study at the p<0.05 significance level (Figure 27). However, the 0 m landscape position was significantly higher than the streamside 28 m position in both the after application and post application phases of the study (Figure 28).







69




BROMIDE BROMIDE
RUNOFF WATER RUNOFF WATER
OM FROM APPLICATION STRIP 8M FROM APPLICATION STRIP
2 2

a a a a a a
aE
0.5f 1.5
-Jcno( i i ;05l


0 0
PRE APP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE




BROMIDE BROMIDE
RUNOFF WATER RUNOFF WATER
18M FROM APPLICATION STRIP 28M FROM APPLICATION STRIP
2 2
LUw
a a a r- a a a
1.5 1.5a1.5

U)-
(W0.5 W 0.5

0 = 0

PREAPP AFTER APP POST APP PRE APP AFTER APP POST APP
PHASE PHASE


Figure 27. Bromide least square means of natural log transformed surface runoff water
data grouped by study phase (pre-application, after application, and post
application) for each of the landscape positions. Differing alphabet

designations indicate study periods were significantly different (p<0.05).





BROMIDE BROMIDE
RUNOFF WATER RUNOFF WATER
AFTER APPLICATION POST APPLICATION
2 2

Li 1.5 a ab ab b" 15 a _ab a b



0W0

0 00M 8M 18M 28M 0M 8M 18M 28M
LANDSCAPE POSITION LANDSCAPE POSITION


Figure 28. Bromide least square means of natural log transformed surface runoff water
data grouped by landscape position for the two study periods following
application. Differing alphabet designations indicate landscape positions were
significantly different (p<0.05).





70


Statistics did show that the riparian system was effective in diminishing

concentrations of the analytes. In comparisons among landscape positions in the three months following application, each analyte exhibited significantly larger concentrations in the 0 m landscape than the final 28 m position. Deethylatrazine, hydroxyatrazine, and bromide concentrations in final 9 months phase of the study also had significantly larger concentrations in the 0 m landscape position than the 28 m position. Though no significant differences were found between landscape positions for atrazine or deisopropylatrazine in the final phase of the study, mass loads of these analyes were much smaller in the final study phase. Mass loads of atrazine and deisopropylatrazine entering the 0 m position in the final phase were 9% and 18% of the total load calculated for the entire year following application.

Surface Runoff Sediment Concentrations

There were measurable concentrations of the triazine compounds associated with the runoff sediments (Table 12). Hydroxyatrazine, which has been shown to be more readily adsorbed onto soil particles (Brouwer et al., 1990, Clay and Koskinen, 1990, Roy and Krapac, 1994), had a higher average concentration (0.77 tag L-) in the initial three month period following application than the parent or two other degradation products. Table 12. Surface runoff sediments concentrations averaged over the immediate after
application period (4/24/03-7/31/03) for each landscape position. Mean
followed by standard deviation (n)
Landscape AT DEA DIA HA
position jig L jig L-1 jig L- jig L'
0 m 0.420.76 (11) 0.140.31 (11) 0.000.00 (10) 0.771.16 (11)
8 m 0.110.29 (15) 0.170.26 (15) 0.030.10 (15) 0.020.09 (15)
18 m 0.000.00 (10) 0.280.31 (10) 0.000.00 (10) 0.000.00 (10)
28 m 0.100.38 (14) 0.190.57 (14) 0.110.40 (14) 0.120.46 (14)





71


Table 13. Surface runoff sediments concentrations averaged over the final post
application period (7/31/03-4/5/04) for each landscape position. Mean
followed by standard deviation (n)
Landscape AT DEA DIA HA
position gg L-1 gg L-' 9g L-1 [g L'
0 m 0.000.00 (23) 0.400.33 (23) 0.140.29 (23) 0.100.25 (23)
8 m 0.000.00 (25) 0.240.29 (25) 0.06-0.14 (25) 0.060.23 (25)
18 m 0.060.23 (14) 0.240.29 (14) 0.060.16 (14) 0.010.05 (14)
28 m 0.000.00 (23) 0.200.27 (23) 0.080.17 (23) 0.050.18 (23)
During the post application period of the study (final 9 months), the average hydroxyatrazine concentration entering the buffer was much lower (0.10 jig U) (Table 13).

Statistical analysis of hydroxyatrazine in runoff sediment showed that

concentrations at the 0 m landscape position during the period immediately after application were significantly higher than the pre or post application periods at the p<0.05 level (Figure 29). No significant differences were found for any of the other landscape positions. Significant differences in hydroxyatrazine concentrations between landscape positions were observed only at the 0 m landscape position during the period immediately after application (Figure 30), thus indicating the grass zone was effective in reducing the surface runoff water sediment loads of hydroxyatrazine.

In contrast to hydroxyatrazine, average deethylatrazine concentrations increased from 0.14 jig L- to 0.40 jig L-1 between the two periods after application for the 0 m landscape position (Table 13). Statistical analysis of deethylatrazine concentrations between periods showed no significant differences (Figure 31). Analysis between landscape positions found significant differences, but only during the last period of the study. Neither atrazine nor deisopropylatrazine were found to exhibit significant differences either between periods or between landscape positions at the p<0.05 level.







72




HYDROXYTRAZINE HYDROXYTRAZINE
RUNOFF SEDIMENT RUNOFF SEDIMENT
OM FROM APPLICATION STRIP 8M FROM APPLICATION STRIP
2 2 .-h a ~41.5




0 _4
PREAPP AFTERAPP POSTAPP PREAPP AFTERAPP POSTAPP
PHASE PHASE


HYDROXYTRAZINE
RUNOFF SEDIMENT
28M FROM APPLICATION STRIP
2
1.5 A. a


S 0.5
0
PREAPP AFTERAPP POSTAPP
PHASE


Figure 29. Hydroxyatrazine least square means of natural log transformed runoff
sediment data grouped by study phase (pre-application, after application, and
post application) for each of the landscape positions. Differing alphabet designations indicate study periods were significantly different (p<0.05).
Insufficient data to analyze 18 m position.



HYDROXYATRAZINE HYDROXYATRAZINE
RUNOFF SEDIMENT RUNOFF SEDIMENT
AFTER APPLICATION POST APPLICATION
2 ----- ---- ---- ---2 --- -------__ -

1.5 _-----b-________ -__ a___ a0.5 U ~0,5OM 8M 18M 28M OM SM 18M 28M
LANDSCAPE POSITION LANDSCAPE POSITION


Figure 30. Hydroxyatrazine least square means of natural log transformed runoff
sediment data grouped by landscape position for the two study periods
following application. Differing alphabet designations indicate landscape
positions were significantly different (p<0.05).





73


DEETHYLATRAZINE DEETHYLATRAZINE
RUNOFF SEDIMENT RUNOFF SEDIMENT
AFTER APPLICATION POST APPLICATION
2 2
1.5 a a a a a h ph bL0.5 o05 __ _ __

OM 8M 18M 28M OM 8M 18M 28M
LANDSCAPE POSITION LANDSCAPE POSITION

Figure 31. Deethylatrazine least square means of natural log transformed runoff
sediment data grouped by landscape position for the two study periods
following application. Differing alphabet designations indicate landscape
positions were significantly different (p<0.05). Dilution Results

The addition of analyte-free rain water to surface flow has the effect of diluting concentrations and enlarging changes in concentration reductions as flow moves downslope. To estimate the true attenuation effects of the buffer system on analyte concentrations, dilution factors were calculated for the grass (Zone 3) and both upper and lower portions of pines (Zones 2a and 2b). For each analyte, the net concentration change for each vegetative zone diminished as time progressed throughout the study.

In the present study, a rainfall event producing a sufficient number of samples for each zone to permit dilution calculations did not occur until approximately two and a half months following application of the atrazine (Table 14). During this event, an atrazine concentration change of 3.90 gg L-1 was measured in surface runoff water passing through Zone 3. In the earlier study, Vellidis et al. (2002) found concentration changes of 27.62 and 15.75 jig L- in the grass zone for rainfall events which occurred 4 and 8 weeks, respectively, after application of atrazine on the same application strip.

The dilution factors calculated for each zone within an event were applied to each analyte, and ranged from 0.22 to 0.77 (Table 14). Although the largest net concentration





74


change (NCC) in percent per meter of flow did occur in the grass zone for atrazine (5.28

to 9.76%) the upsiope portion the pines (Zone 2a) had the most consistent changes (4.31

to 7.34% per meter of flow). Aside from a net concentration change of 0.80 [tg L-1during

the first rainfall event shown in Table 14, the lower pine zone had no atrazine

concentrations above the detection limits from which to calculate a zone 2b average for

dilution calculations. For atrazine and the other analytes, an asterisk in the concentration

change column indicates there were no measurable values above the detection limits with

which to calculate concentration changes.

Table 14. Atrazine concentration changes, dilution factors, and estimated concentration
changes due to dilution and other factors for six runoff events
Date Buffer Concentration Dilution Dilution Net Concentration
Zone a Change Factorb Concentration Chgnge
([tg L-') Change (g~g L-')c (% mI')d
(gtg U')
7 Jul 3 3.90 0.22 0.85 3.05 9.76
2003 2a 0.33 0.26 0.09 0.24 7.34
2b 1.49 0.46 0.69 0.80 5.36
23 Jul 3 0.59 0.58 0.34 0.25 5.28
2003 2a 0.66 0.30 0.20 0.47 7.01
2b *0.51
24 Jul 3 1.07 0.53 0.57 0.50 5.88
2003 2a 1.21 0.60 0.73 0.48 4.02
2b *0.71
5 Aug 3 0.04 0.31 0.01 0.02 8.58
2003 2a 0.52 0.32 0.17 0.35 6.78
2b *0.38
8 Aug 3 0.81 0.23 0.19 0.62 9.59
2003 2a 0.80 0.59 0.46 0.35 4.31
2b *0.77
9 Feb 3 *0.33
2004 2a *0.35
2b *0.44
laJ Zone 3 is 8 m of rass, zone 2a -is the upper 10 m of pines, and zone 2b is the lower 10 m of pines Eb Rainfall input (m )/(Rainfall input (m-3) + Surface runoff entering zone(in3)) E'l Concentration change not due to dilution
[d] Percent concentration change not due to dilution per meter of zone width, 8m for zone 3 and 10 m for zones 2a and 2b
* No sample measurements above detection limits





75


Table 15. Deethylatrazine concentration changes, dilution factors, and estimated
concentration changes due to dilution and other factors for six runoff events. Date Buffer Concentration Dilution Dilution Net Concentration
Zone'a Change Factorb Concentration Change
(pgg L-') Chang Ie (gg U 1)c (% m-)d
(g.g L)
7 Jul 3 3.42 0.22 0.75 2.67 9.76
2003 2a 2.07 0.26 0.55 1.52 7.34
2b 0.62 0.46 0.28 0.33 5.36
23 Jul 3 -0.45 0.58 -0.28 -0.20 5.28
2003 2a 0.58 0.30 0.17 0.41 7.01
2b 0.39 0.51 0.20 0.19 4.92
24 Jul 3 0.82 0.53 0.44 0.39 5.88
2003 2a 1.26 0.60 0.76 0.51 4.02
2b *0.71
5 Aug 3 0.78 0.31 0.24 0.53 8.58
2003 2a *0.32
2b -1.01 0.38 -0.38 -0.62 6.2o
8 Aug 3 1.56 0.23 0.36 1.19 5.98
2003 2a *0.59
2b *0.77
9 Feb 3 -0.36 0.33 -0.11 -0.24 8.36
2004 2a *0.35 0.13 0.23 6.47
2b *0.44
LFa] Zone 3 is 8 m of grass, zone 2a is the upper 10 mn of pines, and zone 2b is the lower 10 m of pines [b] Rainfall input (in) / (Rainfall input (Mn) + Surface runoff entering zone (in3)) E'I Concentration change not due to dilution
[d] Percent concentration change not due to dilution per meter of zone width, 8m for zone 3 and l1in for zones 2a and 2b
*No sample measurements above detection limits
The dilution concentration information for deethylatrazine is provided in Table 15.

As with the hydroxyatrazine and bromide data, there are a few of the rainfall events

during which downslope concentration were greater the upsiope concentrations thus

producing negative concentration changes in the tables. While surface topography and

flow paths most likely caused variations in the volumes of surface runoff water measured

in each collector, it is probable that variations in the rates of formation and movement of

the degradation products were also responsible for the downslope increases in

concentration. Of the degradation products, deethylatrazine exhibited the largest net

concentration changes in the first rainfall event. For the July 7, 2003, event, the average





76


concentration was diminished by 2.67 j tg L-lin the grass zone, 1.52 tg L-1 in Zone 2a,

and 0.33 [tg LU1 in Zone 2b (Table 15).

Because deisopropylatrazine concentrations were below detection limits in Zone

2b, there were no calculated concentration changes in Zone 2b for the first rainfall event

(Table 16). However the grass (Zone 3) produced a 1.42 tg L-1 decrease and the upper

pine zone showed a 0.45 [tg LU' decrease. Deisopropylatrazine was present in the lowest

concentration of the degradation products (Table 10), but the diminished concentrations

Table 16. Deisopropylatrazine concentration changes, dilution factors, and estimated
concentration changes due to dilution and other factors for six runoff events.
Date Buffer Concentration Dilution Dilution Net Concentration
Zone a Change Factorb Concentration Change
(mg U ) Change (mg U1)' (% m1-)d
(mg L-')
7 Jul 3 1.81 0.22 0.40 1.42 9.76
2003 2a 0.61 0.26 0.16 0.45 7.34
2b *0.46
23 Jul 3 *0.58
2004 2a *0.30
2b 0.62 0.51 0.31 0.30 4.92
24 Jul 3 0.54 0.53 0.29 0.25 5.87
2003 2a 0.56 0.60 0.34 0.23 4.02
2b *0.71 0
5 Aug 3 0.37 0.31 0.12 0.26 8.58
2003 2a *0.32
2b *0.38
8 Aug 3 0.66 0.23 0.15 0.510 9.58
2003 2a *0.59
2b *0.77
9 Feb 3 *0.33
2004 2a *0.35
2b *0.44
[a] Zone 3 is 8 m of grass, zone 2a is the upper 10 m of pines, and zone 2b is the lower 10 m of pines [b] Rainfall input (mn) / (Rainfall input (mn) Surface runoff entering zone (in3)) ['] Concentration change not due to dilution [d) Percent concentration change not due to dilution per meter of zone width, 8 mn for zone 3 and 10 mn for zones 2a and 2b
* No sample measurements above detection limits





77


and reductions in concentrations in the buffer system are important to be recognized.

Deisopropylatrazine retains the phytotoxic properties of the parent atrazine.

Concentration changes in hydroxyatrazine were found which were attributed more

to the buffer system than rainfall dilution (Table 17). However, two cases of increased

concentrations in Zone 2a made comparisons of maximum net concentrations with the

other analytes impossible. Net concentrations for all rain events combined ranged from

0.06 to 1.21 [tg L- for zones resulting in a positive concentration change.

Table 17. Hydroxyatrazine concentration changes, dilution factors, and estimated
concentration changes due to dilution and other factors for six runoff events. Date Buffer Concentration Dilution Dilution Net Concentration
Zonea Change Factorb Concentration Change
(jig L-') Change (jig U 1)c (% m-')
(jig L-')
7 Jul 3 -0.98 0.22 -0.21 -0.76 9.76
2003 2a 1.65 0.26 0.44 1.21 7.34
2b *0.46
23 Jul 3 0.14 0.58 0.08 0.06 5.28
2003 2a 0.66 0.30 0.20 0.47 7.01
2b 0.54 0.51 0.28 0.26 4.92
24 Jul 3 0.11 0.53 0.06 0.50 5.88
2003 2a 1.56 0.60 0.93 0.62 4.02
2b *0.71
5 Aug 3 -0.22 0.31 -0.07 -0.15 8.58
2003 2a 1.24 0.32 0.40 0.84 6.78
2b *0.38 0.00 0.00 0.00
8 Aug 3 0.43 0.23 0.10 0.33 9.59
2003 2a 0.88 0.59 0.50 0.38 4.31
2b *0.77
9 Feb 3 *0.33
2004 2a *0.35
2b *0.44
[aJ Zone 3 is 8 m of grass, zone 2a is the upper 10 m of pines, and Ozone 2b is the lower 10 m of pines [bI Rainfall input (in) / (Rainfall input (mn) + Surface runoff entering zone (in3)) [ci Concentration change not due to dilution
[d] Percent concentration change not due to dilution per meter of zone width, 8 in for zone 3 and 10 in for zones 2a and 2b
* No sample measurements above detection limits





78


The processes involved in producing surface runoff water concentration changes for atrazine in buffer systems are most likely infiltration and adsorption to soil and plant material. Barfield et al. (1998) found that infiltration followed by adsorption in the surface layer were the major removal mechanisms in natural fescue filter strips. Barfield et al. (1998) also found increasing the flow width from 4.57 to 13.72 did improve the trapping efficiency of the grass strip. The present study clearly illustrates the importance of degradation of the parent compound prior to, or during transport in the surface runoff water, and the importance and effectiveness of these riparian buffer systems in reducing concentrations of the degradation products. It is not possible based on the results of the study to determine which degradation processes are occurring while in the system as each degradation product was present in the upslope landscape position immediately following application.

Surface Runoff Mass Loads

Examination of the mass load of each analyte summed for the two periods following application stresses the importance of the degradation products in surface runoff water transport (Table 18). The measured load of atrazine entering the 0 m upslope position was 769 mg in surface runoff water. Approximately half as much deethylatrazine (445 mg) and hydroxyatrazine (318 mg) and about 20% as much deisopropylatrazine (147 mg) entered the system compared with the parent atrazine. Removal efficiency (%) of the analytes from the surface runoff water was between 97.5% and 100.0 % for atrazine, deisopropylatrazine, and hydroxyatrazine. Deethylatrazine exhibited the lowest removal efficiency of the triazines with an estimated 92.0%.






79


Table 18. Surface runoff water mass loads summed over the combined two phases of the
study following application for each landscape position. Mean followed by
(n)
Landscape BR AT DEA DIA HA
position g mg mg mg mg
0Om 1084.85 769.51 445.49 146.96 318.09
(32) (34) (34) (34) (34)
8 m 851.51 6049.94 184.88 18.55 78.78
(38) (40) (40) (40) (40)
18 m 156.82 426.141 13.80 2.70 2.37
(24) (24) (24) (24) (24)
28 mn 157.66 19.10 35.76 0.63 0.00
(37) (37) (37) (37) (37)
Removal 85.5 97.5 92.0 99.6 100.0
Efficiency %
Removal Efficiency
(Mass load entering at 0m-nMass load Leaving 38 rn/Mass load entering at 0 rn)* 100
1-ydroxyatrazine removal efficiency associated with the sediments portion of the

surface runoff (67.4%) was much lower than the ground water or surface runoff water

(Table 19). Hydroxyatrazine removal (%) from runoff water was comparable to the

removal from ground water with 92% of the mass entering the 0 mn landscape position

being removed by the 28 mn position. Atrazine and deethylatrazine removal efficiencies

for the surface runoff sediments loads were 81.3 and 75.7%, respectively, and were

roughly 10- 15 % lower than removal efficiency estimates for water.

Table 19. Surface runoff sediment loads summed over the combined two periods of the
study following application for each landscape position. Mean followed by
(n)
Landscape AT DEA DIA HA
position mg mg mg mg
0 m 62.46(24) 176.46 (34) 64.60 (34) 234.99 (34)
8 m 6.04 (40) 94.35 (40) 43.28 (40) 13.78 (40)
18 m 59.50 (24) 48.10 (24) 2.64 (24) 1.20 (24)
28 m 11.70 (37) 42.92 (37) 21.07 (37) 18.69 (37)
Removal 81.3 75.7 67.4 92.0
Efficiency %
Removal Efficiency
(Mass load entering at 0 rn- Mass load Leaving 38 in/Mass load entering at 0 m)* 100





80


Deisopropylatrazine exhibited the poorest removal efficiency (%) with an estimated 67.4% of the analyte associated with the sediments load being removed by the system.

Surface runoff sediment concentrations have been shown to be reduced greatly in vegetative buffer zones. Assmussen et al. (1977) found 94% sediment removal during wet antecedent conditions and 98% during dry antecedent conditions in 24.4 m grassed waterways while studying 2, 4-D runoff. The 2, 4-D was removed (30%) by infiltration, reduction in sediments load, and adsorption to vegetation and organic matter. Sediment loads were reduced by 72% in 4.6 m and 75% in 9.1 m buffer strips studied by Mickelson and Baker (1993). The larger buffer strip also removed 23% more of the atrazine applied to simulate surface runoff. Arora et al. (1996) found sediment removal efficiencies of 40-100% using 20.1 m bromegrass buffer strips, and atrazine retention by sediment removal in the buffer strip accounted for only 5% of the total atrazine retention by the buffer strip.

Historically, studies of pesticides in surface runoff have shown that the majority of pesticide lost in runoff is associated with the water phase (Burgoa and Wauchope, 1995, Wu et al., 1983). In the current study, this was true for the atrazine in comparing total mass loads in runoff water and runoff sediments entering the 0 m position summed for the entire year following application. In contrast, the deethylatrazine and deisopropylatrazine loads were roughly 40% of the runoff water loads entering the system and the hydroxyatrazine sediments mass load was approximately 75% as large as the runoff water load. Surface adsorption thus played a predominant role in removal of the degradation products from the surface runoff.





81


The H-flume at the wetland outlet captured all surface flow leaving the Dairy Wetland

and consequently can be used to quantify the overall efficiency of the forested buffer

system. Table 20 gives the estimated total mass loads exiting the H-flume for each analyte. The vast majority of the mass of atrazine occurred in the H-flume sample

collected immediately after the first rainfall event following application of the atrazine.

Though the buffer system was extremely efficient in reducing mass loads of the analytes,

the timing and intensity of the first rainfall event is clearly of the greatest importance.

The atrazine load leaving through the H-flume was almost entirely due to the one event

April 28, 2003 (98.6%), while 47% and 45%, respectively, of the deethylatrazine and

deisopropylatrazine loads were due to the first event following application.

Comparatively, 16.3 and 24.0% of the hydroxyatrazine and bromide loads were attributed

to the event on April 28, 2003. It is suggested that the hydroxyatrazine load was

diminished due to its increased affinity for soils.

Table 20. Mass loads of each analyte and the total water volume exiting the forested
buffer system through the H-flume into the farm pond.
Parameter Water BR AT DEA DIA HA
3 g mg mg mg mg
Total summed for entire study following application Water 2993.11 1610.00 7908.98 1000.44 264.41 116.94
Sediments 106.91 328.97 136.77 595.63
Total Leaving 2993.11 1610.00 8015.89 1329.41 401.18 712.57

H-flume sample from 4/28/03
Water 217.47 387.09 7799.48 626.31 177.24 116.34
Sediments 102.21 0.00 0.00 0.00
Total for sample 387.09 7901.69 626.31 177.24 116.34
Percent of total 24.0 98.6 47.1 44.2 16.3


For this one composite sample (April 28, 2003) the estimated concentration of

atrazine being discharged into the farm pond over the course of the initial rainfall event





82


was 36 jig L-1. The same sample produced estimated concentrations of 2.8 pig L-1 deethylatrazine, 0.8 jig L-1 deisopropylatrazine, and 0.5 jug L- hydroxyatrazine being discharged into the pond during this event. Field studies have shown that concentrations of pesticides in runoff normally decreases exponentially with the days following application (Gaynor, et al., 1995, Wauchope and Leonard, 1980, Leonard et al., 1979) and that substantial losses are always associated with heavy rain events which occur within 14 days of application (Wauchope, 1978).

The quantities measured leaving the H-flume during the first rain event were a very small portion of the total 2680 g of active ingredient applied. Approximately 8.2 g of atrazine, or 0.3% of the total applied, flowed out of the buffer system in the first rain event. Combining the mass loads for all analytes and comparing to the total mass of atrazine applied estimates less than 0.03% of the total applied left in the form of degradation products during the first rain event.

Aside from losses of the analytes in the initial rain event following application, the restored riparian buffer system was extremely efficient in reducing mass loads being transported through the system (Table 2 1). Atrazine and each of its degradation products exhibited 84.2% to 100% removal, with deethylatrazine being the least efficiently removed in both ground water and surface runoff water. Removal of the analytes from Table 21. Removal efficiency (%) from the groundwater, surface runoff water, and
surface runoff sediment matrices complied from mass loads summed for the
entire year following application.
Matrix AT DEA DIA HA
Ground water 95.4 84.2 99.5 94.7
Surface runoff water 97.5 92.0 99.6 100.0
Surface runoff sediment 81.3 75.7 67.4 92.0
Removal Efficiency =
(Mass load entering at 0mn- Mass load Leaving 3 8 or 28 rn/ Mass load entering at 0m) *100





83


the sediment portion of the runoff was less than from the water matrices, though the mass loads were still substantially reduced in the riparian buffer system (67.4% to 92.0%).

While the presence of the degradation products in large fractions compared to the parent compound clearly indicates the importance of the degradation pathway, it was not possible to distinguish between degradation which occurred in the application strip (or in a field under conventional production conditions) and degradation of atrazine which occurred within the buffer. Significant concentrations and loads of metabolites were measured entering the system as well as at several landscape positions within the buffer system. Each of the analyses was present in all matrices in the first samples following application.

Based on the findings of this work, it is clear that the major attenuation processes are further degradation in conjunction with plant uptake from groundwater and surface runoff water. The presence of deethylatrazine as the predominant metabolite in both water matrices denotes the importance of biotic degradation in the system (Mills and Thurman, 1994, Erickson and Lee, 1989). Adsorption of the degradation products to surface soils during overland transport was important in removal of the degradation products during runoff events.













CHAPTER 5
SOIL STUDY

Materials and Methods

Field Configuration and Sampling Strategy

Soil samples were first collected on April 16, 2003, to establish background concentrations for the compounds. Following application, soil cores were taken approximately every 6 weeks for a total of eight sampling dates which produced 478 samples. Samples were collected at the upslope edge of Zones 1, 2, 3, and the mid-point of Zone 2 in three transects which corresponded to the position of the LIFE samplers (Figure 7).

Samples were collected in close proximity to LIFE samplers. The first set of

samples was collected approximately 0.5 in away from the samplers. The second set was collected approximately 0.5 in away from the opposite side of the samplers. Subsequent samples were collected 0.5 in away from previous sampling locations alternating from left to right.

Samples were collected in 121.9 x 6.56 cm core sleeves which were PETG plastic, zero contamination, and inert to chemicals/contaminants (Giddings Machine Company, Fort Collins, CO). The sleeves were placed inside a steel soil probe (Giddings Machine Company). The probe was 128.2 cm long with a diameter of 6.56 cm. It was modified locally to provide a surface for placing a drive head used to pound the probe into the earth. The drive head was placed on top of the probe and the probe was pounded into the




84





85


earth using a gas powered jackhammer until the upper handle bracket touched the soil surface.

To account for any compaction associated with driving the probe into the ground or with loss of a portion of the soil core during extraction from the ground, depths from the soil surface to the top of the soil probe and to the top of the sleeve were measured before the probe was extracted from the ground, Distance from the LIFE sampler was also recorded. The probe was extracted by attaching a manually operated winch cable to the handle at the top of the probe. The winch was attached to a metal tripod which was placed over the probe prior to extraction. The probe containing the soil core was gradually winched up as a vacuum was applied to the soil core with a small hand pump to minimize losses from the bottom of the core as it was extracted. The sleeve was then removed from the soil probe. To prepare the probe for the next sample, soil residue was removed with a steel brush. If needed, clean water was used to rinse the probe and paper towels used to dry it.

Holes from the extract cores were filled using a 50:50 mixture of silica sand and

bentonite clay. Samples were returned to the lab within 2 hours and stored at 4' C. Over the following 2 days the soil cores were divided into 5 segments beginning at the soil surface end. The segments corresponded to depth from the soil surface and were 0-12.5 cm, 12.5-25.0 cm, 25.0-50.0 cm, 50.0-75.0 cm, and 75.0-100 cm below the ground

surface. On a few occasions a section of the bottom segmented was lost during extraction. In that case, the remaining section was used. Each section was issued a sample number, sieved to 2 mm, and total weight recorded. A 25.00.1 g sample was placed into an aluminum planchette for moisture determination and another 25.00.1 g





86


portion weighed into a glass jar for chemical extraction and frozen until extracted and analyzed. Any remaining soil was also frozen separately. Sample Preparation

The method for extraction of soil samples was adapted from a method provided by the USDA-ARS-SEWRL Pesticide Residue Laboratory (Tifton, Georgia) and the mixed mode extractant method developed by Lerch and Yong-Xi (2001). Preliminary low recoveries of the chlorinated species using only the mixed extractant mandated incorporating two 100% acetonitrile extraction steps prior to one mixed mode extraction process. Lerch and Yong-Xi (2001) perfected the mixed mode extractant for the hydroxylated atrazine species only, and were not attempting to recover chlorinated species.

Extraction began by allowing the sieved soil samples (25 g), which had been

weighed and frozen, to equilibrate to room temperature. Samples were then spiked with 200 pl of surrogate (10 mg L-1 terbuthylazine) or surrogate plus analyte (10 mg Ldeisopropylatrazine, deethylatrazine, hydroxyatrazine, and atrazine) for matrix recovery checks. Samples were extracted 2 times with 50 ml of acetonitrile and once with 50 ml of the mixed mode extractant. For each extraction, the samples were placed on a rotary bed shaker at 210 rpm for 2 hours. Samples were then centrifuged at 5000 rpm for 10 minutes and the liquid was vacuum filtered through a 7 cm Whatman GF/B glass fiber filter into a 500 ml glass bottle. The extraction was repeated twice. A third extraction was performed following the addition of 50 mL of the mixed mode extractant.

Following the final extraction, the mixed mode extractant and soil were centrifuged and the liquid decanted onto the filtration device. The soil was resuspended in 10 ml of acetonitrile and poured onto the filtration device. The sample jars were rinsed with