Kamrun Nahar Doctor of Philosophy November 2015

Home , Kamrun Nahar

A Low Cost Method for Glyphosate Analysis, and Site Investigation and Modelling of Glyphosate Fate and Transport from Genetically Modified Canola Farmland in Parkes, NSW, Australia

Kamrun Nahar

A dissertation in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Engineering and Information Technology The University of New South Wales Canberra, Australia

November 2015 ORIGINALITY STATEMENT

'I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person , or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.'

Signed ...... _. ~ --· .. -: ......

Date ...... -~ .T./ _1 _1 / ~Jj /.~ ...... COPYRIGHT STATEMENT

'I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ...... ~ ......

Date ...... R,.?./t?l

AUTHENTICITY STATEMENT

'I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.'

Signed • •••• •• • • ••••••••• fO'f ...... ~......

Date ...... ~ . !.!.~.~. .!~.~-~?...... PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: NAHAR

First name: KAMRUN Other name/s:

Abbreviation for degree as given in the University calendar: PhD

School: School of Engineering and Information Technology Faculty:

Title: A Low Cost Method for Glyphosate Analysis, and Site Investigation and Modelling of Glyphosate Fate and Transport from Genetically Modified Canola Farmland in Parkes, NSW, Australia

Abstract 350 words maximum: (PLEASE TYPE) Since 2000, the use of a glyphosate, a broad-spectrum herbicide, has increased dramatically due to the cultivation of genetically modified crops in Australia. The research conducted for this thesis analysed glyphosate in environmental samples obtained from an agricultural area in Parkes, NSW, and its fate and transport in an agricultural system. Firstly, a low-cost analytical method for glyphosate analysis was established in soils and waters, to avoid the need for much more expensive analytical methods such as High Performance Liquid Chromatography. The method was calibrated against a commercially available enzyme-linked immunosorbent assay (ELISA) method, for the analysis of glyphosate which involves derivatisation with 9-Fluorenylmethyl chloroformate (FMOC-chloride) and measurement of the emissions acquisition at 310 nm with excitation wavelengths of 268 nm. Using linear regression techniques, the calibration curves showed linearity for ranges from 10 µg/L to 25000 µg/L and 10 µg/L to 1,000 µg/L (Models I and III respectively) with good reproducibility (recovery 108.3±20.8; R2=0.997 and 115.230±25.852; R2=0.998 respectively). Secondly, a site investigation was conducted of farmland cultivated with GM canola in the Parkes region, New South Wales, Australia, which involved the sampling of surface waters and soils and glyphosate analysis. The sampled soils are loams under the USDA system (clay 13.8-15.8%, silt 39-43%, sand 41.2-47.2%) and were collected according to the Australian National Environmental Protection Measure sampling protocol. The found glyphosate levels ranged between I 0 and 67 µg/L in waters and 0.10 and 0.575 mg/kg in soils, with aqueous levels lying below Australian and international drinking water guidelines. However, some exceeded the Canadian Freshwater Guideline for the Protection of Aquatic Life of 65 µg/L and most exceeded the South African chronic guideline of 0.3 µg/L. Thirdly, glyphosate sorption isotherms were constructed by batch tests on several soils from the study site and fitted with linear and non-linear Langmuir, Freundlich and Redlich-Peterson isotherms and desorption tests performed using a O. IM KH2P04 solution. Column-leaching experiments were also conducted on glyphosate-dosed and were modelled using the one-dimensional transport model HYDRUS- 1D with sorption and degradation. Finally, the findings are incorporated into a conceptual model of glyphosate transport pathways and environmental receptors.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

27 Nov 2015 ~- ...... ~ ...... : ...... Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and re uire the a roval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

This thesis is dedicated to my family

Acknowledgements

The author wishes to express her sincere gratitude and appreciation to her research supervisor Dr Robert Niven for his valuable guidance, inspiration, suggestions and continuous support through the course of the study. The author also wishes to express her sincere gratitude to her research co-supervisor, Associate professor Dr Tapabrata Ray for the valuable advices given during the course of her study.

The Laboratory facilities provided by the School of Physical, Environmental and

Mathematical Sciences at The University of New South Wales, Canberra are gratefully acknowledged.

The author wishes to express her sincere gratitude to Associate Professor Dr

Stuart Pearson for his guidance and invaluable help from the beginning of the research.

The author gratefully acknowledges Professor Dr Hans Riesen, for his valuable time and contribution towards the experimental part of the research.

A special ‘thanks’ goes to Ms Kate Badek for her assistance in the Laboratory and

Field works from the beginning of the research.

Many thanks to the following people for their assistance and support during the course of this study and preparation of this thesis

Dr David Paull, for his good and helpful suggestions and enormous help and support during the author’s candidature.

Dr A.F.M. Mokhlesur Rahman for his good suggestions and support

Denise Russell - this work presented in this thesis would not have been possible without her help and support

Mr Jim Baxter, David Sharp, Matthew Barrett and Peter Palmer for their assistance during the laboratory testings

The author would like to thank her colleagues Ms Yasmin A. M. A. Abdelraouf,

Mohammad Shakhaout H. Khan, and Md Sayem Uddin for their helpful suggestions and sharing knowledge so willingly.

The most important “thank you” goes to my dear husband, Md. Wahid Ferdous.

Thank you for your love, for your endless patience, for comforting and encouraging me during the challenging periods.

The author wishes to thank her parents who lived far away from here, for their love and support throughout the period of research.

This acknowledgement will not be complete without extending sincere gratitude and appreciation to the School of Engineering and Information Technology at

University of New South Wales, Canberra for providing me with financial support and assistance to carry out this study.

III

Abstract

To reduce the world’s dependency on fossil fuels and greenhouse gas emissions, the implementation of biofuel policies has fostered the expansion of the biofuel industry, with the global production of biofuels, particularly crop-based ones increasing dramatically over the last few years. In this regard concerns have been raised in the scientific community regarding the environmental consequences of land being converted to irrigated agriculture and the associated increase in herbicide use. The current insufficient facilities for second-generation biofuel production have encouraged many countries to continue crop-based biofuel production and to look for alternative genetically modified (GM) crops which are suggested by some previous studies. In Australia the cultivation of this new varieties of GM crops which are usually produced by the insertion of genes from soil bacterium into crops to make them tolerant to one of the most commonly used herbicides, glyphosate (N-

(phosphonomethyl) glycine) has started since their introduction in 2000.The use of glyphosate, a broad-spectrum herbicide, for genetically modified (GM) crops production has increased considerably with approximately 15 Mt of glyphosate are applied annually in Australia to control agricultural, urban and roadside weeds.

Nevertheless, the prohibitive cost of analysing glyphosate in environmental samples often creates difficulties in monitoring studies with a lack of information on its effects both in Australia and overseas.

The research conducted for this thesis analysed glyphosate in environmental samples obtained from an agricultural area in Parkes, NSW, which has been used for GM canola cultivation for the last 8 years, and its fate and transport in an agricultural system. Firstly, a low-cost analytical method (fluorescence spectrometry) was IV established to determine the glyphosate in waters and soils, and avoid the need for much more expensive analytical methods, such as high-performance liquid chromatography (HPLC). For this analysis, the method was compared with, calibrated against and used in conjunction with the commercially available enzyme-linked immunosorbent assay (ELISA) method. This proposed analytical method involved glyphosate extraction, derivatisation with a 9-fluorenylmethyl chloroformate (FMOC- chloride) and measurements of emission acquisitions at 310 nm with excitation wavelengths of 268 nm. Using linear regression techniques, the calibration curves showed linearity for ranges from 10 µg/L to 25000 µg/L and 10 µg/L to 1,000 µg/L

(Models I and III respectively) with good reproducibility (recovery 108.3±20.8;

R2=0.997 and 115.230±25.852; R2=0.998 respectively). The method’s appropriate limits of detection and quantification were 10 µg/L and 0.063 mg/kg, and 29 µg/L and

0.194 mg/kg in waters and soils respectively. The samples were also compared with an established UV-visible spectrophotometric method developed in this study and

HPLC/MS-MS analyses conducted by the National Measurement Institute, Sydney,

Australia. Cost estimates indicate that the new method can be conducted for a cost of

AU$47.42 per sample (in 2014 Australian dollars, excluding GST) at a commercial laboratory, which compares favourably with quoted commercial rates of AU$150-280 per sample (excluding GST) using existing methods.

Secondly, a site investigation was conducted in a GM canola cultivated farmland located in the Parkes region of New South Wales, Australia. GM canola growers from different parts of NSW were surveyed first and then a family-operated farm in the agricultural district of Parkes was chosen for the case study which involved sampling of surface waters and soils, and the analysis of glyphosate concentrations. All the

V samples were collected according to the Australian National Environmental Protection

Measure (NEPM) sampling protocol. The soil samples were classified as loams under the USDA system (clay 13.8‒15.8%, silt 39‒43%, sand 41.2‒47.2%). The found glyphosate levels ranged between 10 and 67 µg/L in waters and 0.10 and 0.575 mg/kg in soils, with aqueous levels lying below Australian and international drinking water guidelines. However, some exceeded the Canadian Freshwater Guideline for the

Protection of Aquatic Life of 65 µg/L and most exceeded the South African chronic guideline of 0.3 µg/L.

Thirdly, to determine the effects of long-term applications of glyphosate on the sorption to three different soils from the site investigation area were examined. In this regard sorption isotherms were constructed by batch tests on several soils from the study site and fitted with linear and non-linear Langmuir, Freundlich and Redlich-

Peterson isotherms, and was compared with the results from UV-visible spectrophotometric analysis. The method of least squares was used for finding the parameters of the isotherms. The experimental values were taken as the initial assumption for determining the model parameters with both linear and non-linearised isotherms. To check variations between the experimental and calculated values for sorption capacity, Chi-square and Average Percentage Errors (APE) were calculated.

Results obtained from the different isotherm models showed a good agreement with the Freundlich isotherm. Glyphosate desorption tests were also performed using a

0.1M KH2PO4 solution on the same soils.

Fourthly, to examine whether the leaching of glyphosate via macro-pore flows can occur in GM canola agricultural soils, column-leaching experiments were conducted

VI on glyphosate-dosed soils, using application and flow rates representative of field conditions with bromide as a non-reactive tracer. The experimental findings with the miscible-displacement experiments of glyphosate and the non-reactive tracer bromide were modelled using the one-dimensional transport model HYDRUS-1D with sorption and degradation. To do so, the physical transport parameters, pore water velocity (ν) and dispersivity (λ) were determined at first, by fitting the experimental bromide breakthrough curves (BTCs) with the analytical solution to the advection-dispersion equation for a pulse boundary condition at the upper and a zero gradient condition at the lower boundaries. Afterwards these parameters along with the parameters from the sorption experiments were used in HYDRUS 1D to describe the transport behaviour of glyphosate, which is well described by a fully kinetic one-site sorption model with sorption and degradation. Finally, glyphosate fate and transport in that agricultural system were presented in a conceptual flow diagram model with an overview of glyphosate transport pathways and environmental receptors.

VII

Table of Contents

Page Dedication I Acknowledgements II Abstract IV Table of Contents VIII List of Figures XV List of Tables XIX Associated Publications XXII Notations XXIII Abbreviations XXIV

Chapter 1 Introduction 1-12 1.1 General 1 1.2 Background and Problem Statement 3 1.3 Objectives of Research 8 1.4 Scope of Study 9 1.5 Organisation of Thesis 10

Chapter 2 Review of Biofuels 13-45 2.1 General 13 2.2 Biofuels and their Classification 15 2.3 Worldwide Trends in Biofuel Production 19 2.3.1 The United States of America (USA) 19 2.3.2 Canada 22 2.3.3 Australia 23 2.3.4 Argentina 23 2.3.5 Brazil 24 2.3.6 The European Union 25 2.3.7 France 26 2.3.8 Germany 26 2.3.9 China 26

VIII

2.3.10 India 27 2.3.11 Indonesia 27 2.3.12 Malaysia 27 2.3.13 Thailand 28 2.4 Environmental Impacts of Large-scale Cultivation 28 2.4.1 Soil 28 2.4.1.1 Soil erosion 29 2.4.1.2 Removal of soil during harvesting 29 2.4.1.3 Soil quality 30 2.4.2. Water 31 2.4.2.1 Water consumption 32 2.4.2.2 Irrigation 32 2.4.2.3 Water quality and aquatic system 33 2.4.3 Terrestrial biodiversity 37 2.5 Genetically Modified (GM) Crops: A New Option for Biofuel 38 2.5.1 GM crops 38 2.5.2 Different varieties of GM crops 39 2.5.2.1 Triazine-tolerant (TT) crops 39 2.5.2.2 Clearfield crops 40 2.5.2.3 Roundup-ready crops 40 2.5.3 Biofuel production from GM crops 42 2.5.4 Impacts of large-scale cultivation of GM crop 43 2.6 Conclusions 44

Chapter 3 Review on Herbicide Glyphosate 46-69 3.1 General 46 3.2 Introduction 47 3.3 Factors Affecting Glyphosate Performance 51 3.4 Mode of Action and Persistence 52 3.4.1 Plants 52 3.4.2 Soils 52 3.4.2.1 Sorption 54

3.4.2.2 Leaching 58 3.5 Impacts of Glyphosate 60 3.5.1 Animals 61 3.5.1.1 Rats 61 3.5.1.2 Rabbits 62 3.5.1.3 Hens and goats 62 3.5.2 Birds 62 3.5.3 Fish 63 3.5.4 Amphibians 64 3.5.5 Humans 64 3.5.6 Plants 66 3.5.7 Marine Organisms 67 3.6 Conclusions 68

Chapter 4 Establishment of Analytical Method for 70-116 Glyphosate Detection 4.1 General 70 4.2 Introduction and Background 71 4.2.1 Gas chromatography 72 4.2.2 Liquid chromatography 74 4.2.3 Capillary electrophoresis 76 4.3 Establishment of the Analytical Method 78 4.3.1 Fluorescence spectroscopy 78 4.3.2 Materials and methods 80 4.3.2.1 Reagents 80 4.3.2.2 Solution 80 4.3.2.3 Procedure 81 4.3.3 Method validation and results 87 4.3.3.1 Calibration curve and linearity 87 4.3.3.2 Limits of detection (LOD) and limits of 88 quantification (LOQ) 4.3.3.3 Precision and accuracy 89

4.3.4 Discussion 89 4.4 Comparative Study 98 4.4.1 General 98 4.4.2 Enzyme-Linked Immunosorbent Assay (ELISA) method 98 4.4.2.1 Principle 99 4.4.2.2 Reagents 100 4.4.2.3 Sample preparation 100 4.4.2.4 Reagent preparation 101 4.4.2.5 Assay procedure 102 4.4.2.6 Results 103 4.4.3 UV-visible spectrophotometric method 105 4.5 Costing of the Analytical Method 111 4.6 Conclusions 115

Chapter 5 Site Investigation 117-159 5.1 General 117 5.2 GM Canola Growers’ Views and Experiences 118 5.2.1 Background on farms 118 5.2.2 Crop sequence used for GM canola plantation 120 5.2.3 Adoption of RR canola 121 5.2.3.1 Positives 121 5.2.3.2 Negatives 122 5.2.4 Performances of different varieties 122 5.2.5 Sowing system(s) used 123 5.2.6 Gross margins and variable costs 123 5.2.7 Problem weeds 124 5.3 Site Investigation 124 5.3.1 Study area: Parkes, NSW 124 5.3.1.1 Background details 124 5.3.1.2 Potential chemical contamination 125 5.3.2 Sowing, herbicide application and rainfall events 131 5.3.3 Sampling and analysis program 132

5.3.4 Sample collections 133 5.3.4.1 Water sampling 133 5.3.4.2 Soil sampling 133 5.4 Soil Characterisations 136 5.4.1 Soil texture (hydrometer test) 136 5.4.1.1 Apparatus and materials 136 5.4.1.2 Reagents 136 5.4.1.3 Method 136 5.4.1.4 Calculations 137 5.4.2 Soil Taxonomy (ST) and World Reference Base (WRB) 138 system 5.4.3 Moisture content 143 5.4.4 pH 143 5.4.5 Electrical conductivity (EC) 144 5.4.6 Organic carbon (Walkely and Black 1934) 144 5.4.7 Cation exchange capacity, trace elements (aluminium and 145 iron) and phosphorus (Colwell) 5.4.8 Metal contents 145 5.5 Quality Assurance (QA) 146 5.5.1 Identification of QC samples 146 5.5.1.1 Field QC samples 146 5.5.1.2 Laboratory QC samples 148 5.5.2 Acceptable criteria for QC samples 149 5.6 Glyphosate Levels in Collected Samples 154 5.7 Discussions 157 5.8 Conclusions 158

Chapter 6 Glyphosate Sorption onto GM Canola 160-189 Cultivated Soils 6.1 General 160 6.2 Introduction and Background 161 6.3 Glyphosate Sorption Experiments 163

XII

6.3.1 Materials and methods 163 6.3.1.1 Reagents 163 6.3.1.2 Solutions 163 6.3.2 Soil properties 164 6.3.3 Procedure 164 6.4 Soil Sorption Isotherms 166 6.4.1 Langmuir isotherm 166 6.4.2 Freundlich isotherm 167 6.4.3 Redlich-Peterson isotherm 168 6.5 Results 169 6.5.1 Linear isotherm 169 6.5.2 Non-linear isotherms 179 6.6 Glyphosate Desorption 186 6.7 Conclusions 188

Chapter 7 Glyphosate Leaching Experiments and 190-220 One-dimensional Transport Model 7.1 General 190 7.2 Background 192 7.3 Materials and Methods 193 7.3.1 Reagents 193 7.3.2 Solutions 193 7.3.3 Soil materials 194 7.4 Column-leaching Experiments 194 7.4.1 Experimental set-up 194 7.4.2 Glyphosate applications 196 7.5 Column Experiments Parameters 197 7.5.1 Bromide transport 198 7.5.2 Glyphosate transport 198 7.5.3 Results 199 7.5.3.1 Bromide breakthrough curves 199 7.5.3.2 Glyphosate breakthrough curves 202

XIII

7.5.3.3 Mass recovery 206 7.5.4 Estimated parameters 208 7.6 Conceptual Site Model 212 7.6.1 Discussion on different processes 215 7.6.1.1 Volatilisation 215 7.6.1.2 Sorption and desorption 216 7.6.1.3 Photolysis 216 7.6.1.4 Degradation 217 7.6.1.5 Transport 218 7.7 Conclusions 218

Chapter 8 Conclusions and Future Recommendations 221-225 8.1 General 221 8.2 Major Conclusions of Thesis 222 8.3 Recommendations for Future Research 224

References 226-251

Appendix A: Quotes from Commercial Laboratory 252

XIV

List of Figures

Chapter 1 Introduction Page Figure Figure title 1.1 Flow diagram of dissertation 12

Chapter 2 Review of Biofuels Figure Figure title 2.1 World annual ethanol and biodiesel production 1975–2009 13 2.2 Overview of routes for conversion of crops to biofuels 17 2.3 Volume requirements of Renewable Fuel Standard (RFS2) 20 2.4 WA average yields of various cropping systems (2000-2007) 40 2.5 GM roundup ready weed management (Wimmera, Victoria) 41

Chapter 3 Review of Herbicide Glyphosate Figure Figure title 3.1 Glyphosate degradation pathway 53 3.2 Distributions of glyphosate species as function of pH (Bjerrum 56 diagram) 3.3 Tentative reaction schemes for sorption of glyphosate and 58 phosphate

Chapter 4 Establishment of Analytical Method for Glyphosate Detection Figure Figure title 4.1 Derivatisation reaction scheme for fluorometric analytical 72 method 4.2 Fluorescence Energy level diagram(Jablonski diagram) 79 4.3 Excitation acquisition of the derivatised glyphosate 84 4.4 Emission acquisition of derivatised glyphosate for water samples 85 at different concentrations

4.5 Standard calibration curves for fluorescence spectrometry of 85 glyphosate in waters (all waters) 4.6 Standard calibration curves for fluorescence spectrometry of 86 glyphosate in waters (all waters up to 1.0 µg/mL) 4.7 Standard calibration curves for fluorescence spectrometry of 86 glyphosate in waters (log-log scale) 4.8 ELISA plate’s 96 test wells 99 4.9 ELISA standard plot 104 4.10 UV-visible Spectrophotometric standard plot 107 4.11 Comparison of ELISA and fluorescence spectrometric results 110 4.12 Comparison of UV-visible spectrophotometric and fluorescence 110 spectrometric results for soil samples

Chapter 5 Site Investigation Figure Figure title 5.1 Location of the site investigation 126 5.2 Glyphosate-contaminated sampling points on Parkes plains 127 5.3 Soil sampling locations 128 5.4 Soil sampling locations 129 5.5 Location of surface water sampling from creek 130 5.6 Precipitations, sowing and herbicide applications in fields (G1 to 131 G5) and sampling times (Episodes 1 to 5) 5.7 Details of soil sampling: (a) point near canola crop; (b) collecting 134 surface soil; (c) sampling up to 30 cm 5.8 Glyphosate concentrations in water samples determined by 154 ELISA method 5.9 Glyphosate concentrations in soil samples determined by ELISA 154 method 5.10 Glyphosate concentrations in water samples determined by 155 fluorescence spectrometric method (Model I)

5.11 Glyphosate concentrations in water samples determined by 155 fluorescence spectrometric method (Model III)

XVI

5.12 Glyphosate concentrations in soil samples determined by 156 fluorescence spectrometric method (Model I) 5.13 Glyphosate concentrations in soil samples determined by 156 fluorescence spectrometric method (Model III)

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils Figure Figure title 6.1 Langmuir linearised isotherm models for glyphosate sorption by 174 three different soils (fluorescence spectrometric, Model I) 6.2 Langmuir linearised isotherm models for glyphosate sorption by 175 three different soils (fluorescence spectrometric, Model III) 6.3 Langmuir linearised isotherm models for glyphosate sorption by 176 three different soils (UV-visible spectrophotometric) 6.4 Non-linear isotherm models for glyphosate sorption by three 182 different soils (Fluorescence spectrometric Model I) 6.5 Non-linear isotherm models for glyphosate sorption by three 183 different soils (Fluorescence spectrometric Model III) 6.6 Non-linear isotherm models for glyphosate sorption by three 184 different soils (UV-visible spectrophotometric)

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model Figure Figure title 7.1 Column-leaching experimental set-up 195 7.2 Bromide concentrations in leachates obtained from column 200 experiments 7.3 Bromide BTCs fitted with experimental data 201 7.4 Glyphosate concentrations in leachates under different 204 application scenarios 7.5 pH and conductivity levels in effluent samples 205 7.6 Percent recovery of glyphosate from the three soils columns 207

XVII

7.7 Glyphosate BTCs fitted with equilibrium model 210 7.8 Glyphosate BTCs fitted with one-site sorption model 211 7.9 Conceptual site model of glyphosate fate and transport in the 214 agricultural system

XVIII

List of Tables

Chapter 2 Review of Biofuels Page Table Table title 2.1 World’s fuel ethanol production (2008–2007) 14 2.2 World’s biodiesel production (2008–2007) 14 2.3 First- and second- generation biofuels 18 2.4 Volumes used to determine proposed 2014 percentage standards 21

Chapter 3 Review on Herbicide Glyphosate Table Table title 3.1 Physical and chemical properties of glyphosate 49 3.2 Toxicity classification - glyphosate 50 3.3 Glyphosate sorption by different pure minerals 56

Chapter 4 Establishment of Analytical Method for Glyphosate Detection Table Table title 4.1 Back calculated concentrations of the calibration standards and 91 calculated precision and accuracy of the fluorescence spectrometric method (water samples using Model I) 4.2 Back calculated concentrations of the calibration standards and 93 calculated precision and accuracy of the fluorescence spectrometric method (water samples using Model II) 4.3 Back calculated concentrations of the calibration standards and 95 calculated precision and accuracy of the fluorescence spectrometric method (water samples using Model III). 4.4 Back calculated concentrations of the calibration standards and 96 calculated precision and accuracy of the fluorescence spectrometric method (water samples using Model IV). 4.5 Method validation parameters 97 4.6 Comparison of different glyphosate detection methods for water 108

XIX

4.7 Comparison on different glyphosate detection method for soil 109 4.8 Cost analysis of the new analytical method (Commercial 113 laboratory)

Chapter 5 Site Investigation Table Table title 5.1 Identification of samples taken from study area 135 5.2 Summary of the Australian Soil Classification (ASC) orders 139 5.3 Rationalised key to the WRB reference soil groups 140 5.4 The orders of soil taxonomy 141 5.5 Australian soil classification and international equivalents 142 5.6 Major soil properties and compositions 145 5.7 QC samples’ acceptable criteria 150 5.8 Results of blind replicates in water samples for different episodes 151 5.9 Results of blind replicates in soil samples for different episodes 151 5.10 Results of rinsates 152 5.11 Results of laboratory duplicates in water samples for different 152 episodes 5.12 Results of laboratory duplicates in soil samples for different 153 episodes

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils Table Table title 6.1 Properties of three different soils from GM canola paddocks 164 6.2 Different linearised forms of isotherm equations 167 6.3 Linearised isotherm parameters for three different soils 171 (fluorescence spectrometric, Model I) 6.4 Linearised isotherm parameters for three different soils 172 (fluorescence spectrometric, Model III) 6.5 Linearised isotherm parameters for three different soils (UV- 173 visible spectrophotometric)

6.6 Non-linearised isotherm parameters from fluorescence 179 spectrometric analysis, Model I) 6.7 Non-linearised isotherm parameters from fluorescence 180 spectrometric analysis, Model III) 6.8 Non-linearised isotherm parameters from UV-visible 181 spectrophotometric analysis 6.9 Correlation coefficient, APE and Chi-square values for 185 linearised and non-linearised isotherms (Fluorescence spectrometric, Model I) 6.10 Correlation coefficient, APE and Chi-square values for 185 linearised and non-linearised isotherms (Fluorescence spectrometric, Model III) 6.11 Correlation coefficient, APE and Chi-square values for 186 linearised and non-linearised isotherms (UV-visible spectrophotometric) 6.12 Desorption of glyphosate from different soils, calculated using 187 Model I 6.13 Desorption of glyphosate from different soils, calculated using 187 Model III

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model Table Table title 7.1 Physical properties of soil columns 202 7.2 Fitting parameters for glyphosate BTCs using equilibrium model 209 7.3 Fitting parameters for glyphosate BTCs using one-site sorption 209 model

XXI

Associated Publications

1. Nahar K, Niven RK, Pearson S, Badek K, Riesen H, Rahman AFM. A simple and low cost direct fluorometric method to quantify the herbicide glyphosate in soils and/or water. (Ready to submit in Environmental Science & Technology).

2. Nahar K, Niven RK, Pearson S, Badek K, Riesen H and Rahman AFM. Sorption and leaching of glyphosate on the agricultural soils from genetically modified canola cultivation in Parkes, NSW, Australia. (Ready to submit in the Journal of Soil Science)

3. Nahar K, and Niven RK, “Glyphosate and its environmental impact – A review on current trend”. (Draft ready for Journal of Environmental Quality).

4. Nahar K, and Niven RK, “Investigation of glyphosate in environmental samples from genetically modified canola farmland in Parks, NSW, Australia” AusBiotech National Conference, Gold Coast, 29-31 Oct 2014.

5. Nahar K, and Niven RK, “Glyphosate contaminated site investigation using a low cost analytical technique” Australia & New Zealand Student & Young Professional Congress 2014 (ANZSCON 2014), Queensland, 3-5 July 2014.

6. Nahar K, Niven RK, Pearson S, Badek K, Riesen H and Rahman AFM. “Glyphosate in waters and soils from genetically modified canola cultivation in Parkes, NSW, Australia” Society for Engineering in Agriculture (SEAg), Perth, Australia; 22-25 Sept 2013; (Full paper in USB).

7. Nahar K, Niven, RK, Pearson S, Badek, K, Riesen H, and Rahman AFM. “Sorption and desorption of glyphosate on Australian soils cultivated with genetically modified canola” Sorption Reversibility session, Environment Division, 245th ACS National Meeting & Exposition, New Orleans, Louisiana, April 7-11, 2013.

XXII

Notations

Notation Description

αk First-order sorption ε Molar absorptivity θ Volumetric water content ρ Soil’s bulk density μ First-order dissipation λ Dispersivity

퐶푒푞 Equilibrium concentration

퐶푠 Sorbed Concentration D Dispersion Coefficient I The transmitted intensity

I0 The intensity of the incident light at a given wavelength jw Irrigation rate

퐾푓 Freundlich sorption coefficients

Krp Redlich-Peterson isotherm constant

푞푚 Maximum sorption capacity sk Sorbed concentration of the kinetic sorption sites k se Sorbed concentration of the liquid phase t1/2 Half-life v Pore-water velocity

XXIII

Abbreviations

Abbreviation Elaboration AMPA Aminomethylphosphonic acid APE Average Percentage Errors BNP National Biofuel Policy BTC Breakthrough Curve ADE Advection-Dispersion Equation ELISA Enzyme-linked immunosorbent assay EPA Environmental Protection Agency EPSP Enzyme 5-Enolpyruvylshikimate-3-phosphate ETBE Ethyl-Tertiary-Butyl-Ether EU European Union FAME Fatty Acid Methyl Esters FAO Food and Agricultural Organisation FMOC-chloride 9-fluorenylmethyl chloroformate GBR Great Barrier Reef GC Gas Chromatography GHG Green House Gas GM Genetically Modified GMHT Genetically Modified Herbicide-Tolerant (GMHT) HPLC High-Performance Liquid Chromatography ICP-MS Inductively Coupled Plasma-Mass Spectrometry IPAD Integrated Pulsed Amperometric Detection IPCC Inter-Governmental Panel on Climate Change ITA Incentive Tax Allowance LC Liquid Chromatography LC/MS Liquid Chromatography Mass Spectrometry LCA Life-Cycle Assessment LCS Laboratory Control Samples LD Lethal Dose LOD Limit of detection LOQ Limit of Quantification

XXIV

LOR Lower Concentration of Reliability MDL Method Detection Limit MTBSTFA N-methyl-N-(tert-butyldimethylsilyl) trifluoroacetamide NBD-F 4-Fluoro-7-nitrobenzofurazan NEPM National Environmental Protection Measure NMI National Measurement Institute NRDC The National and Development Reform Commission (NDRC) OP Open Pollinated OPA Ophthaldialdehyde PME Palm Methyl-Esters PNPB Program on Biodiesel Production and Usage POEA Polyoxyethyleneamine QA Quality Assurance QC Quality Control RFS Renewable Fuel Standard RPD Relative Percentage Difference RR Roundup-Ready SD Standard Deviation SOM Soil Organic Matter SSA Specific Surface Area TBDMCS Tertbutyldimethyl Chlorosilane t-BDMS t-Butyldimethylsilyl TFAA Trifluoracethicanhidride TFE Trifluorethanol TLC Thin Layer Chromatography TMS Trimethylsilyl TT Triazine-Tolerant UPLC Ultra Performance Liquid Chromatography USDA United States Department of Agriculture VEETC Volumetric Ethanol Exercise Tax Credit (VEETC)

XXV

Chapter 1 Introduction

1.1. General

As the global production of biofuels, particularly crop-based ones, increases [1-4], the environmental consequences of land being converted to irrigated agriculture [5] and the associated increase in herbicide use have emerged as prominent issues in the scientific community. As suggested in the literature, genetically modified (GM) crops can be used for biofuel production [6, 7], although there has been considerable controversy concerning the possible impacts of their widespread cultivation on the environment. Australia, as a major agricultural exporter, with around two-thirds of its total production exported [8], has cultivated new varieties of GM crops since their introduction in 2000 [9, 10]. These crops are usually produced by the insertion of genes from soil bacterium into crops to make them tolerant to one of the most commonly used herbicides, glyphosate (N-(phosphonomethyl) glycine). This has meant that due to the increase in GM crop agriculture during the past few years, the use of glyphosate has also increased, with studies finding that in Australia, approximately 15, 000 tonnes of glyphosate are applied annually to control agricultural, urban and roadside weeds

[11, 12]. However, due to the high cost of analysing glyphosate in environmental samples, there is a lack of information on its effects both in Australia and overseas

[13].

Chapter 1 Introduction

The research for this thesis investigates glyphosate and its fate and transport in the agricultural system. In this regard at first, a low-cost analytical method (direct fluorescence spectrometry) was established for determining glyphosate in waters and soils was established to avoid the need for much more expensive analytical methods such as high-performance liquid chromatography (HPLC). This was calibrated against, and used in conjunction with, the commercially available enzyme-linked immunosorbent assay (ELISA) method, for the analysis of glyphosate in soils and waters. A site investigation was then conducted of farmland cultivated with GM canola in the Parkes region of New South Wales, Australia, which involved the sampling of surface waters and soils and analysis of glyphosate concentrations. Glyphosate sorption isotherms were then constructed by batch tests on several soils from the study site and fitted with linear and non-linear Langmuir, Freundlich and Redlich-Peterson isotherms, and showed good agreement with the Freundlich isotherm. Desorption tests performed using a 0.1M KH2PO4 solution on the same soils. Column-leaching experiments were also conducted of glyphosate-dosed soils using application and flow rates representative of field conditions with bromide as a non-reactive tracer. The results of the glyphosate sorption and leaching experiments were then incorporated in a one-dimensional transport model (HYDRUS 1D) to determine if the observed behaviour of glyphosate was in agreement with commonly recognised transport processes. A fully kinetic, one-site convective-dispersive model with degradation and

Freundlich sorption and an equilibrium model were considered in transport experiments. Later, the findings are incorporated into a conceptual model of glyphosate fate and transport in an agricultural soil system, including an overview of its transport pathways and environmental receptors.

Chapter 1 Introduction

1.2. Background and Problem Statement

As previously discussed, the worldwide increase in crop-based biofuel production has led to an increasing concern regarding the use of herbicides as the amount of cropland increases, particularly land converted to agriculture [5]. In Australia, biofuel production increased between 2009–10 and 2010–11, continuing the trend of recent years [1] and reaching a total of 419 mega litres (ML) in 2010–11, an increase from

354 ML in 2009–10, while ethanol and biodiesel production increased from 269 ML to 319 ML and from 85 ML to 100 ML respectively [2]. In Canada, the federal government projected that ethanol production would increase from 42 million litres in

2006 to approximately 799 million litres by 2010 [3]. In the United States of America

(USA), the target production of 132billion litres of renewable fuel by 2017 has led to an increase in ethanol production [4] and, in 2006, ethanol accounted for 3.5% of total

U.S. fuel consumption. About 98% of all U.S. ethanol production is corn-based which has gradually been accomplished by converting existing cropland to corn and planting it in consecutive growing seasons rather than following a different crop rotation plan.

Furthermore, farmers now rely on different insecticides and GM crop varieties, closely related to Roundup-ready corn, to protect their crops against large-scale failures.

Although there has been considerable controversy concerning the possible impacts of

GM crop cultivation on the environment, previous literature has recommended that

GM crops be used as new options for biofuel production [6, 7]. In Australia, new varieties of GM crops account for an increasing amount of total agricultural land use and the Australian Government plans to increase their cultivation [10]. These crops are usually produced to make plants tolerant to glyphosate, which is one of the most

Chapter 1 Introduction

widely used herbicides in Australia, with approximately 15, 000 tonnes applied annually to control agricultural, urban and roadside weeds [11, 12]. Studies to detect glyphosate in a diversity of water bodies have been conducted; for instance, one in the mid-western U.S. involving extensive sampling of aquatic systems found that 36% and

69% of samples contained detectable concentrations of glyphosate and AMPA respectively [14, 15]. An Australian field study conducted on irrigated sugarcane production reported 54 µg/L as the peak glyphosate concentration in samples [16] while a similar study in Canada found glyphosate concentrations as high as 40.8 µg/L

[17]although field dissipation studies reported 1700 µg/L [18, 19]. To date, several farm-level case studies have been conducted worldwide [20-22] and, recently, a few in different parts of Australia [16, 23] but none in the Parkes region of NSW.

The toxicities of glyphosate and its formulations has led to world environmental organisations setting safe limits for its use in order to protect human life and marine and aquatic systems [13, 24]; for example, an Australian health-based guideline of

1000 µg/L of glyphosate in drinking water was first scheduled in the 6th edition of the

Australian Drinking Water Guidelines [19]. Also, the Australian and New Zealand

Environment and Conservation Council (ANZECC) reports statistical percentile trigger values for glyphosate in freshwaters of 99% = 370µg/L, 95% = 1200µg/L, 90%

= 2000µg/L and 80% = 3600 µg/L [25]. However, it does not provide trigger values for glyphosate in marine waters. The World Health Organization calculated a human health-based criterion of 900 µg/L [26] although it did not establish a formal guideline

[27]. The U.S. Environmental Protection Agency lists a Maximum Contaminant Level of 700 µg/L in its National Drinking Water Standards [28], with a Reference Dose of

Chapter 1 Introduction

2000 µg/L (the amount a person can consume each day over a lifetime without incurring an ‘appreciable risk’ of negative effects) [29]. The Canadian Government established a health-based Maximum Acceptable Concentration of 280µg/L [30], which is equal to the Livestock Water Guideline in the Canadian Water Quality

Guidelines for the Protection of Agricultural Water Uses [31], as well as a Fresh Water

Guideline of 65 µg/L in the Canadian Water Quality Guidelines for the Protection of

Aquatic Life [32]. Several studies in South Africa have detected glyphosate in rural water sources [33, 34] and a recent study formulated a chronic guideline for glyphosate in water of 0.3µg/L [35]. South African legislation has recognised the pollution of environmental water by glyphosate to be an important public health problem [36-38].

The costs of analysing glyphosate in environmental samples can be very expensive depending on the analytical method used; for example, those quoted for laboratory analyses in Australia are of the order of $150-$280 for each water sample and $200-

$280 for each soil sample (see Appendix A and discussion in chapter 4). Such costs impose major constraints on the inclusion of glyphosate in regular monitoring studies, especially in the university and agricultural sectors, leading to a lack of information on glyphosate in global environmental datasets [13, 39]. However, over the last few years, the development of analytical techniques for glyphosate quantification has increased and several alternative methods have been proposed [22, 40-43]. Most reported in the literature for quantifications of glyphosate in water, soils, fruits, crops, vegetables and other samples are based on chromatographic separation and determination [44]. Depending on the detection technique, glyphosate determination usually needs a derivatisation step; for example, analysis by gas chromatography is

Chapter 1 Introduction

performed after derivatisation of glyphosate into volatile and thermally stable derivatives [45, 46]. UV–visible and fluorescence detection methods are normally used for glyphosate derivatives in the liquid chromatographic method [47-49] and, for glyphosate quantification by LC-MS, a quite complicated derivatisation step has to be conducted to improve the chromatographic performance [50, 51]. Also, the sorption

14 of glyphosate on soils using C-labelled glyphosate can be performed in laboratory experiments [52] which allow the quantification of glyphosate at low concentrations but require special, expensive equipment to detect 14C-labelled molecules. Therefore, a simple, fast and low-cost method is required for analyses of glyphosate in water and soil samples, which can be used in field monitoring and also be suitable for glyphosate sorption and transport studies on agricultural soils, is required.

This research explores a new method for glyphosate analysis, as well as glyphosate’s fate and transport in a GM canola agricultural system. Firstly, a low-cost analytical method (direct fluorometric spectrometry) is established for the determination of glyphosate in soils and waters. This is calibrated against, and used in conjunction with, a commercially available enzyme-linked immunosorbent assay (ELISA) method for the analysis of glyphosate in soils and waters. It involves derivatisation using 9- fluorenylmethyl chloroformate (FMOC-chloride) and measurement of the emission acquisitions at 310 nm. This proposed analytical method exhibits a linear calibration curves for ranges from 10 µg/L to 25000 µg/L with good reproducibility (recovery

108.3±20.8%; R2=0.997 for waters and aqueous extracts), and also over the lower range 10 µg/L to 1000 µg/L (recovery 115.2±25.9%; R2=0.998). The appropriate limits of detection for waters and soils were 10 µg/L and 0.063 mg/kg respectively and

Chapter 1 Introduction

limits of quantification were 29 µg/L and 0.194 mg/kg respectively. Secondly, a site investigation was conducted of farmland cultivated with GM canola in the Parkes region of New South Wales, Australia, which involved the sampling of surface waters and soils, and the analysis of glyphosate concentrations. The sampled soils were classified as loams under the USDA system (clay 13.8-15.8%, silt 39-43%, sand 41.2-

47.2%), and the soil and water samples were collected according to the Australian

National Environmental Protection Measure (NEPM) sampling protocol[53]. The determined glyphosate levels ranged between 0.01 and0.067 µg/mL in waters and 0.10 and 0.575 mg/kg in soils, with aqueous levels lying below Australian and international drinking water guidelines although some exceeded the Canadian freshwater guideline of 65 µg/L and most exceeded the South African trigger value of 0.3 µg/L. Thirdly, glyphosate sorption isotherms were constructed by batch tests on several soils from the study site and fitted with linear and non-linear Langmuir, Freundlich and Redlich-

Peterson isotherms. Desorption tests performed using a 0.1M KH2PO4 solution on the same soils. Also, column-leaching experiments on glyphosate-dosed soils, using application and flow rates representative of field conditions with bromide as a non- reactive tracer. These were modelled using the one-dimensional transport model

HYDRUS-1D. At first, the physical transport parameters, pore water velocity (ν) and dispersivity (λ) were determined by fitting the experimental bromide breakthrough curves (BTCs) with the analytical solution to the advection-dispersion equation with a pulse boundary condition at the upper and a zero gradient condition at the lower boundaries. Then, these parameters and those parameters from the sorption experiments were used in HYDRUS 1D to describe the transport behaviour of glyphosate. Finally, a conceptual model of glyphosate fate and transport in an

Chapter 1 Introduction

agricultural system was constructed with an overview of glyphosate transport pathways and environmental receptors.

1.3. Objectives of Research

 To establish a low-cost analytical method (direct fluorescence spectrometry)

for determining glyphosate in waters and/or soils to avoid the need to use

expensive analytical methods such as High-Performance Liquid

Chromatography (HPLC).

 To conduct a site investigation of farmland in the Parkes region of New South

Wales, Australia, cultivated with genetically modified (GM) canola to

determine glyphosate concentrations in surface waters and soils.

 To study the glyphosate sorption and desorption characteristics of soils from

the case study site and compare the results with those from established

analytical method [54].

 To conduct column-leaching experiments on glyphosate-dosed soils using

application and flow rates representative of field conditions with bromide as a

non-reactive tracer.

 To incorporate the sorption and transport of glyphosate in a one-dimensional

transport model (HYDRUS 1D) to determine if the observed behaviour of

glyphosate was in agreement with commonly recognised transport processes.

 Finally, to construct a conceptual model of glyphosate fate and transport in an

agricultural system.

Chapter 1 Introduction

1.4. Scope of Study

This study conducted a glyphosate analysis using the new, simple, fast and low-cost established method and its fate and transport in the GM canola agricultural system, focusing on:

(a) reviewing biofuels and their relevant worldwide policies, cultivation and

potential environmental impacts on agricultural land use;

(b) surveying the literature on the main herbicide, glyphosate, used in the

cultivation of genetically modified (GM) crops in Australian agriculture;

main herbicide, glyphosate;

(d) examining the problems associated with analysing glyphosate samples using

high-cost analytical methods;

(e) developing a low-cost analytical method (direct fluorometric spectrometry) for

the determination of glyphosate in waters and soils;

(f) conducting a site investigation on farmland in the Parkes region of New South

Wales, Australia, cultivated with GM canola.

(g) determining the glyphosate sorption and desorption characteristics of soils

from the case study site;

(h) conducting column-leaching experiments with glyphosate-dosed soils using

application and flow rates representative of field conditions with bromide as a

non-reactive tracer; and

(i) Incorporating results from sorption and leaching experiments to an one

dimensional transport model (HYDRUS 1D).

Chapter 1 Introduction

1.5. Organisation of Thesis

This dissertation consists of eight chapters which describe the different investigations conducted in this study.

Chapter 1 presents a general introduction to the current research, including its problem statement, and objectives and scope, and thesis organisation.

Chapter 2 reviews the general understanding of biofuels, in particular, their resurgent popularity due to recent rises in oil prices and growing concerns about the environmental impacts on soils and waters of widespread biofuel crop cultivation. This chapter also provide information on genetically modified crops as alternatives for biofuel production, as suggested by some previous literature and the impact on the environment if their large-scale cultivation was undertaken.

Chapter 3 provides a survey of the literature on the major herbicide, glyphosate, which is used for GM crops in Australian agriculture. It includes information on glyphosate’s formulation, sorption, degradation, toxicological impact and potential threat to both human life and marine and aquatic systems.

Chapter 4 presents a simple, fast and low-cost direct fluorescence spectrometric method which covers the range of glyphosate concentrations well below the maximum acceptable concentrations determined by different organisations, e.g., the Australian

Drinking Water Guidelines (1µg/mL) [19, 55]. This chapter describes all the procedure of preparing samples, standards, development of the standard calibration curves,

Chapter 1 Introduction

laboratory Quality Control (QC) and analytical method validation with the calculation of the acceptable operating limits (limits of detection and limits of quantification). It also provides a comparative study with previously established direct low cost enzyme- linked immunosorbent assay (ELISA) method, UV-visible spectrophotometric method and HPLC method

Chapter 5 reports on a site investigation of farmland in the Parkes region of New South

Wales, Australia, cultivated with GM canola. This incorporated firstly, a survey on

GM canola growers from different parts of NSW and then a site investigation on a family-operated farm in the agricultural district of Parkes in the central west of NSW.

All the information regarding sample analysis program and quality assurance (QA) assessments, including a detailed analysis of quality control (QC), are also provided.

Chapter 6 includes studies of glyphosate sorption and desorption on three different soils from the case study site, with their results compared with those from an established UV-visible spectrophotometric method [54]. It includes information regarding the soil sorption isotherm models, application of different linear and non- linear isotherm model to the experimental data and their results compared and discussed, with the perfect isotherm match explaining the data well.

Chapter 7 incorporates column-leaching experiments of glyphosate-dosed soils using application and flow rates representative of field conditions with bromide as a non- reactive tracer. The first section discusses the experiments, experimental set-up and different glyphosate application scenarios. Then, the sorption and transport are

Chapter 1 Introduction

incorporated in a one-dimensional transport model (HYDRUS 1D) to determine if the observed behaviour of glyphosate is in agreement with commonly regarded transport processes. In the later section a conceptual site model were also discussed with a flow diagram of glyphosate application and its fate and transport pathways

Chapter 8 summarises the main findings of this study, presents its conclusions and provides recommendations for future research.

The flow diagram of the thesis is shown in Fig. 1.1.

Fig. 1.1: Flow diagram of dissertation

Chapter 2 Review of Biofuels

2.1. General

Driven by the agenda to reduce the world’s dependency on fossil fuels and greenhouse gas emissions, the implementation of biofuel policies has fostered the expansion of the biofuel industry, with biofuel production increasing dramatically over the last few years with those of fuel ethanol and biodiesel increasing from 16.9 to 72 billion litres and from 0.8 to 14.7 billion litres respectively between 2000 and 2009 [56] (Fig. 2.1).

As the world’s largest fuel ethanol producer, the United States of America (USA) has provided strong financial incentives for biofuel manufacturers. Also, in the European

Union, the world’s largest biodiesel producer, biofuel consumption is mostly motivated by a combination of mandates from France and Germany (Table 2.1 and

Table 2.2).

Fig. 2.1: World’s annual ethanol and biodiesel production 1975–2009* [56]

(Note: original data in US gallons changed to litres in 2010 [57]; 2009 values projections; available from www.earthpolicy.org)

Chapter 2 Review of Biofuels

Table 2.1: World’s fuel ethanol production (2008–2007) in million litres [58]

Country 2008 2007 USA 33, 737 24, 360 Brazil 24, 261 18, 815 EU 27 2748 2138 China 1882 1822 Canada 892 830 Thailand 337 297 Colombia 296 281 India 247 198 Australia 97 99 Other 480 311 World 64, 981 49, 112

(Note: original version in US gallons converted to litres in 2010 [57])

Table 2.2: World’s biodiesel production (2008–2007) in million litres [59]

Country 2008 2007 EU 27 9164 7377 US 3078 2733 Argentina 1550 522 Brazil 1238 457 Australia 1051 524 Malaysia 609 240 Indonesia 405 327 India 227 114 Canada 114 99 Other 1036 895 World 18, 472 13, 060

(Note: original version in tonnes converted in litres in 2010 [57])

Chapter 2 Review of Biofuels

Most commercially available biofuels are from food crops, such as sugar cane and sugar-beet, corn and oily seeds [60], which have been criticised in different studies as being prime reasons for increases in food prices [61-65]. According to data provided by the Food and Agricultural Organisation of the United Nations (FAO), the indices for cereal grains, oils and sugar prices were 70%, 69% and 276% higher respectively in 2010 than the 2002–2004 average values [66] despite substantial drops in 2009.

Also, a life-cycle assessment (LCA) conducted on specific types of feedstock biofuel crops and processing techniques reported a negative net contribution as well as criticism of the increased linkage between the food and energy sectors [67-70].

Although changes were seen in some cases, corn-based ethanol production proved to be the most harmful LCA profile [71].

This chapter presents information on two aspects of biofuels: the first provides a general understanding of them and their classifications, and an overview of the current output targets and support schemes around the world; and the second reviews genetically modified (GM) crops as alternatives for biofuel production, as suggested by previous studies [6, 7] and the impact on the environment if their large-scale cultivation was undertaken.

2.2. Biofuels and Their Classification

For modern biofuel production, the quickest, most efficient and often the only, way to convert plants to biofuel feedstocks is biotechnologically. Biofuels are fuels produced from renewable organic sources and can be derived from sources such as wheat, sugar, tallow, oil crops, waste vegetable oils, woody (cellulose) parts of plants and wood

Chapter 2 Review of Biofuels

wastes [72, 73]. In the past, the term ‘biofuels’ has usually meant ethanol and biodiesel made from crops, including canola, corn, sugarcane and rapeseed. But later, second- generation ethanol and biodiesel production has paid greater attention to reducing the impact of biofuel production on food commodities and improving its balance of greenhouse gas (GHG) emissions.

Bioethanol, an alcohol, is usually mixed with petrol while biodiesel is used either on its own or in a mixture. Pioneers in the biofuel field, such as Henry Ford and Rudolph

Diesel, designed cars and engines to run on biofuels and, before World War II, the UK and Germany both sold biofuels mixed with petrol or diesel made from crude oil but the later availability of cheap oil ensured its market dominance [74]. In either liquid form, such as bioethanol and biodiesel, or gaseous form, such as biogas and hydrogen, biofuels are simply transportation fuels derived from biological sources such as the following.

 Cereals, grains, sugar crops and other starches are fermented to produce

ethanol which is used as either a motor fuel in pure (‘neat’) form or as a

blending component in gasoline (as ethanol or after being converted to ethyl-

tertiary-butyl-ether- (ETBE)) [75].

 Cellulosic materials, including grasses, trees and various waste products from

crops, wood-processing facilities and municipal solid waste, can also be

converted to alcohol although their processing is more complex than that of

sugars and grains

Chapter 2 Review of Biofuels

 Oil-seed crops (e.g., rapeseed, soybean and sunflower) can be converted into

methyl esters, a liquid fuel which can be either blended with conventional

diesel fuel or burnt as pure biodiesel [75].

 Organic waste material can be converted into energy forms which can be used

as automotive fuel, for example, waste oil (e.g., cooking oil) into biodiesel;

animal manure and organic household wastes into biogas (e.g., methane); and

agricultural and forestry waste products into ethanol [76].

The routes for biofuel production can be categorised as: extraction of vegetable oils, fermentation of sugars to alcohol, gasification and chemical synthesis, and direct liquefaction. Many of their resultant fuels are methanol, ethanol, hydrogen, synthetic diesel, biodiesel and bio oil (Fig. 2.2).

Fig. 2.2: Overview of routes for conversion of crops to biofuels [77, 78]

Chapter 2 Review of Biofuels

Biofuel production is classified into two groups: first-generation biofuel feedstocks from sources such as sugarcane and cereal grains which produce bioethanol and biobutanol and oilseeds which produce biodiesel; and second-generation bioethanolic/biobutanolic biofuels that come from cultivated lignocellulosic crops or straw wastes. Mounting concerns regarding the large-scale cultivation of first- generation biofuel crops and increasing food prices led biotechnologists to discover second–generation ones. Table 2.3 list first- and second-generation biofuels.

Table 2.3: First- and second- generation biofuels First-generation biofuels Second-generation biofuels (from seeds, grains or sugars) (from lignocellulosic biomasses, such as crop residues, woody crops and energy grasses) Petroleum-gasoline substitutes– Biochemically produced petroleum ethanol or biobutanol by fermentation gasoline substitutes – ethanol or of starches (corn, wheat, potato) or butanol by enzymatic hydrolysis. sugars (sugar beets, sugar cane). Petroleum diesel substitutes – biodiesel Thermochemically produced by transesterification of plant oils, also petroleum gasoline substitutes– called fatty acid methyl esters methanol, Fischer-Tropsch gasoline (FAMEs) and fatty acid ethyl esters and mixed alcohols (FAEEs).

From rapeseed (RME), soybean Thermochemically produced (SME), - sunflower, coconut, palm, petroleum-diesel substitutes – jatropha, recycled cooking oil and Fischer-Tropsch diesel, dimethyl animal fats. ether (also a propane substitute) and green diesel From pure plant oils (straight vegetable - oils).

Chapter 2 Review of Biofuels

2.3. Worldwide Trends in Biofuel Production

Attempts have been made by governments around the world to approve legislative instruments that promote the biofuel industry with the aims of reducing oil dependency, increasing the use of renewable energies and contributing to reductions in declining farm incomes. Due to such government interventions and increasing investment in second-generation technologies (e.g., in the USA), concerns have been raised regarding rising food prices which have led some countries to temporarily halt or reduce support programs (e.g., China) while Germany has established a direct link between biofuel consumption and GHG savings. The dangerous linkage between energy needs and food consumption as well as the goal of cost-competitive manufacture remain key issues on the political agendas of biofuel-producing nations.

However, the governments of some countries are considering/have considered decreasing the support which has, to date, sustained development of the biofuel sector.

In spite of the huge financial aid granted for biofuel production (one study suggested that, for the USA alone, financial backing is projected to reach US$16 billion annually by 2014 [79], the current production rate is limited compared with the total consumption of transport fuels and the small numbers of vehicles able to run on blends with high biofuel contents (e.g., E85). An overview of the output targets and support schemes for different biofuels around the world is provided below.

2.3.1. The United States of America (USA)

In the USA, the updated Renewable Fuel Standard (RFS2) went into effect in 2010 after proposals were finalised according to the Energy Independence and Security Act of 2007 [80]. Its target was set at a total of 136 billion litres of renewable fuel

Chapter 2 Review of Biofuels

consumption in transport fuels by 2022, with not less than 80 billion litres from advanced biofuels and 60 billion litres from cellulosic biofuels, as shown in Fig. 2.3.

The projected 2014 volumes of biofuels used to determine the proposed percentage standards are shown in Table 2.4.

Fig. 2.3: Volume requirements of Renewable Fuel Standard (RFS2) [80]

One major reason for the interest in biofuels is their potential to reduce GHG emissions for which the RFS2 stipulates at least 50% reductions from advanced biofuels while only 20% are required from standard biofuels. The criteria also states that all renewable fuels should certify the type of feedstock used and comply with regulations relating to land use. These new regulations (RFS2) have further increased biofuel targets hence concern for the environment gradually increased [57].

Chapter 2 Review of Biofuels

Table 2.4: Volumes used to determine proposed 2014 percentage standards [81]

Category Volumea Range Cellulosic biofuel 64.35 million litres 30.28-113.56 million litres Biomass-based diesel 4.85 billion litres 4.85 billion litres Advanced biofuel 8.33 billion litres 7.57-9.50 billion litres Renewable fuel 57.58 billion litres 56.78-58.75 billion litres a: All volumes ethanol-equivalent except biomass-based diesel which is actual

Note: all volumes converted from US gallons to litres

In terms of financial support schemes, USA biofuel policies have focused mainly on the ethanol industry, with the Energy Tax Act establishing tax credits for ethanol blenders. Ethanol production is dependent mainly on the corn industry and it is estimated that 20% of the USA’s corn supply was employed as ethanol feedstock in

2006 [82]. In the past few years, the other source of financial support for biofuels has been the Volumetric Ethanol Exercise Tax Credit (VEETC) which was endorsed by the American Jobs Creation Act in 2004 and expired in 2011. Another study revealed that the total annual financial support for ethanol production (inclusive of local and state subsidy programs, especially present in the mid-western region) ranged from

US$5.8 to US$7 billion in 2006 and reached US$11billion in 2008 [83]. For biodiesel production, the 2004 American Jobs Creation Act provided tax credits under the

VEECT with subsidies of up to US$1 per gallon for biodiesel produced from virgin oils (the main feedstock) [57].

Chapter 2 Review of Biofuels

2.3.2. Canada

Interest in biofuels, including bioethanol and biodiesel, has increased over the last few years in Canada, with mandates of a 5% renewable content in gasoline by 2010 and a

2% renewable content in diesel fuel and heating oil by 2012. As reported in a previous study, to achieve those targets, 1.9 billion litres of ethanol and 520 million litres of biodiesel were procured by 2012 [84]. Prior to 2009, ethanol production was almost uniquely based on cereal grains, with corn and wheat accounting for 69% and 30% of ethanol output respectively. In 2009, biodiesel was obtained primarily from animal fats, with tallow grease (49%) the leading biodiesel feedstock followed by yellow grease (37%), while its production from canola crops had increased and accounted for

14% of total biodiesel production [85].

Annually, the production of renewable fuels has a positive impact on the Canadian economy of CAN$2 billion [86]. In addition to funding for the federal mandates, several funding schemes have been established by different organisations, e.g. Eco-

ENERGY for Biofuels Program, econ-AGRICULTURE Biofuels, to support research clusters, e.g., the Agricultural Bio-Products Innovation Program and Sustainable

Development Technology Canada to accelerate the commercialisation of innovative agricultural products and services (Agri-Opportunities Program). In this sense, ethanol and biodiesel manufacturers have obtained maximum incentive rates of CAN$0.10 and

CAN$0.20 per litre respectively from 2008 to 2010 through the Eco-ENERGY for

Biofuels Program [57].

Chapter 2 Review of Biofuels

2.3.3. Australia

From the perspective of Australian biofuel production, the Australian government set a non-binding target in 2001 of obtaining 350 million litres by 2010. In addition, state governments set different targets, for example, in 2006, the state of New South Wales

(NSW) targeted a 10% binding share of ethanol in gasoline by 2011 and the state of

Queensland required 5% ethanol content in gasoline by 2011 [87]. As reported,

Australia’s ethanol and biodiesel production capacities in 2007 were 140 million and

323 million litres respectively, with planned future capacities in excess of 1 billion litres for both [87]. However, another study revealed that actual production was much lower and amounted to 83 million litres of ethanol and 77 million litres of biodiesel in the 2006–2007 fiscal year [88].

Compared with other Australian industries, biofuels are highly subsidised despite their low outputs. The most financially supportive scheme is a tax rebate (the Ethanol

Production Grant and Energy Grants — Cleaner Fuels Scheme) that exactly offsets the fuel excise duty of AU$0.38143 per litre for both ethanol and biodiesel until 2011.

On July 1 2011, the Ethanol Production Grant was eliminated, with an alternative set of payments introduced and made via the Energy Grants—Cleaner Fuels Scheme beginning at AU$0.1 per litre. Biodiesel production received similar treatment.

2.3.4. Argentina

According to the Argentine Biofuel Law 26.093 of 2006, a substitution of fossil fuels by 5% biofuels was to begin in 2010. Although, like biofuel manufacturers in the USA who export their products, they do not receive extra tax incentives, prices are set by

Chapter 2 Review of Biofuels

the government and financial incentives reviewed annually [89]. As the world’s third- largest soybean producer and leading exporter of soybean meals and soybean oil,

Argentina’s biofuel production is based on soybean and its processing cluster in

Rosario. Biodiesel production was increased considerably from 2007 with an output increment of 433% in 2007-2008 and an estimated output of 880 million litres in 2009

[90]. However, ethanol production is substantially less developed, producing 45 million litres in 2009, and is linked to the sugar industry.

2.3.5. Brazil

The oil crisis in 1970 encouraged Brazil to construct the most developed and integrated biofuel program in the world. Following the National Alcohol Program begun in 1975, which focused on producing ethanol from sugar cane, it succeeded in commercialising biofuels over the next decade and, in 1985, approximately 96% of automobiles sold in

Brazil were powered by ethanol. However, by the late 1990s, sales of ethanol-powered vehicles plummeted to 1% and the over-valuation of the Brazilian currency (1994–

1999) increased ethanol production costs. To limit these drawbacks, the government tried to implement legislation in 1993 that required a 22% ethanol content to be added to gasoline and, in 2003, this percentage was raised to 25%. From 2003 to 2008, ethanol regained its initial success and, once again, became a cheap and sought-after alternative to oil. The success of the National Alcohol Program reflected the importance of sugar and ethanol production to the Brazilian economy. Previous studies have reported that the sugar and ethanol industries account for 3.6 million jobs and

3.5% of GDP while ethanol production alone consumes 50% of the total supply of sugar cane [91]. Brazil’s ethanol is recognised as the most price-competitive biofuel

Chapter 2 Review of Biofuels

in the world, with its estimated cost ranging between US$0.23 and US$0.29 per litre

[67, 92].

Although Brazil’s government did not provide direct subsidies for ethanol production until 2010 [57], nonetheless, it afforded the ethanol industry more preferential treatment than gasoline producers. Its successful experience regarding bioethanol encouraged Brazil to investigate biodiesel and it initiated The National Program on

Biodiesel Production and Usage (PNPB) in 2005 [93]. Biodiesel industries are based mainly on soybeans with other important vegetable oil plants being castor beans, palm trees and jathropas. However, in comparison with ethanol production, as that of biodiesel is not cost-competitive, it is subsidised.

2.3.6. The European Union

The European Union Directive 2009/28/EC [94] on renewable energy endorsed a minimum binding target of 10% of biofuels to be in transport fuels by 2020, with a minimum 35% reduction in GHG emissions to be achieved by biofuels during their life cycles. Although each country was asked to aim at an indicative 2% share by 2005, only 1% was achieved from that year until 2010. To improve production rates and achieve the targets set by this common agenda, several countries implemented tax reductions or exemptions through the directive 2003/96/EC on Energy Taxation.

Chapter 2 Review of Biofuels

2.3.7. France

The mandatory blending of bioethanol and biodiesel is included in French biofuel policies although the target of 10% biofuel production by 2010 was stopped by the government and, currently, France does not have specific biofuel goals.

2.3.8. Germany

The German government approved a revision of its mandatory biofuel targets for transport to 5.25% in 2009 and 6.25% from 2010 to 2014. Additionally, the minimum biodiesel content in transport diesel was retroactively set at 4.4% for 2009 and kept constant until 2014.

2.3.9. China

The National and Development Reform Commission (NDRC) in China announced medium- and long-term development plans for the share of total primary energy consumption by renewable energy to rise to 15% by 2020. Biofuels were expected to play an important role in achieving its targets, with projected ethanol and biodiesel productions of 10 million tonnes and 2 million tonnes respectively by 2020 [95].

China’s biofuel policies have focused mainly on ethanol production, with biodiesel only marginally promoted by the government regardless of the fact that diesel is

China’s predominant transport fuel. Due to the lack of government support and permission not being granted to new factories employing corn or maize feedstocks

[95], attention has switched to non-grain crops, such as cassava, sweet sorghum and sweet potatoes. However, the limited non-grain feedstock supplies constrain the

Chapter 2 Review of Biofuels

growth of the biofuel sector, with ethanol production forecast to reach 1.7 million tonnes in 2009 [96].

2.3.10. India

A national policy on biofuels in India was approved in September 2008, with a target of obtaining 20% shares of biodiesel and bioethanol blended with mineral diesel and gasoline by 2017. A feature of it was that biodiesel production should use non-edible oilseeds cultivated on waste and marginal land. Although a minimum purchase price

(MPP) for bioethanol and biodiesel was implemented [97], there were no direct financial tax incentives offered for ethanol production.

2.3.11. Indonesia

In Indonesia, mandatory levels of biofuel consumption were introduced in October

2008 with a target of 20% by 2025. The ethanol component of gasoline is projected to increase to 15% by 2025 [98] while the government’s subsidisation of fuel prices was estimated to be more than US$14.5 billion by October 2008 [98].

2.3.12. Malaysia

As Malaysia is the leading producer of palm oil, it has shown increasing interest in biodiesel and, in 2005, launched the National Biofuel Policy (BNP) which has a mandated target of 5% of biodiesel (B5). The types of biodiesel it produces are envodiesel and palm methyl esters (PME) and, in 2007, it exported 95, 000 tonnes of

PME, equivalent to around 75% of its total biodiesel production [99].Biodiesel producers can obtain financial incentives by claiming either the Pioneer Status (PS) in

Chapter 2 Review of Biofuels

the biofuels sector or by applying for an Incentive Tax Allowance (ITA) supported by the Promotion of Investment Acts of 1986 [100].

2.3.13. Thailand

Thailand’s biofuel policy was implemented successfully, with an estimated consumption level of 970 million litres in 2009 [101]. Biodiesel blending was mandatory from 2008 while ethanol consumption was incentivised through tax exemptions that allowed ethanol blends to be significantly cheaper than regular gasoline. Gasohol E10 consumption increased from 4.8 to 9.3 million litres per day between 2007 and 2008, and reached 12.3 million litres per day in 2009 [101]. Due to the government requirement that B100 be blended with standard diesel in order to reach the required biofuel share, the demand for pure biodiesel (B100) increased substantially and the target was 935 million litres by 2011 [102]. Moreover, the Thai government intends to import approximately 200, 000 tonnes of palm oil by 2015 in order to meet its envisaged biodiesel share [102].

2.4. Environmental Impacts of Large-scale Cultivation

2.4.1. Soil

Although the direct environmental impacts on soils of biofuel crops processing are relatively few, with the exception of localised impacts associated with the construction and operation of the biofuel crops processing facilities, indirect impacts may arise from the disposal of its waste products on land. The problems associated with large scale biofuel crops cultivation is mostly same as problem arising from crops cultivation.

Some of the associated problems in soils are discussed below:

Chapter 2 Review of Biofuels

2.4.1.1. Soil erosion

In agronomic terms, soil erosion is problematic as it, results in the total loss of a fundamental resource or redistributes organic matter and nutrient-rich materials on a landscape scale along slopes [103]. In tropical agriculture, soil erosion is recognised as a major problem, in a number of areas under biofuel crop cultivation, as evidence in general reviews [104, 105]. Cultivation on slopes is undesirable, as it tends to increase rates of runoff and erosion. It has been suggested that biofuel crops should not be grown on slopes greater than 8% although some of 20–30% are used, for example, in parts of the Caribbean and South Africa [106]. As a result, biofuel crop production on marginal land has led to higher production costs and lower productivity due to soil erosion.

2.4.1.2. Removal of soil during harvesting

The crops harvested from the field always contain some amount of soil which may be

3-5% of soil by weight, as estimated by the UNEP (1982). However, the amount of extraneous material (soil, trash) removed from a field depends on the harvesting method used and the soil and weather conditions during a harvest [107] and, generally, the increased mechanisation of harvesting results in higher levels of extraneous material. The main factors determining whether biofuel crops are cut by hand or machine are the cost of labour and the size of the cultivated area. In the particular case of a recumbent crop for which its growing cycle is lodging at an early stage and developing stalks horizontally, the most straightforward harvesting methods can result there being in around 40% of extraneous material in loads delivered to the factory, half of which is soil [107].

Chapter 2 Review of Biofuels

2.4.1.3. Soil quality

Few studies have reported details of the relationship between the yields of biofuel crops and soil quality on a within-field scale. In previous studies [108, 109] provide a concise synthesis of world literature on the impact of the cultivation of biofuel crops

(sugar cane) on soil quality, with the main effects identified being:

• loss of soil organic matter;

• soil acidification;

• changes in soil nutrient levels;

• soil salinization and sodification;

• compaction of topsoil; and

• pre-harvest burning of biofuel crops

Every biofuel crops has some tolerance limit of pH, such as a range of 5-8 for sugar cane and some studies have suggested an optimum pH of about 6 although some cane fields on very acidic soils are known to produce high yields [110, 111].A decrease in the soil pH when virgin land is brought under sugarcane cultivation is commonly observed for most soil types at a range of depths. Such effects have been specifically recorded in Australia [112-114], Brazil [115], Fiji [116], Florida [111], Papua New

Guinea [117], The Philippines [118], Puerto Rico [119] and South Africa [120].

The changes in soil nutrient levels under biofuel crops depend primarily on whether fertiliser inputs are greater or less than the nutrients removed with harvested cane or lost by other means [109]. Net losses of nutrients with for high removal rates in no-till corn, suggesting that increased fertilisation rates are needed to maintain soil fertility

Chapter 2 Review of Biofuels

as reported in [121] . Conversely, it was found that [122] increasing residue return rates, increase the total N uptake (immobilisation) from soil, suggesting that it is necessary to add N fertiliser when residues remain to avoid soil mining for residue decomposition.

Poorly managed irrigation is the main cause of human-induced soil salinization [109,

123, 124]. Naturally, shallow water tables are generally associated with areas of low elevation in the landscape and there is evidence that they are common throughout the sugar industry [125]. Excessive salt concentrations in soil under sugarcane have been reported from many areas, particularly where rainfall is relatively meagre and irrigation practised[109], including those from Australia [126], India [127], Iran [128,

129], Swaziland [130], the USA and Venezuela [131].

Most biofuel crops are cultivated on large-scale plantations with a high degree of mechanization, and heavy vehicles are often used for in-field operations such as tilling and harvesting [108]. Particularly when the soil is wet, this results in very important associated compaction problems, including breakdown of the soil structure and direct damage to crop stools [106, 132].

2.4.2. Water

With increasing populations and, food requirements, industrialisation and urbanisation, the world is on the brink of an unprecedented water crisis. While this can be partially attributed to the uneven geographic distribution of water, it has been exacerbated by the absence of appropriate national and international policies for

Chapter 2 Review of Biofuels

ensuring the sustainable use of water. However the current biofuel development strategies may aggravate it and access to water could become a primary factor in the development of biofuel feedstock production.

2.4.2.1. Water consumption

Biofuel production requires water inputs at two stages: while growing the feedstock and during the production process in biofuel plants. Where irrigated agriculture is practised (drawing on surface and groundwater reserves), a major objective should be to maximise the crop yield per unit of water consumed, although different frameworks for maximising water productivity may operate on larger farm scales. Enhancing the efficiency of water use and water productivity in agriculture is particularly important as, in many parts of the world, this activity is a dominant consumer of water; for example, irrigated agriculture is the largest user of water in South Africa [133] and

Pakistan [134] and there are increasing demands for it from the agricultural sectors in many countries [135].

2.4.2.2. Irrigation

The need for irrigation in the cultivation of biofuel crops varies between localities, according to factors such as topography and soil type, and particularly climatic factors such as rainfall. As well as influencing the need for irrigation, local climatic conditions affect the precise pattern of the development of biofuel crops and the most appropriate timing for particular operations (planting, harvesting, etc.). A number of detailed examples of the influence of climatic patterns on the nature and timing of cane cultivation operations, citing parts of Kenya and Papua New Guinea are given in

Chapter 2 Review of Biofuels

literature [106], where sugar cane was grown without irrigation, parts of Sri Lanka where irrigation was necessary and parts of Zambia where irrigation was required. The availability of water is one of the principal constraints on expansion of the sugar industry in Australia, where 60% of the crop is fully or partially irrigated as this, is essential for consistently profitable sugar cane cultivation and accounts for around a third of production costs, and where the irrigation of sugar cane increased by 6% between 1970 and 1990. Examples of the scales of irrigation of sugar cane, specific localities illustrate the huge quantities of water involved, with reportedly, 3.8

ML/ha/year consumed in the Bundaberg region , 10 ML/ha/year the Burdekin region

[136] and from 15.3 to 53.8 ML/ha/year in the Ord irrigation area in the north-west

Western Australia where commercial sugarcane production only started in 1995 [137].

2.4.2.3. Water quality and aquatic system

The environmental impacts of conventional biofuel crop production are no different from those of other farm crops, such as the potential for nutrient enrichment of excessive fertiliser applications and sedimentation from land conversions. It is generally recognised that agriculture is a significant source of water pollution with fertilisers and livestock effluent accounting for as much as 40% of nitrogen and 30% of phosphate emissions in surface water which, results in oxygen depletion and eutrophication, insecticide and herbicide runoff impairing water and wildlife habitat systems. The production of biofuel crops can have an impact on water quality and aquatic ecosystems through both the cultivation and processing of different crops. In relation to cultivation, the main considerations arise from runoff and leaching, which can lead to the pollution of groundwater (which may include sources of human

Chapter 2 Review of Biofuels

drinking- water), surface water (including natural watercourses such as rivers and streams) and ultimately coastal environments. The main polluting agents are nutrients

(notably nitrates and phosphates derived from fertilisers, which can cause eutrophication), agrochemicals such as insecticides, and sediments arising from soil erosion. Irrigation can increase runoff and deep drainage, with its impacts on water quality potentially arising from the use of waste or saline water, which can also contribute to the salinisation of soils [124]. Although based on observations in

Australian sugar cane growing areas, this probably holds true for most other

(particularly tropical) regions. In addition, the inputs associated with intensive agriculture in general (rather than management strategies for particular crops) result in increased risks of pollution.

Runoff

Surface drainage (runoff) is closely linked to soil factors, including compaction (which tends to increase runoff rates) and erosion (which tends to be promoted by runoff).

Surface drainage water can carry with it dissolved nutrients, soluble insecticide, herbicides residues and soil sediments. Soil erosion can facilitate the movement of nutrients from the agricultural fields, with studies conducted in Australia suggesting that 50% of N and 80% of P transported (in a flood event in the Herbert River catchment) were bound to sediments. The amounts of nutrients lost from fields depend on application rates of inorganic fertilisers [138].

Glyphosate-based herbicides, glyphosate and its major metabolite AMPA have frequently been detected in water bodies. Both glyphosate and AMPA are water

Chapter 2 Review of Biofuels

soluble and can persist in aquatic environments for several weeks [24]. Previous studies focused on organochlorine insecticides, with those conducted from 1999 along most of the Great Barrier Reef (GBR) coast discovering insecticide residues, particularly herbicides such as diuron and atrazine, in coastal sediments, sea-grass and irrigation tailwaters [139, 140]. Further studies showed that all the rivers examined which discharged into the GBR and drained agricultural and urban areas showed extensive insecticide contamination, frequently in quantities greater than those specified in water quality guidelines [16, 141-149] while GBR lagoon waters also showed widespread insecticide contamination with levels occasionally above water quality guidelines [145, 150, 151]. Also, glyphosate was detected in about 30% and

AMPA in 50% of samples collected across Europe in 1993–2009 [152].

Leaching

This subsurface drainage water may carry soluble chemical residues with it (leaching) which affect the chemical composition of deeper soil strata or groundwater are flushed out into waterways. Other environmental factors may influence the specific composition of leachates [109]; for example, soil acidification promotes the leaching of certain nutrients (such as Ca and Mg) with that of nitrate contributing to the eutrophication of freshwater and coastal ecosystem, and the accumulation in the soil of others (notably Al). Although most leached nitrate appears to originate from mineralised soil organic nitrogen, excessive application of inorganic fertilisers also contributes to the loss of nitrate from soils.

In Australia, a study [153] of fertiliser N dynamics in recently harvested, trash- blanketed sugar cane fields, found no evidence of N leaching (although 20–30% of

Chapter 2 Review of Biofuels

applied N was lost to NH3 volatilisation). While it has also been suggested that some of the groundwater pollution that they report may have derived from natural, geological sources, most came from alluvial aquifers close to sites associated with intensive horticulture and sugarcane cultivation [154]. Studies in the USA have shown that the leaching of fertiliser N under sugar cane cultivation can be considerable. In

Florida, [155] found that losses of 6–24% of N applied as urea, depending on the rate of fertiliser application and irrigation level, was mostly as nitrate but, when irrigation took place before the N hydrolysed from the urea was completely nitrified, the leaching of ammonium was also considerable. In Louisiana, [156] found the leaching of nitrate, amounting to 15–60% of applied N in relation to the leaching of insecticide residues, with studies also show a mixed pattern of results, although in many cases levels of pollution are below those that would cause immediate concern.

Off-farm transport was monitored for atrazine, diuron, hexazinone and acetochlor in runoff water on a 500 m2 plot scale and across a 40 ha catchment in Valetta (Mauritius)

[157] showed that mean herbicide concentrations are low and do not exceed existing drinking-water standards and that the total masses of herbicides lost by runoff over one growing season represent very low proportions of the quantities applied (not more than 0.02% of atrazine, 0.32% of hexazinone, 0.07% of diuron and 0.19% of acetochlor). On a plot scale, these herbicide losses occurred mainly as sediment-bound residues, but, on a catchment scale, 70–95% occurred as dissolved residues. In the measured USA, leaching of atrazine and metribuzin from soils under sugar-cane [156,

158] showed that for both, it was greatest immediately following application, but decreased over a period of weeks. In Brazil, residues of atrazine, simazine and

Chapter 2 Review of Biofuels

ametryne were measured in surface and groundwater, in an area where sugarcane is intensively grown and from which the water-table of an important aquifer is recharged.

Ametryne residues were detected in a small number of surface water samples, but almost always at levels below internationally recommended environmental limits. In

Australia, of insecticide contamination events arising from sugar cane cultivation is scarce, although low levels of atrazine and heptachlor have been detected in the

Burdekin Delta aquifer system [154].

2.4.3. Terrestrial biodiversity

Increased biofuel production, as is expected to occur in the coming decades, will have large impacts on biological diversity, with negative be mostly a result of habitat loss, increased invasive species, and pollution resulting from the use of fertilisers and herbicides. Direct competition occurs when biofuels are grown in what was previously conservation land and indirect competition when biofuels displace food crops that in turn displace conservation land [159]. Under one of its scenarios (Global

Orchestration), the Millennium Ecosystem Assessment predicts that by the year 2050, the demand for food crops will increase by 3321 million tonnes (Mt) over the current

3906 Mt [160], with that for cereals alone increasing by 73% and that for livestock by

63%. The total area devoted to agriculture, depends on demand and yield with the latter varying according to the degree of development and transfer of technology. From this point of view the most optimistic scenario predicts increases in agricultural land of

0.01% per year from 1997 to 2050 and the most pessimistic increases of 0.34% per year for the same period which means an increase of agricultural land of 137 million hectares (Mha) by 2050.It has been suggested in many previous studies that a large

Chapter 2 Review of Biofuels

increase in the total amount of agricultural land will be required for foodstocks in the future and if the land needed for biofuel feedstock production is in addition to that a large net loss in the total area remaining in conservation is likely.

2.5. Genetically Modified (GM) Crops: A New Option for Biofuel

2.5.1. GM crops

Over the past few decades, the global increase in agricultural activities and land-use changes have introduced transformations in farming practices and landscape structures

[161, 162] to achieve better adaptability and increased productivity which have been shown to impact on farmland biodiversity [163-165]. In this regard, introducing GM

(or transgenic) crops technology is also likely to change modern agriculture.

GM crops (GMCs, GM crops or transgenic crops) are plants used in agriculture, the

DNA of which have been modified using genetic engineering techniques. In most cases, the aim is to add a unique characteristic or attribute to the genetic makeup of a crop using recombinant DNA (gene-splicing) technology which uses identification, isolation and manipulation followed by the introduction of the desired gene(s) from one organism (for example, a plant or bacteria) to another, thus giving rise to a transgenic or GM organism.

After the first development and commercialisation of GM crops in 1996, the amount of land being cultivated with them has increased throughout the world. Between 2004 and 2005, the global area occupied by GM crops increased by 11% and, by the end of

2005, covered 90 million hectares. Commercialised GM crops were planted in 23

Chapter 2 Review of Biofuels

countries in 2007, a total cultivated area of 114.3 million hectares, a 67-fold increase compared with that in 1996 [166]. Over 3000 field trials of about 25 GM plant species have been conducted, as documented in OECD lists (OECD, 2003; releases until

2000).

In the Australian market, GM foods have been commercially available since 2000 when soybean, canola, corn and cotton were approved by the Food Standards Australia and New Zealand [9]. GM canola was first grown commercially in NSW in 2008 by

108 growers with, in that year, approximately 9600 hectares planted in NSW and

Victoria [167]. In 2009, the uptake of GM canola in NSW increased four-fold, with over 41, 000 hectares planted [167], mostly with Roundup-ready canola. 93% of GM canola growers rated the weed control achieved by Roundup-ready as excellent compared with much lower ratings for the Clearfield®, Triazine-tolerant and conventional canola systems [168].

2.5.2. Different varieties of GM crops

2.5.2.1. Triazine-tolerant (TT) crops

TT crops are those in which the gene inserted provides tolerance to the triazine herbicides (i.e., atrazine and simazine) obtained from a mutant weedy brassica and, in comparison with non-TT crops, they reduce the capability of the plant to use sunlight.

In Australia, the first trial of TT GM canola was conducted in 1970 in rotation with winter cereals [169], with the variety adopted different from that used in Canada due to the concern raised regarding cruciferous weeds such as wild radish

(Raphanusraphanistrum) [170]. The increased use of TT canola varieties in crop

Chapter 2 Review of Biofuels

rotations in Western Australia (WA) has been reported in the literature [171]. Fig. 2.4 provides information regarding the different varieties of TT canola used in average long-term trials conducted by the national Grains Research and Development

Corporation.

TT: Triazine-tolerant varieties Hybrid: Conventional hybrid varieties Conv: Conventional varieties (i.e., non- CF: Clearfield imidazolinone- herbicide-tolerant) resistant varieties

Fig. 2.4: WA average yields of various cropping systems (2000-2007) [171]

2.5.2.2. Clearfield crops

The Clearfield variety of GM crops was commercially released in Australia in 2007.

Developed through irradiation and selection breeding methods, it is tolerant to the imidazolinone group of herbicides which can limit crop choices for the following year depending on the soil’s chemistry and rainfall.

2.5.2.3. Roundup-ready crops

Roundup-ready crops (Fig. 2.5) are genetically engineered crops that have had their

DNA altered to allow them to withstand the herbicides glyphosate or glufosinate ammonium. Of the different varieties of GM crops, they have the dominant trait for

Chapter 2 Review of Biofuels

commercial crops (71% of cultivated area) followed by insect resistance (18%) [172].

Only four GM herbicide-tolerant (GMHT) plant species with tolerances to either glufosinate or glyphosate have been widely grown as approved varieties: corn (Zea mays), cotton (Gossypiumhirsutum), soybean (Glycine max) and oilseed rape or canola, with the last covering 5% of the global biotech area. According to previous studies [173, 174] and a survey of growers of different GM crops in NSW, Australia

(discussed in the case study chapter), there has been an increase in the choice of

Roundup-ready crops.

Fig. 2.5: GM roundup ready weed management (Wimmera, Victoria) [171]

Note:

 Roundup Ready hybrid sown on 9 May 2008 and photograph taken 2 July 2008

 Paddock sprayed with herbicide trifluralin 480 @ 1.6 litres/hectares giving good early grass control.

(circle 1)

 Glyphosate-tolerant crop sprayed with herbicide glyphosate (Roundup) at 900 grams/hectare around 23

May (14 days after planting);.photograph taken when crop at six- to seven-leaf stage showing early broad

leaf dying (circle 2)

 Too late to spray crop with second glyphosate application and late-germinating weeds (circle 3) will

receive selective control, if required and available (not for radish or turnip), but unlikely to be required

as canola should out-compete weeds

Chapter 2 Review of Biofuels

2.5.3. Biofuel production from GM crops

The use of plant oils for biofuels can be a sustainable alternative to non-renewable petroleum [175]. Globally, a majority of total petroleum consumption is used for transportation and electricity generation, leaving only a small amount (10–20%) for use as raw materials in various industries. Transportation is a major source of daily increases in GHG emissions [176], to reduce which organisations worldwide have set a target of substituting 10% of the fossil fuels used in the transportation sector with renewable sources by 2020 [94], with some countries aiming to achieve this almost entirely by first-generation biofuels [177]. However, concerns regarding the decrease in available agricultural area due to biofuel crop cultivation and increasing food prices has led the scientific community to look for alternative options [178-181]. Some research suggests that, to compensate for the loss of area for food production, biofuel production could trigger indirect land-use changes by recruiting new farmland through partially converting carbon-rich natural ecosystems which could balance the GHG benefits of biofuels over fossil fuels [159, 177, 182]. However, the efficiency of GHG could be improved if on-site yields could be substantially increased, i.e., if food, feed and fuels could be produced on essentially the same agricultural land. In this respect,

GM plants have been recommended as a new option for biofuel production [6, 7].

More than 95%of the world’s biodiesel is produced by edible oils which are first- generation biofuels. However, their low yields and the major controversy regarding food versus fuel has prevented them being used for purposes other than primarily as food [183]. In some previous studies, to reduce the utilisation of edible oils, non-edible oils were tested as second-generation biofuels, with the most important the ratanjot

Chapter 2 Review of Biofuels

(Jatrophacurcas), karanja (Pongamiapinnata), rubber seed tree (Hevcabrasiliensis), desert date (Balanitesaegyptiaca), sea mango (Cerberaodollam), terminaliabelerica, neem (Azadirachtaindica), koroch seed oil (Pongamiaglabra vent.), mahua

(Madhucaindica and Madhucalongifolia), tobacco seed (Nicotianatabacum L.) and jojoba (Simmondsiachinensis) [184]. Other studies have also reported that animal fats, waste oils and grease can be considered second-generation feedstocks and microalgae cultured in bioreactors third-generation biofuels. As suggested in the literature, first- and second-generation biofuels can be substantially improved by genetic engineering through the introduction of traits advantageous to biodiesel production from oil- producing microalgae and fungi into conventional and non-conventional oil crops

[185].

2.5.4. Impacts of large-scale cultivation of GM crop

The increased use of GM crops in agricultural systems as alternatives to biofuels has raised concerns about their anticipated benefits and harm. The long-term implications of introducing GM crops into farming practices are that current management regimes may be altered and new cropping techniques [186, 187], such as weed control, soil tillage and crop rotations, required which farmers may find difficult to accept [188,

189]. Of several studies conducted to investigate the positives and negatives of GM crops. some have found that they have decreased the intensity and use of chemicals, providing evidence that, from 1996 to 2004, the estimated total global volume of active ingredients applied to GM crops fell by 6%, a decline attributed to the commercialisation of GM crops [190, 191]. However, other studies have found a dramatic increase in the use of herbicides and raised concerns about the impact this

Chapter 2 Review of Biofuels

might have on weed management strategies. The variety of herbicides applied to herbicide-resistant GM crops has decreased as these crops are resistant to certain herbicides such as glyphosate and glufusinate [192]. One study argued that GM crops were responsible for an increase of 383 million pounds of herbicide applied in the USA by 2009, while herbicide-resistant crop technology has led to a 239 million kilogram

(527 million pound) increase in herbicide use in the USA between 1996 and 2011.

The issue regarding the cultivation of GM crops that it increases herbicide use has also become part of the debate concerning changes in tillage practices increasing carbon sequestration. As estimated by the Inter-governmental Panel on Climate Change

(IPCC) (2007) [193], agriculture contributes 10–12% of anthropogenic GHG emissions. Also, estimates by the World Bank (2010) [194] show that 30% of the world’s GHG emissions comes from agriculture, deforestation, land-use changes and forest degradation [195]. Additionally, the widespread adoption of conservation tillage

(including zero tillage) could sequester 24, 000–40, 000 kg of carbon per year [196] while it is suggested in the literature that the carbon sink capacity of the world’s agricultural and degraded soils is 50–66% of the world’s historic loss of 42–78 × 1012 kg of carbon [196].

2.6. Conclusions

The increase in worldwide biofuel production has been driven by different government policies, with the main issues blending targets, tax exemptions and subsidies. Political debate and environmental concerns over the impact of biofuels on food prices and

GHG emissions has fostered the implementation of new rules for supporting

Chapter 2 Review of Biofuels

bioethanol and biodiesel production. In this regard, the USA has already begun focussing on the development of second-generation technologies and has set significant production targets for cellulosic sources. Worldwide, manufacturers have taken into consideration the indirect land use for bioethanol and biodiesel production and rules have been imposed on them to certify the origins of their feedstock. Recently, with regard to its future biofuel targets, Germany considers GHG reductions in terms of biofuel production rather than output volumes;

Nevertheless, as there are currently no production facilities for second-generation biofuels, most countries are continuing to produce biofuels from grains, sugar and vegetable oils. It has become a great challenge for policy makers and industry executives alike to continue to expand the biofuel sector while taking into consideration ecologically sustainable production requirements. In this respect, GM plants have been recommended as new options for biofuel production [6, 7].

The contents of this chapter were in two sections: the first described biofuels and their production worldwide, and the impacts of the large-scale cultivation of biofuel crops; and the second provided information regarding GM crops as alternatives for biofuel production, as recommended by several different studies in the literature, and environmental concerns regarding their large-scale cultivation.

Chapter 3 Review of Herbicide Glyphosate

3.1. General

There has been increasing concern worldwide about the use of herbicides as the amount of cropland increases, particularly that converted to agriculture [5]. Australia is a major agricultural exporter, with around two-thirds of its total annual production exported [8] and, since its introduction of genetically modified (GM) crops in 2000, new varieties of them now account for a large amount of its total agricultural land use

[9]. These crops are usually produced by genes from soil bacterium being inserted into them to make them tolerant to one of the most commonly used herbicides, glyphosate.

Due to this increase in GM crop agriculture, the use of glyphosate has also increased during the past few years, with studies finding that around 15 Mt of glyphosate are applied annually in Australia to control agricultural, urban and roadside weeds.

The movement of herbicides off-site into aquatic ecosystems has become a major issue for marine ecosystems and natural resource management organisations [197-199].

Over the last few years, several studies have reported potential threats to near-shore water quality [16, 145, 200, 201] and the loss of corals and seagrass on the world heritage-listed Great Barrier Reef (GBR) [202, 203], with significant amounts of herbicides found in river mouths and GBR lagoons as run-offs from nearby agricultural areas [16, 145, 204]. Also, information in the literature shows that there are possibly leaching and toxicity problems associated with using glyphosate [205-211].

Chapter 3 Review of Herbicide Glyphosate

The aim of this chapter is to review the literature on the major herbicide used for GM crops in Australian agriculture, glyphosate, including its formulation, sorption, degradation, toxicological impact and potential threat to animals, birds, fish, amphibians and human life through its widespread use. Firstly, glyphosate and its formulation products, together with factors affecting its performance, are introduced.

Then its sorption to soil and potential risk of leaching, followed by its toxicological impact and formulation products, are fully discussed.

3.2. Introduction

Glyphosate (N-(phosphonomethyl) glycine) is a white, odourless crystalline solid, broad-spectrum, non-selective, post-emergence herbicide used in agricultural, forestry and aquatic weed control. It was first introduced in 1971 by the Monsanto Company as an unwanted weedkiller for agricultural purposes [212, 213], and its physical and chemical properties are given in Table 3.1. It is the active ingredient in several commercial products, with the most well-known Roundup which is widely applied in

GM crop agriculture. Its capability to control perennial weeds has made glyphosate the essential, biggest-selling, fastest-growing agrochemical in the world [212]. Also, its low-cost generic formulations manufactured in China are widely available.

Glyphosate is commonly used in salt form, most typically as an isopropyl amine salt, but sometimes commercial manufacturers add to it to facilitate movement of its active ingredient into plants [214]. It is very soluble in dilute bases and strong acids, and forms soluble salts. It is amphoteric (that is, it has the characteristics of both a base and acid, and can react as either) and can exist as different ionic forms depending on

Chapter 3 Review of Herbicide Glyphosate

the pH of the soil or water, with its toxicity dependent on the pH and higher at high pH values .

Owing to the widespread adoption of glyphosate in Australia over the last few years and the Australian Government’s plans to increase cultivation of GM crops [10], there has been considerable controversy concerning the possible impacts of glyphosate on the environment. It toxicity classification is provided in Table 3.2 and the safe limits set by different world environmental organisations to protect human life and marine and aquatic systems have already discussed in section 1.2.

Table 3.1: Physical and chemical properties of glyphosate

Active Form Vapor pressure, Henry’s constant, Molecular Solubility in Log Kow Koc Ingredient [215, 216] mPa (25oC) or (atm-m3/mol) weight’ (g/mol) water (mg/L) [215-217] [218] mmHg (45oC) [217] [215-217] [215, 216] [215-217] Glyphosate Odourless, 1.31 x 10-2or 4.08 x10-19 169.07 pH 1.9: 10500 <-3.2 300-20100 acid white 1.84 x 10-7 pH 7: 157000 solids Glyphosate Odourless, 2.1 x 10-3or 6.27 x 10-27 228.19 pH 4.06: 786000 -3.87 or -5.4 300-20100 isopro- white 1.58 x 10-8 pylamine salt solids Glyphosate Odourless, 9 x10-3or 1.5 x 10-13 186.11 pH 3.2: 144000 -3.7 or -5.32 300-20100 ammonium white 6.75 x 10-8 salt solids

Table 3.2: Toxicity classification- glyphosate Parameter High Toxicity Moderate Toxicity Low Toxicity Very Low Toxicity

Acute Oral LD50 ≤ 50 mg/kg > 50 – 500 mg/kg > 500 – 5000 mg/kg > 5000 mg/kg

Inhalation LC50 ≤ 0.05 mg/L (aerosol) >0.05 – 0.5 mg/L > 0.5 – 2.0 mg/L > 2.0 mg/L (dust)

Dermal LD50 ≤ 200 mg/kg > 200 - 2000 mg/kg >2000 – 5000 mg/kg > 5000 mg/kg Primary Eye Corrosive (irreversible Corneal involvement or Corneal involvement or Minimal effects clearing Irritation destruction of ocular tissue) or other eye irritation other eye irritation in less than 24 hours corneal involvement or clearing in 8 –21 days clearing in 7 days or less irritation persisting for than 21 days Primary Skin Corrosive (tissue destruction Severe irritation at 72 Moderate irritation at 72 Mild or slight irritation Irrigation into the dermis and/or scarring) hours (severe erythema hours (moderate at 72 hours (no irritation or edema) erythema) or erythema)

3.3. Factors Affecting Glyphosate Performance

Up to two glyphosate applications can be performed on the tops of GM crops through a 6 leaf stages prior to bud formation. The capability of glyphosate to reliably control a broad range of grass and broadleaf weeds in crops offers growers a robust, alternative weed control option. The first trial of it was conducted in Australia in 1997 and GM food has been commercially available since 2000 [219]. A rainfall event is an important factor regarding the application of glyphosate because, if it occurs after an application, the glyphosate can be washed off before it has a chance to enter the leaf and, also, the plant will not receive a lethal dose of the herbicide as rain water will reduce its activity by dilution. An application is normally performed after checking the weather conditions to establish that there is no expectation of rain within the next 24 hours [220, 221]. Glyphosate products are normally mixed with water before application to facilitate a smooth process [214]. Another factor affecting glyphosate performance is the water quality (soft or hard) as hard water contains large amounts of dissolved salts, such as calcium and magnesium, which have positive charges and may associate with the negatively charged glyphosate molecules to displace the isopropylamine or other salt used in the formulated product. Also, plants will absorb less glyphosate bound with calcium or magnesium salts than with glyphosate’s formulated salt [214]. In addition, a delay in glyphosate application can cause weed control failures as it allows weeds to reach sizes that are difficult to kill consistently.

Therefore, better performances can be obtained by minimising its application time and using the proper rate [214].

Chapter 3 Review of Herbicide Glyphosate

3.4. Modes of Action and Persistence

3.4.1. Plants

In plants, glyphosate disrupts the shikimic acid pathway (a seven-step metabolic route used by bacteria, fungi, algae, etc.) through inhibition of the enzyme 5- enolpyruvylshikimate-3-phosphate (EPSP) syntheses. The resulting deficiency in

EPSP production leads to reductions in the aromatic amino acids vital for protein synthesis and plant growth [215, 216]. Glyphosate is absorbed by the leaves and stems of a plant and transported throughout the plant via its phloem [215, 218] accumulating in its meristems, immature leaves and underground tissues [216, 222]. The sodium salt of glyphosate acts as a plant growth regulator which accelerates fruit ripening [223] while very small amounts of it are metabolised in plants, with AMPA its only significant degradation product [218]. Lettuce, carrots and barley have been shown to contain glyphosate residues up to one year after the relevant soil was treated with 3.71 pounds of glyphosate per acre [224, 225] while glyphosate had a median half-life of 8 to 9 days in the leaf litter of red alder and salmonberry sprayed with Roundup® [226].

Plants exposed to glyphosate display stunted growth, loss of green colouration, leaf wrinkling or malformation, and tissue death, with the death of a plant possibly taking from 4 to 20 days to occur [216, 222].

3.4.2. Soils

Studies have revealed that glyphosate is biologically degraded mainly by soil microorganisms and has a minimal effect on soil micro-flora [227], with some also stating that doses of glyphosate of <10000 µg/L stimulate soil micro-flora, including actinomycetes, bacteria and fungi, while concentrations > 10000 µg/L have

Chapter 3 Review of Herbicide Glyphosate

detrimental impacts on micro-flora populations [228]. The major metabolite of glyphosate is AMPA (CH6NO3P) which causes its microbial biodegradation in soils

[229], aquatic sediments and waters while its photo- and chemical degradations are negligible compared with those of other herbicides [222].

When applied directly to a soil, glyphosate shows little herbicidal activity due to its inactivation as a result of sorption [229, 230]. It is considered to be a moderately persistent herbicide with a half-life (t1/2) ranging between 10 and 100 days (average 47 days) [231]. The main route of its biodegradation appears to be by the splitting of the

C-N bond to produce AMPA (Fig. 3.1) while a secondary one of splitting the C-P bond can also occur. AMPA is biologically degradable, with its end-product carbon dioxide, and degradation occurs more rapidly in aerobic than anaerobic conditions. Half-lives for biodegradation in soil are between a few days and several months [232].

Fig. 3.1: Glyphosate degradation pathway [233]

Chapter 3 Review of Herbicide Glyphosate

3.4.2.1. Sorption

Glyphosate is a small molecule with three polar functional groups (carboxyl, amino and phosphonate) which is strongly sorbed by soil minerals after application [234-

240], with a usual half-life of 3-174 days on soil minerals and 5-91 days in water [241].

The sorption of glyphosate to soil usually occurs through the phosphonic acid group

(organic compounds containing C-PO(OH)2 or C-PO(OR)2) in its phosphonate anion

3− form, similar to that of phosphate (PO4 ) in soil [229, 234, 236, 242, 243] even though the carboxylic group can also participate. Glyphosate soil sorption has also been interpreted as occurring by ion exchange and hydrogen bonding [244]. Previous studies in the literature suggest that the degree of sorption of glyphosate by soils or clays is a function of the soil’s cation exchange capacity (CEC), clay content [230] and concentrations of organic matter, iron and aluminium amorphous oxides [245]. It has been reported that glyphosate forms mono- and divalent anions which have high affinities for, in particular, trivalent cations such as Al3+ and Fe3+ within the pH range of 4-8 [246, 247] and, as will be shown, its sorption characteristics explain its mobility in soil. Also, microbial biodegradation of glyphosate occurs in soils, aquatic sediments and waters, mainly by splitting of its C–N bond (Fig. 3.1), and produces aminomethylphosphonic acid (AMPA), glyphosate’s principal degradation product

[26].

Features affecting glyphosate sorption onto soil

Specific surface area (SSA) and mineral group

Glyphosate sorption is only possible for variable-charge surfaces not permanent- charge (negative) sites on layer silicates as it is an anion in the relevant pH range of soils (Fig. 3.2). According to the information provided in the literature, soils with good

Chapter 3 Review of Herbicide Glyphosate

portions of variable-charge minerals have proven to be more effective than those enriched with permanent-charge minerals such as illite, smectite and vermiculite [248-

251]. The sorption sites mainly used by glyphosate, especially for sorption by goethite

(α-FeOOH), are surfaces of aluminium and iron oxides, poorly ordered aluminium silicates (allophane/imogolite) and edges of layer silicates [236, 246, 247, 251-253].

Different amounts of glyphosate sorbed by different pure minerals, depending on the surface area and mineral group, are presented in Table 3.3.

Chapter 3 Review of Herbicide Glyphosate

Table 3.3: Glyphosate sorption by different pure minerals

SSAa Glyphosate sorption Mineral References (m2g-1) (mmol kg-1) (µmol m-2) Aluminium and iron oxides Gibbsite 45 72 1.60 [252] Ferrihydrite 343 635 1.85 [254] Goethite 85 125 1.45 [246] Hematite 33 86 2.61b [254] Clay silicates Kaolinite 1 12 3.9 0.33 [252, 255] Kaolinite 2 22 6.9 0.31 [252, 255] Illite 43 5.2 0.12 [252, 255] Montmorillonite 32 6.5 0.20 [252, 255] a: SSA determined by applying BET equation to N2 sorption b: This value is undoubtedly too high owing to inappropriate hematite synthesis, as discussed by Gimsing and Borggaard [254].

Fig. 3.2: Distributions of glyphosate species as function of pH (Bjerrum diagram) (acid dissociation constants pKa1 = 2.22, pKa2 = 5.44 and pKa3 = 10.13[246]; and zwitterionic structures of carboxyl and amino groups shown for entire pH range)

Chapter 3 Review of Herbicide Glyphosate

Glyphosate sorption onto soil surfaces depends not only on the SSA and mineral group but is also greatly influenced by the pH. As found in some studies, glyphosate sorption by goethite decreases at increasing pH [246, 247], which is in accordance with the decreased sorption of other phosphonates and phosphate at increasing pH [249], while many agree that glyphosate sorption onto soil can be decreased at increasing pH [133,

256]. On the other hand, increasing the pH level by liming may cause the opposite effect as, although glyphosate sorption will increase, this is attributable to the formation of glyphosate-sorbing aluminium and iron oxides at higher pH [250].

Soil organic matter (SOM)

In contrast to all other insecticides which are moderately to weakly sorbed in soils by organic matter as their molecules are dominated by a polar groups, i.e., aliphatic and/or aromatic carbon and often have only one functional group, glyphosate, a small molecule with three polar functional groups (carboxyl, amino and phosphonate ), is strongly sorbed by soil minerals [249]. While some studies have revealed very high glyphosate sorption by four different purified humus samples, some have also suggested that the soil sorption of glyphosate is not, or sometimes negatively, correlated with the SOM content which seems to play a controversial dual role in it.

Phosphate

Phosphate reacts similarly to glyphosate as both are sorbed by ligand exchange on variable-charge Al-OH and Fe-OH surface sites with the formation of strong Al-O-P or Fe-O-P bonds [136, 253]. Therefore, phosphate may influence the sorption of

Chapter 3 Review of Herbicide Glyphosate

glyphosate on soil surfaces by competing against it for sorption sites. This competition reaction mechanism was first shown in 1974 [229, 257], which described phosphate as the most important factor for glyphosate sorption, and was later discussed in several studies [133, 258-261]. Its occurrence could severely impact glyphosate bonding and, thus, leachability, especially on many agricultural soils worldwide which are saturated or nearly saturated with phosphate because of surplus phosphorus fertilisation over many years worldwide [262-264]. This competitive reaction mechanism on an iron oxide surface is shown in Fig. 3.3.

Fig. 3.3: Tentative reaction schemes for sorption of glyphosate and phosphate (A - competitive sorption; B - sorption in two planes (glyphosate on phosphate), with both glyphosate and phosphate assumed to form binuclear, bidentate surface complexes, and zwitterionic structures of carboxyl and amino groups on glyphosate omitted [265])

3.4.2.2. Leaching

Glyphosate transport from terrestrial to aquatic environments, as a potential sorbable compound can occur in solution or in suspension, i.e. the compound can be transported

Chapter 3 Review of Herbicide Glyphosate

as solutes or co-transported bonded to soil colloids (colloid-facilitated or particle- bonded transport) [265]. The dissolved form can be moved by leaching through the soil (subsurface runoff) ending up in drainage and groundwater. On the other hand the particle-bonded form can be moved by overland flow (surface runoff) forming open waters such as streams and lakes [265]. Movement of water through the soil can be described as a matrix flow (piston flow) in case of uniform, non-structured soils e.g. many sandy soils, while in structured soils, e.g. many clayey soils, preferential flow bypassing more or less the soil matrix is common. The pathways for the preferential flow are usually considered as macrospores, including biopores and cracks between aggregates but can also be bands of higher hydraulic conductivity such as sand bands in a clay matrix [266, 267].

For uniform non-structured soils e.g. sandy, oxide-poor soils with high hydraulic conductivity that receive high precipitation rates, can prohibits higher glyphosate leaching with lower sorption capacity. In a previous study on sandy soil with a very low mineralisation rate [249] indicated the high risk of glyphosate leaching having lower sorption capacity. However, as the topsoil contain high contents of poorly ordered aluminium and iron oxides and hence high glyphosate sorption capacity, a limited translocation of glyphosate from top- to subsoil, well above 1m need to considered as well [207, 249, 268]. However, other studies suggested no leaching of glyphosate (or AMPA) over two years of filed study which occurred due to the absence of macropores, i.e. the soil is non-structured and water moves through the soil matrix as piston flow [207]. The lysimeter (column) experiments showed no or very little glyphosate leaching in sandy soils without macropores, and hence no bypass flow

Chapter 3 Review of Herbicide Glyphosate

showing the concentration of glyphosate below the detection limit (< 0.04 μgL-1) in drainage water samples taken 1m below the soil surface in a sandy, non-vegetated soil sprayed with roundup [209, 269, 270]. Glyphosate translocation from topsoil to subsoil

(down to 50–70 cm below the soil surface) was found in one sample in a sandy soil

[271].

In case of structural soils the transport rates of insecticides in soils can be enhanced with preferential flow pathways through macropores (sandy loam, clayey soil etc.)

[272]. Studies with subsurface leaching of glyphosate (and AMPA) in structured soils and soil materials with macropores and bypass flow are reported in literature [205-

207]. Glyphosate leaching can be severe on gravelly materials and very coarse- textured soil materials such as under railway embankments. Samples having glyphosate concentration above the European threshold (0.1 μgL-1 (under railway embankments) and up to 1300 μgL-1 (short columns packed with gravel) were reported in literature [209, 210]. However long-term uses of glyphosate over any kind of soil may lead to glyphosate pollution to groundwater. Heavy rainfall shortly after glyphosate which is an important factor for this case while vegetation, tillage and phosphate concentration seem to have little or no effect on glyphosate leaching [265].

3.5. Impacts of Glyphosate

Owing to the widespread use of glyphosate throughout the world and studies reporting its adverse effects on the human body, aquatic life and terrestrial species [273, 274], concern has arisen regarding its associated potential environmental and toxicological risks. Studies suggest that the glyphosate-based formulation product Roundup has

Chapter 3 Review of Herbicide Glyphosate

strong toxicity [275-278] and glyphosate’s other formulation products contain a number of adjuvants and surfactants which enhance its activity. One of its well-known surfactants is polyoxyethyleneamine (POEA) which accelerates adsorption and increases the efficiency of the formulation’s penetration into plant tissue. However, as the literature suggests, these additional substances are more toxic and considered key components for glyphosate toxicity [279]. As its surface-active compounds change the properties of glyphosate, its commercial formulations are of higher toxicities than the active substance itself. A survey of the literature on glyphosate-dosed toxicological studies of animals, birds, fish, amphibian and human life are discussed in the following sub-sections:

3.5.1. Animals

3.5.1.1. Rats Animal studies have indicated that 30-60% of glyphosate is absorbed after ingestion

[280-282]. In a study on rats given a single oral dose of 10 mg/kg to 1000 mg/kg of glyphosate found that, after seven days, the absorbed dose had distributed throughout the body although was primarily concentrated in the bone [280-282]. The acute oral lethal dose (LD) of LD50 and acute inhalation lethal concentration (LC) of LC50 in rats were found to be greater than 4320 mg/kg [283] and greater than 4.43 mg/L based on a 4-hour nose-only inhalation study [284] respectively. Animal studies of the effect of glyphosate on animal metabolism indicate that it is primarily excreted through the urine and faeces [218, 226, 281]. A rat given a single oral dose of glyphosate eliminated 0.27% of it as carbon dioxide and excreted 97.5% as glyphosate in urine and faeces, with all of it cleared from its body 168 hours after administration [280].

Chapter 3 Review of Herbicide Glyphosate

However, a previous study of a glyphosate formulation product found that it provoked apoptosis and necrosis of a rat’s testis cells [285].

3.5.1.2. Rabbits

Glyphosate showed comparatively less toxicity for rabbits when applied to the skin, with their acute dermal LD50 greater than 2 g/kg [283] and, in studies using glyphosate manufactured products, researchers observed mild eye irritations in rabbits that cleared in seven days [286, 287]. Also, as isopropylamine and ammonium salts are low in toxicity via the dermal route, the LD50 in rabbits was greater than 5000 mg/kg for both salts which are considered slight eye irritants but not skin irritants [215].

3.5.1.3. Hens and goats

Several previous studies found that, after feeding hens and goats, glyphosate and its major metabolite AMPA were found in eggs, milk and the animals’ body tissues [281,

282, 288-290]

3.5.2. Birds

An acute oral toxicity study found that a single dose of technical-grade glyphosate was practically non-toxic to bobwhite quail as their LD50 were greater than 2000 mg/kg

[291] while studies with technical-grade glyphosate found 8-day dietary LC50 greater than 4000 µg/mL for mallard ducks and bobwhite quail which indicated slight toxicity

[291]. Although glyphosate at dietary levels of up to 1× 106 µg/L is not expected to cause reproductive impairment in birds [292], an ecological risk assessment concluded

Chapter 3 Review of Herbicide Glyphosate

that the greatest risk posed by glyphosate and its formulated products to birds and other wildlife results from its altering their habitats [24].

3.5.3. Fish

The median half-life of glyphosate in water varies from a few to 91 days [215] and it was shown that it does not undergo hydrolysis in buffered solutions with pH values of

3, 6, or 9 at 35 °C. Its photodegradation in water was insignificant under natural light in buffered solutions with pH values of 5, 7 and 9 [293, 294] while its concentrations in two ponds declined rapidly, although its binding to bottom sediments depended heavily on the metals in the sediments [295].

For studies conducted on fish with exposure to Roundup, the LC50 found were 140

µg/L for rainbow trout (Onchorynchusmykiss), 97 µg/L for fathead minnows

(Pimephalespromelas), 130 mg/L for channel catfish (Icaluruspunctatus) and 150

µg/L for bluegill sunfish (Lepomismacrochirus) [296]. From a previous study involving 48-hour exposures to the formulated product Roundup®, theLC50 for daphnids and midge larvae were 3µg/L and 16 µg/L respectively [296].

However, the literature reveals a 17-32 times more toxicity in fish using a glyphosate formulation product (Roundup) than only glyphosate because of the presence of amine salt and its surfactants. The LC50 values for formulated glyphosate products after a 96- hour exposure ranged from 1.3 µg/L to greater than 1 µg/L [292]

Chapter 3 Review of Herbicide Glyphosate

3.5.4. Amphibians

Studies were also conducted on the calculated LC50 values for four species of amphibians, the northern leopard frog (Rana pipiens), wood frog (R. sylvatica), green frog (R. clamitans) and American toad (Bufo americanus)), exposed to the original

Roundup® formulation of glyphosate, with the 24-hour LC50 values for these different species ranging from 6.6 to 18.1 µg/L [297].

On the other hand, the surfactant in the formulated glyphosate proved most toxic to R. clamitans with 24- and 96- hour exposures, showing LC50 of 1.1 µg/L (95% CI 1.1-

1.2) and 1.1 µg/L (95% CI 1.0-1.1) respectively. A chronic toxicity study with technical-grade glyphosate reported reduced reproductive capacity in Daphnia magna, with a maximum acceptable toxicant concentration of 50 to 96 µg/L [298].

3.5.5. Humans

Experiments conducted on human skin resulted in a maximum of 2.2% of 2.6 ug/cm2 glyphosate being absorbed across it, with absorption peaking 8 hours after administration [299]. Studies conducted on 346 volunteers with glyphosate applied on their skins concluded that this did not lead to photo irritation or photosensitisation

[300]. In some previous literature, it was revealed that glyphosate undergoes little metabolism and is excreted primarily unchanged in the faeces and, secondarily, in the urine [218, 226, 281]. High ratios of glyphosate to AMPA were detected in a human patient’s blood serum 8 hours (22600 μg/L glyphosate to 180 μg/L AMPA) and 16 hours (4400 μg/L glyphosate to 30 μg/L AMPA) post-ingestion, as well as in the patient’s total amount of urine which indicates that glyphosate metabolism was

Chapter 3 Review of Herbicide Glyphosate

minimal [301]. Also, researchers have detected AMPA in urine (0.2-0.3% of administered dose) and faeces (0.2-0.4% of administered dose) [295, 302]. Two human patients who were poisoned with glyphosate had peak plasma glyphosate concentrations within 4 hours of ingestion but, after 12 hours, it was almost undetectable [303].

However, stronger themolysis and lipid peroxidation was observed in human erythrocytes incubated with the glyphosate formulation Roundup which caused the development of many diseases associated with reduced glucose-6-phosphate dehydrogenase activity [304, 305]. It was also found that herbicide formulations caused stronger oxidative damage and LC50 to the human liver, with that in the HepG2 cell significantly lower for herbicide formulations such as Roundup than glyphosate alone [306]. Studies of the human body have demonstrated that the surfactant POEA is very toxic for the hepatic (HepG2), embryonic (HEK293) and placental (JEG3) cell lines. According to previous literature, glyphosate formulations can act as typical hormonal modulators and be the major cause of death of human umbilical, embryonic and placental cells [307], can induce alterations in cephalic and neural crest development and can shorten the anterior-posterior axis on Xenopus laevis embryos

[274]. A very recent study reported on the exposure-response relationship between glyphosate exposure and the incidence of some types of cancer [308]. As the literature suggests, long-term exposure to glyphosate in contaminated soil and water, even at low concentrations, can lead to health problems whereas the use of formulation-based glyphosate can cause liver damage and amenia by promoting haematological and hepatic alterations [309]. Although studies have pointed out a link between

Chapter 3 Review of Herbicide Glyphosate

Parkinson’s disease and exposure to glyphosate, the actual mechanism remains unclear. In addition, glyphosate-induced cell death via autophagy pathways has been shown in the literature [310]. Glyphosate impacts on the human body are also associated with a prooxidant–antioxidant imbalance in human erythrocytes which may lead to increased levels of met-Hb seriously affecting the lives of erythrocytes. Also, oxidative stress contributes to alterations in the shapes of red blood cells and changes the viability of erythrocytes [311]. Studies conducted of human erythrocytes with a glyphosate metabolite and impurities showed slightly stronger damage to human erythrocytes than that of glyphosate.

3.5.6. Plants

Glyphosate is a broad-spectrum, post-emergent herbicide which inhibits the growth of plants through interfering with the biosynthetic pathway of the essential aromatic amino acids phenylalanine, tyrosine and tryptophan [312]. Because of its assumed plant-specific mode of action, it is generally considered of low toxicity to animals [24] and other non-target organisms [280]. However, some previous studies revealed that, based on relevant chemical safety data sheets, glyphosate-based herbicides are classified as hazardous to the aquatic environment (toxic to aquatic life with long- lasting effects) [312]. Questions regarding the safety of glyphosate-based herbicides are periodically raised, with a few recent independent studies indicating that it may not be as safe as previously assumed [274, 313-315].

Chapter 3 Review of Herbicide Glyphosate

3.5.7. Marine Organisms

Due to the extensive use of glyphosate-based herbicides, glyphosate and its major metabolite AMPA have frequently been detected in water bodies. Both glyphosate and

AMPA are water soluble and can persist in aquatic environments for several weeks

[24]. The GBR lies along the north-eastern coast of Australia and, due to its World

Heritage status, many more research and monitoring studies of insecticides have been carried out on it and the GBRCA (Great Barrier Reef Catchment Area) than on most equivalent areas in Australia. While previous studies focused on organochlorine insecticides, from 1999, many extensive studies along most of the GBR coast discovered insecticide residues, particularly herbicides such as diuron and atrazine, in coastal sediments, sea-grass and irrigation tailwaters [139, 140]. Further studies showed that all rivers discharging to the GBR which were examined and which drained agricultural and urban areas showed extensive insecticide contamination frequently above water quality guidelines [16, 141-149] while studies in the GBR lagoon waters also showed widespread insecticide contamination with levels occasionally above water quality guidelines [145, 150, 151].

According to a previous study [152], glyphosate was detected in about 30% and

AMPA in 50% of samples collected across Europe in 1993–2009. The highest observed concentrations were glyphosate 50µg/L and AMPA – 49µg/L. AMPA has usually been detected in higher concentrations and in a larger proportion of samples than glyphosate. Another study revealed that high concentrations of glyphosate (up to

328µg/L) have been recently reported in USA [316]. As previously mentioned, proposed environmental quality standards for freshwater ecosystems vary greatly (10–

Chapter 3 Review of Herbicide Glyphosate

27, 000µg/L). According to the Water Framework Directive [317], the permitted level of chemicals in surface water may be evaluated using a safety factor of 1000 for the lowest acute EC50 value from the applied test set.

In another study [24] reported that the short-term toxicities of glyphosate and its formulated products for aquatic organisms vary remarkably: (1) micro-organisms (3–

7 days’ EC50 0.64–590 mg Acid Equivalent (AE/L)); (2) macrophytes (7–14 days’

EC50 1.6–25.5 mg AE/L); (3) invertebrates (2–10 days’ EC50 7 – >1000 mg AE/L);

(4) fish (2–4 days’ LC50 5.8 – >1000 mg AE/L). The large variation in toxicity values may/can be explained mainly by the wide variety of glyphosate-based herbicides tested but also by the different test organisms, test conditions (temperature, test media) and test designs used. In addition, there are several different formulations for glyphosate which have analogous brand names (e.g., Roundup) and exhibit varying degrees of toxicity [318].

3.6. Conclusions

Currently, due to the increased adoption of GM crops in agriculture in Australia, the propagation of glyphosate in the environment continues to increase. Considering the worst-case scenario of a ten-fold increase in glyphosate usage in subsequent years, there are environmental concerns that it creates a potential threat to aquatic and marine ecosystems and human life, who are constantly exposed to it [319] which will not only affect users of preparations containing glyphosate but also those who do not have direct contact with it. Considering the widespread and frequent use of glyphosate, environmental monitoring studies are important.

Chapter 3 Review of Herbicide Glyphosate

This chapter provided an extensive review of the literature on the herbicide glyphosate, and its sorption and degradation. The toxicological impacts of it and its formulation products on the human body, and aquatic and marine systems were also discussed.

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.1. General

Worldwide concerns have been raised regarding the environmental implications of a dramatic increase in crop-based biofuel production and associated increases in herbicides use. The use of glyphosate, a broad-spectrum herbicide, for genetically modified (GM) crops production has increased considerably in recent years. The cost of measuring concentrations of organic contaminants using standard methods such as

High Pressure Liquid Chromatography (HPLC) can be expensive especially when large numbers of samples are involved. For example, those quoted for laboratory analyses in Australia are of the order of $150-$280 for each water sample and $200-

$280 for each soil sample (see later discussion and Appendix A). This cost can impose major constraints on the design of monitoring studies, especially in the university and agricultural sectors [39]. In the last few years, the development of analytical techniques for glyphosate quantification has been the suggest of considerable research and several alternative methods have been proposed [22, 40, 41, 43, 44, 320].

The first section of this chapter discusses the established analytical methods for glyphosate quantification. Experimental research is then reported on the establishment of a simple, fast and low cost direct fluorescence spectrometry method for glyphosate measurement in environmental samples. This section examines the sample extraction and analytical methods, development of the standard calibration curves and statistical

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

calculations including the acceptable operating limits (limits of detection and limits of quantification).The new fluorescence method is based on glyphosate quantification by direct fluorescence detection, involving sample extraction, derivatisation of glyphosate using 9-fluorenyl methoxycarbonyl chloride (FMOC-Cl), and measurement of emissions acquisition at 310 nm wavelength with 268 nm excitation wavelength. A reported low-cost UV-visible spectrophotometric method [54]was also developed for the determination of glyphosate in waters and soils but was found to have much lower precision. Comparative studies of the fluorescence method to previously established enzyme-linked immunosorbent assay (ELISA) method, UV- visible spectrophotometric method and also to high performance liquid chromatography mass spectrometry (HPLC-MS/MS) method was then reported, including of their accuracy and precision for the detection and quantification of glyphosate-spiked waters and soil samples.

4.2. Introduction and Background

Due to the lack of chromophore or fluorophore groups in the glyphosate structure it needs to be derivatised before analysis which causes glyphosate one of the most difficult herbicide to analyse. Glyphosate has three chemical groups: phosphonate, amine and carboxylate. The amine group reacts with FMOC-Cl in acetonitrile at pH 9 which gives the derivatised glyphosate as reaction product (Fig. 4.1).The analysis also exhibits great difficulties due to its chemical properties, highly solubility of water and polar nature [41, 321].

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

Fig. 4.1: Derivatisation reaction scheme for fluorometric analytical method [50]

To date, many different analytical methods for determining glyphosate residues in soil and water have been applied [22, 40, 41, 43, 320]. Those for analysing glyphosate include thin-layer chromatography (TLC) [222], gas chromatography (GC) [322], liquid chromatography (LC) [323], capillary electrophoresis [324, 325], electrochemiluminescence [324], conductivity detection [326], inductively coupled plasma–mass spectrometry (ICP–MS) [327], integrated pulsed amperometric detection (IPAD) [328], detection by immunosensors [329], UV-visible spectroscopy

[54, 330, 331] and the enzyme-linked immunosorbent assay (ELISA) method [332].

These methods are discussed below.

4.2.1. Gas chromatography

Gas chromatography (GC) is one of the most sensitive methods for glyphosate analysis. It involves a procedure whereby a sample is vaporised and injected onto the head of the chromatographic column and is then transported through the column by the flow of an inert, gaseous mobile phase. The column itself contains a liquid stationary phase which is adsorbed onto the surface of an inert solid. Glyphosate, being a very polar and non-volatile compound, has to be derivatised for GC analyses. In the literature, trifluoroacetic acid (TFAA), a trifluoroacetic anhydride–

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

trimethylorthoacetate reagent [333], N-methyl-N-(tert-butyldimethylsilyl) trifluoroacetamide (MTBSTFA) and 1% tertbutyldimethylchlorosilane (TBDMCS)

[301] have been used for this purpose. In most cases, derivatisation is performed to convert polar N-H, O-H and S-H groups into thermally stable, non-polar groups [45,

334]. GC derivatisation methods can be classified into four groups according to the reagents used and the reaction achieved, that is, silylation, acylation, alkylation and esterification.

Silylation is the most widely used derivatisation procedure for sample analysis by GC which involves replacement of active hydrogen by an alkylsilyl group, such as trimethylsilyl (TMS) or t-butyldimethylsilyl (t-BDMS). Silylation reagents are popular because they are easy to use and readily form derivatives.

Acylation reagents offer the same types of advantages available from silylation reagents: creating less polar, more volatile derivatives. The process however includes more readily targeting highly polar, multi-functional compounds, such as carbohydrates and amino acids. In addition, acylating reagents provide the distinct advantage of introducing electron capturing groups, thus enhancing detectability during analysis.

Alkylation derivatisation technique involves reduction of the molecular polarity by replacing active hydrogen with an alkyl group. These reagents are used to modify compounds having acidic hydrogen, such as carboxylic acids and phenols. Alkylation reagents can be used alone to form esters, ethers, and amides—or they can be used in

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

conjunction with acylation or silylation reagents. A two-step approach is commonly used in the derivatisation of amino acids, where multiple functional groups on these compounds may necessitate protection during derivatisation.

Esterification, the reaction of an acid with an alcohol in the presence of a catalyst to form an ester is the most popular method of alkylation. Alkylation reagents are available in several configurations that enable the formation of a variety of esters.

Alkyl esters are stable, and can be formed quickly and quantitatively.

4.2.2. Liquid chromatography

Liquid chromatography (LC) is a method in which the compounds are separated by different partitioning between stationary and mobile phases. Liquid chromatography generally performed on a column with a fritted bottom that holds a stationary phase in equilibrium with a solvent. The mixture to be separated is loaded onto the top of the column followed by more solvent. The different components in the sample mixture pass through the column at different rates due to differences in their partitioning behaviour between the mobile liquid phase and the stationary phase. The compounds are separated by collecting aliquots of the column effluent as a function of time. LC can be subdivided into different forms which are discussed below.

High-performance liquid chromatography (HPLC) is a form of liquid chromatography to separate compounds that are dissolved in solution. The underlying principles are governed by the van Deemter equation, which is an empirical formula that describes the relationship between linear velocity (flow rate) and plate height (HETP or column

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

efficiency). The particle size is one of the main variables, a van Deemter curve is normally used to investigate chromatographic performance. By using smaller particles, speed and peak capacity (number of peaks resolved per unit time in gradient separations) are extended to new limits, termed Ultra Performance Liquid

Chromatography, or UPLC. The instrumentation consists of a reservoir of mobile phase, a pump, an injector, a separation column, and a detector. Compounds are separated by injecting a plug of the sample mixture onto the column and the mixture pass through the different components of the column at different rates due to differences in their partitioning behaviour between the mobile liquid phase and the stationary phase. HPLC analysis of herbicides is most often performed on reversed phase columns [335] with fluorescence detection using precolumn derivatisation with

FMOC-Cl (9-fluorenylmethyl-chloroformate) to give the fluorescent derivative [336,

337]. 4-Fluoro-7-nitrobenzofurazan (NBD-F) [338], Phthaldialdehyde (OPA),

Trifluorethanol (TFE) and Trifluoracethicanhidride (TFAA) [339] are used for glyphosate derivatisation.

Liquid Chromatography Mass Spectrometry (LC/MS) is an analytical technique for identification, quantitation and mass analysis of a wide variety of non-volatile or semi- volatile organic or inorganic compounds in a mixture. Similar to HPLC, LC/MS allows a compound's intrinsic affinity for both a "mobile phase" (typically a buffered solvent) and a "stationary phase" (porous solid support with specialized coating). Normally, a pump is used to provide a continuous flow of a solvent into which a dissolved sample is introduced and when the sample is in the solvent flow, it travels through an analytical column. The separation of the compounds present in the sample mixture is dependent

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

on their affinity to the coated particles in the column. After the separation, the compound passes through a mass detector. For glyphosate quantification by LC-MS, a quite complicated derivatisation step has to be conducted to improve the chromatographic performance [50, 340].

4.2.3. Capillary electrophoresis

Capillary electrophoresis is an analytical technique that separates ions based on their electrophoretic mobility with the use of an applied voltage. This method is used mostly because of its faster results, provides high resolution separation and a useful technique with a large range of detection methods available The rate at which the particle moves is directly proportional to the applied electric field--the greater the field strength, the faster the mobility. The system consists of a high-voltage power supply, a sample introduction system, a capillary tube, a detector and an output device. Some instruments include a temperature control device as the separation of the sample depends on the electrophoretic mobility and the viscosity of the solutions decreases as the column temperature rises. Some previous studies were also conducted for determining glyphosate and AMPA using capillary electrophoresis with indirect detection method [341, 342].

The methodologies mentioned above for glyphosate quantification in environmental samples are not only expensive but can involve complicated, time consuming procedures. It would therefore be useful to have a simpler and faster method which is also of lower cost analysis than the above-mentioned chromatographic methods or capillary electrophoresis method. For developing a simple, fast and lower cost

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

glyphosate detection method attempts were made and reported in previous literature.

Several studies have examined glyphosate quantification by UV-visible spectroscopy

[54, 330, 331]. The colorimetric method proposed in 1981 [330] involves oxidation of glyphosate to orthophosphate with hydrogen peroxide and further measurement at 830 nm as the phosphomolybdate heteropoly blue complex. The method provides good results with the glyphosate concentration range of 1 to 20µg/mL but the major concern was using hydrogen peroxide which occasionally causes explosions during an evaporation step. Another method was developed in 2009 [331] based on the reaction of glyphosate with carbon disulphide to form dithiocarbamic acid, which was further followed by complex formation with copper in the presence of ammonia. The method can be applied successfully to the environmental samples with limits of detection and quantification of 1.1 and 3700 µg/L respectively with the requirement of a preconcentration step before measurement. In an recent developed method [54] using UV- visible spectrophotometry, glyphosate was first derivatised and the absorbance measured at 265 nm. The method involves a simple procedure but mostly suitable for high concentration in soil.

The aim of this chapter is to report the development of a simple, fast and low cost direct fluorescence spectrometric method is developed herein covering a range of glyphosate concentrations extending below the maximum acceptable concentration by different organizations e.g. Australian Drinking Water Guidelines (1µg/mL) [19],

National Environment Protection Measure (NEPM) guideline for fresh water

(0.37µg/mL). The method involves glyphosate extraction, derivatisation using 9- fluorenyl methoxycarbonyl chloride (FMOC-Cl), and measurement of fluorescence

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

emissions at 310 nm with 268 nm excitation wavelength. Details of the method are discussed below:

4.3. Establishment of the Analytical Method

4.3.1. Fluorescence spectroscopy

Fluorescence spectroscopy is a type of electromagnetic spectroscopy which measures the intensity of photons emitted from a sample after it has absorbed photons.

Fluorescence occurs when a fluorescent capable material (a fluorophore) is excited into a higher electronic state by absorbing an incident photon and cannot return to the ground state except by emitting a photon. The emission usually occurs from the ground vibrational level of the excited electronic state and goes to an excited vibrational state of the ground electronic state. Thus fluorescence signals occur at longer wavelengths than absorbance. The energies and relative intensities of the fluorescence signals give information about structure and environments of the fluorophores. Most fluorescent molecules are aromatic. Fluorescence is an important investigational tool in many areas of analytical science, due to its high sensitivity and selectivity. It can be used to investigate real-time structure and dynamics both in solution state and under microscopes, particularly for bio-molecular systems.

The necessary instrumentation required within a typical Fluorescence Spectrometer

(Spectrofluorometer) are a sample holder, incident photon source (typically a xenon lamp), monochromators used for selecting particular incident wavelengths, focussing optics, photon-collecting detector (single, or preferably multiple channel) and finally

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

a control software unit. An emission monochromator or cut-off filters are also usually employed. The detector is usually set at 90 degrees to the light source. The intrinsic sensitivity of fluorescence is also its biggest problem and care must be taken to record a true fluorescence signal of the analyte of interest.

A fluorescence emission spectrum (Fig. 4.2) is recorded when the excitation wavelength of light is held constant and the emission beam is scanned as a function of wavelength. An excitation spectrum is the opposite, whereby the emission light is held at a constant wavelength, and the excitation light is scanned as a function of wavelength. The excitation spectrum usually resembles the absorbance spectrum in shape. Most materials are not naturally fluorescent and those need to derivatise to make it fluorosent molecule. Several previous studies showed derivatised glyphosate quantification followed by chromatographic separation and further fluorescence detection [47-49].

Fig.4.2: Fluorescence Energy level diagram (Jablonski diagram)

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.3.2. Materials and methods

4.3.2.1. Reagents

Glyphosate (PESTANAL, analytical standard) and the derivatisation reagent, 9- fluorenyl methoxycarbonyl chloride (FMOC-Cl) (HPLC grade), were obtained from

Fluka (Germany). HPLC grade acetonitrile and diethyl ether were purchased from

Sigma-Aldrich Australia. All other chemicals including KCl, HCl, KOH, NaOH, disodium tetraboratedecahydrate (Na2B4O7.10H2O), potassium phosphate monobasic

(KH2PO4) and diethylether were analytical grade.

4.3.2.2. Solutions

The standard calibration curves for glyphosate quantification in environmental samples were prepared by using standard solutions of various concentration (Blank,

0.010, 0.020, 0.050, 0.065, 0.125, 0.180, 0.200, 0.450, 0.500, 0.600, 0.700, 0.800,

0.900, 1.000, 2.000, 2.500, 5.000, 10.000, 12.000, 15.000, 20.000, 22.000, 25.000

μg/mL) with analytical-standard glyphosate (PESTANAL, 99.729%) using distilledwater.0.1M KH2PO4 solution was prepared for the extraction procedure involved in soil sample preparation. The derivatisation reagent FMOC-Cl solutions of

1 g/L were prepared by dissolving the reagent in acetonitrile and were always prepared just before the experiments. Buffer solution (pH =9) was prepared by dissolving

15.255 g of Na2B4O7.10H2O in 1000mL of distilled water. 0.1 M EDTA solution was used for pre-treatment of samples to correct the sensitivity of the fluorogenic reagent to divalent ions in the amino-acid coupling [343].

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.3.2.3. Procedure

Water samples

The quantification of glyphosate in water samples was performed on samples filtered through a 0.45 µm membrane (cellulose acetate) using the filtration unit. After that the derivatisation procedure was performed as: 0.5 mL of borate buffer and 0.5 mL of

FMOC-Cl in acetronitrile were added in succession to each 3 mL sample. The mixture was then shaken in a mechanical shaker for 2 hours at room temperature. The resulting solution mixed with 4 mL of diethyl-ether and was placed in a separating funnel, shaken and the diethyl-ether separated from the sample. Different solvents were tested for their ability to remove excess reagent (FMOC-Cl) from the solution, which otherwise would have interfered with the fluorescence measurements. Diethylether gave the best results for efficiency and decantation facilities than CH2Cl2 [54] and ethylacetate [344]. The extraction of the FMOC-Cl with diethylether was performed six times in succession to obtain reliable results without the interference of organic matter present in the samples.

Soil samples

For the soil samples, glyphosate was determined after extraction from control soil samples of 20 g of each with 0.1 M KH2PO4; agitation (1 h agitation+ 0.5 h break + 2 h agitation +0.5 h break + 1 h agitation); centrifugation (4000 rpm; 17 min) and filtration through the filtration unit. To avoid the interference from the impurities in soil 0.1 M KH2PO4 solution has been used here as alternative extraction solvents rather than NaOH. As the literature suggests induced phosphorus provide leaching of glyphosate from soil as phosphorus and glyphosate compete for the same sorption site

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

[265]. Also the solution was made in water which was used as a solvent for glyphosate extraction in previous studies [345].The extraction was repeated three times on solid residue to obtain a 20 mL extract from each sample, and the extracts filtered (0.45 µm membrane). The derivatisation procedure performed was similar as for water samples with 2 hours of mechanical agitation. In case of soil samples an additional 0.1 M EDTA solution was used for pre-treatment of samples to correct the sensitivity of the fluorogenic reagent to divalent ions in the amino-acid coupling. The resulting solution then mixed with 4 mL of diethyl-ether, shaken, and separated from the sample. For calculating the glyphosate concentration in soil samples the results in aqueous solutions were multiplied by an appropriate factor:

Conc. in solution (µg/mL) × volume of extractant (mL) Glyphosate conc. in soil, mg/kg = Mass of soil (kg) × 1000 (µg/mg)

Filtration

The filtration procedure was conducted using a filtration unit. At first about 2 mL of the water/soil/extraction solution mixture was poured into the bottom container of the filter device then the plunger rod containing the filter was attached to the bottom holder of the filter device. For the filtrate to flow more quickly gentle pressure was continuously applied to the plunger handle. The filtrate containing the extracted glyphosate was transferred to glass vials with Teflon stoppers.

Method blank

All the results were quantified after the blank correction. For water samples, the blank solution was prepared by treating 3 mL of distilled water with 0.5 mL of borate buffer,

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

0.5 mL of FMOC-Cl and 4 mL of diethylether, shaken and centrifuged as described above. For soil samples the blank solution was prepared by using 3 mL of the extract from analyte free soil treating with 0.5 mL of borate buffer, 0.5 mL of FMOC-Cl and

4 mL of diethylether, shaken and centrifuged.

Spectrum recordings

The aqueous phase, which contained the derivatised product was quantified by fluorescence spectrometry in a Spectrofluorometer (Fluoromax-3) with a fluorescence quartz cuvette. The derivatisation step by fluorescent maker molecule FMOC-Cl made the reaction product to emit the absorbed energy at certain excitation wavelength. In this experiment the excitation wavelength was chosen from the excitation acquisition plot (shown in Fig. 4.3) in which the glyphosate derivatised product showed the maximum peak at 268 nm wavelengths. Afterwards the emission acquisition was recorded with scan starting from 290 nm to 450 nm with the excitation wavelength of

268 nm.

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

Fig.4.3: Excitation acquisition of the derivatised glyphosate

The characteristics of the fluorescence spectra of the derivatised glyphosate are illustrated in Fig. 4.4 which shows its emission acquisition of extracts prepared by the derivatisation procedure. The results correspond to solutions of different glyphosate concentrations, specifically 0.010, 0.020, 0.050, 0.065, 0.125, 0.180, 0.200, 0.450,

0.500, 0.600, 0.700, 0.800, 0.900, 1.000, 2.000, 2.500, 5.000, 10.000, 12.000, 15.000,

20.000, 22.000 and 25.000 μg/mL. For glyphosate quantification, the emission acquisitions (Fig. 4.4) of the derivatised product were taken from measurements at peaks of 310 nm. The first peak was not considered since it arose from the Raman band for water due to Raman light scattering and not fluorescence. The peak intensities at 310 nm of the data are plotted against glyphosate concentrations in Figures 4.5 to

4.7 in both linear and log-log plots.

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

Fig. 4.4: Emission acquisition of derivatised glyphosate for water samples at

different concentrations

Fig. 4.5: Standard calibration curves for fluorescence spectrometry of glyphosate in waters (all waters)

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

Fig. 4.6: Standard calibration curves for fluorescence spectrometry of glyphosate in waters (all waters up to 1.0 µg/mL)

Fig. 4.7: Standard calibration curves for fluorescence spectrometry of glyphosate in

waters (log-log scale)

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.3.3. Method validation and results

To assure that a new analytical method generates reliable and interpretable information about the sample, it must undergo an evaluation denominated validation. Previous studies were conducted regarding the validation of analytical methods [346-349], which describe definitions, procedures, parameters and strategies of validation. The parameters for method validation have been defined in different working groups of national and international committees and are described in the literature [350-356].

The analytical parameters normally involves are: calibration curve and linearity, detection limit, quantification limit, precision and accuracy. To quantify the effectiveness of the fluorescence spectrometric method for the analysis of water and soil samples the calculated validation parameters are discussed below.

4.3.3.1. Calibration curve and linearity

Different calibration curves were prepared using the following four different models

(where I represents intensity × 107 and C the glyphosate concentration in µg/mL).

Full range: 0.010 to 25.000 µg/mL

Model I 퐼 = 2.9437퐶 + 0.2369

Model II 퐼 = 3.7038 퐶0.8521

Low range: 0.010 to 1.000 µg/mL

Model III 퐼 = 3.4661퐶 + 0.0899

Model IV 퐼 = 2.989 퐶0.7551

The parameters for these models were obtained by either linear or power-law regressions, with their curves illustrated in Figures 4.5 to 4.7 by both linear and log- log plots. The calibration standard concentrations were then back-calculated by

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

inserting their peak intensities into each model to determine the mean accuracy of each standard (Table 4.1 to 4.4) which produced the following reproducibilities and correlation coefficients:

Model 1: recovery 108.316±20.758; R2=0.997

Model II: recovery 105.536±27.447; R2=0.986

Model III: recovery 115.230±25.852; R2=0.998

Model IV: recovery 114.711±27.904; R2=0.970

As evident from the figures and correlation coefficients, Model I was a more acceptable fit with the calibration data over their entire range of 0.010 to 25.000 µg/mL while Model III was more acceptable over the lower range of 0.010 µg/mL to 1.000

µg/mL.

4.3.3.2. Limits of detection (LOD) and limits of quantification (LOQ)

Estimation of LOD and LOQ was made according to the approach of standard deviation of the response [357]. This approach is based on the standard deviation of a response (s) and the slope of the analytical curve (b). The detection and quantification limits are, thus:

푠 퐿푂퐷 = 3.3 × ⁄푏 (4.1) 푠 퐿푂푄 = 10 × ⁄푏 (4.2)

The standard deviation of the response was determined based on the standard deviation of the y-intercept of the regression line. The LODs and LOQs for the proposed method calculated using four different models are presented in Table 4.5. As is evident, Model

III provided the lowest LOD and LOQ for glyphosate analysis in water of 0.010 µg/mL and 0.029 µg/mL respectively and those from Model I were 0.632 µg/mL and 0.194

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

µg/mL respectively. Since the latter model is more appropriate for a soil analysis, applying the main conversion factor of 1/10 used in the current study (refer to the discussion in Chapter 5) yields a LOD and LOQ for the analysis of glyphosate in soils of 0.063 mg/kg and 0.194 mg/kg respectively.

4.3.3.3. Precision and accuracy

The precision and accuracy of the method were determined by analysing replicates of the control sample containing concentrations of standard solutions. Six determinations of the same concentrations were conducted over three runs on six different days to determine the between-run precision and accuracy of the method (Tables 4.1 to 4.4).

The accuracy and precision of the method are reported as the percentage recovery and standard deviation (SD) respectively in Tables 4.1, 4.2, 4.3, 4.4 and 4.5. In addition, laboratory control samples (LCS) with analyte-free matrices (water samples with de- ionised water and clean soil samples with no glyphosate) to analyse spikes in those with known concentrations of glyphosate. An assessment of each LCS was undertaken by calculating its percentage recovery (%) from the spike. The results are provided in the comparative study sections (Table 4.6 and Table 4.7).

4.3.4. Discussion

From the results obtained from the four different models, it was considered that Model

I could be used with good results in terms of accuracy and precision for samples with higher aqueous concentrations (e.g, extracts obtained from soil analyses and sorption isotherms) and Model III for samples with lower ones (e.g., environmental water

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

samples). Therefore, these two models were selected for the subsequent analyses conducted and reported in Chapters 5 and 6.

Table 4.1: Back calculated concentrations of the calibration standards and calculated precision and accuracy of the fluorescence

spectrometric method (water samples using Model I)

Glyphosate Fluorescence spectrometric method Concentration, Intensity × 107 Calculated concentration % Recovery µg/mL Determination Mean Determination Mean Determination Mean 0.010 0.108, 0.106, 0.110, 0.131 -0.044, -0.045, -0.043 -0.036 438.368, 445.156, 359.348 0.157, 0.146, 0.160 -0.027, -0.031, -0.026 430.664, 271.815, 309.272, 260.81 0.020 0.142, 0.137, 0.147, 0.167 -0.032, -0.034, -0.031 -0.024 160.756, 169.928, 118.847 0.198, 0.183, 0.195 -0.013, -0.018, -0.014 152.943, 65.804, 91.748, 71.902 0.050 0.229, 0.2265, 0.232 0.267 -0.003, -0.004, -0.002 0.010 5.260, 7.484, 3.622 26.211 0.305, 0.313, 0.300 0.023, 0.026, 0.021 46.036, 51.957, 42.908, 0.065 0.305, 0.303, 0.307 0.355 0.023, 0.023, 0.024 0.040 35.413, 34.740, 36.665 61.609 0.347, 0.346, 0.521 0.037, 0.037, 0.096 57.677, 56.898, 148.262 0.125 0.480, 0.426, 0.521 0.515 0.083, 0.064, 0.096 0.094 66.067, 51.328, 77.096 75.500 0.486, 0.588, 0.589 0.084, 0.119, 0.119 67.579, 95.354, 95.576 0.180 0.626, 0.675, 0.652 0.694 0.132, 0.149, 0.141 0.155 73.408, 82.649, 78.341 86.311 0.766, 0.705, 0.742 0.180, 0.159, 0.172 99.830, 88.311, 95.326 0.200 0.721, 0.737, 0.700 0.769 0.164, 0.170, 0.157 0.181 82.147, 84.872, 78.703 90.400 0.821, 0.837, 0.800 0.198, 0.204, 0.191 99.132, 101.857, 95.689, 0.450 1.650, 1.668, 1.630 1.668 0.480, 0.486, 0.473 0.486 106.681, 108.040, 105.166 107.997 1.680, 1.688, 1.689 0.490, 0.493, 0.493 108.946, 109.550, 109.601 0.500 1.806, 1.785, 1.820 1.820 0.533, 0.526, 0.538 0.538 106.584, 105.185, 107.559 107.575 1.866, 1.825, 1.820 0.553, 0.540, 0.538 110.660, 107.903, 107.559 0.600 2.216, 2.226, 2.232 2.224 0.672, 0.676, 0.678 0.675 112.045, 112.613, 112.934 112.531 0.700 2.531, 2.455, 2.588 2.524 0.779, 0.753, 0.799 0.777 111.312, 107.627, 114.116 111.019 0.800 2.820, 2.876, 2.945 2.880 0.878, 0.896, 0.920 0.898 109.688, 112.051, 114.998 112.246

Glyphosate Fluorescence spectrometric method Concentration, Intensity × 107 Calculated concentration % Recovery µg/mL Determination Mean Determination Mean Determination Mean 0.900 3.269, 3.269, 3.266 3.268 1.030, 1.030, 1.029 1.030 114.440, 114.454, 114.330 114.408 1.000 3.484, 3.409, 3.503 3.545 1.103, 1.077, 1.110 1.110 110.307, 107.742, 110.962 112.388 3.584, 3.549, 3.503 1.137, 1.125, 1.110 113.704, 112.498, 110.962 2.000 6.970, 5.877, 6.653 6.499 2.287, 1.916, 2.180 2.128 110.307, 107.742, 110.962 106.379 2.500 7.068, 7.1085, 6.980 7.052 2.321, 2.334, 2.291 2.315 114.365, 95.796, 108.978 92.605 5.000 14.136, 13.968, 14.980 14.361 4.722, 4.665, 5.008 4.798 92.821, 93.365, 91.628 95.964 10.000 28.693, 29.762, 27.650 28.702 9.667, 10.030, 9.312 9.670 94.431, 93.294, 100.167 96.697 12.000 38.776, 39.884, 35.725 38.128 13.092, 13.468, 12.056 12.872 96.667, 100.300, 93.125 107.267 15.000 44.862, 42.937, 40.260 42.686 15.160, 14.506, 13.596 14.421 109.101, 112.237, 100.463 96.137 20.000 56.736, 58.645, 60.600 58.660 19.193, 19.842, 20.506 19.847 95.966, 99.209, 102.529 99.235 22.000 65.678, 62.368, 64.750 64.265 22.231, 21.107, 21.916 21.751 101.049, 95.939, 99.617 98.868 25.000 79.696, 71.003, 74.590 75.096 26.993, 24.040, 25.258 25.430 107.972, 96.159, 101.034 101.722 Mean = 108.316±20.758a a = Standard deviation

Table 4.2: Back calculated concentrations of the calibration standards and calculated precision and accuracy of the fluorescence

spectrometric method (water samples using Model II)

Glyphosate Fluorescence spectrometric method Concentration, Intensity × 107 Calculated concentration % Recovery µg/mL Determination Mean Determination Mean Determination Mean 0.010 0.108, 0.106, 0.110 0.131 0.016, 0.015, 0.016 0.020 157.389, 153.972, 161.281, 198.601 0.157, 0.146, 0.160 0.024, 0.022, 0.025 244.356, 224.320, 250.290 0.020 0.142, 0.137, 0.147 0.167 0.022, 0.021, 0.023 0.026 108.914, 104.077, 113.060, 131.726 0.198, 0.183, 0.195 0.032, 0.029, 0.031 160.720, 146.276, 157.307 0.050 0.229, 0.2265, 0.232 0.267 0.038, 0.037, 0.039 0.046 76.250, 74.972, 77.192, 91.612 0.305, 0.313, 0.300 0.053, 0.055, 0.052 106.521, 110.107, 104.634 0.065 0.305, 0.303, 0.307 0.355 0.053, 0.053, 0.054 0.064 81.939, 81.533, 82.696 98.413 0.347, 0.346, 0.521 0.062, 0.062, 0.100 95.548, 95.067, 153.694 0.125 0.480, 0.426, 0.521, 0.515 0.091, 0.079, 0.100 0.099 72.657, 63.117, 79.921 78.969 0.486, 0.588, 0.589 0.092, 0.115, 0.115 73.647, 92.161, 92.311 0.180 0.626, 0.675, 0.652 0.694 0.124, 0.135, 0.130 0.140 68.898, 75.268, 72.287 77.854 0.766, 0.705, 0.742 0.157, 0.143, 0.151 87.323, 79.211, 84.137 0.200 0.721, 0.737, 0.700, 0.769 0.146, 0.150, 0.141 0.158 73.158, 75.074, 70.747 79.020 0.821, 0.837, 0.800 0.170, 0.174, 0.166 85.217, 87.176, 82.750 0.450 1.650, 1.668, 1.630, 1.668 0.387, 0.392, 0.382 0.392 86.008, 87.111, 84.782, 87.078 1.680, 1.688, 1.689 0.395, 0.398, 0.398 87.847, 88.339, 88.381 0.500 1.806, 1.785, 1.820, 1.820 0.430, 0.424, 0.434 0.434 86.045, 84.895, 86.848 86.863 1.866, 1.825, 1.820 0.447, 0.436, 0.434 89.411, 87.132, 86.848 0.600 2.216, 2.226, 2.232 2.224 0.547, 0.550, 0.552 0.550 91.185, 91.669, 91.444 91.600 0.700 2.531, 2.455, 2.588 2.525 0.639, 0.617, 0.657 0.638 91.346, 88.137, 93.800 91.095 0.800 2.820, 2.876, 2.945 2.880 0.726, 0.743, 0.764 0.744 90.763, 92.870, 95.507 93.047

Glyphosate Fluorescence spectrometric method Concentration, Intensity × 107 Calculated concentration % Recovery µg/mL Determination Mean Determination Mean Determination Mean 0.900 3.269, 3.269, 3.266 3.268 0.864, 0.864, 0.863 0.863 95.953, 95.965, 95.852 95.923 1.000 3.484, 3.409, 3.503, 3.505 0.931, 0.907, 0.937 0.937 93.070, 90.706, 93.675 93.739 3.584, 3.549, 3.503 0.962, 0.951, 0.937 96.213, 95.096, 93.675 2.000 6.970, 5.877, 6.653 6.500 2.101, 1.719, 1.989 1.936 105.035, 85.970, 99.446 96.817 2.500 7.068, 7.1085, 6.980 7.052 2.135, 2.150, 2.104 2.130 85.415, 85.982, 84.169 85.189 5.000 14.136, 13.968, 14.980 14.361 4.818, 4.751, 5.158 4.909 96.363, 95.026, 103.155 98.181 10.000 28.693, 29.762, 27.650 28.702 11.062, 11.548, 10.592 11.067 110.620, 115.475, 105.915 110.670 12.000 38.776, 39.884, 35.725 38.128 15.754, 16.283, 14.309 15.449 131.281, 135.695, 119.238 128.738 15.000 44.862, 42.937, 40.260 42.686 18.695, 17.756, 16.464 17.638 124.631, 118.376, 109.759 117.589 20.000 56.736, 58.645, 60.600 58.660 24.628, 25.604, 26.609 25.614 123.142, 128.021, 133.046 128.070 22.000 65.678, 62.368, 64.750 64.265 29.245, 27.523, 28.761 28.510 132.934, 125.105, 130.732 129.590 25.000 79.696, 71.003, 74.590 75.096 36.702, 32.048, 33.957 34.236 146.809, 128.194, 135.830 136.944 Mean =105.536±27.447a a = Standard deviation

Table 4.3: Back calculated concentrations of the calibration standards and calculated precision and accuracy of the fluorescence

spectrometric method (water samples using Model III).

Glyphosate Fluorescence spectrometric method Concentration, Intensity × 107 Calculated concentration % Recovery µg/mL Determination Mean Determination Mean Determination Mean 0.010 0.157, 0.146, 0.160 0.154 0.019, 0.016, 0.020 0.019 193.260, 161.448, 202.606 185.771 0.020 0.198, 0.183, 0.195 0.192 0.031, 0.027, 0.030 0.029 156.168, 134.134, 150.989 147.097 0.050 0.305, 0.313, 0.300 0.306 0.062, 0.064, 0.061 0.062 123.919, 128.948, 121.263 124.710 0.065 0.347, 0.346, 0.521 0.405 0.074, 0.074, 0.124 0.091 114.232, 113.570, 191.164 139.655 0.125 0.486, 0.588, 0.589 0.554 0.114, 0.144, 0.144 0.134 91.322, 114.911, 115.100 107.111 0.180 0.766, 0.705, 0.742 0.738 0.195, 0.177, 0.188 0.187 108.345, 98.563, 104.520 103.809 0.200 0.821, 0.837, 0.800 0.819 0.211, 0.215, 0.205 0.210 105.397, 107.711, 102.472 105.193 0.450 1.680, 1.688, 1.689 1.686 0.459, 0.461, 0.461 0.460 101.950, 102.463, 102.507 102.307 0.500 1.866, 1.825, 1.820 1.837 0.512, 0.501, 0.499 0.504 102.464, 100.122, 99.830 100.805 0.600 2.216, 2.226, 2.232 2.224 0.613, 0.616, 0.618 0.616 102.227, 102.708, 102.982 102.639 0.700 2.531, 2.455, 2.588 2.525 0.704, 0.682, 0.721 0.702 100.594, 97.565, 102.976 91.600 0.800 2.820, 2.876, 2.945 2.880 0.788, 0.804, 0.824 0.805 98.457, 100.464, 102.967 100.345 0.900 3.269, 3.269, 3.266 3.268 0.917, 0.917, 0.916 0.917 101.904, 101.916, 101.811 100.630 1.000 3.584, 3.549, 3.503 3.545 1.008, 0.998, 0.985 0.997 100.808, 99.784, 98.479 101.877 Mean =115.230±25.852a a = Standard deviation

Table 4.4: Back calculated concentrations of the calibration standards and calculated precision and accuracy of the fluorescence

spectrometric method (water samples using Model IV)

Glyphosate Fluorescence spectrometric method Concentration, Intensity × 107 Calculated concentration % Recovery µg/mL Determination Mean Determination Mean Determination Mean 0.010 0.157, 0.146, 0.160 0.154 0.020, 0.018, 0.021 0.020 202.002, 183.423, 207.543 197.656 0.020 0.198, 0.183, 0.195 0.192 0.028, 0.025, 0.027 0.026 137.600, 123.736, 134.309 131.882 0.050 0.305, 0.313, 0.300 0.306 0.049, 0.050, 0.048 0.049 97.275, 100.976, 95.334 97.862 0.065 0.347, 0.346, 0.521 0.405 0.058, 0.058, 0.099 0.071 88.985, 88.479, 152.099 109.854 0.125 0.486, 0.588, 0.589 0.554 0.090, 0.116, 0.116 0.108 72.125, 92.880, 93.051 86.019 0.180 0.766, 0.705, 0.742 0.738 0.165, 0.148, 0.158 0.157 91.569, 82.034, 87.347 87.138 0.200 0.821, 0.837, 0.800 0.819 0.181, 0.185, 0.175 0.180 90.289, 92.634, 87.347 90.090 0.450 1.680, 1.688, 1.689 1.686 0.466, 0.469, 0.470 0.468 103.639, 104.293, 104.349 104.094 0.500 1.866, 1.825, 1.820 1.837 0.536, 0.520, 0.518 0.525 107.155, 104.080, 103.697 104.977 0.600 2.216, 2.226, 2.232 2.224 0.673, 0.677, 0.679 0.676 112.138, 112.810, 113.191 112.713 0.700 2.531, 2.455, 2.588 2.525 0.802, 0.770, 0.827 0.800 114.596, 110.066, 118.074 114.245 0.800 2.820, 2.876, 2.945 2.880 0.926, 0.950, 0.981 0.952 115.729, 118.763, 122.573 119.022 0.900 3.269, 3.269, 3.266 3.268 1.126, 1.126, 1.124 1.125 125.087, 125.105, 124.939 125.044 1.000 3.584, 3.549, 3.503 3.545 1.272, 1.255, 1.234 1.254 127.171, 125.506, 123.393 125.356 Mean =114.711±27.904a a = Standard deviation

Table 4.5: Method validation parameters

Model I Model II Model III Model IV Recovery % 108.316 ± 20.758 105.536 ± 27.447 115.230 ± 25.852 114.711 ± 27.904 Slope 2.9437 - 3.4661 - Intercept 0.2369 - 0.0899 - Base - 3.7038 - 2.989 Exponent - 0.8521 - 0.7551 Regression range, (µg/mL) 0.010-25.000 0.010-25.000 0.010-1.000 0.010-1.000 Correlation coefficient 0.997 0.986 0.998 0.970 Total error (µg/mL) 0.564 174.7 0.010 5.608 Limit of detection, LOD in waters (µg/mL) 0.632 155.6 0.010 6.191 Limit of quantification, LOQ in waters (µg/mL) 1.914 471.7 0.029 18.761 Limit of detection LOD in soil, (mg/kg)a 0.063 15.6 0.001 0.619 Limit of quantification LOQ in soil, (mg/kg) a 0.194 47.2 0.003 1.876 a = based on conversion factor of 1/10

4.4. Comparative Study

4.4.1. General

In order to validate the method, the analytical results obtained with the proposed method were calibrated against and compared with a commercially available enzyme- linked immunosorbent assay (ELISA) method, for the analysis of glyphosate in soils and waters. This was conducted under controlled laboratory condition using the procedure in Abraxis Glyphosate Plate Kit purchased from ABRAXIS, LLC. A second method for the analysis of water and soil samples with higher concentration of glyphosate, involving a spectrophotometric method [54] was developed and is reported here. For the comparative study samples were spiked with known concentration of glyphosate and then analysed with three different methods.

4.4.2. Enzyme-Linked Immunosorbent Assay (ELISA) method

Several previous studies reported the ELISA method as a cost-effective method capable of detecting insecticides in water at concentrations below certain established water quality guidelines [332, 358, 359], and has been shown to compare quantitatively and qualitatively with wet chemistry analysis [332, 359, 360]. However,

ELISA techniques can suffer from false positives based on cross reactivity with other organic contaminants [361-363]. ELISA method uses antibodies and colour change to identify a substance. Antigens from the sample are attached to a surface. Then, a further specific antibody is applied over the surface so it can bind to the antigen. This antibody is linked to an enzyme, and, in the final step, a substance containing the enzyme's substrate is added. The subsequent reaction produces a detectable signal, most commonly a colour change in the substrate.

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.4.2.1. Principle

The Abraxis Glyphosate Plate Kit (Fig. 4.8) applies the principles of the ELISA to the determination of glyphosate. The sample tested was derivatised, with an antibody specific for glyphosate to microtitre wells coated with Goat Anti-rabbit Antibody added and incubated for 30 minutes. Then, the glyphosate enzyme conjugate was added, at which point a competitive reaction occurred between the glyphosate in the sample and the enzyme-labelled glyphosate analogue for the antibody binding sites on the microtitre well. This reaction was allowed to continue for sixty (60) minutes and, after a washing step and the addition of a substrate (coloured solution), a coloured signal (blue) was generated.

Fig. 4.8: ELISA plate’s 96 test wells

The presence of glyphosate was detected by adding the ‘coloured solution’ containing the enzyme substrate (hydrogen peroxide) and chromogen (3, 3', 5, 5’- tetramethylbenzidine). The enzyme-labelled glyphosate bound to the glyphosate antibody to catalyse the conversion of the substrate/chromogen mixture to a coloured

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

product. After an incubation period, the reaction was stopped and stabilised by the addition of a diluted acid (stopping solution). Since the labelled glyphosate (conjugate) was in competition with the unlabelled glyphosate (sample) for the antibody sites, the colour developed was inversely proportional to the concentration of glyphosate in the sample.

4.4.2.2. Reagents

The Abraxis Glyphosate Plate Kit contains; microtitre plate coated with goat-anti rabbit antibody, glyphosate antibody solution, glyphosate enzyme conjugate, glyphosate standards of six concentrations (0.000, 0.075, 0.200, 0.500, 1.000, 4.000 ppb), control(approximately 0.750 ppb of glyphosate in distilled water with a non- mercury preservative and stabilizers), diluent/zero standard (sample diluent), color solution, stopping solution, washing buffer 5X concentrate, assay buffer, derivatisation reagent, derivatisation reagent diluents.

4.4.2.3. Sample preparation

Water

Determination of this herbicide in water was done on samples filtered through 0.45µm pore size filter paper to remove particles. For samples having glyphosate concentration greater than 4 ppb repeat testing were performed using a diluted sample. The dilution of the samples was done with an appropriate amount of zero standards. Afterward these dilution factors (1:100 and 1:500) were multiplied to the actual results found from the experiment.

100

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

Soil

For soil samples the extraction procedure was performed using 1M 20 mL NaOH solutions into the soil collection bottle, vigorously and continuously shake for 30 minutes and waited at least 5 minutes for the particles to settle down. The filtration procedure performed here is the same as discussed in the fluorescence spectrometric method. For samples having concentration higher than 4ppb were diluted with appropriate amount of zero standards and results were further multiplied by the dilution factor (1:100, 1:500 and 1:5000). Glyphosate concentration in soil was determined by multiplying the assay results with an appropriate factor provided in the previous section.

4.4.2.4. Reagent preparation

All reagents were allowed to come to room temperature.

Wash buffer

In a 1000 mL container, the wash buffer concentrate 1:5 was diluted by the addition of deionised or distilled water (i.e. 100 mL of wash buffer 5 × concentrate plus 400 mL of water.

Derivatisation of standards, control, and samples

 The derivatisation reagent was diluted using the glass syringe with 3.5 mL of

derivatisation reagent diluent and was thoroughly mixed. ((Diluted reagent was

prepared prior to use within the same day)

 All the test tubes were labelled for standards, control, and samples.

101

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

 By using the precision pipette capable of delivering 250 µL solution, standard,

control, and sample were taken into separate disposable tubes.

 1.0 mL of assay buffer was added prior to vortex for mixing.

 100 µL of the diluted derivatisation reagent was added to each tube followed

by an immediate vortex after addition of reagent until no swirling lines were

present.

 All the tubes were incubated at room temperature for 10 minutes.

 Afterward the ELISA was performed by the following the assay procedure.

4.4.2.5. Assay procedure

 At first, 50 μL of the appropriate derivatised standard, control, or sample was

added in to the plate.

 Anti-Glyphosate antibody solution (50 μL) was then successively added to

each well and covered with parafilm. The mixing of the contents was done by

moving the strip holder in a circular motion on the bench top. After that the

solution was incubated at room temperature for 30 minutes.

 After the incubation, the covering was removed and 50 μL of enzyme

conjugate solution was added to the individual wells successively. Again the

wells were covered with parafilm and mixed by moving the strip holder in a

circular motion on the bench top. After that the solution was incubated at room

temperature for 60 minutes.

 The covering was again removed and the contents of the wells were vigorously

shaken into a waste container. All the strips were washed 3 times using the 1X

wash solution with a volume of at least 250 μL per each wash step. The

102

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

remaining buffer in the wells was removed by patting the plate on a dry stack

of paper towels.

 Afterwards, 150 µL of colour solution was added successively to each well and

incubated for 20-30 minutes followed by the addition of Stopping Solution

(100 µL).

 Finally, the solution was then transferred to a micro cuvette with 0.1cm optical

length and the absorbance reading was measured using a UV-visible

spectrophotometer at 450 nm within 15 minutes after adding the stopping

solution.

4.4.2.6. Results

The mean absorbance value for each standard (B) was obtained from the UV-visible spectrum and the percentage ratio of it to the mean absorbance value of the diluent

(B0) was calculated and plotted against the concentration of glyphosate to obtain a five-point calibration curve (Fig. 4.9) which was then used to calculate the concentrations in different samples (soil and water). The method detection limit

(MDL) as defined by Abraxis (0.1 µg/L for water and 0.038 µg/mL for soil) was simply used for reporting data. The results obtained from the spiked samples are provided in table 4.6 and 4.7 and the comparison graphs are provided in Fig. 4.11.

103

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

a) Absorption spectra of glyphosate by ELISA method at different concentrations

(b) constructed calibration curve

Fig. 4.9: ELISA standard plot

104

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.4.3. UV-visible spectrophotometric method

Ultraviolet–visible spectroscopy

Ultraviolet–visible spectroscopy or ultraviolet-visible spectrophotometry (UV-Vis or

UV/Vis) refers to absorption spectroscopy or reflectance spectroscopy in the ultraviolet-visible spectral region. It involves measuring the amount of ultraviolet or visible radiation absorbed by a substance in solution. The absorption or reflectance in the visible range directly affects the perceived colour of the chemicals involved. In this region of the electromagnetic spectrum, molecules undergo electronic transitions.

Instrument which measure the ratio, or function of ratio, of the intensity of two beams of light in the UV-Visible region are called Ultraviolet-Visible spectrophotometers.

The fundamental law that governs the quantitative spectrophotometric analysis is the

Beer -Lambert law.

Beer’s law

It states that the intensity of a beam of parallel monochromatic radiation which decreases exponentially with the number of absorbing molecules. In other words, absorbance is proportional to the concentration.

Lambert’s law

It states that the intensity of a beam of parallel monochromatic radiation which decreases exponentially as it passes through a medium of homogeneous thickness. A combination of these two laws yields the Beer-Lambert law

105

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

Beer-Lambert law

When beam of light is passed through a transparent cell containing a solution of an absorbing substance, reduction of the intensity of light may occur. Mathematically,

Beer- Lambert law is expressed as

퐼 퐴 = 푙표푔 ( 0) = 휀푐퐿 (4.3) 10 퐼 where

A = the measured absorbance

I0 = the intensity of the incident light at a given wavelength

I = the transmitted intensity, L the path length through the sample

c = concentration of the absorbing species

ε = a constant known as the molar absorptivity or extinction coefficient.

Previous studies were reported for glyphosate quantification by UV-visible spectroscopy [54, 330, 331]. A recent established direct UV-visible spectrophotometric method [54] was also used in conjunction with the ELISA method for comparison with the fluorescence spectrometric results. The method involves a derivatisation step and further measurement of the absorbance at 265 nm. The calibration curve was constructed using glyphosate concentrations of 0.010, 0.020,

0.050, 0.065, 0.125, 0.180, 0.200, 0.450, 0.500, 1.000, 2.000, 2.500, 5.000, 10.000,

12.000, 15.000, 20.000, 22.000 and 25.000 µg/mL (shown in Fig. 4.10). The LOD and

LOQ of were calculated and the values obtained are 2.54µg/mL and 7.70 µg/mL.

Results obtained from the samples spiked with known concentration and analysed with this is provided in Table 4.6 and Table 4.7.

106

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

(a) UV-visible absorbance spectra of derivatised glyphosate at different

concentrations

(b) constructed calibration curve

Fig. 4.10: UV-visible Spectrophotometric standard plot

107

Table 4.6: Comparison of different glyphosate detection methods for water

Glyphosate Recovery by ELISA method Recovery by Fluorometric method Recovery by Spectrophotometric method Spiked con, Peak %B/B0 Mean Found con. % Recovery Found conc. % Recovery Peak mean Found con. % Recovery µg/mL Mean, x 107 0.010 92.391 1.700 0.012 120.000 0.114 0.007 70.406 ND - - 0.020 85.379 1.600 0.022 110.000 0.152 0.018 89.667 ND - - 0.050 64.034 1.200 0.053 105.000 0.311 0.064 127.582 ND - - 0.065 59.765 1.120 0.065 100.000 0.359 0.078 119.446 ND - - 0.090 49.626 0.930 0.129 143.000 0.549 0.132 147.176 ND - - 0.180 44.824 0.840 0.183 101.500 0.687 0.172 95.708 ND - - 0.200 41.622 0.780 0.210 105.000 0.727 0.184 91.907 ND - - 0.450 54.429 1.020 0.454 100.956 1.689 0.461 102.526 ND - - 0.500 53.362 1.000 0.513 102.540 1.849 0.508 101.506 ND - - 0.600 50.160 0.940 0.580 96.667 2.216 0.613 102.236 ND - - 0.700 48.026 0.900 0.695 99.271 2.447 0.680 97.152 ND - - 0.800 46.958 0.880 0.809 101.125 2.587 0.720 90.057 ND - - 0.900 44.824 0.840 0.886 98.444 2.969 0.831 92.297 ND - - 1.000 41.622 0.780 1.000 99.970 3.486 0.980 97.983 ND - - ND = Not Detected

108

Table 4.7: Comparison on different glyphosate detection method for soil

Recovery by HPLC/MS-MS Glyphosate Recovery by ELISA method Recovery by Fluorescence method Spectrophotometric method by NMI lab Found Found Peak Found Spiked con, con. In % Found con. In Peak con. In % % %B/B0 Mean Mean, x % Recovery conc. µg/mL soilb recovery soilb(mg/kg) Mean soilb Recovery Recovery 107 (mg/kg) (mg/kg) (mg/kg) 0.010 129.348 ND ND ND ND - - ND - - 0.020 120.598 ND ND ND ND - - ND - - 0.050 66.702 1.250 0.005 96.000 ND - - ND - - 0.065 57.631 1.080 0.007 100.000 ND - - ND - - 0.125 49.626 0.930 0.011 88.480 0.464 0.011 86.347 ND - - 0.180 42.156 0.790 0.018 101.389 0.650 0.016 89.777 ND - - 0.200 41.089 0.770 0.021 103.050 0.681 0.017 85.271 ND - - 0.450 55.496 1.040 0.045 99.689 1.350 0.036 80.791 0.500 53.362 1.000 0.054 108.980 1.510 0.041 81.945 ND - - 1.000 42.689 0.800 0.099 99.000 2.720 0.084 84.353 ND - - 0.110 110.000 2.000 30.950 0.580 0.193 96.500 5.450 0.178 88.547 ND - - 0.180 90.000 2.500 65.101 1.220 0.246 98.400 7.095 0.233 93.190 ND - - 5.000 54.429 1.020 0.510 102.000 13.561 0.453 90.526 0.375 0.447 91.000 10.000 42.689 0.800 0.994 99.400 27.102 0.913 91.263 0.788 0.920 91.300 12.000 39.488 0.740 1.175 97.917 31.128 1.049 87.450 0.885 1.031 87.400 15.000 34.685 0.650 1.498 99.867 41.686 1.408 93.871 1.205 1.397 93.800 20.000 27.855 0.522 2.049 102.450 55.693 1.884 94.195 1.673 1.933 94.000 22.000 27.962 0.524 2.293 104.227 64.299 2.176 98.920 1.870 2.159 98.700 25.000 27.428 0.514 2.589 103.560 65.096 2.203 88.133 1.900 2.193 87.900

ND = Not Detected; NMI = National Measurement Institute b = Glyphosate concentration in soil = [(Assay result × Volume of extractant (mL)) / Weight of soil (kg)] × Dilution factor Note: Initial soil samples taken was 200g and which was used all through the calculations to match the same quantity of spiked soil samples send to the NMI lab

109

Water sample analyses Soil sample analyses

Fig.4.11: Comparison of ELISA and fluorescence spectrometric results

Fig.4.12: Comparison of UV-visible spectrophotometric and fluorescence

spectrometric results for soil samples

110

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

4.5. Costing of Analytical Method

Two quotes from a commercial laboratory for analysis of glyphosate in soil and water samples by the standard (HPLC) method, obtained during the period of this thesis, are included in Appendix A. These respectively indicate the cost of analysis to be AU$150 to AU$280 per sample for water samples, and AU$200 to AU$280 per soil sample. As indicated throughout this thesis, these costs are substantial, severely limiting the ability of interested parties – especially in the agricultural sector – to conduct comprehensive environmental studies of glyphosate in soils and waters.

For comparative purposes, a detailed cost estimate of the fluorescence spectrophotometric analytical method at a commercial analytical laboratory, allowing for all estimated costs, is summarised in Table 4.8. The reported figures are based on

Australian labour rates (Professionals Australia [364]; ABS [365]) in 2014 Australian dollars, excluding Goods and Services Tax (GST), with all other assumptions stated in

Table 4.8. The calculation suggests that the underlying cost per sample (before profit margin) for a commercial laboratory is AU$47.42. This is substantially lower than the above-quoted prices for analysis of glyphosate in soils and waters at a commercial laboratory, of AU$150-280 per sample. Even if allowance is made for a higher-than- the-Australian-average rate of labour on-costs of, say, 25% (as against the 12% figure implicit in the calculations in Table 4.8), and if allowance is also made for a gross profit margin of 50%, the costs to the end user of the method described in this dissertation are well below those of the lower of the two quotes provided. The lower cost of fluorescence spectrophotometry is primarily a consequence of its substantially lower equipment capital costs, maintenance costs and reagent costs compared to

111

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

HPLC. The fluorescence spectrophotometric method developed in this thesis for glyphosate analysis therefore offers substantial cost savings over the standard commercial method.

112

Table 4.8: Cost analysis of the new analytical method (Commercial Laboratory)

COST ESTIMATE FOR ANALYSIS BY COMMERCIAL LABORATORY

Costs are calculated on a 2 day per week basis, assuming 40 samples per week, based on 48 working weeks per year. All cost estimates in 2014 Australian dollars, excluding GST.

ITEM Total per Item total Subtotal year (AU$) (AU$) Salary Component Laboratory technician, salary $75,500 per year (reference: Professionals Australia [364], 75500.00 629.17 median salary for 2014), pro rata for 2 days, assuming 48 weeks per year Laboratory technician, on-costs (including superannuation contributions, payroll tax, 9149.78 76.25 worker’s compensation insurance, fringe benefits tax; reference ABS [365], from lower half of Table 4, entitled Costs per Employee, for Professional, Scientific and Technical Services, given by ($71112-$62494)/$71112, applied pro rata.) Subtotal 705.41 Cost of Building Cost of laboratory space [366] (assuming 5 m2) at AS$500 per m2 per week, pro rata for 2 1000.00 days Cost of electricity and water (assuming 10% of the total) 100.00 Subtotal 1100.00 Operating Equipment and Experimental Fluoromax-3 ( AU$10,000 with life span of 10 years), pro rata for 2 days 1000.00 8.33 Refrigerator ( AU$450 with life span of 5 years), pro rata for 2 days 90.00 0.75 Glassware, pro rata for 2 days 300.00 2.50

113

Computer ( AU$1500 with life span of 3 years), pro rata for 2 days 500.00 4.17 Holder (AU$600 with life span of 3 years), pro rata for 2 days 200.00 1.67 Vacuum unit( AU$450 with life span of 5 years), pro rata for 2 days 90.00 0.75 Allowance for breakage and maintenance at 10% 1.82 Subtotal 19.98 Material cost Glyphosate (PESTANAL, analytical standard) (for 1000 sample), pro rata for 2 days or 40 1500.00 60.00 samples (FMOC-Cl) (HPLC grade) (for 1000 sample), pro rata for 2 days or 40 samples 140.00 5.60 HPLC grade acetonitrile (1000 samples), pro rata for 2 days or 40 samples 45.00 1.80 Other chemicals (total) (for 1000 samples), pro rata for 2 days or 40 samples 100.00 4.00 Subtotal 71.40

GRAND TOTAL, per 2 days or 40 samples 1896.80 GRAND TOTAL, per sample 47.42

114

4.6. Conclusions

The simple, fast and low cost fluorescence spectrometric method developed herein may allow others in future to conduct broad-scale monitoring programs to assess whether levels of herbicides such as glyphosate are increasing in soils and surface waters, due to the increased cultivation of genetically modified crops over the last few years in Australia. This method also provides a cost-effective alternative for improving temporal and spatial monitoring, which would probably allow greater flexibility for monitoring agencies to develop broad scale projects for determining the environmental fate and transport of glyphosate. Also the experimentation technique involved is relatively simpler and faster than the other chromatographic methods. It is also substantially cheaper: the cost estimate for the new method is AU$47.42 per sample at a commercial laboratory in 2014 Australian dollars, excluding gross profit margins and excluding GST, but including all capital depreciation, maintenance, reagent and labour costs. This can be compared to commercial rates of AU$150-280 excluding

GST using the standard method. For illustration, for a sampling program involving the collection of 100 samples for glyphosate analysis over the course of the growing season, use of the standard chromatographic method would cost of the order of

AU$15000-28000, whereas analysis using the fluorescence spectrometric method would cost approximately AU$4742. This allows a substantial cost saving over the existing method.

The validation of the analytical method with commercially available enzyme-linked immunosorbent assay (ELISA) method show good agreement in the found accuracy and precision. The percentage recoveries for water and soil analysis (Table 4.6 and

Table 4.7) for both methods suggest there was no significant difference between the

115

Chapter 4 Establishment of Analytical Method for Glyphosate Detection

values of the new low cost fluorescence spectrometric method and the previously established ELISA. Also the graphical (Fig. 4.11 and Fig. 4.12) presentation for both water and soil analysis on shows a linear line 450 insect. The comparison with the established UV-visible spectrometric method represents that concentration below

2.5µg/mL was not detectable by this method and suitable only for soil-sorption studies with much higher concentrations. Because of the funding problem only few soil samples were sent to outside National Measurement Institute, Sydney for HPLC-

MS/MS analysis.

To start with the content of this chapter the main section involves establishment of a simple, fast and low cost fluorescence spectrometric method for glyphosate detection in soil and water samples. In this sense some background studies were also discussed in the very first section. The quantification of the analytical method was conducted with prepared standard calibration graphs and calculating the different operating limits. Later on a comparison study has been conducted for the validation of the new method with ELISA, direct UV-visible spectrometric and HPLC-MS/MS method.

116

Chapter 5 Site Investigation

5.1. General

Following the worldwide introduction of genetically modified (GM) crops in the mid-

1990s, more than 100m hectares (ha) of land have been planted with different varieties, with around 5% of it covered with canola (Brassica napus) [367]. The widespread cultivation of GM crops has been adopted mainly by farmers in countries which allow these crops to be grown commercially and it has been reported that, due to their simpler and more effective weed control, of the different varieties of GM canola, herbicide- tolerant ones are the most popular [173, 368]. Previous studies showed that, in Canada, more than 90% of canola production is of herbicide-tolerant varieties and it covers an area of nearly 6 Mha [173, 174, 369].

In Australia, canola has been grown since the 1970s in rotation with winter cereals

[169] and, before the introduction of glyphosate-tolerant varieties, its production was heavily reliant on triazine-tolerant (TT) ones. Later, Roundup-ready (RR) varieties provided an alternative technology for incorporation in farming systems. In New South

Wales (NSW), GM canola was first planted in 2008 by 108 growers and approximately

9600 ha were cultivated in NSW and Victoria in that year [167] while, in 2009, its uptake in NSW increased four-fold with over 41, 000 ha planted [167]. Most of these areas are planted with RR canola and 93% of GM canola growers have rated its weed control as excellent compared with much lower ratings for the Clearfield, TT and conventional canola systems [168].

117

Chapter 5 Site Investigation

This chapter provides information regarding a case study of GM canola cultivation for which, firstly, GM canola growers from different parts of NSW were surveyed and then a family-operated farm in the agricultural district of Parkes in the central west of

NSW selected for a site investigation. The first section presents the views and experiences of different GM canola growers in NSW and then the site investigation conducted to determine the amounts of the herbicide glyphosate in the soils and surface waters forming the agricultural land is described. Sample analysis program and quality assurance (QA) assessment, including a detailed analysis of quality control (QC), are also discussed. All investigations were conducted in accordance with the consent and ethics approval procedures of the University of NSW.

5.2. GM Canola Growers’ Views and Experiences

GM canola has been grown in NSW since 2008. To perform this farm-level case study, growers were first contacted through different farmers’ organisations and then questionnaire surveys of different NSW farms growing GM canola, which included farmers’ views and experiences regarding the planting of GM canola, were conducted.

This information and those provided by the Birchip Cropping Group GM Canola grower case studies are given below.

5.2.1. Background on farms

The first GM canola farm surveyed is located in southern NSW and covers approximately 4000 ha of farmed land on which the farmer and his family have worked for a long time (approximately 50 years). Prior to 1970s, this farm was split equally between sheep and crop production but is now about 20 per cent grazed and 80 per

118

Chapter 5 Site Investigation

cent cropped. It was first planted with canola in the late 1980s and produces yields of between 1.9 and 2.5 tonnes per ha. At the time the survey was conducted, according to the farmer, canola was valued at around AU$400 per tonne compared with wheat at

AU$150 per tonne. This farm consists of wheat, canola, 1000 merino ewes and Angus cattle, and has an average annual rainfall of 450 mm, with information from the nearest post office indicating that the average rainfall in the growing season is 278 mm. Its soil classification ranges between red loam and grey self-mulching cracking clays, with pHwater of between 4.5 and 6.

The second farm surveyed is located in the Riverina district of NSW and has been farmed by the family for about 90 years, with the first 60 to 70 years focused on grazing. On it, grain production began in the 1970s, with its intensity increasing in the late 1980s to 1990s and reaching the current cropped area of 85% of its 3,200 ha, of which about one-third is canola. The farming system is no-till farming using press wheels, narrow points, direct drilling and more powerful machinery. This farm includes wheat, canola, triticale and a small flock of Dohne sheep (African merinos) and has an average annual rainfall of 425 mm, with the average growing season rainfall

290 mm. Its soil types vary from predominantly red-brown earth to brown-grey clay interspersed with black cracking clay with a small area of loam, and its pH is around

4.8 to 5.5. It has no major subsoil constraints for canola as no effects appear on yield maps while around 15-20% of the land, which is on red soils, is increasingly acidic at increasing depths.

Located in the central west of NSW, the third farmer works in his wife’s family farming business which is further considered for a site investigation of glyphosate

119

Chapter 5 Site Investigation

contamination from January to October 2012 based on different glyphosate applications interval. GM canola plantation has been occurring from the last 8 years.

As a young farmer, he has farmed over 2000ha of GM canola during the last five years after starting his business cropping 34000ha of winter cereals, canola and legumes in a continuous cropping rotation utilising no-till and control traffic management practices. In addition, he and his wife run a small prime lamb business consisting of

400 Dorper ewes. According to a USDA textural classification chart this farm’s soil types vary from loam to brown-grey clay or red-brown earth, and its average annual rainfall is584.6 mm, with an average growing season rainfall of 200 mm.

5.2.2. Crop sequence used for GM canola plantation

Canola is grown as both a break crop and for its profitability. Although lupin has been the only other break crop to be successfully grown and produce similar yields to canola, as its lower prices have made it far less profitable, often providing only about half the gross margin of canola, it has not been grown since the 1990s. According to one farmer, he rotates between TT and RR canola, with two wheat crops between (TT canola-wheat-wheat-RR canola-wheat-wheat) which allows for more effective rotations of herbicide groups. The other two usually rotate between RR canola-wheat- wheat or RR canola-wheat-canola which depend on drought conditions.

120

Chapter 5 Site Investigation

5.2.3. Adoption of RR canola

5.2.3.1. Positives

Regarding their preference for RR rather than TT canola or Clearfield canola, they explained that RR canola is effective in controlling weeds, including group A-resistant ryegrass. They find it provides more flexibility than TT canola in dry seasons as a herbicide can be applied after rather than before rain, and it also reduces their use of residual herbicides which have caused problems in the past. The farmers find that this system allows for early sowing with no-till farming and that the yields of GM varieties are roughly 10% higher than those of TT varieties. Although late germination of ryegrass can occur after the six-leaf stage in RR canola, it is possible to spray these weeds which one farmer explained he was able to do without any obvious yield loss.

However, this is not recommended as results can be unpredictable and, according to the herbicide label “applications may be made in RR canola from crop emergence to the six-leaf stage”.

They explained that they adopted RR canola because of its enhanced vigour, and higher yields and profit than TT canola which is another useful tool in weed management. They suggested that undecided growers should not be afraid to try a small area but advised that they pay attention to detail. According to their perspective, one of the main changes that should be made when farming with RR canola is to avoid using glyphosate the year after the canola is planted by using gramoxone or spray-seed instead and a group B herbicide to clean up volunteers.

121

Chapter 5 Site Investigation

Another very important reason they raised for growing RR canola was that it allows them to reduce their reliance on simazine and atrazine which some of them dislike due mainly to their residues; for example, following a TT canola crop in the dry year of

1992, one farmer lost half of one of his wheat crops through poor establishment due to triazine residues. By contrast, glyphosate does not leave any residues and can be applied after rain. The TT system carries additional risk in that, as simazine and atrazine need soil moisture to work, in dry years, the efficacy of the triazine herbicides is compromised if forecast follow-up rains do not occur. Also, the yields of RR and conventional canola are generally higher than those of TT canola.

5.2.3.2. Negatives

According to one farmer, dust compromises the efficacy of glyphosate in a dry year and weather-damaged GM grain is more difficult to sell than non-GM grain, and receiving sites are limited. As one farmer explained, 1000 tonnes of his second-grade

(CAN-02) canola received an additional AU$40 to AU$50 penalty because it was mixed with much lower-quality canola (CAN-03) at the site. For him, the cost of freight to the CAN-01 receiving sites was AU$2 to AU$3 a tonne more than to the closer non-GM sites. In 2009 and 2010, GM canola attracted lower prices than equivalent non-GM canola, with the penalty in 2009 about AU$5 per tonne and, in

2010, about AU$30 to $35 per tonne.

5.2.4. Performances of different varieties

To test the performances of different varieties of canola, the farmers tried them for different time durations. As they explained, they first grew RR hybrid canola GT61 in

122

Chapter 5 Site Investigation

2008 when it was the only open-pollinated (OP) RR variety available. According to one farmer, its yield exceeded those of the TT varieties and, in one year, was exactly double although there were differences in maturity which would have been a factor.

They also grew other varieties for different time periods, such as the early maturing

OP varieties, GT Scorpion, GT Taipan and GT Cougar, and hybrids Hyola-404RR,

Hyola-502RR and Hyola-505RR.

5.2.5. Sowing system(s) used

Regarding the system they used to sow canola plants, the farmers explained that most of the canola planting was conducted at a rate of 1.7-3 kg per ha in a no-till system with row spacing varying between 15 and 30 cm. They normally used a Jaenchke bar with narrow points and press wheels, with the seeds dropped onto the surface and inserted into the soil by the press wheels. They usually started canola planting in the middle of April although this depended on the soil’s moisture content as sowing usually requires dry conditions.

5.2.6. Gross margins and variable costs

One farmer provided information that RR canola gave him a gross margin of AU$96 to $104 per ha which was more than that of TT canola in 2010, although it depended on the number of applications of RR herbicide, and was due mainly to the 10 per cent higher yield of RR canola but also lower herbicide costs (AU$8.50 per ha for one spray compared with AU$20 per ha for simazine and atrazine). However, this was somewhat countered by the higher seed costs (AU$9/ha), lower grain price (AU$45/ha) and end- point royalty and grain technology fee (about AU$39/ha). According to another

123

Chapter 5 Site Investigation

farmer, the gross margin for RR canola in the years it was grown averaged AU$226/ha compared with that of TT canola (excluding the worst drought years) of AU$168/ha.

5.2.7. Problem weeds

The most prevalent weed found in canola plantations is annual ryegrass while there are others, such as wild oats, wild radish, volunteer cereals, black oats and brome grass and the broadleaf weed group of which the main ones are Indian hedge mustard (wild mustard), shepherd’s purse and wild turnip.

5.3. Site Investigation

5.3.1. Study area: Parkes, NSW

5.3.1.1. Background details

After the survey of different growers of GM canola in NSW was conducted, a family- operated farm in the agricultural district of Parkes in the central west of NSW was selected for detailed study (Figs. 5.1, 5.2, 5.3, 5.4 and 5.5). The objective of this investigation was to assess the environmental contamination across the site caused by the main herbicide used in GM canola cultivation, glyphosate. The farmer works in his wife’s family farming business, and the land had been used for the previous five years for continuous cropping rotations with wheat, winter cereals, legumes and GM canola utilising no-till and control traffic management practices. Over 1500 ha of GM canola had been planted in that time, and about 410 ha last year, with waste associated with the applications of herbicides, insecticides and fertilisers during both preparation of the land and the growing season. The region is characterised by very slight undulations cut by shallow creeks, with characterisation studies finding that 0.93 - 1.1% of organic

124

Chapter 5 Site Investigation

matter is present in the surface soils and a pH (CaCl2) of 4.15 - 5.61. The average annual rainfall is 584.6 mm and the average growing season rainfall 200 mm. The scope of the investigation included sampling and analysing soils and surface waters from the GM canola paddock.

5.3.1.2. Potential chemical contamination

Based on discussions with the farmer, the main source of potential contamination is from the past and present use of different herbicides, insecticides and fertilisers for land preparation and during the overall growing season. Of the different chemicals applied, this study focused mainly on contamination from glyphosate which is the major herbicide used in GM canola agriculture.

125

Fig. 5.1: Location of the site investigation

126

Fig. 5.2: Glyphosate-contaminated sampling points on Parkes plains

127

Fig. 5.3: Soil sampling locations

128

Fig. 5.4: Soil sampling locations

129

Fig. 5.5: Location of surface water sampling from creek

130

5.3.2. Sowing, herbicide application and rainfall events

Fig. 5.6 shows the sowing event and herbicide applications as well as daily rainfall during the period studied. The sowing of GM canola was conducted on 27/04/12 after three applications of glyphosate which was then applied twice more during the overall crop rotation. The dates of the herbicide applications and sowing were obtained from the farmer and precipitation data from the Bureau of Meteorology.

Fig. 5.6: Precipitations, sowing and herbicide applications in fields (G1 to G5) and

sampling times (Episodes 1 to 5)

131

Chapter 5 Site Investigation

5.3.3. Sampling and analysis program

The sampling and analysis program was designed after a quick survey of the land which incorporated an overview of existing site conditions, the surrounding environment and areas of potential concern regarding glyphosate contamination. The purpose of the program was to characterise the soils and surface waters across the site in accordance with the relevant NEPM sampling schedule guideline, with particular emphasis on glyphosate contamination.

The sampling program consisted of collecting:

 soil samples from the glyphosate applied area (Parkes plains) (Fig. 5.1 and Fig.

5.2) after different episodes at different times of the year (Episodes 1 to 5 in

Fig. 5.6 and Table 5.1), to allow take into account glyphosate applications and

rainfall events, from sampling depths of up to 30 cm.

 surface water samples from three different locations (L1, L2 and L3) of

drainage lines during different episodes to allow for glyphosate applications

and rainfall events (Fig. 5.2)

 surface water samples from creeks where glyphosate runoff from the land can

occur (Fig. 5.5); and

 soil samples from two other locations in addition to the Parkes plains for

glyphosate sorption and leaching studies conducted in Chapter 6 and Chapter

7 (Fig. 5.3 and Fig. 5.4).

132

Chapter 5 Site Investigation

5.3.4. Sample collection

5.3.4.1. Water sampling

Water samples were collected during different episodes, to take account of glyphosate applications and rainfall events, from three locations: 1 (L1), adjacent to the GM canola cultivation area; 2 (L2), where drainage lines formed on the paddock; and 3

(L3), at the end of the paddock where a wetland was formed by the water streams flowing through the cultivated area (Fig. 5.2). Also, water samples from a few centimetres below the water surface were collected from creeks using 50 mL clear glass vials (COSPAK) which were filled approximately half-full, immediately packed in an Esky to which ice had been added and stored in the dark in order to inhibit degradation [370]. It appeared that there would be no issue of sorption to glass as, according to the literature, there is no difference between collecting glyphosate in glass or plastic [371]. At the end of each sampling day, the samples in the Esky were transported to the University laboratory where more ice was added, and then analysed within one working day. All containers were labelled with the sample and study numbers, and date collected.

5.3.4.2 Soil sampling

A random sampling pattern was chosen for soil sampling as the site is very large and other sampling patterns could create complexities. In accordance with the NEPM

(1999) Schedule B(2) Guideline on Data Collection, Sample Design and Reporting

[53], samples were collected from different locations at different times of the year (Fig.

5.7) to allow for glyphosate applications and rainfall events. All samples were collected from depths of up to 30 cm directly from the hand auger using new nitrile

133

Chapter 5 Site Investigation

gloves and immediately placed in acid-washed glass jars. Care was taken to ensure that representative samples were obtained from the depths required and sample integrity was maintained. The samples were each labelled with a unique sample identification number, together with the date of collection and study number, before being placed on ice within an Esky. At the end of each sampling day, the Eskies were returned to the University laboratory where they remained overnight.

For environmental samples (waters and soils) analysis the sampling for different episodes were conducted on the same targeted portion of the glyphosate contaminated land. For glyphosate fate and transport analysis soil samples were also collected from two other locations in addition to this land.

Also, separate soil samples were collected and analysed to determine the major soil properties, including texture, major cations and other constituents, exchangeable ions, total Fe and total Al.

(a) (b) (c)

Fig. 5.7: Details of soil sampling: (a) point near canola crop; (b) collecting surface

soil; (c) sampling up to 30 cm

134

Chapter 5 Site Investigation

Table 5.1: Identification of samples taken from study area

Episode Description Month Description Episode 1 Before crop During Water samples from L1, L2, L3, Creek 1 rotation January and Creek 2 Soil samples from different points in paddock (Soil 1) Episode 2 After second End of Water samples from L1, L2, L3, Creek 1 application of February and Creek 2 glyphosate Soil samples from different points in paddock (Soil 2) Episode 3 After rainfall Beginning Water samples from L1, L2, L3, Creek 1 of March and Creek 2 Soil samples from different points in paddock (Soil 3) Episode 4 After fourth Middle of Water samples from L1, L2, L3, Creek 1 application of July and Creek 2 glyphosate Soil samples from different points in paddock (Soil 4) Episode 5 After rainfall Beginning Water samples from L1, L2, L3, Creek 1 of August and Creek 2

135

Chapter 5 Site Investigation

5.4. Soil Characterisations

The soil used for chemical analysis was sieved through a 2 mm sieve and air- or oven- dried (1050 C), with its basic physico-chemical properties determined before it was treated with glyphosate.

5.4.1. Soil texture (hydrometer test)

The soil was sieved through a 2mm sieve and oven-dried at 105°C for 24 hours, with its texture determined by the hydrometer method described below.

5.4.1.1. Apparatus and materials

Glass cylinders, each with a 1000 mL capacity, and a thermometer, hydrometer (A-

060), electric mixer with a dispersing cup, plunger and balance were used.

5.4.1.2. Reagents

The dispersing solution was 5%: 50 g of sodium hexametaphosphate, with Na6(PO3)6 dissolved in de-ionised water and diluted to 1 litre.

5.4.1.3. Method

 The blank was prepared by mixing 100 mL of the 5% dispersing solution and

880 mL of de-ionised water in a 1000 mL cylinder, and was not diluted to 1000

mL as the other 20 mL was the volume occupied by 50 g of soil.

 50 g of the soil was transferred to the dispersing cup along with 100 mL of the

5% dispersing solution.

 The dispersing cup was attached to the mixer and mixed for 30–60 sec.

136

Chapter 5 Site Investigation

 The suspension was transferred quantitatively from the dispersing cup to a

1000 mL cylinder and filled to the 1000 mL mark with de-ionised water and

equilibrated to room temperature or allowed to stand overnight to equilibrate.

 For each set of samples, the temperature and hydrometer readings of the blank

were recorded using the procedure described below.

 To determine the density, the plunger was inserted into the suspension and

carefully mixed for 30 sec until a uniform suspension was obtained. Then, the

plunger was removed (a 40-sec timer begun) and the hydrometer gently

inserted into the suspension.

 The hydrometer reading was recorded at 40 sec, by which time the sand had

settled to the bottom of the cylinder, which indicated the amount of silt and

clay suspended. This procedure was repeated for each sample.

 The hydrometer reading was recorded again after 6 hours, 52 minutes, by

which time the silt had settled to the bottom of the cylinder, which indicated

the amount of clay in the suspension.

5.4.1.4. Calculations

Corrections for meniscus, dispersing agent and temperature effects were taken into account for each of the sample hydrometer tests.

Particle size (hydrometer)

% clay = (corrected hydrometer reading at 6 hrs, 52 min × 100) / mass of sample

% silt = [(corrected hydrometer reading at 40 sec × 100) /mass of sample] - % clay

% sand = 100% - % silt - % clay

137

Chapter 5 Site Investigation

5.4.2. Soil Taxonomy (ST) and World Reference Base (WRB) system

There are two modern soil classification systems generally regarded as having worldwide application, the World Reference Base (WRB) [372] and USDA Soil

Taxonomy (ST) [373]. Neither has been widely used in Australia although some state authorities have identified major soils in soil survey reports in terms of ST [374] and there have been various attempts to correlate ST with Australian soils [375].

As a step towards stimulating interest in utilising international soil classification systems in Australia, particularly the WRB, soil data collected during soil surveys within the Parkes region of New South Wales have been utilised to correlate the

Australian Soil Classification (ASC) (Table 5.2) with the WRB (Table 5.3) and ST

(Table 5.4). As per the information provided in different tables (Tables 5.2, 5.3, 5.4 and 5.5), the soils are characterised as Anthrosols (WRB system), Chromosols (ASC) and Alfisols or Mollisols (ST).

138

Chapter 5 Site Investigation

Table 5.2: Summary of the Australian Soil Classification (ASC) orders [376]

Description Order

‘Human-made’ soils Anthroposols Organic soil material Organosols Negligible pedological organisation Rudosols Weak pedological organisation Tenosols Bs, Bh or Bhs horizons Podosols Clay > 35%, cracks, slicken sides Vertosols Prolonged seasonal saturation Hydrosols

Strong texture-contrast pH<5.5 in B horizon Kurosols pH ≥5.5 in B horizon Chromosols pH ≥5.5 in B horizon Sodosols

Lacking strong texture contrast Calcareous throughout Calcarosols High free-iron B horizon Ferrosols Structured B horizon Dermosols Massive B horizon Kandosols

139

Chapter 5 Site Investigation

Table 5.3: Rationalised key to the WRB reference soil groups [372]

Description Reference soil 1 Soils with thick organic layers: Histosols 2 Soils with strong human influence Soils with long and intensive agricultural use: Anthrosols Soils containing many artefacts: Technosols 3 Soils with limited rooting due to shallow permafrost or stones Ice-affected soils: Cryosols Shallow or extremely gravelly soils: Leptosols 4 Soils influenced by water Alternating wet–dry conditions, rich in swelling clays: Vertisols Floodplains, tidal marshes: Fluvisol Alkaline soils: Solonetz Salt enrichment upon evaporation: Solonchaks Groundwater-affected soils: Gleysols 5 Soils set by Fe/Al chemistry Allophanes or Al-humus complexes: Andosols Cheluviation and chilluviation: Podzols Accumulation of Fe under hydromorphic conditions: Plinthosols Low activity clay, P fixation, strongly structured: Nitisols Dominance of kaolinite and sesquioxides: Ferralsols 6 Soils with stagnating water Abrupt textural discontinuity: Planosols Structural or moderate textural discontinuity: Stagnosols 7 Accumulation of organic matter, high base status Typically mollic: Chernozems Transition to drier climate: Kastanozems Transition to more humid climate: Phaeozems 8 Accumulation of less soluble salts or non-saline substances Gypsisols Gypsum: Durisols Silica: Calcisols Calcium carbonate: 9 Soils with a clay-enriched subsoil Albeluvic tonguing: Albeluvisols Low base status, high-activity clay: Alisols Low base status, low-activity clay: Acrisols High base status, high-activity clay: Luvisols High base status, low-activity clay: Lixisols 10 Relatively young soils or soils with little or no profile development With an acidic dark topsoil: Umbrisols Sandy soils: Arenosols Moderately developed soils: Cambisols Soils with no significant profile development: Regosols

140

Chapter 5 Site Investigation

Table 5.4: The orders of Soil Taxonomy [373]

Description Order

Indicators of translocation of silicate clays without Alfisols excessive leaching of bases Dominance of short-range-order minerals or aluminium- Andisols humus complexes that result from weathering and mineral transformation with a minimum of translocation— generally form on recent volcanic ejecta Soils of dry areas (lack of water available to mesophytic Aridisols plants for very extended periods) Absence of distinct pedogenic horizons Entisols Presence of permafrost and other features associated with Gelisols freezing and thawing Very high content of organic matter Histosols Soils with adequate water available for plants and one or Inceptisols more pedogenic horizons—can grade towards any other soil order Very dark brown to black, structured surface horizon, and Mollisols high base saturation Extreme weathering of most minerals other than quartz to Oxisols kaolin and iron oxides B horizon consisting of an accumulation of black or Spodosols reddish amorphous materials that have a high cation exchange capacity Indicators of translocation of silicate clays and intensive Ultisols leaching Clay soils that shrink and swell Vertisols

141

Chapter 5 Site Investigation

Table 5.5: Australian Soil Classification and international equivalents [377]

(Values in parentheses are the number of profiles classified; total is 175)

Australian soil World Reference Base Soil Taxonomy classification

Chromosols (5) Luvisols (3), Lixisols (1), Alfisols (3), Phaeozems (1) Mollisols (2) Dermosols (18) Luvisols (5), Acrisols (3), Mollisols (10), Phaeozems (6), Cambisols (2), Ultisols (4), Lixisols (2) Inceptisols (3), Alfisols (1) Ferrosols (11) Nitisols (8), Acrisols (3) Oxisols (6), Ultisols (3), Alfisols (2) Hydrosols (24) Gleysols (8), Fluvisols (2), Inceptisols (10), Planosols (2), Acrisols (1), Alfisols (8), Lixisols (2), Solonchaks (6), Mollisols (3), Stagnosols (2), Regosols (1) Entisols (3)

Kandosols (13) Acrisols (2), Lixisols (3), Alfisols (6), Gleysols (2), Cambisols (2), Mollisols (3), Phaeozems (1), Ultisols (2), Luvisols (1), Planosols (1), Inceptisols (2) Plinthosol (1) Kurosols (51) Acrisols (27), Lixisols (9), Alfisols (28), Planosols (6), Luvisols (5), Ultisols (21), Solonetz (3), Mollisols (2) Plinthosol (1)

Organosols (5) Histosols (5) Podosols (18) Arenosols (11), Podzols (6), Entisols (12), Phaeozem (1) Spodosols (6) Rudosols (5) Regosol (1), Fluvisols (2), Entisols (5) Arenosol (1), Leptosol (1) Sodosols (4) Solonetz (2), Lixisols (2) Alfisols (4) Tenosols (11) Regosols (3), Lixisols (1), Inceptisols (3), Arenosols (4), Phaeozems (1), Entisols (6), Leptosols (1), Mollisols (1), Cambisols (1) Alfisols (1)

Vertosols (9) Vertisols (9) Vertisols (9) Anthroposol (1) Arenosol (1) Entisol (1)

142

Chapter 5 Site Investigation

5.4.3. Moisture content

As the results for the soil analysis were calculated on the basis of an oven-dried sample weight, the moisture analysis was executed before any other analysis. An aluminium dish was placed in an oven at a temperature of 105°C, left for 2 hours and then cooled to room temperature in a desiccator, with the weight of the empty dish recorded. Then, a soil sample was weighed and put in the dish, the dish placed for 12 hours in the oven at 105°C and then cooled to room temperature in a desiccator and weighed again. The moisture content was then calculated using the following equations [378].

(퐵−퐶)×100 M (moisture content) = (5.1) (퐵−퐴) where

A: empty crucible weight

B: sample + dish weight before drying

C: sample + dish weight after drying

5.4.4. pH

The pH of the soil was measured potentiometrically in a 1:5 soil–water suspension and

CaCl2 suspension. Twenty grams of an oven-dried sieved soil were weighed and transferred into a 200 mL beaker, with 100 mL of distilled water and CaCl2 added separately while stirring for one hour using an electric magnetic stirrer. The pH conductivity meter (TPS Lab chem.) was calibrated using pH buffers of 4.0, 7.0 and

9.0, and the pH of suspension measured.

143

Chapter 5 Site Investigation

5.4.5. Electrical conductivity (EC)

An EC measurement gives the concentrations of soluble salts in a soil at any particular temperature and is performed in a 1:5 soil-water suspension using a conductivity meter.

5.4.6. Organic carbon (Walkely and Black 1934)

The organic carbon in each sample was oxidised with potassium dichromate and sulphuric acid, and the excess potassium dichromate titrated against ferrous ammonium sulphate. One gram of soil was weighed and transferred into a 500 mL conical flask, with 10 mL of 1N K2Cr2O7 and 20 mL of conc. H2SO4 added and swirled carefully before the mixture was left to stand for 30 min. Then, 200 mL of distilled water and 10 mL of H3PO4 were slowly added followed by 1 mL of a diphenylamine indicator, with the resultant suspension titrated against 0.5 N ferrous ammonium sulphate solutions until a green colour started appearing which indicated the end point.

The blanks were run simultaneously (Walkely and Black, 1934 & Jackson, 1962) and the carbon content was calculated using the following equation.

10(퐵−푆)×0.39 Organic carbon = (5.2) (퐵×푊) where

퐵 = mL of ferrous ammonium sulphate solution used for blank

푆 = mL of ferrous ammonium sulphate solution used for sample

푊 = sample weight (g)

0.39 = conversion factor (correction factor for 70% oxidation of organic carbon)

% of organic matter = 1.72 × % of organic carbon

144

Chapter 5 Site Investigation

5.4.7. Cation exchange capacity, trace elements (aluminium and iron) and phosphorus (Colwell)

Using standard methods for each element, these parameters were determined by the

National Measurement Institute in Sydney, Australia, with the Cation Exchange

Capacity of pH 7obtained from the ammonium acetate method and trace elements from inductively coupled plasma mass spectrometry (ICP-MS).

5.4.8. Metal contents

The metal contents were determined by analysis by X-ray fluorescence (XRF). The result for the basic physico-chemical properties and the compositions of the soil are given in Table 5.6.

Table 5.6: Major soil properties and compositions

Properties Value Properties Value Sand (%) 41.20 - 47.20 Total solids (%) 86.90 - 98.40 Soil Silt (%) 39.00 - 43.00 P-Colwell (mg/kg) 20-140 texture Clay (%) 13.80 - 15.80 pH, (CaCl2) 4.15 - 5.61

SiO2 (%) 53.80 -78.33 Conductivity (µs/cm) 12.90 - 114.5

Fe2O3(%) 5.63-17.60 Organic matter (%) 0.96 -1.50

Al2O3 (%) 9.02 -14.27 Moisture content (%) 4.37 - 19.94 CaO (%) 0.25-9.21 CEC (mEq/100g) 3.80 - 7.50

Metal K2O (%) 1.81 -3.55 Ex-Mg (mEq/100g) 0.36 - 1.60

content TiO2 (%) 1.79-2.13 Ex-K (mEq/100g) 0.97 -1.60

SO3 (%) 0.83-0.95 Ex-Na (mEq/100g) 0.04-0.18 MnO (%) 0.23-0.28 Ex-Al (mEq/100g) 0.06-0.25

MoO3 (%) 0.02 -0.03 Ex-Ca (mEq/100g) 1.40 – 4.00 ZnO 0.03-0.04 Aluminium (mg/kg) 11600-16500

Rb2O 0.03-0.04 Iron (mg/kg) 18100- 21500

145

Chapter 5 Site Investigation

5.5. Quality Assurance (QA)

The QA assessment detailed below, which included analyses of QC samples, was carried out to determine the suitability and reliability of the field procedures and analytical results.

5.5.1. Identification of QC samples

5.5.1.1. Field QC samples

Blind replicates

Blind replicates are collected at the same time and from the same sampling point, and preserved, stored, transported and analysed in the same manner, as the corresponding primary samples. Their assessments are undertaken by calculating the relative percentage difference (RPD) defined as

Result No.1−Result No.2 RPD (%) = 100 × (5.3) Mean Result

Blind replicates are used to evaluate the precision of the total sampling and analysis and, in the case of soil samples, sample variability. In the AS4482.1 [379] and NEPM

[380], it is recommended that 1 blind replicate sample is collected for every 20 primary samples. Their results are provided in Table 5.8 and Table 5.9.

Split duplicates

Split duplicates are collected at the same time and from the same sampling point, and preserved, stored, transported and analysed with a separate method, as the corresponding primary samples. Their assessments are undertaken by calculating the

RPD defined as

146

Chapter 5 Site Investigation

ResultNo.1−ResultNo.2 RPD (%) = 100 x (5.4) Mean Result

A split duplicate is used to provide a check on the analytical proficiency of laboratories and, hence precession and comparability. In the AS4482.1 [379] and NEPM [380], it is recommended that 1 split duplicate sample is collected for every 20 primary samples. All the blind replicates were analysed with two different methods reported in

Table 5.8 and Table 5.9 so another analysis was not required.

Rinsates

A rinsate is a sample of distilled or de-ionised water poured over the surface of a decontaminated piece of sampling equipment and collected in an appropriate sample container. It is then analysed for contaminants of concern relating to the investigation and provides an assessment of the potential cross-contamination of chemicals from sampling equipment caused by inadequate decontamination procedures. According to the AS4482.1 [379], one rinsate sample should be collected each day per piece of sampling equipment. Its results are provided in Table 5.10.

Trip blanks

A trip blank is used to measure cross-contamination during sampling, transport, sample preparation and analysis, with the industry standard 1 trip blank per batch of primary samples. As, according to the USEPA, cross-contamination occurs only with volatile organics [381] and glyphosate is non-volatile [292] with a low vapour pressure of 9.3×10-3 MPa (at 25°C), trip blanks are not required for the analysis of glyphosate samples.

147

Chapter 5 Site Investigation

Trip spikes

Trip spikes are samples of analyte-free media (either washed sand or de-ionised water) which remain in the Eskies during sampling activities and are returned to the laboratory unopened. However, as the sample media are spiked with chemicals, the concentrations of which need to be determined, trip blanks are normally used to monitor volatile organic contamination losses during transit. The industry standard is

1 trip spike per batch of primary samples of which volatile concentrations are being measured but, as glyphosate is non-volatile, this test is also not required..

5.5.1.2. Laboratory QC samples

Laboratory duplicates

Laboratory duplicates were field samples prepared and analysed in the same manner twice to provide analytical precision for a sample batch. Their assessments are undertaken by calculating the RPD defined as

ResultNo.1−ResultNo.2 RPD (%) = 100 x (5.5) Mean Result

The NATA specifies 1 per 10 samples for trace element and inorganic analysis. The results obtained for different episodes are provided in Table 5.11 and Table 5.12.

Laboratory control samples (LCS)

LCSs are analyte-free matrices (de-ionised water or clean sand) spiked with known concentrations of target analytes and are carried through the entire preparation and analysis. Their assessments are undertaken by calculating the percent recovery (%R) of the spike defined as

Spiked Sample Result (SSR)− Sample Results (SR) %R = 100 x (5.6) Concentration of Spiked Added (SA)

148

Chapter 5 Site Investigation

This determines the analytical accuracy and precision for a sample batch, and is specified by the NATA as 1 per batch of up to 20 samples, performed in Chapter 4.

Method blanks

Method blanks are analyte-free matrices carried through the entire preparation and analysed to establish that laboratory contamination does not cause false positives and are prepared with every batch of up to 20 samples for all organic and inorganic analyses, as in Chapter 4.

5.5.2. Acceptable criteria for QC samples

The QC acceptable criteria adopted for this investigation are provided in Table 5.7 and are in general agreement with Table 4 in AS4482.1 [379] and the NEPM [380].

149

Chapter 5 Site Investigation

Table 5.7: QC samples’ acceptable criteria

QC sample Acceptable range Field QC samples RPD <50% when average concentration > 10 x (LOR) Blind replicate and split RPD <75% when average concentration 5 to 10 x duplicate (LOR) RPD <100% when average concentration < 5 x (LOR) Rinsate Analytical result < LOR Trip blank Analytical result < LOR Trip spike ± 30% Laboratory QC samples Laboratory duplicate RPD <30% when average concentration > 10 x (LOR) RPD <50% when average concentration 4 to 10 x (LOR) RPD <100% when average concentration < 4 x (LOR) Laboratory control %R of 70-130% (general analytes) sample Matrix spike %R of 70-130% (general analytes) Method blank Analytical results < LOR Note: LOR = lowest concentration of reliability

150

Chapter 5 Site Investigation

Table 5.8: Results of blind replicates in water samples for different episodes

Glyphosate Unit LOR Result no. 1 Result no.2 RPD % Fluorescence Spectrometry Episode 1 µg/mL 0.022 0.025 0.025 0.000 Episode 2 µg/mL 0.022 0.039 0.039 0.000 Episode 3 µg/mL 0.022 0.060 0.060 0.000 Episode 4 µg/mL 0.022 0.028 0.016 54.55 Episode 5 µg/mL 0.022 0.040 0.040 0.000 ELISA Episode 1 µg/mL 0.0001 0.029 0.029 0.000 Episode 2 µg/mL 0.0001 0.042 0.042 0.000 Episode 3 µg/mL 0.0001 0.067 0.067 0.000 Episode 4 µg/mL 0.0001 0.027 0.024 11.77 Episode 5 µg/mL 0.0001 0.039 0.039 0.000

Table 5.9: Results of blind replicates in soil samples for different episodes

Glyphosate Unit LOR Result no. 1 Result no.2 RPD % Fluorescence Spectrometry Episode 1 mg/kg 0.070 0.250 0.250 0.000 Episode 2 mg/kg 0.070 0.450 0.450 0.000 Episode 3 mg/kg 0.070 0.118 0.118 0.000 Episode 4 mg/kg 0.070 0.170 0.070 83.33 ELISA Episode 1 mg/kg 0.038 0.273 0.273 0.000 Episode 2 mg/kg 0.038 0.445 0.445 0.000 Episode 3 mg/kg 0.038 0.125 0.125 0.000 Episode 4 mg/kg 0.038 0.163 0.070 11.77

151

Chapter 5 Site Investigation

Table 5.10: Results for rinsates

Glyphosate Unit LOR Rinsate Fluorescence spectrometry Episode 1 µg/mL 0.022 < 0.022 Episode 2 µg/mL 0.022 < 0.022 Episode 2 µg/mL 0.022 < 0.022 Episode 4 µg/mL 0.022 < 0.022 Episode 5 µg/mL 0.022 < 0.022 ELISA Episode 1 µg/mL 0.0001 < 0.0001 Episode 2 µg/mL 0.0001 < 0.0001 Episode 2 µg/mL 0.0001 < 0.0001 Episode 4 µg/mL 0.0001 < 0.0001 Episode 5 µg/mL 0.0001 < 0.0001

Table 5.11: Results of laboratory duplicates in water samples for different episodes

Glyphosate Unit LOR Result no. 1 Result no.2 RPD % Fluorescence Spectrometry Episode 1 µg/mL 0.022 0.028 0.028 0.000 Episode 2 µg/mL 0.022 0.045 0.045 0.000 Episode 3 µg/mL 0.022 0.067 0.067 0.000 Episode 4 µg/mL 0.022 0.023 0.012 62.86 Episode 5 µg/mL 0.022 0.048 0.048 0.000 ELISA Episode 1 µg/mL 0.0001 0.024 0.024 0.000 Episode 2 µg/mL 0.0001 0.042 0.042 0.000 Episode 3 µg/mL 0.0001 0.065 0.065 0.000 Episode 4 µg/mL 0.0001 0.024 0.020 18.182 Episode 5 µg/mL 0.0001 0.049 0.049 0.000

152

Chapter 5 Site Investigation

Table 5.12: Results of laboratory duplicates in soil samples for different episodes

Glyphosate Unit LOR Result no. 1 Result no.2 RPD % Fluorescence Spectrometry Episode 1 mg/kg 0.070 0.300 0.300 0.000 Episode 2 mg/kg 0.070 0.575 0.575 0.000 Episode 3 mg/kg 0.070 0.133 0.133 0.000 Episode 4 mg/kg 0.070 0.158 0.060 89.66 ELISA Episode 1 mg/kg 0.038 0.313 0.313 0.000 Episode 2 mg/kg 0.038 0.525 0.525 0.000 Episode 3 mg/kg 0.038 0.138 0.138 0.000 Episode 4 mg/kg 0.038 0.153 0.060 87.32

153

Chapter 5 Site Investigation

5.6. Glyphosate Levels in Collected Samples

From the samples collected from the field after the different episodes of glyphosate application previously discussed, glyphosate levels were determined using the new fluorescence spectrometric and ELISA methods (Figs. 5.8, 5.9, 5.10, 5.11, 5.12 and

5.13).

0.080 0.070 0.060 L1 0.050 L2 0.040 L3 0.030 C1

Glyphosate, (mg/L) Glyphosate, 0.020 C2 0.010 0.000 Episode 1 Episode 2 Episode 3 Episode 4 Episode 5

Fig. 5.8: Glyphosate concentrations in water samples determined by ELISA method

0.70

0.60

0.50

0.40

0.30

0.20 Glyphosate, (mg/kg) Glyphosate, 0.10

0.00 Episode 1 Episode 2 Episode 3 Episode 4

Fig. 5.9: Glyphosate concentrations in soil samples determined by ELISA method

154

Chapter 5 Site Investigation

0.080 0.070 0.060 L1 0.050 L2 0.040 L3 0.030 C1

0.020 C2 Glyphosate, (mg/L) Glyphosate, 0.010 0.000 Episode 1 Episode 2 Episode 3 Episode 4 Episode 5

Fig. 5.10: Glyphosate concentrations in water samples determined by fluorescence

spectrometric method (Model I)

0.080 0.070 0.060 L1 0.050 L2 0.040 L3 0.030 C1

0.020 C2 Glyphosate, (mg/L) Glyphosate, 0.010 0.000 Episode 1 Episode 2 Episode 3 Episode 4 Episode 5

Fig. 5.11: Glyphosate concentrations in water samples determined by fluorescence

spectrometric method (Model III)

155

Chapter 5 Site Investigation

0.60

0.50

0.40

0.30

0.20

Glyphosate, (mg/kg) Glyphosate, 0.10

0.00 Episode 1 Episode 2 Episode 3 Episode 4

Fig. 5.12: Glyphosate concentrations in soil samples determined by fluorescence

spectrometric method (Model I)

0.60

0.50

0.40

0.30

0.20

Glyphosate, (mg/kg) Glyphosate, 0.10

0.00 Episode 1 Episode 2 Episode 3 Episode 4

Fig. 5.13: Glyphosate concentrations in soil samples determined by fluorescence

spectrometric method (Model III)

156

Chapter 5 Site Investigation

5.7. Discussions

As seen in Figs 5.8 to 5.13, the glyphosate concentrations in most samples were detected by both the ELISA and new fluorescence spectrometric methods although some water samples with glyphosate concentrations lower than the detection limit were not detectable by the latter. Using the newly developed method, it was also observed that most samples with lower concentrations were detected by Model III rather than Model I (Fig. 5.10 and Fig.5. 11 respectively) and, for samples with higher glyphosate concentrations, Model I provided better results than Model III (Fig. 5.12 and Fig. 5.13 respectively).

The water samples collected after the high rainfall event in March showed the highest concentrations of glyphosate which suggests that the freshly applied glyphosate was washed out by that event. The glyphosate concentrations in water samples collected after the fourth application were not so pronounced probably due to the lower dosage and lower intensities of the preceding rainfall events. Clearly, rainfall events play a notable role as they transport newly applied glyphosate into surface waters through the mechanisms of dilution or drift of the surface material by runoff.

For soil samples, higher concentrations were observed in those collected after the second application of glyphosate (1.2 L/ha + 1.5 L/ha) which appeared to reflect the applied doses as well as lower rainfall events while those collected after the rainfall event in March showed a decline in glyphosate concentrations. Also, the samples collected before crop rotation showed some base-level glyphosate concentrations which could be related to the higher doses of herbicide applied at that time and before

157

Chapter 5 Site Investigation

sowing started which were usually aimed at washing the herbicide and preparing the land for cultivation.

Glyphosate concentrations in the water samples ranged from 10 to 67 µg/L and, in the soil samples, from 0.1 mg/kg to 0.575 mg/kg. According to the Australian Drinking

Water Guideline of 1000 µg/L [19] and ANZECC guidelines for freshwater in

Australia (see section 1.2), this study revealed glyphosate levels within human health- based limits. However, they approach the Canadian Water Quality Guideline for the

Protection of Aquatic Life of 65 µg/L and also exceed South African safety guidelines value of maximum 0.3 µg/L [35]. In addition, water quality issues for the GBR arising from discharge of pollutants from the adjacent catchment areas (the Great Barrier Reef

Catchment Area, GBRCA) have long been recognised [201, 382-388] and considered as one of the main contributing factors to the loss of coral cover on the GBR. Due to the weight of scientific evidence of the severity of the issue and better protection of the human health and the World Heritage status (The GBR) [383, 388] many more research and monitoring studies need to be carried out within Australia. Also much work is required for continuous updating and improvement of the available guidelines

[151, 389].

5.8. Conclusions

In this chapter, details of discussions held with some GM canola growers in NSW regions were reported and a site investigation involving collections of soil and water samples after different episodes of glyphosate application described. The levels of glyphosate in the collected samples were determined using a new low-cost

158

Chapter 5 Site Investigation

fluorescence spectrometric approach and the ELISA method, with a QA conducted and the different properties of the soils examined.

159

Chapter 6

Glyphosate Sorption onto GM Canola Cultivated Soils

6.1. General

As discussed in Chapter 3N-(Phosphonomethyl)glycine (PMG), also known as glyphosate, is a broad-spectrum, non-selective, post-emergence herbicide that is the most commonly used herbicide worldwide, mainly for weed control in agriculture, horticulture, forestry, aquaculture and urban areas. Its use has rapidly increased within the last couple of years due to the introduction of genetically modified (GM) glyphosate-resistant crops [10, 209, 390] which is raising concerns regarding its safety, i.e., its absence of any harmful environmental effect rather than its on-target organisms

(for only undesirable weeds). Regardless of information in the literature which shows that glyphosate is relatively environmentally safe [24, 297, 391-394], some investigations have pointed out possible leaching and toxicity problems associated with its use [103, 132, 395-399]. Glyphosate, when applied directly to soils it shows little herbicidal activity due to its inactivation in soils as a result of its sorption to soil[229, 230] and, compared with all other insecticides, possesses strong sorption.

While most insecticides are moderately to weakly sorbed in soils by soil organic matter

(SOM) as their molecules are dominated by a polar groups, i.e., aliphatic and/or aromatic carbon, and often have only one functional group, on the other hand, glyphosate, a small molecule with three polar functional groups (carboxyl, amino and phosphonate) is strongly sorbed by soil minerals[249].

160

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

In this chapter, the effects of the sorption and desorption of glyphosate on three different soils from the site investigation area (location map shown in Chapter 5) are examined. It includes an evaluation of sorption isotherm models using a new low-cost fluorescence spectrometric method, presentation of relevant statistical correlations of sorption properties and discussion of the sorption mechanism and implications of

Roundup application in these three soils. Different linear and non-linear isotherm models were applied to the experimental data and their results compared and discussed, with the isotherm match explaining the data well. Comparative results are also provided with established UV-visible spectrophotometric.

6.2. Introduction and Background

Glyphosate has a distinct tendency to adsorb to soil constitutes. There is evidence in the literature that the soil sorption and degradation of glyphosate vary greatly depending on the soil’s properties, such as pH, iron and aluminium oxides, and structure, and the application time and microbial activity of glyphosate [190, 250, 268,

272, 400-402]. As glyphosate is bound to a soil through its phosphonic acid moiety, it competes with inorganic phosphate for sorption sites and it was found in a study that adding 98 or 196 kg/ha of phosphate decreased glyphosate inactivation in the soil

[229]. As reported in the literature, there is a correlation between glyphosate sorption and excess phosphorus sorption sites because inorganic phosphate excludes glyphosate from sorption sites [130, 131, 246, 251, 252, 260]. Several papers have reported that the sorption of glyphosate by a soil or clay depends on its cation exchange capacity

(CEC) (its capacity for the ion exchange of cations between it and its solution), clay content, organic matter, and iron and aluminum amorphous oxides [230, 245].

161

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Regarding its CEC, it increases in the order of Na+< Mg2+< Ca2+ < Zn2+< Mn2+ Fe3+<

Al3+ [230, 403] which suggests that coordination bonding of the phosphonic acid moiety of glyphosate with surface-exchanged polyvalent cations, especially Fe3+ and

Al3+, plays an important role within the pH range from 4 to 8 [246, 247].

Some previous studies have also provided information regarding glyphosate’s chelation with metals in a solution [157], polyvalent cations in the inter-layers of montmorillonite [404] and coordination bonding with surface Al3+ and Fe3+[157]. As suggested by some authors, hydrogen bonding is the primary sorption mechanism of glyphosate to humic substances [234, 405], from which it can be concluded that the actual sorption of glyphosate in soils may include mechanisms such as precipitation reactions, coordination bonding and less specific sorption interactions. This might cause difficulties for predicting the degree to which long-term phosphorus and lime applications in agricultural soils will influence glyphosate’s sorptive strength.

Glyphosate sorption onto soil is usually described by the sorption isotherm which is a graph of the equilibrium surface excess or amount of a compound adsorbed (Cs) plotted against the equilibrium solution concentration of the compound (Ceq, ) at a fixed temperature, pressure and solution chemistry, e.g., pH and ionic strength [406]. To describe the data, equations such as the Langmuir, Freundlich and Redlich-Peterson isotherms are commonly used, while some problems arise regarding insufficient data and more complicated expressions [153]. As discussed in the literature, to determine the parameters for isotherm equations, regression methods are normally used

[407].Describing the sorption data involves two steps: choosing an equation which

162

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

describes the data; and using the regression method to find an optimal parameter set.

Therefore, it is very important to determine the type of isotherm equation which best fits the given data.

6.3. Glyphosate Sorption Experiments

6.3.1. Materials and methods

6.3.1.1. Reagents

Glyphosate (PESTANAL, analytical standard) and the derivatisation reagent 9- fluorenyl methoxycarbonyl chloride (FMOC-Cl) were obtained from Fluka (Germany) and HPLC-grade acetonitrile and diethyl ether from Sigma-Aldrich Australia. All other chemicals, including KCl, HCl, KOH, NaOH, disodium tetraborate decahydrate

(Na2B4O7.10H2O), potassium phosphate monobasic (KH2P04) and diethylether, were analytical grade.

6.3.1.2. Solutions

Seven glyphosate solutions of various concentrations (5.000, 10.000, 12.000, 15.000,

20.000, 22.000 and 25.000 µg/mL) were prepared with analytical-standard glyphosate

(PESTANAL, 99.729%) using distilled water and FMOC-Cl solutions of 1 gL-1 by dissolving the reagent in acetonitrile just before conducting the experiments. 0.1M

KH2PO4 solution was prepared for glyphosate desorption test. A buffer solution (pH

=9) was prepared by dissolving 15.255 g of Na2B4O7.10H2O in 1000 mL of distilled water, and a 0.1 M EDTA solution used to pre-treat samples to correct the sensitivity of the fluorogenic reagent to divalent ions in the amino-acid coupling [343].

163

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

6.3.2. Soil properties

Sorption isotherms were performed on three different soils from three different paddocks (location map provided in Chapter 5), with the soil properties characterised and presented in Table 6.1. Soil A, from a field in which no GM canola cultivation occurred during the previous five years (Inglewood); for the soil B, from a field in which canola harvesting had just finished (Woodbine); and for the soil C, from a field in which the land was fully prepared for that year’s canola cultivation (Parkes plains)

Table 6.1: Properties of three different soils from GM canola paddocks

Soil pH OC CEC Fe (%) Al (%) Sand (%) Silt (%) Clay (%) A 4.5 0.94 6.81 5.633 9.018 47.2 39 13.8 B 4.8 1.1 7.4 17.597 14.271 43.2 29 27.8 C 5 0.9 4.75 7.649 11.03 41.2 43 15.8

6.3.3. Procedure

A batch technique was used to measure glyphosate sorption onto soils with seven glyphosate solutions of 5.000, 10.000, 12.000, 15.000, 20.000, 22.000 and 25.000

µg/mL. Twenty-four hours before the start of the sorption experiment, a 25 mL aqueous glyphosate solution prepared in 0.1 M KCL was added to 2 g of air-dried soil weighed into 50 mL centrifuge tubes to ensure uniform moisturising of the samples which were rotated end over end for another 24h. Small amounts of either a KOH or

HCl solution were added to reach constant pH of 4.5, 4.8 and 5 for soils A, B and C respectively. The tubes were then centrifuged at 4000 rpm for 17 min and the supernatants withdrawn for derivatisation and further analysis. The sorbed amount (Cs) was calculated from the difference between the initial glyphosate concentration (Ci)

164

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

and the concentration remaining in the supernatant solution (Ce), and experiments carried out at the ambient temperature (20 ± 2oC).

Glyphosate concentrations were quantified using the new fluorescence spectrometric method established in Chapter 4 and with the already established- visible direct spectrophotometric method. The derivatisation procedure involved was the same as in

Chapter 4. After derivatisation, the mixture was then shaken in a mechanical shaker for 2 h at room temperature, and the resulting solution mixed with 4 mL of diethylether, shaken and centrifuged at 4000 rpm for 15 min to separate the diethyl-ether. The aqueous phase, which contained the derivatisation product, was then extracted and quantified by both the fluorescence spectrometric and UV-visible direct spectrophotometric methods. The recordings of the emission acquisition for the new fluorescence spectrometric method were performed on the same way as discussed in

Chapter 4.For the UV-visible spectrophotometric method the absorbance spectra of derivatised glyphosate were measured on the 265nm wavelength. To avoid the organic matter interference from the soils always a blank solution was used a base line correction. The blank solution was prepared by 3 mL of the extract obtained from the analyte free soil samples and 0.1M KCl solution, 0.5 mL of the borate buffer, 0.5 mL of FMOC-Cl and 4 mL of diethylether, and being shaken and centrifuged, as previously described.

The samples of soil containing sorbed glyphosate were then subjected to desorption using a 0.1M KH2PO4 agitation (17 min), centrifugation (4000 rpm for 10 min) and a filtration trough 0.45 µm filter paper. This extraction process was repeated twice on

165

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

the solid residue and a 25 mL extract from each sample obtained, with the equilibrium time between each desorption step 24 h.

6.4. Soil Sorption Isotherms

In the present study, the sorption capacity and equilibrium isotherm for glyphosate sorption onto soil were estimated using three equilibrium models, the Langmuir, the

Freundlich, and the Redlich-Peterson isotherm equations

6.4.1. Langmuir isotherm

The theoretical equilibrium isotherm developed by Irving Langmuir [158], an

American chemist, relates the amount of solute sorbed on a surface to its concentration.

Its equation is derived from simple mass-action kinetics assuming chemisorption and its model based on the two assumptions that the forces of interaction between sorbed molecules are negligible and, once a molecule occupies a site, no further sorption takes place. Therefore, theoretically, a saturation value is reached beyond which no further sorption takes place and this equation canals reduce to Henry’s law at lower initial concentrations. Alternatively, at higher concentrations, it predicts a mono-layer sorption capacity which can be represented by:

푞푚푏퐶푒푞 퐶푠 = (6.1) 1+푏퐶푒푞 where 퐶푒푞 is the equilibrium concentration (µg/mL), 퐶푠 the amount of glyphosate sorbed (mg/kg), 푞푚 the maximum sorption capacity (mg/kg) and 푏 a constant related to the free energy of sorption (L/ mg).

166

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Eq. (6.1) can be linearised to at least four different forms [408] which are given in

Table 6.2.

Table 6.2: Different linearised forms of isotherm equations

Isotherm type Linear regression Plot

퐶푒 1 퐶푒 Langmuir type 1 = ( ) + ( ) 퐶푒⁄퐶푠 vs 퐶푒 퐶푠 푏푞푚 푞푚 1 1 1 Langmuir type 2 = ( ) + ( ) 1⁄퐶푠vs 1⁄퐶푒 퐶푠 푞푚 푏푞푚 퐶푒

퐶푠 Langmuirtype3 퐶푠 = 푞푚 − ( ) 퐶푠 vs 퐶푠⁄퐶푒 푏퐶푒

퐶푠 Langmuir type 4 = 푏푞푚 − 푏퐶푠 퐶푠⁄퐶푒 vs 퐶푠 퐶푒 1 Freundlich 푙표푔(퐶 ) = 푙표푔(퐾 ) + 푙표푔 (퐶 ) 푙표푔(퐶푠) vs 푙표푔 (퐶푒) 푠 푓 푛 푒

퐶푒 푙푛 (퐾푟푝 − 1) 푙푛(퐾푟푝퐶푒/퐶푠 − 1)vs Redlich-Peterson 퐶푠 푙푛(퐶푒) = 푙푛(µ) + 훽 푙푛(퐶푒)

6.4.2. Freundlich isotherm

An empirical sorption isotherm for non-ideal sorption on a heterogeneous surface, as well as multi-layer sorption, presented by Herbert Max Finley Freundlich [156], a

German physical chemist, is expressed by

1 ⁄푛 퐶푠 = 퐾푓퐶푒푞 (6.2) where 퐶푒푞 (µg/mL) is the solute concentration at equilibrium, 퐶푠 (mg/kg) the amount of glyphosate sorbed at equilibrium, 퐾푓a constant which indicates the relative sorption

1-(1/n) 1/ n 1 capacity of the adsorbent (mg L ) /kg) and ⁄푛a constant, and퐶푠 can be calculated from the simple mass balance equation as:

(퐶 −퐶 )푉 퐶 = 0 푒 (6.3) 푠 푀

167

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

where 퐶0 is the initial glyphosate concentration (µg/mL), 퐶푒 the glyphosate concentration at equilibrium (mg/L), 푉 the volume of the solution (L) and 푀 the mass of the sorbent (kg). The linearised form of Eq. (6.2) can be obtained by taking the log on each side as:

1 푙표푔(퐶 ) = 푙표푔 (퐾 + 푙표푔(퐶 ) (6.4) 푠 푓) 푛 푒

The Freundlich isotherm, which was derived by assuming an exponentially decaying sorption site’s energy distribution, has the limitation that it does not follow the fundamental thermodynamic basis because it does not reduce to Henry’s law at lower concentrations.

6.4.3. Redlich-Peterson isotherm

The Redlich-Peterson isotherm has three parameters and the features of both the

Freundlich and Langmuir isotherm equations, and can be described as [8]:

퐾푟푝퐶푒 퐶푠 = 훽 (6.5) 1+ µ퐶푒

where Krp is the Redlich-Peterson isotherm constant L/kg, µ the Redlich-Peterson isotherm constant L/mgβ, where β is the exponent which lies between 0 and 1.

However, it has the following two limitations.

When β = 1, the Langmuir equation results are given by:

퐾푟푝퐶푒 퐶푠 = (6.6) 1+µ퐶푒

When β = 0, the Redlich-Peterson isotherm equation transforms to the Henry’s law equation as:

168

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

퐾 퐶 퐶 = 푟푝 푒 (6.7) 푠 1+µ

Eq. (6.5) can be rearranged as:

퐶푒 퐾푟푝 − 1 = µ퐶푒 (6.8) 퐶푠

Then, Eq. (6.8) can be transformed to a linear equation by taking the logarithms:

퐶푒 푙푛 (퐾푟푝 − 1) = 푙푛(µ) + 훽 푙푛(퐶푒) (6.9) 퐶푠

The Redlich-Peterson constants can be obtained by plotting the left-hand side of Eq.

(6.9) but its isotherm contains three unknown parameters,Krp, µ and β, which cannot be obtained by linearising it. Therefore, these were obtained using a minimisation procedure to maximise the coefficient of determination푅2 between the theoretical data for Cs predicted from Eq. (6.9) and the experimental data.

6.5. Results

6.5.1. Linear isotherms

Using linear regression analysis for the best-fit equation has been the most commonly used technique for determining the best-fit isotherm. In this study, the method of least squares was used to find the parameters for the isotherms. To determine the best models for describing the experimental results for sorption isotherms by comparing the parameters derived from the linear and non-linearised isotherms, and also to calculate the theoretical sorption isotherms, different linear and non-linear models were applied. The experimental values of Ce and Cs were the initial assumptions for determining the model parameters from/for both the linear and non-linearised isotherms. The linear coefficient (R2) indicated a fit between the experimental and

169

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

calculated values for the linearised isotherm models. To check variations between the experimental and calculated values for sorption capacity, the Chi-square values were determined using Excel and the average percentage errors calculated using:

⃒(퐶푒,푒푥푝−퐶푠,푐푎푙)⃒ 퐴푃퐸 = ∑푁 퐶푠,푐푎푙 × 100 (6.10) 1 푁

Three different isotherm models (Langmuir, Freundlich and Redlich-Peterson) were used with the non-linear and linear forms of their corresponding isotherms, as shown in Table 6.2. The Langmuir coefficients for the four linearised Langmuir equations,

Langmuir-1, -2, -3 and -4, were obtained by plotting graphs of 퐶푒⁄퐶푠 versus 퐶푒, 1⁄퐶푠 versus 1⁄퐶푒, 퐶푠 versus 퐶푠⁄퐶푒 and 퐶푠⁄퐶푒 versus 퐶푠 respectively. The parameters calculated from the fluorescence spectrometric and UV-visible spectrophotometric method are shown in Tables 6.3, 6.4 and 6.5 respectively, and plots of the experimental and observed data in Figs. 6.1, 6.2 and 6.3 respectively.

170

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.3: Linearised isotherm parameters for three different soils (fluorescence

spectrometric, Model I)

Linearised isotherm Soil A Soil B Soil C Langmuir-1 qm 312.5 400.0 357.1 b 0.182 0.313 0.131 푅2 0.834 0.892 0.921 Langmuir-2 qm 212.8 333.3 400.0 b 0.402 0.429 0.112 푅2 0.970 0.994 0.984 Langmuir-3 qm 235.8 353.3 325.4 b 0.337 0.396 0.156 푅2 0.688 0.825 0.757 Langmuir-4 qm 281.4 393.3 386.7 b 0.232 0.327 0.118 푅2 0.688 0.825 0.757 Freundlich

퐾푓 57.15 93.54 45.39 1 ⁄푛 0.562 0.623 0.686 푅2 0.979 0.981 0.963 Redlich-Peterson

퐾푟푝 300.0 300.0 300.0 훽 0.507 0.517 0.348 휇 4.224 2.158 5.652 푅2 0.966 0.962 0.846

171

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.4: Linearised isotherm parameters for three different soils (fluorescence

spectrometric, Model III)

Linearised isotherm Soil A Soil B Soil C Langmuir-1 qm 270.3 357.1 312.5 b 0.203 0.326 0.145 푅2 0.833 0.882 0.914 Langmuir-2 qm 188.7 312.5 357.1 b 0.417 0.405 0.120 푅2 0.974 0.995 0.983 Langmuir-3 qm 207.1 320.0 281.3 b 0.357 0.397 0.175 푅2 0.698 0.826 0.735 Langmuir-4 qm 247.2 358.1 342.0 b 0.249 0.328 0.128 푅2 0.698 0.826 0.735 Freundlich

퐾푓 51.80 84.96 42.09 1 ⁄푛 0.576 0.648 0.698 푅2 0.980 0.979 0.962 Redlich-Peterson

퐾푟푝 300.0 300.0 300.0 훽 0.488 0.472 0.335 휇 4.761 2.490 6.164 푅2 0.964 0.945 0.825

172

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.5: Linearised isotherm parameters for three different soils (UV-visible

spectrophotometric)

Linearised isotherm Soil A Soil B Soil C Langmuir-1 qm 322.6 400.0 243.9 b 0.158 0.212 0.232 푅2 0.782 0.863 0.782 Langmuir-2 qm 222.2 303.0 196.1 b 0.321 0.340 0.392 푅2 0.968 0.977 0.968 Langmuir-3 qm 239.6 323.5 197.67 b 0.289 0.313 0.409 푅2 0.665 0.647 0.647 Langmuir-4 qm 298.9 409.6 249.0 b 0.187 0.208 0.240 푅2 0.665 0.647 0.647 Freundlich

퐾푓 53.30 73.19 56.47 1 ⁄푛 0.578 0.643 0.491 푅2 0.963 0.966 0.963 Redlich-Peterson

퐾푟푝 300.0 300.0 300.0 훽 0.479 0.443 0.479 휇 4.640 3.085 4.640 푅2 0.934 0.912 0.934

173

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

300

250

200

150

(a) Soil A Cs,(mg/kg) 100 Experimental Langmuir 1 50 Langmuir 2 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150 (b) Soil B Cs,(mg/kg) Experimental 100 Langmuir 1 Langmuir 2 50 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150 (c) Soil C Cs,( mg/kg) Cs,( Experimental 100 Langmuir 1 Langmuir 2 50 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

Fig. 6.1: Langmuir linearised isotherm models for glyphosate sorption by three

different soils (fluorescence spectrometric, Model I)

174

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

300

250

200

150

(a) Soil A Cs,(mg/kg) 100 Experimental Langmuir 1 50 Langmuir 2 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

250

200

150

(b) Soil B Cs,(mg/kg) 100 Experimental Langmuir 1 50 Langmuir 2 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

(c) Soil C Cs,( mg/kg) Cs,( 100 Experimental Langmuir 1 Langmuir 2 50 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

Fig. 6.2: Langmuir linearised isotherm models for glyphosate sorption by three

different soils (fluorescence spectrometric, Model III)

175

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

300

250

200

150

(a) Soil A Cs,(mg/kg) 100 Experimental Langmuir 1 50 Langmuir 2 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

(b) Soil B Cs,(mg/kg) 100 Experimental Langmuir 1 50 Langmuir 2 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

(c) Soil C Cs,(mg/kg) 100 Experimental Langmuir 1 50 Langmuir 2 Langmuir 3 Langmuir 4 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

Fig. 6.3: Langmuir linearised isotherm models for glyphosate sorption by three

different soils (UV-visible spectrophotometric)

176

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

From Tables 6.5 and 6.6, it can be inferred that the different linear Langmuir equations show different Langmuir constants, as indicated by the variations in errors which correspond specifically to the modes of linearization [209]. For the three different soils, A, B and C, it is observed that, from both applied methods, the Langmuir-2 linearised equation showed higher values of the correlation coefficient (R2 = 0.970,

0.994 and 0.984 for Model I, R2 = 0.974, 0.995 and 0.983 for Model III and R2 = 0.968,

0.977 and 0.968 for UV method) than the other three (Langmuir-1, -3 and -4). The sorption capacities of the three soils, A, B and C, from fluorescence spectrometric using Model I were found to be 312.5 mg/kg, 400.0 mg/kg and 357.1 mg/kg, 212.8 mg/kg, 333.3 mg/kg and 400.0 mg/kg, 235.8 mg/kg, 353.3 mg/kg and 325.4 mg/kg, and 281.4 mg/kg, 393.3 mg/kg and 386.7 mg/kg for the Langmuir-1, -2, -3 and -4 equations respectively and for Model III were found as 270.3 mg/kg, 357.1 mg/kg and

312.5 mg/kg, 188.7 mg/kg, 312.5 mg/kg and 357.1 mg/kg, 207.1 mg/kg, 320.0 mg/kg and 281.3 mg/kg, and 247.2 mg/kg, 358.1 mg/kg and 342.0 mg/kg respectively . From the UV visible direct spectrophotometric analysis were found to be 322.6 mg/kg, 400.0 mg/kg and 243.9 mg/kg, 222.2 mg/kg, 303.0 mg/kg and 196.1 mg/kg, 239.6 mg/kg,

323.5 mg/kg and 197.7 mg/kg, and 298.9 mg/kg, 409.6 mg/kg and 249.0 mg/kg for the

Langmuir-1, -2, -3 and -4 equations respectively. Due to linearisation, errors in computing the parameters may be responsible for variations in the adsorption capacity

(qm) and adsorption constant (b) (from the Langmuir-1 -4 linearised isotherms). On the other hand, as discussed in the literature, the transformation of a non-linear to linear isotherm model seems to implicitly alter the error functions as well as error variances and normality assumptions of the least-squares methods [406, 407]. To verify the

177

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Langmuir models, their average percentage errors and Chi-square values for both methods were calculated (Tables 6.9, 6.10 and 6.11).

The Freundlich linearised equation is presented in Table 6.2 and the corresponding coefficients estimated by plotting graphs of log(퐶푠) versus log(퐶푒) using both the experimental and observed data. Also, the coefficients (viz. 퐾푓and1⁄푛) for both the linearised and non-linearised isotherms are shown in Tables 6.3 to 6.8. The magnitude of the Freundlich adsorption capacity (푛) gives an indication of the favourability of sorption, with its values ranging from 2 to 10 indicating good sorption capacity, from

1 to 2 moderate adsorption capacity and less than one poor adsorption capacity [409,

410]. The sorption capacities of soils A, B and C for both the linear and non-linear isotherms presented in Tables 6.3 to 6.8 are within the range of 1 to 2, that is, moderate.

In order to verify the validity of the Freundlich isotherm model, the average percentage errors and Chi-square values were calculated (Tables 6.9, 6.10 and 6.11).

Using experimental data, the Redlich-Peterson isotherm was plotted between ln (퐾푟푝퐶푒⁄퐶푠 − 1) and ln (퐶푒). As it contained three unknown parameters (viz. 퐾푟푝, 휇 and 훽) which it was not possible to obtain by linearisation, they were obtained by minimising the isotherm equation (maximising the correlation coefficient) (Table 6.3,

6.4 and 6.5). The average percentage errors and Chi-square values were also calculated and provided in Tables 6.9, 6.10 and 6.11.

178

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

6.5.2. Non-linear isotherms

In the present study, a trial and error procedure was used to study non-linear isotherm models developed to determine their isotherm parameters by minimising the respective coefficient of determination between the experimental data and their isotherms using the solver add-in with Microsoft’s spreadsheet Excel.

Their coefficients determined by analyses are shown in Tables 6.6, 6.7 and 6.8, with two corresponding plots drawn for 퐶푒 versus 퐶푠 (Figs. 6.4 to 6.6) using the experimental data and values predicted by them. The average percentage error Chi- square values are provided in Table 6.9, 6.10 and 6.11.

Table 6.6: Non-linearised isotherm parameters from fluorescence spectrometric

analysis, Model I)

Non-linearised isotherm Soil A Soil B Soil C Langmuir

qm 465.8 557.9 387.6 b 0.083 0.155 0.109 푅2 0.995 0.995 0.991 Freundlich

퐾푓 44.90 79.79 47.28 1 ⁄푛 0.682 0.702 0.639 푅2 1.000 1.000 1.000 Redlich-Peterson

퐾푟푝 2497 1523 35.72 훽 0.319 0.310 1.412 휇 555.2 18.12 0.030 푅2 1.000 1.000 0.984

179

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.7: Non-linearised isotherm parameters from fluorescence spectrometric

analysis, Model III)

Non-linearised isotherm Soil A Soil B Soil C Langmuir qm 465.8 557.9 387.6 b 0.083 0.155 0.109 푅2 0.996 0.996 0.993 Freundlich

퐾푓 44.90 79.79 47.28 1 ⁄푛 0.682 0.702 0.639 푅2 1.000 1.000 1.000 Redlich-Peterson

퐾푟푝 2497 1523 35.72 훽 0.319 0.310 1.412 휇 555.2 18.12 0.030 푅2 1.000 1.000 0.989

180

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.8: Non-linearised isotherm parameters from UV-visible spectrophotometric

analysis

Non-linearised isotherm Soil A Soil B Soil C Langmuir qm 449.7 433.9 255.6 b 0.087 0.184 0.209 푅2 0.995 0.987 0.980 Freundlich

퐾푓 46.32 77.69 56.70 1 ⁄푛 0.663 0.600 0.494 푅2 1.000 1.000 1.000 Redlich-Peterson

퐾푟푝 280.4 300.0 199.8 훽 0.340 0.443 0.507 휇 601.0 3.085 351.4 푅2 1.000 0.999 1.000

181

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

300

250

200

150

(a) Soil A Cs,(mg/kg) Experimental 100 Freundlich

50 Langmuir Redlich-Peterson 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

(b) Soil B Cs,(mg/kg) Experimental 100 Freundlich Langmuir 50 Redlich-Peterson

0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

Fig. 6.4: Non-linear isotherm models for glyphosate sorption by three different soils

(Fluorescence spectrometric Model I)

182

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

300

250

200

150

(a) Soil A Cs,(mg/kg) 100 Experimental Freundlich 50 Langmuir Redlich-Peterson 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

250

200

150

100 (b) Soil B Cs,(mg/kg) Experimental Freundlich 50 Langmuir Redlich-Peterson 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

Fig. 6.5: Non-linear isotherm models for glyphosate sorption by three different soils

(Fluorescence spectrometric Model III)

183

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

300

250

200

150

(a) Soil A (mg/kg) Cs, Experimental 100 Freundlich Langmuir 50 Redlich-Peterson

0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

(b) Soil B Cs,(mg/kg) 100 Experimental Freundlich 50 Langmuir Redlich-Peterson 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

300

250

200

150

(c) Soil C (mg/kg) Cs, 100 Experimental Freundlich 50 Langmuir Redlich-Peterson 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Ce, (mg/L)

Fig.6.6: Non-linear isotherm models for glyphosate sorption by three different soils

(UV-visible spectrophotometric)

184

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.9: Correlation coefficient, APE and Chi-square values for linearised and non- linearised isotherms (Fluorescence spectrometric, Model I)

Isotherm APE Chi -Square Soil A Soil B Soil C Soil A Soil B Soil C Linearised Langmuir-1 11.29 7.571 6.286 0.047 0.193 0.559 Langmuir-2 10.57 5.857 7.000 0.0006 0.061 0.528 Langmuir-3 11.57 6.000 7.428 0.008 0.124 0.519 Langmuir-4 12.00 7.428 7.143 0.043 0.190 0.556 Freundlich 6.000 8.571 8.428 0.699 0.269 0.196 Redlich-Peterson 6.429 5.857 7.857 0.471 0.454 0.296 Non-linearised Langmuir 15.00 15.00 5.714 0.0001 0.0002 0.306 Freundlich 11.00 11.86 8.429 0.111 0.010 0.209 Redlich-Peterson 11.00 12.00 7.571 0.110 0.010 0.193

Table 6.10: Correlation coefficient, APE and Chi-square values for linearised and non-linearised isotherms (Fluorescence spectrometric, Model III)

APE Chi -Square Isotherm Soil A Soil B Soil C Soil A Soil B Soil C Linearised Langmuir-1 11.00 6.571 6.571 0.132 0.349 0.654 Langmuir-2 9.143 5.428 7.142 0.006 0.228 0.605 Langmuir-3 11.14 5.714 7.286 0.037 0.273 0.609 Langmuir-4 11.14 7.000 7.286 0.121 0.349 0.640 Freundlich 16.43 16.57 17.43 0.004 0.0002 0.0005 Redlich-Peterson 6.000 6.571 8.000 0.625 0.551 0.377 Non-linearised Langmuir 11.14 6.857 8.000 0.004 0.028 0.488 Freundlich 7.143 8.714 10.00 0.453 0.322 0.378 Redlich-Peterson 7.143 8.571 7.286 0.452 0.316 0.329

185

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.11: Correlation coefficient, APE and Chi-square values for linearised and non-linearised isotherms (UV-visible spectrophotometric)

Isotherm APE Chi -Square Soil A Soil B Soil C Soil A Soil B Soil C Linearised Langmuir-1 12.57 8.140 15.14 0.034 0.088 0.002 Langmuir-2 10.00 9.710 12.57 0.001 0.007 0.001 Langmuir-3 10.42 9.285 13.29 0.007 0.026 0.001 Langmuir-4 12.00 9.285 15.00 0.034 0.091 0.003 Freundlich 7.000 6.857 10.86 0.407 0.080 0.020 Redlich-Peterson 8.000 6.571 10.86 0.275 0.098 0.015 Non-linearised Langmuir 13.57 9.143 15.57 0.002 0.084 0.001 Freundlich 9.143 8.286 11.43 0.251 0.078 0.016 Redlich-Peterson 9.143 6.571 11.43 0.260 0.098 0.016

6.6. Glyphosate Desorption

Although glyphosate sorption onto soil has been reported as capable of retaining residual activity towards some plant species [411], and despite its well-known high water solubility [412], the literature suggests studying glyphosate desorption from soil.

Since glyphosate is not only sorbed onto soils through strong interactions but also through relatively weak bonds, such as hydrogen bonding on soil humic material [234], it is important to evaluate its possible desorption.

Glyphosate desorption from the three soils using 0.1M KH2PO4 solution, agitation (17 min), centrifugation (4000 rpm for 10 min) and a filtration with a 0.45 µm filter paper in the filtration unit. All the process of analysis and quantification were same as discussed above. The results are listed in Table 6.12 and 6.13.

186

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

Table 6.12: Desorption of glyphosate from different soils, calculated using Model I

Glyphosate Desorption (%) added, mg/L Soil A Soil B Soil C 5 63.38 72.41 85.68 10 54.84 73.59 68.12 12 56.97 62.48 62.16 15 56.74 62.51 61.27 20 66.29 60.86 65.13 22 67.43 56.07 63.52 25 73.36 54.75 61.11

Table 6.13: Desorption of glyphosate from different soils, calculated using Model III

Glyphosate Desorption (%) added, mg/L Soil A Soil B Soil C 5 74.63 85.26 90.50 10 64.57 86.64 80.21 12 67.08 73.56 73.19 15 66.81 73.60 72.14 20 78.05 71.65 76.69 22 79.39 66.02 74.79 25 86.37 64.47 71.96

It must be noted that this study showed that, despite the common belief that glyphosate is readily and permanently sorbed onto soils, there may still be a large part of the fixed herbicide that can be easily returned to the soil solution. In particular, soils lacking the chemical characteristics conducive to extensive adsorption of glyphosate (amorphous hydrous oxides and vermiculite) are prone to release adsorbed glyphosate in the desorbing solution. In Models I and II, the least-sorbing C soils (Tables 6.12 and 6.13) desorbed 85.68% and 90.50% respectively of the adsorbed herbicide and even the

187

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

high-adsorbing soils, A and B, released around 50% and 85 % respectively. Our findings provide information on both glyphosate desorption from soils and soil sorption. Moreover, this study also indicates that this herbicide can be extremely mobile in a soil environment if applied on soils unable to retain its molecules sufficiently long for their microbial degradation. This may also lead to the herbicide leaching to lower soil layers in which limited biological activity occurs.

6.7. Conclusions

Estimation of sorption isotherm models using a new low-cost fluorescence spectrometric method with relevant statistical correlations of sorption properties were discussed in this chapter. Discussions were also provided on the sorption mechanism and implications of glyphosate application for glyphosate sorption in three different soils. Different linear and non-non-linear isotherm models were applied to the experimental data and their results compared and discussed, with the perfect isotherm match explaining the data well. Comparisons results also provided with an already established UV-visible spectrophotometric method.

The results for glyphosate sorption measured by fluorescence spectrometric and UV- visible spectrophotometric analyses of three soils with different glyphosate concentrations showed that the sorbed herbicide decreased in the order of B > A > C for both cases. This order of sorption was approximately parallel to the amounts of iron and aluminium (Table 6.1) present in the soils which suggested the strong sorption of glyphosate mainly on the Fe oxides through a ligand exchange mechanism which confirms the previous findings [229]. The glyphosate desorption results for the three

188

Chapter 6 Glyphosate Sorption onto GM Canola Cultivated Soils

soils after a 0.1 M KH2PO4 treatment indicated a highest of 90.50% of soil glyphosate desorption from soil C having the lowest sorption capacity (Table 6.13).

From the isotherm models in a summary from Tables 6.3 to 6.11, it can be seen that for the Langmuir model (linearised and non-linearised form) the correlation coefficients, R2 ranged between 0.647 - 0.995 which is lower than those of the other two models. The average percentage error which is in between 5.428-15.570 found higher than the other models (APE for Freundlich model in between 6.000-16.57 and

APE for Redlich-Peterson in between 5.857-11.43). This indicates that the Langmuir model is not able to perfectly describe the equilibrium data.

As, for the Freundlich model, the values of the correlation coefficients (0.963-1.000) were higher and Chi-square values (0.0002-0.699) also lower than those of the compared models. Also the magnitude of the Freundlich adsorption capacity (n) gives an indication of the favourability of sorption, with its values ranging from 2 to 10 indicating good sorption capacity. All these shows that there was good agreement between the experimental and calculated data for both analysis methods applied with the Freundlich model.

The correlation coefficient values (0.817-1.000) with the Redlich-Peterson model were found to be as good as those of the Freundlich model. Also its average percentage errors (5.714-11.430) and Chi-square values (0.010-0.625) as shown in Tables 6.9,

6.10 and 6.11 respectively were in the same sequence as those of the Freundlich model, they were also able to perfectly describe the experimental data.

189

Chapter 7 Glyphosate Leaching Experiments and One- Dimensional Transport Model

7.1. General

As revealed in the previous studies and the sorption trials in Chapter 6, the non- selective, post-emergence herbicide glyphosate [N (phosphonomethyl)glycine] shows a tendency for strong sorption to soil materials (caused mainly by its aluminium and iron oxides and low pH values [229, 230, 234, 256, 261, 413, 414]). Due to this, and its fast microbial degradation and low toxicity for non-target organisms, the risk of it contaminating surface or groundwater is generally assumed to be low [248, 258].

Nevertheless, a few previous studies have reported that the transport of glyphosate in soils is very variable and depends on the soils’ properties, such as their pH, iron and aluminium oxides, and structures, and the glyphosate’s application time and microbial activity [229, 258, 265, 414]. In addition, some field studies have demonstrated a potential risk to the aquatic environment as they found glyphosate and its degradation products up to depths of 1m [205-207]. The worldwide introduction of gene technology in the form of genetically modified (GM) glyphosate-resistant crops, especially in Australia, the USA and Canada [15, 415, 416], has led to the rapid increase in use of glyphosate which raises concerns regarding its effects in terrestrial and aquatic environments.

The leaching of glyphosate through agricultural soils associated with GM canola cultivation with a higher glyphosate sorption capacity (as discussed in the previous

190

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

chapter) is examined in this chapter. It involves column-leaching experiments on glyphosate-dosed soils using application and flow rates representative of field conditions with bromide as a non-reactive tracer. The results obtained are then incorporated in a one-dimensional transport model (HYDRUS 1D) to determine whether the observed behaviour of glyphosate is in agreement with commonly recognised transport processes.

The first section includes background studies relating to glyphosate’s fate and transport which is followed by the experimental section which discusses the column- leaching experiments, experimental set-up and different application scenarios. Three different glyphosate applications, considering soil-glyphosate contact time variations and supposing that the phosphate accumulation in soil might result in a reduced glyphosate retention capacity and, thus, increased leaching risk, were performed.

In a later section, miscible-displacement experiments with glyphosate and a non- reactive tracer bromide were modelled using the one-dimensional transport model

HYDRUS-1D. At first, the physical transport parameters, the pore-water velocity (ν) and dispersivity (λ), were determined by fitting the experimental bromide breakthrough curves (BTCs) with the analytical solution to the advection-dispersion equations (ADEs) for the pulse boundary condition at the upper and zero gradient condition at the lower. Then, these parameters and those from the sorption experiments were used in HYDRUS 1D to describe the transport behaviour of glyphosate.

Finally, a conceptual model of glyphosate fate and transport in the agricultural system including an overview of glyphosate transport pathways and environmental receptors.

191

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

7.2. Background

Transport of glyphosate from terrestrial to aquatic environments as a potential sorbable compound can occur in either solution or suspension [265]. The dissolved form can be moved through the soil by leaching (sub-surface runoff) and end up as a: (i) surface flow in dissolved form; (ii) surface flow bound to sediment; or (iii) sub-surface flow after drainage [265].

For glyphosate transport in a non-structured soil, e.g., sandy, an oxide-poor soil with high hydraulic conductivity that receives high amounts of precipitation, can allow higher glyphosate leaching with a lower sorption capacity. In this regard previous study reported a sandy soil [249] showed the high risk of glyphosate leaching having a lower sorption capacity. Nevertheless, no leaching of glyphosate (or its degradation product AMPA) was found due to the absence of macropores, i.e., the soil was non- structured and water moved through its matrix as a piston flow in another two years of field study [207]. Other studies also reported no or very little glyphosate leaching in sandy soils without macropores [209, 269, 270]. For structured soils, the transport rates of insecticides can be enhanced by preferential flow pathways through macropores

(loam, clayey soil, etc.) [272] and the sub-surface leaching of glyphosate (and AMPA) are reported in the literature [205-207]. Glyphosate leaching can be severe on gravelly and very coarse-textured soil materials as reported in the literature [209, 210].

The aim of this study is to examine whether the leaching of glyphosate via macropore flows can occur in GM canola agricultural soils with higher sorption capacities. To do this, glyphosate transport experiments involving the column leaching of glyphosate-

192

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

dosed soils using application and flow rates representative of field conditions with bromide as a non-reactive tracer are conducted. The experimental findings are then incorporated in a one-dimensional transport model to determine whether the observed transport behaviour is in agreement with commonly recognised transport processes.

7.3. Materials and Methods

7.3.1. Reagents

HPLC grade glyphosate (PESTANAL, HPLC grade) and the derivatisation reagent 9- fluorenyl methoxycarbonyl chloride (FMOC-Cl) were obtained from Fluka (Germany) and HPLC-grade acetonitrile and diethylether from Sigma-Aldrich Australia. All the other chemicals, including KBr, KCl, HCl, KOH, NaOH, disodium tetraboratedecahydrate (Na2B4O7.10H2O) and potassium phosphate monobasic

(KH2P04), as well as the diethylether were analytical grade.

7.3.2. Solutions

The solutions used in the experiments were prepared by mixing appropriate amounts of an analytical-standard glyphosate (PESTANAL, 99.729%) stock solution and distilled water, with background electrolyte (0.1M KCl). FMOC-Cl solutions of 1 g L-

1 prepared just before the experiments by dissolving the reagent in acetonitrile. The buffer solution (pH =9) was prepared by dissolving 15.255 g of Na2B4O7.10H2O in

1000 mL of distilled water, and a 0.1 M EDTA solution used to pre-treat the samples to correct the sensitivity of the fluorogenic reagent to divalent ions in the amino-acid coupling [343]. The KBr (0.01M) solutions prepared in HPLC-grade water were used as non-adsorbed tracers in the soil column experiments.

193

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

7.3.3. Soil materials

The soils used to fill the columns were collected from three different locations: for the soil A column, from a field in which no GM canola cultivation occurred during the previous five years; for the soil B column, from a field in which canola harvesting had just finished; and, for the soil C column, from a field in which the land was fully prepared for that year’s canola cultivation (location map is provided in Chapter 5). The physico-chemical soil properties were determined using the standard methods presented in Table 6.1 in Chapter 6.

7.4. Column-leaching Experiments

7.4.1. Experimental set-up

In the column-leaching experiments, three glass columns (5 cm inner diameter and 20 cm height), each with a glass sinter plate attached to its bottom and filled with air-dried and sieved top-soil (up to 15 cm in depth which maintained a bulk density of 1.5 g/cm3) originating from one of three different agricultural sites (with the soils termed A, B and C. Each outlet was connected to a Teflon tube (inner diameter 3 mm) attached to a leachate collector below the glass sinter disc. On the top of each column, an arrangement for continuous irrigation was made by fitting a disc with 30 holes to a peristaltic pump, with speed control to control the flow rate, which was connected to a reservoir with water, as shown in Fig. 7.1.

194

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Fig. 7.1: Column-leaching experimental set-up

195

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

The columns were preconditioned for 3 days before application of the glyphosate solution to establish steady-state hydrodynamic conditions which resulted in a constant outflow. Pulse applications of glyphosate were performed with syringes (50 mL) and spray applications applied directly on the soil’s surface using a conventional spray head. Throughout the experiments, effluent samples were taken from the bottom of the column as they passed through the Teflon tube and then analysed for glyphosate, bromide, pH and electrical conductivity. Glyphosate concentrations were quantified following the same procedure as discussed in Chapter 6 for sorption experiments with these experiments performed at a temperature of 20 (±5) oC.

7.4.2. Glyphosate applications

Three different glyphosate applications, considering soil-glyphosate contact time variations and supposing that phosphate accumulation in the soil might result in a reduced glyphosate retention capacity and, thus, increase the leaching risk were performed. Before each glyphosate application, the relevant column was pre-saturated with artificial rain (0.01M CaCl2), as recommended by the EPA guidelines [417]. In application 1, glyphosate was applied as a pulse input and then the soil irrigated using artificial rain (0.01M CaCl2) while, in applications 2 and 3, it was sprayed over a soil column’s surface and left for three days before irrigation was started. In application 3, as well as the pre-saturation solution, additional phosphorus (KH2PO4 (0.008M)) was applied to the column.

The potassium bromide (0.01M) and glyphosate solutions were applied to each column as pulse inputs using 50 mL syringes (Application 1). The glyphosate pulse application

196

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

occurred after one hour of artificial rain, was followed by irrigation with rainwater for

4 hours and then the bromide pulse application was started with 0.01M KBr and KOH as a background electrolyte. The glyphosate application rate was 4.5 kg/ha, as derived from the equation given in the EPA guidelines [417] and chosen based on realistic values, with the amount applied on each column 10 L ha-1 Roundup, the highest of the recommended application rates for farmland (2 to 10 L ha-1). Three hours later, irrigation was continued with its rate (4±0.2 mm/h) chosen according to the highest rainfall occurring during the whole year of GM canola cultivation from sowing to harvesting. The leachates of the columns were collected at different time intervals and analysed for bromide, glyphosate, pH and electrical conductivity.

7.5. Column Experiments Parameters

The miscible-displacement experiments with glyphosate, and bromide as a non- reactive tracer, were modelled using the computer program HYDRUS 1D, a numerical model that can simulate the movement of water and solutes under both equilibrium and non-equilibrium flow and transport conditions in soil. It numerically solves the

Richards equation by finite element methods for variably saturated water flows, and advection-dispersion types of equations for heat and solute transports. To simulate the solute transport using this numerical one-dimensional transport model, it was assumed that the column experiments were conducted under steady-state conditions with homogeneous distributions of water content as the columns were homogeneously packed and irrigated at a constant rate. The aim of using this transport model was to determine whether the observed transport behaviour of glyphosate was in agreement with commonly recognised transport processes. Also, an inverse modelling technique

197

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

was used to fit the model solution to the observed data in order to estimate the reaction and transport parameters, an approach which uses a least-squares optimisation routine to obtain the best-fit model solution by iteratively changing the model parameters until a best fit is achieved.

7.5.1. Bromide transport

The physical transport parameters, pore-water velocity (ν) and dispersivity (λ), were determined by fitting the experimental bromide BTCs with the analytical solution to the ADEs for the pulse boundary condition at the upper and zero gradient condition at the lower. To do this, the CXTFIT code [418] was used in the inverse mode with the

ADE for an inert tracer and rewritten as

휕∁ 휕2∁ 휕∁ = 퐷 − 푣 (7.1) 휕푡 휕푧2 휕푧 where ∁ is the liquid phase concentration [M L-3], 푡the time [T], D the dispersion coefficient [L² T-1], z the depth [L] and vthe pore-water velocity [L T-1]. The parameters vand D were estimated and the dispersivity (λ) and volumetric water

j content (θ)calculated respectively as D/v and w, where j is the irrigation rate. v w

7.5.2. Glyphosate transport

The glyphosate transport was described with a fully kinetic, one-site, convective- dispersive model with degradation and Freundlich kinetic sorption and a chemical equilibrium model based on equations (7.2) to (7.4).The equilibrium model was used to merely predict the glyphosate breakthrough, with the parameters derived from the tracer experiment and theKfandn values from the batch experiments. Additionally, this

198

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

model was used in its inverse mode withKf and n fitted to the glyphosate-equivalent

BTC. The one site sorption model is described by the following system of equations

휕∁ 휕푠푘 휕2∁ 휕∁ 휃 + 휌 = 휃퐷 − 휃휈 − 휇∁ (7.2) 휕푡 휕푡 휕푧2 휕푧

휕푠푘 휌 = 훼 휌[푠푘 − 푠푘] − 휇퐶 (7.3) 휕푡 푘 푒

푘 푛 푠푒 = 퐾푓∁ (7.4)

k where se is the sorbed concentration that would be reached at equilibrium with the liquid phase concentration [MM-1], sk the sorbed concentration of the kinetic sorption

-1 -3 3 -3 sites [MM ], ρ the soil’s bulk density [M L ], θ the water content [L L ], αk a first-

-1 -1 order sorption rate constant [T ], µ a first-order dissipation rate [T ], 퐾푓the Freundlich sorption coefficients [M1-(1/n)LnM-1] and n the Freundlich exponent [-]. These four unknown parameters, αk, µ, Kf and n, were estimated by fitting the curve to the experimental data.

For simulation of the solute transport, the boundary conditions chosen were the flux concentration at the top and zero concentration gradient at the bottom. The physical transport parameters, dispersivity (λ) and pore-water velocity (ν) (provided in Table

7.1), were taken from the solution of the bromide BTC assuming the same physical transport processes as those for bromide.

7.5.3. Results

7.5.3.1. Bromide breakthrough curves

The hydrodynamic properties of each column were found by applying the non-reactive tracer (bromide), the BTCs of which for the three different soil columns are shown in

199

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Fig. 7.2. Faster transport of the bromide was observed in the column with soil C which may have been due to its higher pore-water velocity and lower water content than those of the other columns. The physical properties of these soil columns are provided in

Table 7.1. The bulk densities (ρ) and irrigation rates (jw) were determined experimentally, the pore-water velocities (v) and dispersion coefficients (D) fitted to the bromide BTCs (Fig. 7.3.), and the water contents (θ) calculated. The BTCs of all three soil columns were well described by the physical equilibrium ADEs with r² values of 0.99.

Fig.7.2: Bromide concentrations in leachates obtained from column experiments.

200

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Soil A

Soil B

Soil C

Fig.7.3: Bromide BTCs fitted to the experimental data

201

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Table 7.1: Physical properties of soil columns

Soil Bulk Irrigation Pore-water Dispersion Soil water column density, rate, velocity coefficient content g cm-3 cm h-1 (ν), cm h-1 (D), cm2d-1 (θ), (cm3cm-3) Soil A 1.5 0.4 2.624 19.73 0.152 Soil B 1.5 0.4 2.488 14.64 0.160 Soil C 1.5 0.4 2.734 24.49 0.146

7.5.3.2. Glyphosate breakthrough curves

After application 1, breakthroughs of the glyphosate were observed after 100 h, 312 h and 103 h in soil columns A, B and C respectively. The highest glyphosate breakthrough was observed in soil column C at a peak of 0.7627µg/mL which was in agreement with the lower sorption capacity of soil C than soils A and B (Fig. 7.4).

For glyphosate application 2, comparatively lower concentrations of glyphosate than for application 1 were found in the leachate samples, with the first arrivals in soils A,

B and C at 205 h, 328 h and 168 h respectively. For glyphosate application 3, it was difficult to assess from the present data the role of the phosphorus in the pre-saturation solution in the leaching of glyphosate. Although the columns with soils C and A had the highest glyphosate leaching rates, they were probably closely linked to their macropore structures because bromide was also leached at higher rates. Also, the column with soil B had similar lower glyphosate leaching rates as its bromide BTC.

Therefore, it seems that the competition between phosphate and glyphosate for sorption sites did not lead to higher leaching rates despite the 3 days of contact time between the soil and solution. This somewhat contradicts the previous literature in

202

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

which batch sorption experiments revealed a strong reduction in the sorption of glyphosate in the presence of KH2PO4 [229].

In a summary of the three soils, soil B showed a lower effluent glyphosate concentration which may have been due to its: (1) lower breakthrough concentration of bromide (Fig. 7.2); (2) lower pH value (Fig. 7.5a); or (3) higher electrical conductivity (Fig. 7.5b). All these factors provide information that the lower pore- water velocity of soil B than soils A and C subsequently led to its lower glyphosate effluent concentration.

203

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Application 1

Application 2

Application 3

Fig.7.4: Glyphosate concentrations in leachates under different application scenarios

204

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

(a)

(b)

Fig.7.5: pH and conductivity levels in effluent samples

205

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

7.5.3.3. Mass recovery

The recovery rates of the applied glyphosate for all three applications are shown in

Fig.7.6 with highest amount of glyphosate recovered from application 1. Considering the peak recovery rates of glyphosate, only 16.52%, 10.82 %, 22.40% for application

1, and 3.98%, 1.58%, 6.58% for application 2 and 9.58%, 5.93%, 14.27%for application 3 were recovered from Soil A, B and C respectively. As glyphosate is non- volatile due to its low vapour pressure of 9.3×10-3 MPa (at 25°C) the loss cannot be counted for volatilisation from the surface. The loss can be mainly attributed to

14 glyphosate mineralisation to CO2 as studies conducted under laboratory and outdoor conditions mentioned that the loss of > 40% after 118 days is in accordance with mineralization rates. As reported in a previous study, 43.5-46.5% of glyphosate

14 mineralisation ( CO2) was found within 42 days in a lab experiment [206]. Another study described mineralisation in soil as the main elimination path of glyphosate with the test conducted under outdoor conditions in a two-chamber-lysimeter-test-system

14 and reported more than 50% of the applied radioactivity was measured as CO2 after

50days[419]. Other laboratory studies also reported half-lives of glyphosate of less than 100 days [265, 268, 414, 420].

With the dissipation rates of glyphosate (Table 7.2 and Table 7.3) obtained from

HYDRUS 1D simulation by using an inverse modelling technique to fit the model solution to the observed data the average half-life of glyphosate for first order kinetics was calculated using the following equation:

0.693 푡 = (7.5) 1/2 µ

206

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Application 1

Application 2

Application 3

Fig.7.6: Percent recovery of glyphosate from the three soils columns

207

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

7.5.4. Estimated parameters

As the effluent concentrations of glyphosate in application 1 showed higher peak values than the other two applications, the numerical simulation was conducted considering only glyphosate application as a pulse input. The results obtained using

HYDRUS-1D showed that the transport behaviour of glyphosate in a loam soil was well described by both the equilibrium and non-equilibrium one-site sorption models.

To describe the data using the equilibrium model, the parameters derived from the

1-(1/n) n -1 batch sorption experiments were 퐾푓= 49.49 mg L kg and 1/n = 0.6289 for soil A,

1-(1/n) n -1 1-(1/n) n -1 퐾푓= 76.71 mg L kg and 1/n = 0.7350 for soil B, and 퐾푓= 39.53 mg L kg and 1/n = 0.7400 for soil C. The fitting of this equilibrium model (Fig. 7.7) to the experimental data, taking 퐾푓 and 1/n as free adjustable parameters, compared well with the results obtained using the parameters in Table 7.1. The parameters obtained are provided in Table 7.2.

The glyphosate BTCs were also well described by a fully kinetic, one-site, convective- dispersive model with degradation and Freundlich kinetic sorption and the fitting parameters (Fig. 7.8) for the three soil columns are provided in Table 7.3. As seen in

Tables 7.2 and 7.3, both models fitted the glyphosate transport experimental data well, with the modelling showing that the values of the Freundlich coefficients for both the equilibrium and one-site sorption models were lower than that obtained from the batch experiments. The sorptions estimated from the column experiments often being lower than those estimated from the batch experiments can be explained as the rate of lateral diffusion of the sorbate to sorbent being limited by the advective transport [421] which

208

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

was not considered in the batch experiments. In addition, batch experiments had possible shortcomings which could limit their transferability to field conditions, such as the breakup of particles and surface ruboff during shaking which apparently lead to larger sorption capacities and low soil to solution ratios [422]. Also, as the sorption of glyphosate decreases with increasing pH and decreasing ionic strength, as the pH value was increased in the column-leaching experiments, the sorption of glyphosate could have decreased [256, 261, 413].

Table 7.2: Fitting parameters for glyphosate BTCs using equilibrium model

-1 -1 a 2 Soil 푲풇 1/n(-) 휶풌 (h ) 흁(h ) Half- RMSE R

column (mg1-(1/n)Lnkg-1) life, t1/2 (days) Soil A 37.210 1.000 0.180 0.397 57.000 0.014 0.968 Soil B 74.713 1.000 0.800 0.449 42.000 0.009 0.991 Soil C 34.633 1.000 0.100 0.516 46.000 0.013 0.979 a= Root Mean Square Error

Table 7.3: Fitting parameters for glyphosate BTCs using one-site sorption model

-1 -1 a 2 Soil 푲풇 1/n(-) 휶풌 (h ) 흁(h ) Half- RMSE R column (mg1-(1/n)Lnkg-1) life, t1/2 (days) Soil A 34.547 1.009 0.694 0.294 42.000 0.016 0.958 Soil B 74.460 1.001 1.620 0.406 37.000 0.006 0.996 Soil C 30.724 1.005 0.383 0.365 33.000 0.015 0.972 a= Root Mean Square Error

209

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Soil A

Soil B

Soil C

Fig. 7.7: Glyphosate BTCs fitted with equilibrium model

210

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

Soil A

Soil B

Soil C

Fig. 7.8: Glyphosate BTCs fitted with one-site sorption model

211

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

7.6. Conceptual Site Model

Conceptual site models are part of the data quality objectives process, which as suggested by United States Environmental protection agency is required for all site activities that involve gathering environmental data. The fate and transport of contaminants in the agricultural system is often complex and depends on the physical/chemical properties of the contaminant and its sorption, degradation, transport, volatilisation and photolysis in different compartments of environment.

The use of glyphosate has increased during the past few years, with studies finding that around 15 Mt of glyphosate are applied annually in Australia to control agricultural, urban and roadside weeds [11, 12]. Glyphosate can be introduced into the environments (soil, water and air) through intentional application (control of agricultural weeds, aquatic weeds, algae, or unwanted invertebrates), spray drift, runoff from agricultural applications or runoff from accidental release. Also concern has raised regarding accidents involving fires at herbicide storage facilities or factories situated near rivers or estuaries as significant amounts of herbicide chemicals can enter the nearby riverine system possibly causing considerable environmental damage. As reported in a previous study, around 1350 Mg of chemicals were combusted, and between 10000 and 15000 m3 of contaminated fire-fighting water were discharged into the Rhine river due to the Syntax accident in Base [423]. Many of those chemicals reported were insecticides, and the impact on the environment was catastrophic. The fate process of glyphosate in soil and water with the interactions between the different compartments of the natural environment: surface, vadose zone and saturated zone are provided in a conceptual model shown in Fig.7.9. Further discussion on the fate

212

Chapter 7 Glyphosate Leaching Experiments and One-Dimensional Transport Model

process of glyphosate involving volatilisation, photolysis, sorption, transport and degradation is provided.

213

Fig. 7.9: Conceptual site model of glyphosate fate and transport in the agricultural system

214

7.6.1. Discussion on different processes

7.6.1.1. Volatilisation

Volatility is a characteristic of vapor pressure. Depending on the temperature and humidity, and in some situations wind, a chemical changes from a liquid or solid state into a gas or vapor. Volatilisation of herbicides from soil is dependent on the vapour pressure of the herbicide, soil and environmental conditions. Incorporation of herbicides versus application to the soil surface, can significantly impact on the extent of volatilisation. Soil partition between the soil water, soil air and the soil solid constituents is normally happen after the herbicide application to soil and the rate is mainly depend on desorption kinetics, water–air and solid–air transfer, and evaporation. All these processes are dependent on the chemical and physical properties of the insecticide (water solubility, vapour pressure, air–water partition coefficient or

Henry’s law constant, diffusivity), the soil solution (temperature, ionic strength), the atmosphere (turbulence, temperature, wind speed), and the soil matrix (soil organic matter, minerals). The partitioning of the insecticide between the soil solution and air is described by the Henry’s law constant given below.

Henry’s law:

p = kX2 where, p = vapor pressure of solute (component 2)

k = Henry’s constant

X2= mole fraction of solute

Concentration of saturated solution So, p = kc and k = Vapor pressure of pure solute

215

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model

Due to the low vapour pressure of glyphosate of 9.3×10-3 MPa (at 25°C), volatilisation from the surface is usually neglected.

7.6.1.2. Sorption and desorption

The ability of herbicide to be sorbed on sediments depends on its polarity; is adsorbed more when its polarity is low. After they are sorbed to soil or sediment, herbicides can be desorbed and be submitted to subsequent diffusion. The effect of sorption- desorption on transport of insecticides is to retard their solute diffusion and the process has been studied in previous[424].As examined in case of the strongly adsorbed cationic herbicides, the adsorbed phase degradation rate was more significant than the degradation rate of herbicide in solution [424]. Reported in a previous study sediment with a high sorptive capacity and a high strength of sorption released herbicides to the water at a slow rate and in low amounts, presenting a smaller hazard to phytoplankton and zooplankton but a greater threat to filter feeding organisms which relied upon organic matter at the sediment-water interface, in the water column [425]. Glyphosate sorption and desorption on soil has conducted in this research by batch test and the results are provided in Chapter 6.

7.6.1.3. Photolysis

The energy required to drive the environmental transformations of organic compounds comes directly from the solar radiation. Herbicides absorb light and decompose spontaneously or by reaction with reagents generated by light. Herbicide photodegradation, or photolysis, depends on different factors such as light quality and

216

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model

quantity, water turbidity, and ph. The overall rate of herbicide photolysis is well described by the first-order kinetics and is expressed as follows[426].

2.303 푘 = 푓 ∑ 퐼 푒 (7.6) 푝 퐽 휆 휆

Where kp = photolysis rate constant

J = conversion constant

푓 = quantum yield

Iλ = sunlight intensity at wavelength X eλ = molar absorbitivity or molar extinction coefficient at wavelength λ.

The photochemical process is divided into two terms: direct and indirect, the direct photolysis of insecticides proceeds through direct sorption of sunlight energy and the indirect photolysis require a sensitizer to initiate light absorption. (Sensitizers include

O3). However, as discussed in previous literature glyphosate is stable to photolysis.

7.6.1.4. Degradation

Degradation is the breakdown of a compound by microorganisms, which can happen due to the alteration of the chemical structure of a substance resulting in loss of specific property of that substance, or the complete breakdown of a compound to either fully

- + oxidised or reduced, simple molecules (e.g., CO2, CH4, H2O, NO3 , NH4 ) and microbial biomass. Glyphosate is completely degraded in soil by microorganisms with its major metabolites of degradation being aminomethylphosphonic acid which is also degraded in soils but at a slower rate than glyphosate. Glyphosate decomposition has been shown under both aerobic and anaerobic conditions. Studies related to degradation of glyphosate showed that glyphosate is first transformed to

217

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model

aminomethylphosphonic acid (AMPA) by cleavage with production of CO2, and

3- AMPA is then mineralized by phosphonates activity that cleaves C-P pond to PO4 ,

+ with production of NH4 , CO2, and H2O. For the first order kinetics the total dissipation rate or decay rate coefficient (휇) can be calculated by the usual formula:

0.693 t = (7.7) 1/2 µ

It follows that herbicides with smallµ values will persist longer in the environment

(large t1/2) than herbicides with larger µ.

The dissipation rate and the half-life of glyphosate have already been calculated in the above section of this Chapter and results are provided in Table 7.2 and Table 7.3.

7.6.1.5. Transport

Rainfall events after the glyphosate application impact the transport of glyphosate through runoff and leaching. For observing glyphosate transport through the soil from the agricultural land column leaching studies have conducted in the earlier section of this Chapter.

7.7 Conclusions

This study revealed leaching experiments involving column-leaching trials under different scenarios of glyphosate application, with the results incorporated in one- dimensional equilibrium and non-equilibrium models with Freundlich sorption and degradation. Background studies relating to glyphosate’s fate and transport followed by the experimental section with the column-leaching experiments, experimental set- up and different application scenarios were discussed in the first section of this chapter.

218

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model

Glyphosate applications were performed considering soil-glyphosate contact time variations and supposing that the phosphate accumulation in soil might result in a reduced glyphosate retention capacity and, thus, increased leaching risk. Results for application 1 demonstrates breakthroughs of the glyphosate after 100 h, 312 h and 103 h in soil columns A, B and C respectively. The highest glyphosate breakthrough was observed in soil column C at a peak of 0.7627µg/mL which was in agreement with the lower sorption capacity of soil C than soils A and B. Glyphosate application 2, revealed comparatively lower concentrations of glyphosate than for application 1 with the first arrivals in soils A, B and C at 205 h, 328 h and 168 h respectively. In case of glyphosate application 3, it was difficult to assess the role of the phosphorus in the pre-saturation solution in the leaching of glyphosate from the results obtained.

From the three soils, soil B showed a lower effluent glyphosate concentration which may have been due to its: (1) lower breakthrough concentration of bromide; (2) lower pH value or (3) higher electrical conductivity. All these factors provide information that the lower pore-water velocity of soil B than soils A and C subsequently led to its lower glyphosate effluent concentration.

The recovery rates of the applied glyphosate for all three applications revealed that considering the peak recovery rates of glyphosate, only 16.52%, 10.82 %, 22.40% for application 1, and 3.98%, 1.58%, 6.58% for application 2 and 9.58%, 5.93%,

14.27%for application 3 were recovered from Soil A, B and C respectively. The loss

14 can be mainly attributed to glyphosate mineralisation to CO2as previous studies provide information in agreement to this.

219

Chapter 7 Glyphosate Leaching Experiments and One-dimensional Transport Model

The miscible-displacement experiments with glyphosate and a non-reactive tracer bromide were modelled in the later section using the one-dimensional transport model

HYDRUS-1D. Physical transport parameters were determined first by fitting the experimental bromide breakthrough curves (BTCs) with the analytical solution to the advection-dispersion equations (ADEs) for the pulse boundary condition at the upper and zero gradient condition at the lower. Afterwards, these parameters were used in

HYDRUS 1D to describe the transport behaviour of glyphosate. The glyphosate transport was described with a fully kinetic, one-site, convective-dispersive model with degradation and Freundlich kinetic sorption and a chemical equilibrium model.

Later glyphosate fate and transport in the agricultural system was presented in a conceptual model with glyphosate transport pathways and environmental receptors.

220

Chapter 8 Conclusions and Future Recommendations

8.1. General

Government policies for increasing the production of biofuels and the current insufficient facilities for second-generation production have encouraged many countries to continue crop-based biofuel production. Consequently, political debates and concerns regarding the environmental impacts of land conversion and the increased use of herbicides have been raised. Although genetically modified (GM) crops have become a new option for biofuel production, they are associated with widespread application of the most commonly used herbicide, glyphosate. The prohibitive cost of analysing glyphosate in environmental samples often creates difficulties for monitoring studies. Therefore, a simple, fast and low-cost method for analysing it in water and soil samples, which can be used for field monitoring and is also suitable for glyphosate sorption and transport studies of agricultural soils, has been developed in this thesis. This method, along with the commercially available enzyme-linked immunosorbent assay (ELISA) method, were used to conduct a site investigation of farmland cultivated with GM canola in the Parkes region of New

South Wales, Australia, which involved sampling the surface waters and soils to analyse their glyphosate concentrations through batch sorption experiments and column leaching trials. The results obtained from the leaching experiments and parameters from the batch experiments were further incorporated in a one-dimensional transport model, HYDRUS-1D, and well described by a one-site sorption model.

Later, a conceptual model of glyphosate’s fate and transport in an agricultural soil

221

Chapter 8 Conclusions and Future Recommendations

system was provided with an overview of glyphosate transport pathways and environmental receptors.

8.2. Major Conclusions of Thesis

A summary of the specific conclusions drawn from this research thesis is provided below.

To avoid the prohibitive cost of analysing the contaminant glyphosate in environmental samples, a simple, fast and low-cost direct fluorescence spectrometry method was developed, with all the procedures involved in preparing samples, establishing standards and developing standard calibration curves, laboratory Quality

Control (QC) and analytical method validations described. The process involvedderivatisationusing9-fluorenylmethyl chloroformate (FMOC-chloride) and measurements of emission acquisitions at 310 nm. The calibration curve showed linearity over the range 10 µg/L to 25000 µg/L using a linear regression (Model I) with good reproducibility (recovery 108.3 ± 20.8; R2=0.997) and also in the lower range 10

µg/L to 1,000 µg/L also using a linear regression (Model III) (recovery 115.230 ±

25.852; R2 = 0.998). The appropriate limits of detection of the method were 10 µg/L and 0.063 mg/kg respectively in waters and soils, and the limits of quantification were

29 µg/L and 0.194 mg/kg respectively. Comparisons with the previously established direct low-cost enzyme-linked immunosorbent assay (ELISA) method in terms of accuracy and precision for the detection and quantification of glyphosate-spiked water and soil samples and the HPLC method were conducted.

222

Chapter 8 Conclusions and Future Recommendations

Cost estimates indicate that the new method can be conducted for a cost of AU$47.42 per sample by a commercial laboratory in 2014 Australian dollars, excluding GST.

This compares favourably with quoted commercial rates of AU$150-280 per sample using the existing method, also excluding GST (Appendix A).

A site investigation of GM canola cultivation in the Parkes region of NSW, in accordance with the consent and ethics approval procedures of the University of NSW, was conducted. Firstly, GM canola growers from different farms in NSW were interviewed and then a family-operated farm in the central west of NSW was selected for a detailed study. The investigation was based on concerns regarding the widespread application of the herbicide glyphosate and its further incorporation in soil and surface waters. A sample analysis program and Quality Assurance (QA) assessment, including a detailed QC analysis, were performed according to the Australian National

Environmental Protection Measure (NEPM) sampling protocol. The glyphosate levels obtained were checked with both Australian and international drinking water guidelines as well as other accepted worldwide guidelines.

The effects of long-term glyphosate applications on the sorption and desorption of glyphosate on three different soils from Parkes region of NSW were examined. This involved an evaluation of sorption isotherm models using the new low-cost fluorescence spectrometric method and comparing the results with those from an established UV-visible spectrophotometric method. Different linear and non-linear form of the Langmuir, Freundlich and Redlich-Peterson isotherms models were applied to the experimental data which agreed with the Freundlich isotherm. A

223

Chapter 8 Conclusions and Future Recommendations

comparison of the linear and non-linear models was conducted. Glyphosate desorption tests were performed using a 0.1M KH2PO4 solution on the same soils.

Column leaching experiments were conducted with glyphosate-dosed soils using application and flow rates representative of field conditions, with bromide as a non- reactive tracer. The experimental setup and different glyphosate application scenarios considering soil-glyphosate contact time variations and supposing phosphate accumulation were also incorporated.

The column leaching experiments with glyphosate and bromide (as a non-reactive tracer) were modelled using a one dimensional transport model HYDRUS-1D with the sorption parameter obtained from the batch experiments. The physical transport parameters were obtained by fitting the experimental bromide breakthrough curves

(BTCs) with the analytical solution to the advection-dispersion equation. To describe the transport experiments, a fully kinetic, one-site convective-dispersive model with degradation and Freundlich sorption, and an equilibrium model were considered, with mass recoveries from the column experiments also provided.

Finally, the findings were incorporated in a conceptual model of glyphosate’s fate and transport in an agricultural soil system, including information regarding its transport pathways and environmental receptors.

8.3. Recommendations for Future Research

The principal degradation product of glyphosate is aminomethylphosphonic acid

(AMPA) to which less attention was paid in this thesis. Therefore, further research to

224

Chapter 8 Conclusions and Future Recommendations

develop a simple, fast and low-cost method which can determine both glyphosate and

AMPA is recommended.

The site investigation for this agricultural land was a small-scale farm and it focused on only one contaminant, glyphosate. As it would be interesting to conduct a site investigation considering all potential contaminants and their exposures, further research on this aspect is recommended

As the leaching experiments conducted in this study considered soils to depths of only up to 20 cm, there is a need for more experiments on this agricultural land for soil depths of >1m.

As environmental samples were collected mainly from the surface waters, different groundwater samples analysis to assess the risk of glyphosate contamination should also be examined.

A complete health risk assessment of the fate of glyphosate and AMPA on this agricultural land is recommended.

As modelling of the transport behaviour of glyphosate in this thesis was limited mainly to a 1D case, further modelling using 3D cases is recommended.

225

References

[1] "Monitoring of the Australian petroleum industry," Australian Competition and Consumer Commission 2011. [2] "Australian Biofuels," APAC biofuels consultants 2011. [3] "Ethanol in Gasoline," Environmental Protection Act., Ontario Regulation. [4] "The potential impacts of increased corn production for ethanol in the Great Lakes-St. Lawrence river region," U.S. army corps of engineers 2007. [5] B. R. Scanlon, I. Jolly, M. Sophocleous, and L. Zhang, "Global impacts of conversions from natural to agricultural ecosystems on water resources: quantity versus quality," Water Resources Research, vol. 43, 2007. [6] J. Gressel, "Transgenics are imperative for biofuel crops," Plant Science, vol. 174, pp. 246-263, 2008. [7] J. R. Park, I. McFarlane, R. H. Phipps, and G. Ceddia, "The role of transgenic crops in sustainable development," Plant Biotechnology Journal, vol. 9, 2010. [8] Australian Agricultural Exports. Available: www.dfat.gov.au [9] Genetically modified crop fact sheet. Available: www.daff.gov.au [10] S. Neales, "Foreign funds to fertilise ‘dining boom’," in The Australian, ed, 18 July 2012. [11] B. R. (Bob), B. K. I, J. G. J, M. Denise, R. R. E, and T. Dennis, "Australian State of the Environment Committee," 2006. [12] J. C. Radcliffe, "Pesticide use in Australia," Parkville, Victoria.2002. [13] D. Barcelo, Trace determination of pesticides and their degradation products in water, 2003. [14] W. A. Battaglin, D. W. Kolpin, E. A. Scribner, K. M. Kuivila, and M. W. Sandstrom, "Glyphosate, other herbicides and tranformation products in Midwestern streams, 2002," Journal of the American Water Resources Association, vol. 41, pp. 323-332, 2005. [15] E. A. Scribner, W. A. Battaglin, R. J. Gilliom, and M. T. Meyer, "Concentrations of glyphosate, its degradation product, aminomethylphosphonic acid, and glufosinate in ground and surface-water, rainfall, and soil samples collected in the United States, 2001-06," U.S. Geological Survey2007. [16] A. M. Davis, P. J. Thorburn, S. E. Lewis, Z. T. Bainbridge, S. J. Attard, R. Milla, et al., "Environmental impacts of irrigated sugarcane production: Herbicide run-offdynamics from farms and associated drainage systems," Agriculture, Ecosystems and Environment, vol. 180, pp. 123-135, 2013. [17] J. Struger, D. Thompson, B. Staznik, P. Martin, T. McDaniel, and C. Marvin, "Occurrence of glyphosate in surface waters of Southern Ontario," Bulletin of Environmental Contamination and Toxicology, vol. 80, pp. 378-384, 2008. [18] H. Mensink and P. Janssen, "Glyphosate," Geneva1994. [19] "National Water Quality Management Strategy, Australian Drinking Water Guidelines -6, Vol 1, 2011." [20] National drinking water database. Available: www.ewg.org [21] L. L. Lin and C. Y. Heong. (2014). Danish water contaminated by RoundUp, ban imposed Available: www.twnside.org.sg

226

References

[22] J. D. Byer, J. Struger, P. Klawunn, A. Todd, and E. Sverko, "Low cost monitoring of glyphosate in surface waters using the ELISA method: an evaluation," Environmental Science & Technology, vol. 42, pp. 6052–6057, 2008. [23] "Birchip Cropping Group GM canola grower case studiesI, BCG/GRDC project – BWD00016." [24] J. P. Giesy, S. Dobson, and K. R. Solomon, "Ecotoxicological risk assessment for roundup herbicide," Reviews of Environmental Contamination and Toxicology, vol. 167, pp. 35-120, 2000. [25] ANZECC and ARMCANZ, "Australian and New Zealand guidelines for fresh and marine water quality," Australian and New Zealand Environment and Conservation Council and Agriculture and Resource Management Council of Australia and New Zealand.2000. [26] "WHO Guidelines for Drinking-water Quality, Glyphosate and AMPA in Drinking-water, WHO/SDE/WSH/03.04/97, updated on June 2005." [27] "WHO guidelines for drinking-water Quality, Vol 1," 2008. [28] Drinking water contaminants, Environmental Protection Agency (EPA), USA, Viewed: 16 April 2013; . Available: http://water.epa.gov/ [29] Technical fact sheet on glyphosate. Available: www.epa.gov [30] "Safe drinking water act," Ontario drinking water quality standards2002. [31] "Canadian water quality guidelines for the protection of aquatic life," 2007. [32] "Surface water quality guidelines for use in Alberta," Environmental Assurance Division1999. [33] H. Davies, "An assessment of the suitability of a series of Western Cape farm dams as waterbird habitats," MSc, Zoology, University of Cape Town, Cape Town, 1997. [34] M. A. Dalvie, E. Cairncross, A. Solomon, and L. London, "Contamination of rural surface and ground water by endosulfan in farming areas of the Western Cape, South Africa," Environmental Health, 2003. [35] P. K. Mensah, C. G. Palmer, and W. J. Muller, "Derivation of south african water quality guidelines for roundups using speciessensitivity distribution," Ecotoxicology and Environmental Safety, vol. 96, pp. 24-31, 2013. [36] "A policy pollution prevention, waste minimization, impact control and remediation," Department of Environmental Affairs and Tourism (DEAT), Pretoria1998. [37] "National water act.," Department of Water Affairs and Forestry (DWAF), Pretoria1998. [38] "South african water quality guidelines," Department of Water Affairs and Forestry (DWAF), Pretoria1997. [39] E. Engelhaupt, "Perspective: biofueling water problems," Environmental Science & Technology, vol. 41, pp. 7593-7595, 2007. [40] S. Kodama, Y. Ito, A. Taga, Y. Nomura, A. Yamamoto, S. Chinaka, et al., "A fast and simple analysis of glyphosate in tea beverages by capillary electrophoresis with on-line copper(II)-glyphosate complex formation," Journal of Health Science, vol. 54, pp. 602-606, 2008. [41] K. Qian, T. Tang, T. Shi, F. Wang, J. Li, and Y. Cao, "Residue determination of glyphosate in environmental water samples with high-performance liquid

227

References

chromatography and UV detection after derivatization with 4-chloro-3,5- dinitrobenzotriﬂuoride," Analytica Chimica Acta, vol. 635, pp. 222-226, 2009. [42] Francisco Sánchez-Bayoa, Ross V. Hyneb, and K. L. Desseilleb, "An amperometric method for the detection of amitrole, glyphosate and its aminomethyl-phosphonic acid metabolite in environmental waters using passive samplers," Analytica Chimica Acta, vol. 675, pp. 125–131, 2010. [43] E. A. Songa, O. A. Arotiba, J. H. O. Owino, N. Jahed, P. G. L. Baker, and E. I. Iwuoha, "Electrochemical detection of glyphosate herbicide using horseradish peroxidase immobilized on sulfonated polymer matrix," Bioelectrochemistry, vol. 75 pp. 117-123, 2009. [44] C. D. Stalikas and C. N. Konidari, "Analytical methods to determine phosphonic and amino acid group-containing pesticides," Journal of Chromatography A, vol. 907, pp. 1-19, 2001. [45] H. Kataoka, S. Ryu, N. Sakiyama, and M. Makita, "Simple and rapid determination of the herbicides glyphosate and glufosinate in river water, soil and carrot samples by gas chromatography with flame photometric detection," Journal of Chromatography A, vol. 726, pp. 253-258, 1996. [46] Su-Hsiang Tseng , Yi-Wen Lo , Pi-Chiou Chang , Shin-Shou Chou , and a. H.- M. Chang, "Simultaneous Quantification of Glyphosate, Glufosinate, and Their Major Metabolites in Rice and Soybean Sprouts by Gas Chromatography with Pulsed Flame Photometric Detector," J. Agric. Food Chem vol. 52, pp. 4057–4063, 2004. [47] Maxim V. Khrolenko and P. P. Wieczorek, "Determination of glyphosate and its metabolite aminomethylphosphonic acid in fruit juices using supported- liquid membrane preconcentration method with high-performance liquid chromatography and UV detection after derivatization with p-toluenesulphonyl chloride," Journal of Chromatography A, vol. 1093, pp. 111-117, 2005. [48] I. D. Merás, T. G. Dı́az, and M. A. Franco, "Simultaneous fluorimetric determination of glyphosate and its metabolite, aminomethylphosphonic acid, in water, previous derivatization with NBD-Cl and by partial least squares calibration (PLS)," Talanta, vol. 65, pp. 7-14, 2005. [49] J. S. Ridlen, G. J. Klopf, and T. A. Nieman, "Determination of glyphosate and related compounds using HPLC with tris(2,2′-bipyridyl)ruthenium(II) electrogenerated chemiluminescence detection," Analytica Chimica Acta, vol. 341, pp. 195-204, 1997. [50] M. Ibáñez, Ó. J. Pozo, J. V. Sancho, F. J. López, and F. Hernández, "Residue determination of glyphosate, glufosinate and aminomethylphosphonic acid in water and soil samples by liquid chromatography coupled to electrospray tandem mass spectrometry," Journal of Chromatography A, vol. 1081, pp. 145- 155, 2005. [51] R.J. Vreeken , P. Speksnijder, I. Bobeldijk-Pastorova , and T. H. M. Noij, "Selective analysis of the herbicides glyphosate and aminomethylphosphonic acid in water by on-line solid-phase extraction–high-performan liquid chromatography–electrospray ionization mass spectrometry," Journal of Chromatography A, vol. 794, pp. 187–199, 1998. [52] C. N. Albers, G. T. Banta, P. E. Hansen, and O. S. Jacobsen, "The influence of organic matter on sorption and fate of glyphosate in soil – Comparing different

228

References

soils and humic substances," Environmental Pollution, vol. 157, pp. 2865- 2870, 2009. [53] N. E. P. Council, "Data collection, sample design and reporting," National Enviornmnetal Protection (Assessment of Site Contamination) Measure, NEPM1999. [54] C. V. Waiman, M. J. Avena, M. Garrido, B. F. Band, and G. P. Zanini, "A simple and rapid spectrophotometric method to quantify the herbicide glyphosate in aqueous media. Application to adsorption isotherms on soils and goethite," Geoderma, vol. 170, pp. 154-158, 2012. [55] N. E. P. A. o. S. C. Measure, "Investigation levels for soil and groundwater," 2013. [56] L. R. Brown, Plan B 4.0: Mobilizing to save civilization. New York: W. W. Norton & Company, 2009. [57] G. Sorda, M. Banse, and C. Kemfert, "An overview o fbiofuel policies across the world," Energy Policy, vol. 38, pp. 6977-6988, 2010. [58] "Annual world ethanol production by country (based on F.O.Licht estimates)," Renewable Fuel Association, RFA2010. [59] "Agricultural outlook 2009-2018," OECD-FAO. [60] E. D. Larson, "Biofuel production technologies: status, prospects and implications for trade and development," Princeton Environmental Institute, Princeton University, 2008. [61] D. Mitchell, "A note on rising food prices " The World Bank2008. [62] "Addressing the global food crisis: Key trade, investment and commodity policies in ensuring sustainable food security and alleviating poverty," presented at the The High-Level Conference on World Food Security: The Challenges of Climate Change and Bioenergy Rome, Italy, 2008. [63] J. Schmidhuber, "Biofuels: An emerging threat to Europe’s Food Security? Impact of an increased biomass use on agricultural markets, prices and food security: A longer-term perspective," Notre Europe2007. [64] S. Johnson, "The (food) price of success," International Monetary Fund2007. [65] V. Mercer-Blackman, H. Samiei, and K. Cheng, "Biofuel demand puschs up food prices," International Monetary Fund 2007. [66] (2010). FAO food price index, Food and Agricultural Organisation of the United Nations. Available: www.fao.org [67] I. d. C. Macedo, M. R. L. V. Leal, and J. E. A. R. d. Silva, "Assessment of greenhouse gas emissions in the production and use of fuel ethanol in Brazil," Government of the State of São Paulo2004. [68] D. Pimentel, "Ethanol Fuels: Energy Balance, Economics, and Environmental Impacts are Negative," Natural Resources Research, vol. 12, 2005. [69] A. E. Farrell, R. J. Plevin, B. T. Turner, A. D. Jones, M. O'Hare, and D. M. Kammen, "Ethanol can contribute to energy and environmental goals," Science, vol. 311, pp. 506-508, 2006. [70] P. J. Crutzen, A. R. Mosier, K. A. Smith, and W. Winiwarter, "N2O release from agro-biofuel production negates global warming reduction by replacing fossil fuels," Atmos. Chem. Phys. Discuss, vol. 7, 2007. [71] "Biofuel support policies: An economic assessment," OECD2008. [72] Biofuels fact sheet. Available: www.industry.gov.au [73] "Biofuels fact sheet," Australian Institute of Petroleum.

229

References

[74] "Quick guide: Biofuels," in BBC News, ed, 2007. [75] (25-02-2014). Biofuels, Biofuel Association of Asutralia. Available: www.biofuelsassociation.com.au [76] (2011). Biofuel production from waste streams, European Biofuels Technology Platform. Available: www.biofuelstp.eu [77] N. Elam, "Automotive fuels survey," International Energy Agency, IEA, Netherlands1996. [78] R. v. d. Broek, M. v. Walwijk, P. Niermeijer, and M. Tijmensen, "Biofuels in the Dutch market: a fact-finding study," Netherlands agency for energy and the environment, NOVEM2003. [79] "Biofuels—at what cost? Government support for ethanol and biodiesel in the United States," The Global Studies Initiative2007. [80] "Renewable fuel standards for Renewable Fuel Standard program (RFS2)," United States Environmnetal Protection Agency2010. [81] "EPA proposes 2014 renewable fuel standards, 2015 biomass-based diesel volume," Environmental Protection Agency (EPA)2013. [82] "Biofuels in the US transportation sector," Annual Energy Outlook2007. [83] D. Koplow and R. Steenblik, "Subsidies to ethanol in the United States," Biofuels, Solar and Wind as Renewable Energy Systems, vol. 4, pp. 79-108, 2008. [84] D. Dessureault, "Canada biofuels annual," USDA Foreign Agricultural Service2008. [85] D. Dessureault, "Canada biofuels annual," USDA Foreign Agricultural Service2009. [86] D. Corporation, "Total economic impact assessment of biofuels plants in Canada," The Canadian Renewable Fuels Association (CRFA)2010. [87] "Biofuels in Australia—An overview of issues and prospects," Australian Government Rural Industries Research and Development Corporation, AGRIRDC2007. [88] D. Quirke, R. Steenblik, and B. Warner, "BIOFUELS – at what cost? Government support for ethanol and biodiesel in Australia " The Global Subsidies Initiative 2008. [89] D. Rutz, R. Janssen, J. Rogat, K. Borch, M. Mittelbach, S. Schober, et al., "Biofuel policies and legislationin latin America,," presented at the 18th European Biomass Conference & Exhibition, Lyon, France, 2009. [90] K. Joseph, "Biofuels Annual " USDA Foreign Agricultural Service2009. [91] E. F. d. Almeida, J. V. Bomtempo, and C. M. d. S. e. Silva, "The performanceof Brazilian biofuels: an economic, environmental and social analysis," Joint Transport Research Centre, Rio de Janeiro, Brazil2008. [92] M. Kojima and T. Johnson, "Potential for biofuels for transport in developing countries," Energy Sector Management Assistance Programme (ESMAP)2006. [93] "Ruling on biodiesel introduction among Brazil national energy sources," Legal Office FAOLEX2005. [94] "On the promotion of the use of energy from renewable sources and amending and subsequently repealing directives 2001/77/EC and 2003/30/EC " Directive 2009/28/EC of the European Parliament and of the Council2009.

230

References

[95] "BIOFUELS – at what cost? Government support for ethanol and biodiesel in China," The Global Subsidies Initiative2008. [96] C. Beckman and J. Junyang, "Peoples Republic of China biofuels annual," USDA Foregin Agricultural Service2009. [97] T. Altenburg, H. Dietz, M. Hahl, N. Nikolidakis, C. Rosendahl, and K. Seelige, "Biodiesel in India: value chain organisation and policy options for rural development," German Development Institute, vol. 43, 2009. [98] H. S. Dillon, T. Laan, and H. S. Dillon, "Biofuels—at what cost? Government support for ethanol biodiesel in Indonesia," The Global Studies Initiative2008. [99] G. P. Lopez and T. Laan, "Biofuels—at what cost? Government support for biodiesel in Malaysia," The Global Studies Initiative,2008. [100] R. Hoh, "Malaysia Biofuels Annual," USDA Foreign Agricultural Service2009. [101] S. Preechajarn and P. Prasertsri, "Thailand Biofuels Annual," USDA Foreign Agricultural Service2009. [102] S. Preechajarn, "Biodiesel demand and supply outlook " USDA Foreign Agricultural Service2010. [103] U. Schwertmann, "Soil erosion: extent, prediction and protection in Bavaria," in Soil Erosion in the European Community – Impact of Changing Agriculture, G. Chisci and R. P. C. Morgan, Eds., ed Rotterdam, Balkema, 1986, pp. 185– 200. [104] R. Lal., Soil erosion in the tropics : principles and management New York: McGraw-Hill, 1990. [105] S. A. El-Swaify, "Factors affecting soil erosion hazards and conservation needs for tropical steeplands," Soil Technology, vol. 11, pp. 3-16, 1997. [106] H. Bakker, Sugar cane cultivation and management: Springer, 1999. [107] J. H. Payne, Cogeneration in the cane sugar industry: Elsevier, 1990. [108] A. E. Hartemink, Soil fertility decline in the tropics with case studies on plantations. Wallingford: CABI, 2003. [109] R. J. Haynes and C. S. Hamilton, "Effects of sugarcane production on soil quality: a synthesis of world literature," in South African Sugar Technologists’ Association, 1999, pp. 46-52. [110] F. Blackburn, Sugar-cane, 1st ed. London and New York, 1984. [111] F. Coale and T. Schueneman, "Sugarcane response to limestone and gypsum application on acidic sandy soils," American Society of Sugar Cane Technologists, vol. 13, pp. 94-97, 1993. [112] A. Garside, R. Bramley, J. Holt, R. Magarey, C. Pankhurst, and J. Skjemstad, "Comparisons between paired old and new land sites for sugarcane growth and yield and soil chemical, physical, and biological properties," Proceedings of the Australian Society of Sugar Cane Technologists, vol. 19, pp. 60-66, 1997a. [113] N. R. Maclean, "Long-term effects of sugar cane production on some physical and chemical properties of soils in the Goondi Mill area," 1975, pp. 123-126. [114] A. W. Wood, "Soil degradation and management under intensive sugarcane cultivation in North Queensland," Soil Use and Management, vol. 1, pp. 120- 124, 1985. [115] J. Caron, C. R. Espindola, and D. A. Angers, "Soil structural stability during rapid wetting: influence of land use on some aggregate properties," Soil Science Society of America vol. 60, pp. 901-908, 1996.

231

References

[116] A. S. Masilaca, R. A. Prasad, and R. J. Morrison, "The impact of sugarcane cultivation on three Oxisols from Vanua Levu, Fiji," Tropical Agriculture, vol. 63, pp. 325-330, 1986. [117] A. E. Hartemink and L. S. Kuniata, "Some factors influencing yield trends of sugar cane in Papua New Guinea," Outlook on Agriculture, vol. 25, pp. 227- 234, 1996. [118] R. A. J. Alaban, F. C. Barredo, and A. L. Aguirre, "An assessment of some indicators and determinants of farm productivity and soil fertility in the VMC district: trends, associations, interactions (1969-1989)," PHILSUTECH Proceedings, vol. 37, pp. 64-83, 1990. [119] F. Abruna-Rodriguez and J. Vicente-Chadler, "Sugarcane yields as related to acidity of a humid tropical Ultisol," Agronomy Journal, vol. 59, pp. 330-332, 1967. [120] B. L. Schroeder, J. B. Robinson, M. Wallace, and P. Turner, "Soil acidification: occurrence and effects in the south african sugar industry," in South African Sugar Technologists' Association, 1994, pp. 70-74. [121] M. J. Lindstrom, "Effects of residue harvesting on water runoff, soil erosion and nutrient loss," Agriculture, Ecosystems & Environment, vol. 16, pp. 103- 112, 1986. [122] J. F. Power and J. W. Doran, "Role of crop residue management in nitrogen cycling and use," ASA special publication - American Society of Agronomy, vol. 51, pp. 101-113, 1988. [123] M. E. Sumner, "Opportunities for amelioration of soil physical and chemical constraints under intensive cropping," in Intensive Sugarcane Production: Meeting the Challenges beyond 2000, Brisbane, Australia, 1997, pp. 305–326. [124] F. Ghassemi, A. J. Jakeman, and H. A. Nix, Salinisation of land and water resources : human causes, extent, management, and case studies. Sydney, NSW: NSW University Press, 1995. [125] C. A. Sweeney, P. J. Thorburn, and K. L. Bristow, "Water table contributions to sugarcane production : impacts on irrigation water use efficiency," in Australian Cane Sugar Technologists, 2001, pp. 116-121. [126] G. J. Ham, A. Z. Cox, and G. G. McMahon, "A review of sodic soils research in the Queensland sugar industry," in Australian Society of Sugar Cane Technologists, 1997. [127] T. N. Tiwari, R. P. Srivastava, and G. P. Singh, "Salinity tolerance in sugarcane cultivars," Sugarcane, vol. 1, pp. 10-14, 1997. [128] J. L. Sehgal, A. I. Allam, R. P. Gupta, and A. Aziz, "The suitability of the soils of the Mesopotamian plain for sugarcane cultivation," in International Society of Sugar Cane Technologists, 1980, pp. 132–151. [129] T. M. Sohrabi, A. Keshavarz, and A. Pazira, "An overview of land drainage in Iran. Drainage in the 21st century: food production and the environment," in Seventh International Drainage Symposium, Orlando, Florida, USA, 1998, pp. 70-81. [130] M. Workman, P. M. Scott, and D. J. Nixon, "A review of the management and amelioration of saline/sodic soils at Mhlume (Swaziland) sugar company," in South African Sugar Technologists' Association, 1986.

232

References

[131] M. Wagner, C. Rincones, G. Medina, and M. Mujica, "Effect of the chisel plough and of irrigation on development of sugarcane in a compacted soil," Agronomia Tropical, vol. 45, pp. 5-26, 1995. [132] SASA, "Standards and guidelines for conservation and environmental management," 2002. [133] E. J. Schmidt, "Improved water management for sugarcane production," in 6th International Micro-irrigation Congress, Cape Town, South Africa, 2000, pp. 1-7. [134] M. Azhar Javed and M. Tariq, "Water crises and production gaps in punjab, risk analysis and mitigation measures," Science International, Lahore2003. [135] D. Yewalekar, "Micro irrigation: a foot step of the 21st century," in International Agricultural Engineering Conference, Bangkok, Thailand, 1998, pp. 684-693. [136] S. Raine, "Increasing the efficiency of furrow irrigation for sugarcane production in the Burdekin," in 12th Irrigation Association of Australia Conference, Tamworth, 1995, pp. 52–58. [137] R. C. Muchow and B. A. Keating, "Assessing irrigation requirements in the Ord Sugar Industry using a simulation modelling approach," Australian Journal of Experimental Agriculture, vol. 38, 1998. [138] R. L. Bengtson, H. M. Selim, and R. Ricaud, "Water quality from sugarcane production on alluvial soils," Transactions of the ASAE, vol. 41, pp. 1331- 1336, 1998. [139] D. Haynes, J. Müller, and S. Carter, "Pesticide and herbicide residues in sediments and seagrasses from the Great Barrier Reef world hreritage area and Queensland coast," Marine Pollution Bulletin, vol. 41, pp. 279-287, 2000. [140] J. F. Müller, S. Duquesne, J. Ng, G. R. Shaw, K. Krrishnamohan, K. Manonmanii, et al., "Pesticides in sediments from queensland irrigation channels and drains," Marine Pollution Bulletin, vol. 41, pp. 294-301, 2000. [141] Z. T. Bainbridge, J. E. Brodie, J. W. Faithful, D. A. Sydes, and S. E. Lewis, "Identifying the land-based sources of suspended sediments, nutrients and pesticides discharged to the Great Barrier Reef from the Tully–Murray Basin, Queensland, Australia.," Marine and Freshwater Research, vol. 60, pp. 1081- 1090, 2009. [142] A. Davis, S. Lewis, Z. Bainbridge, J. Brodie, and E. Shannon, "Pesticide residues in waterways of the lower Burdekin region: challenges in ecotoxicological interpretation of monitoring data," Australasian Journal of Ecotoxicology, vol. 14, pp. 89-108, 2008. [143] A. M. Davis, S. E. Lewis, Z. T. Bainbridge, L. Glendenning, R. D. R. Turner, and J. E. Brodie, "Dynamics of herbicide transport and partitioning under event flow conditions in the lower Burdekin region, Australia," Marine Pollution Bulletin, vol. 65, pp. 182-193, 2012. [144] G. Ham, "Water quality of the inflows/outflows of the Barratta Creek system," Proc. Aust. Soc. Sugar Cane Technology, vol. 29, 2009. [145] S. E. Lewis, J. E. Brodie, Z. T. Bainbridge, K. W. Rohde, A. M. Davis, B. L. Masters, et al., "Herbicides: A new threat to the Great Barrier Reef," Environmental Pollution, vol. 157, pp. 2470-2484, 2009. [146] K. McMahon, S. B. Nash, G. Eaglesham, J. F. Muller, N. C. Duke, and S. Winderlich, "Herbicide contamination and the potential impact to seagrass

233

References

meadows in Hervey Bay, Queensland, Australia," Marine Pollution Bulletin, vol. 51, pp. 325-334, 2005. [147] C. Mitchell, J. Brodie, and I. White, "Sediments, nutrients and pesticide residues in event flow conditions in streams of the Mackay Whitsunday Region, Australia," Marine Pollution Bulletin, vol. 51, pp. 23-36, 2005. [148] R. Packett, C. Dougall, K. Rohde, and R. Noble, "Agricultural lands are hot- spots for annual runoff polluting the southern Great Barrier Reef lagoon," Marine Pollution Bulletin, vol. 58, pp. 976-986, 2009. [149] R. Smith, R. Middlebrook, R. Turner, R. Huggins, S. Vardy, and M. Warne, "Large-scale pesticide monitoring across Great Barrier Reef catchments – paddock to reef integrated monitoring, modelling and reporting program," Marine Pollution Bulletin, vol. 65, pp. 117-127, 2012. [150] K. Kennedy, M. Devlin, C. Bentley, K. Lee-Chue, C. Paxman, S. Carter, et al., "The influence of a season of extreme wet weather events on exposure of the world heritage area Great Barrier Reef to pesticides," Marine Pollution Bulletin, vol. 64, pp. 1495-1507, 2012. [151] S. E. Lewis, B. Schaffelke, M. Shaw, Z. T. Bainbridge, K. W. Rohde, K. Kennedy, et al., "Assessing the additive risks of PSII herbicide exposure to the Great Barrier Reef," Marine Pollution Bulletin, vol. 65, pp. 280-291, 2012. [152] H. Horth and K. Blackmore, "Survey of glyphosate and AMPA in groundwaters and aurface waters in Europe," Monsanto, WRc Swindon, England UC8073.02, 2009. [153] K. Wei-Ping, P. G. Saffigna, A. W. Wood, and J. R. Freney, "Evaluation of methods for control of ammonia volatilization from surface-applied nitrogen fertilizers to sugarcane trash in North Queensland," Pedosphere, vol. 3, pp. 321-330, 1993. [154] B. A. Keating, J. Bauld, J. Hillier, R. Ellis, K. L. Weier, F. Sunners, et al., Downstream Effects of Land Use. Brisbane: Department of Natural Resources, 1996. [155] A. M. O. E. Wali, F. l. Grand, and G. J. Gascho, "Nitrogen leaching from soil and uptake by sugarcane from various urea based fertilizers," Soil Science Society of America Journal, vol. 44, pp. 119-122, 1980. [156] L. M. Southwick, G. H. Willis, D. C. Johnson, and H. M. Selim, "Leaching of nitrate, atrazine, and meribuzin from sugarcane in southern Louisiana," Journal of Environmental Quality, vol. 24, pp. 684-690, 1995. [157] G. Umrit and K. F. N. K. Kwong, "Herbicide dissipation and run-off from soils under sugarcane in Mauritius," presented at the South African Sugar Technologists' Association, 1999. [158] L. M. Southwick, G. H. Willis, and H. M. Selim, "Leaching of atrazine from sugarcane in southern Louisiana," Journal of Agricultural and Food Chemistry, vol. 40, pp. 1264–1268, 1992. [159] T. Searchinger, R. Heimlich, R. A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, et al., "Use of U.S. croplands for biofuels increases greenhouse gases through emissions from land-use change," Science, vol. 319, pp. 1238-1240, 2008. [160] P. Falloon, D. Powlson, and P. Smith, "Managing field margins for biodiversity and carbon sequestration: a Great Britain case study," Soil Use and Management vol. 20, pp. 240-247, 2004.

234

References

[161] C. Boutin and B. Jobin, "Intensity of agricultural practices and effects on adjacent habitats," Ecological Applications, vol. 8, 1998. [162] M. H. Schmidt, I. Roschewitz, C. Thies, and T. Tscharntike, "Differential effects of landscape and management on diversity and density of ground- dwelling farmland spiders," Journal of Applied Ecology, vol. 42, pp. 281-287, 2005. [163] A. McLaughlina and P. Mineau, "The impact of agricultural practices on biodiversity," Agriculture, Ecosystems & Environment, vol. 55, pp. 201-212, 1995. [164] J. R. Krebs, J. D. Wilson, R. B. Bradbury, and G. M. Siriwardena, "The second Silent Spring?," Nature, vol. 400, pp. 611-612, 1999. [165] T. G. Benton, D. M. Bryant, L. Cole, and H. Q. P. Crick, "Linking agricultural practice to insect and bird populations: a historical study over three decades," Journal of Applied Ecology, vol. 39, pp. 673-687, 2002. [166] C. James, "Global status of commercialized biotech/GM crops," ISAAA, Ithaca, NY. 2007. [167] GM canola 2010 - progress, New South Wales,Victoria and Western Australia. Available: www.afaa.com.au [168] Ripper roundup ready results: GM canola a 2008 success. Available: www.afaa.com.au [169] R. Norton, J. Kirkegaard, J. Angus, and T. Potter, "Canola in Australia - The first 30 years," Australian Oilseed Federation1999. [170] I. J. Senior and A. D. Bavage, "Comparison of genetically modified and conventionally derived herbicide tolerance in oilseed rape: a case study," Euphytica, vol. 132, pp. 217-226, 2003. [171] "Information paper on genetically modified canola," Ministerial GMO Industry Reference Group2009. [172] C. James, "Global status of commercialized biotech/GM crops:," The International Service for the Acquisition of Agri-biotech Applications (ISAAA), Ithaca, NY2005. [173] J. Buth, "Herbicide-resistant canola - the first ten years," Topics in Canadian Weed Science, vol. 4, 2007. [174] L. P. Gianessi, "Economic and herbicide use impacts of glyphosate-resistant crops," Pest Management Science, vol. 61, pp. 241-245, 2005. [175] T. P. Durrett, C. Benning, and J. Ohlrogge, "Plant triacylglycerols as feedstocks for the production of biofuels," The Plant Journal vol. 54, pp. 593- 607, 2008. [176] S. Solomon, D. Qin, Z. C. M. Manning, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, "Climate change 2007: The physical science basis. Contribution of working group I to the fourth assessment " The Intergovernmental Panel on Climate Change2007. [177] C. Bowyer, "Anticipated indirect land use change associated with expanded use of biofuels and bioliquids in the EU – An analysis of the national renewable energy action plans," Institute for European Environmental Policy2010. [178] R. Rathmann, A. Szklo, and R. Schaeffer, "Land use competition for production of food and liquid biofuels: An analysis of the arguments in the current debate," Renewable Energy, vol. 35, pp. 14-22, 2010.

235

References

[179] R. Doornbosch and R. Steenblik, "Biofuels: is the cure worse than the disease?," Organisation for Economic Co-operation and Development Paris2007. [180] M. Banse, P. Nowicki, and H. v. Meijl, "Why are current world food prices so high? ," Wageningen2008. [181] M. Banse, H. v. Meijl, A. Tabeau, and G. Woltjer, "Will EU biofuel policies affect global agricultural markets?," Agricultural Economics Research Institute2008. [182] H. J. Croezen, G. C. Bergsma, M. B. J. Otten, and M. P. J. v. Valkengoed, "Biofuels: indirect land use change and climate impact " Delft2010. [183] M. J. Haas, "Improving the economics of biodiesel production through the use of low value lipids as feedstocks: vegetable oil soapstock," Fuel Processing Technology, vol. 86, pp. 1087-1096, 2005. [184] A. E. Atabani, A. S. Silitonga, I. A. Badruddin, T. M. I. Mahlia, H. H. Masjukia, and S. Mekhilefd, "A comprehensive review on biodiesel as an alternative energy resource and its characteristics," Renewable and Sustainable Energy Reviews, vol. 16, pp. 2070-2093, 2012. [185] B. R. Moser, "Biodiesel production, properties, and feedstocks," In Vitro Cellular & Developmental Biology - Plant, vol. 45, pp. 229-266, 2009. [186] G. T. Champion, M. J. May, S. Bennett, D. R. Brooks, S. J. Clark, R. E. Daniels, et al., "Crop management and agronomic context of the farm scale evaluations of genetically modified herbicide-tolerant crops," Philosophical Transactions of the Royal Society B: Biological Sciences, vol. 358, pp. 1801- 1818, 2003. [187] K. R. Hayes, P. C. Gregg, V. V. S. R. Gupta, W. M. R. Jessop, B. S. Lonsdale, and C. K. Williams, "Identifying hazards in complex ecological systems. Part 3: Hierarchical Holographic Model for herbicide tolerant oilseed rape," Environmental Biosafety Research vol. 3, pp. 109-128, 2004. [188] I. J. Senior and P. J. Dale, "Herbicide-tolerant crops in agriculture: oilseed rape as a case study," Plant Breeding vol. 121, pp. 97-107, 2002. [189] J. v. Buskirk and Y. Willi, "Enhancement of farmland biodiversity within set- aside land," Conservation Biology, vol. 18, pp. 987-994, 2004. [190] N. Hesketh, P. C. Brookes, and T. M. Addiscott, "Effect of suspended soil material and pig slurry on the facilitated transport of pesticides, phosphate and bromide in sandy soil," European Journal of Soil Science, vol. 52, pp. 287- 296, 2001. [191] G. Brookes and P. Barfoot, "GM crops: the global socio-economic and environmental impact – the first nine years 1996-2004," PG Economics Ltd2005. [192] B. G. Young, "Changes in herbicide use patterns and production practices," Weed Technology, vol. 20, pp. 301-307, 2006. [193] "Climate Change," Intergovernmental Panel on Climate Change, IPCC2007. [194] "Development and Climate Change," World Bank2010. [195] T. O. West and W. M. Post, "Soil organic carbon sequestration rates by tillage and crop rotation: a global data analysis," Soil Science Society of America Journal, vol. 66, pp. 1930-1946, 2002. [196] R. Lal, "Soil carbon sequestration impacts on global climate change and food," Science, vol. 304, pp. 1623-1627, 2004.

236

References

[197] G. M. Clark, D. A. Goolsby, and W. A. Battaglin, "Seasonal and annual load of herbicides from the mississippi river basin to the gulf of Mexico," Environmental Science & Technology, vol. 33, pp. 981-986, 1999. [198] M. Graymore, F. Stagnitti, and G. Allinson, "Impacts of atrazine in aquatic ecosystems," Environment International, vol. 26, pp. 483-495, 2001. [199] M. J. Cerejeira, P. Viana, S. Batista, T. Pereira, E. Silva, M. J. Valério, et al., "Pesticides in Portuguese surface and ground waters," Water Research, vol. 37, pp. 1055-1063, 2003. [200] J. Brodie, E. Wolanski, S. Lewis, and Z. Bainbridge, "An assessment of residence times of land-sourced contaminants in the Great Barrier Reef lagoon and the implications for management and reef recovery," Marine Pollution Bulletin, vol. 65, pp. 267-279, 2012a. [201] J. Brodie and J. Waterhouse, "A critical review of environmental management of the ‘not so Great’ Barrier Reef," Estuarine, Coastal and Shelf Science, vol. 104-105, pp. 1-22, 2012. [202] G. De’ath, K. E. Fabricius, H. Sweatman, and M. Puotinen, "The 27–year decline of coral cover on the Great Barrier Reef and its causes," in The National Academy of Sciences of the United States of America, 2012. [203] R. J. Orth, T. J. B. Carruthers, W. C. Dennison, C. M. Durate, J. W. Fourqurean, K. H. JR., et al., "A Global Crisis for Seagrass Ecosystems," BioScience, vol. 56, pp. 987-996, 2006. [204] J. E. Brodie, F. J. Kroon, B. Schaffelke, E. C. Wolanski, S. E. Lewis, M. J. Devlin, et al., "Terrestrial pollutant runoff to the Great Barrier Reef: An update of issues, priorities and management responses," Marine Pollution Bulletin, vol. 65, pp. 81-100, 2012b. [205] F. Veiga, J. M. Zapata, M. L. F. Marcos, and E. Alvarez, "Dynamics of glyphosate and aminomethylphosphonic acid in a forest soil in Galicia, north- west Spain," Science of The Total Environment, vol. 271, pp. 135-144, 2001. [206] D. Landry, S. Dousset, J.-C. Fournier, and F. Andreux, "Leaching of glyphosate and AMPA under two soil management practices in Burgundy vineyards," Environmental Pollution, vol. 138, 2005. [207] J. Kjaer, P. Olsen, M. Ullum, and R. Grant, "Leaching of glyphosate and amino-methylphosphonic acid from danish agricultural field sites," Journal of Environmental Quality, vol. 34, pp. 608-620, 2005. [208] R. A. Relyea, "The lethal impact of Roundup on aquatic and terrestrial amphibians," Ecological Applications, vol. 15, pp. 1118-1124, 2005. [209] R. Strange-Hansen, P. E. Holm, O. S. Jacobsen, and C. S. Jacobsen, "Sorption, mineralization and mobility of N-(phosphonomethyl)glycine (glyphosate) in five different types of gravel," Pest Management Science, vol. 60, pp. 570-578, 2004. [210] L. Torstensson, E. Börjesson, and J. Stenström, "Efficacy and fate of glyphosate on Swedish railway embankments," Pest Management Science, vol. 61, pp. 881-886, 2005. [211] K. Siimes, S. Rämö, L. Welling, U. Nikunen, and P. Laitinen, "Comparison of the behaviour of three herbicides in a field experiment under bare soil conditions," Agricultural Water Management, vol. 84, pp. 53-64, 2006. [212] A. D. Baylis, "Why glyphosate is a global herbicide: strengths, weaknesses and prospects," Pest Management Science, vol. 56, pp. 299-308, 2000.

237

References

[213] B. Zhao, J. Zhang, J. Gong, H. Zhang, and C. Zhang, "Glyphosate mobility in soils by phosphate application: Laboratory column experiments," Geoderma, vol. 149, pp. 290-297, 2009. [214] B. Hartzler, C. Boerboom, G. Nice, and P. Sikkema, "Understanding glyphosate to increase performance," U.S. Department of Agriculture2006. [215] The Pesticide Manual: A World Compendium, 14th ed. Hampshire, UK: British Crop Protection Council, 2006. [216] W. K. Vencill, "Herbicide Handbook," 8th ed: Weed Science Society of America, 2002. [217] SRC PhysProp database: glyphosate; gyracuse research corporation. Available: www.syrres.com [218] T. R. Roberts, Metabolic pathways of agrochemicals: Part 1: Herbicides and plant growth regulators. Cambridge, UK: Royal Society of Chemistry, 1998. [219] Information on Genetically Modified Crops. Available: https://www.agric.wa.gov.au [220] C. Britt, A. Mole, F. Kirkham, and A. Terry, The herbicide handbook: guidance on the use of herbicides on nature conservation sites. West Yorkshire, UK: English Nature 2003 in Association with FACT, 2003. [221] Roundup biactive: Quick guide to weed control. Monsanto UK Ltd. 2011. [222] J. E. Franz, M. K. Mao, and J. A. Sikorski, Glyphosate: a unique global herbicide: An American Chemical Society Publication, 1997. [223] "RED Facts: Glyphosate; EPA-738-F-93-011," U.S. Environmental Protection Agency, Washington DC.1993. [224] R. Nicholls, "Confined rotational crop study of glyphosate Part I: In-field portion," U.S. Environmental Protection Agency, Washington DC.1993. [225] P. McMullan, J. Honeggar, and E. Logusch, "Confined rotational crop study of glyphosate Part II. Quantitation, characterization and identification of glyphosate and its metabolites in rotational crops," U.S. Environmental Protection Agency, Washington DC.1993. [226] "Environmental health criteria 159, toxicological evaluations - glyphosate," International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland1988. [227] Pesticide background statements vol. 1, Herbicides: USDA Forest Service, 1984. [228] P. Chakravarty and S. S. Sidhu, "Effect of glyphosate, hexazinone and triclopyr on in vitro growth of five species of ectomycorrhizal fungi," European Journal of Forest Pathology, vol. 17, pp. 204-210, 1987. [229] P. Sprankle, W. F. Meggitt, and D. Penner, "Adsorption, mobility, and microbial degradation of glyphosate in the soil," Weed Science, vol. 23, pp. 229-234, 1975. [230] R. L. Glass, "Adsorption of glyphosate by soils and clay minerals," Journal of Agricultural and Food Chemistry, vol. 35, pp. 497-500 1987. [231] A. G. Hornsby, R. D. Wauchope, and A. E. Herner, Pesticide properties in the environment. New York: Springer, 1996. [232] "Canadian water quality guidelines," Canadian Council of Ministers of the Environment (CCME), Ottawa, Ontario, Canada1989.

238

References

[233] J. Schuette, "Environmental fate of glyphosate," Environmental Monitoring & Pest Management, Department of Pesticide Regulation, Sacramento, California1998 [234] A. Piccolo, G. Celano, M. Arienzo, and A. Mirabella, "Adsorption and desorption of glyphosate in some European soils," Journal of Environmental Science and Health Part B, vol. B29, pp. 1105-1115, 1994. [235] H. DeJonge and L. W. DeJonge, "Influence of pH and solution Composition on the Sorption of Glyphosate and Prochloraz to a Sandy Loam soil.," Cheraosphere, vol. 39, pp. 753-763, 1999. [236] A. L. Gimsing and O. K. Borggaard, "Effect of KCl and CaCl2 as background electrolytes on the competitive adsorption of glyphosate and phosphate on geothite," Clays and Clay Minerals, vol. 49, pp. 270–275, 2001. [237] S. Autio, K. Siimes, P. Laitinen, S. Ramo, S. Oinonen, and L. Eronen, "Adsorption of sugar beet herbicides to Finnish soils," Chemosphere vol. 55, pp. 215–226, 2004. [238] A.L.Gimsing, O.K.Borggaard, and M. Bang, "Influence of Soil Composition on Adsorption of Glyphosate and Phosphate by Contrasting Danish Surface Soils," European Journal of Soil Science, vol. 55, pp. 183–191, 2004. [239] A. L. Gimising and O.K.Borggaard, "Competitive Adsorption and Desorption of Glyphosate and Phosphate on Clay Silicates and Oxides," Clay Minerals vol. 37, pp. 509–515, 2002. [240] A.L. Gimsing, C. Szilas, and O. K. Borggaard, "Sorption of glyphosate and phosphate by variable-charge tropical soils from Tanzania," Geoderma, vol. 138 pp. 127–132, 2007. [241] M. A. Attaallah, "The Kinetic Study of Glyphosate Leachate in Palestinian Soil at Different Concentrations," Master of Science in Chemistry, An-Najah National University Faculty of Graduate Studies, 2011. [242] R. G. Gerritse, J. Beltran, and F. Hernandeg, "Adsorption of atrazine, simazine, and glyphosate in soils of the Gnangara Mound, Western Australia," Australian Journal of Soil Research, vol. 34, pp. 599-607, 1996. [243] W. Yu-jun，, Z. Dong-mei，, and S. Rui-juan, "Effects of phosphate on the adsorption of glyphosate on three different types of Chinese soils," Journal of Environmental Sciences vol. 17, pp. 711-715, 2005. [244] C. J. Miles and H. A. Moye, "Extraction of glyphosate herbicide from soil and clay minerals and determination of residues in soils," Journal of Agricultural and Food Chemistry, vol. 36, pp. 488-491 1988. [245] E. Morillo, T. Undabeytia, C. Maqueda, and A. Ramos, "Glyphosate adsorption on soils of di€ erent characteristics. Infuence of copper addition," Chemosphere vol. 40 pp. 103-107, 2000. [246] J. Sheals, S. Sjoberg, and P. Persson, "Adsorption of glyphosate on goethite: molecular characterization of surface complexes," Environmental Science & Technology, vol. 36, pp. 3090-3095, 2002. [247] B. C. Barja and M. D. S. Afonso, "Aminomethylphosphonic acid and glyphosate adsorption onto goethite: a comparative study," Environmental Science & Technology, vol. 39, pp. 585-592, 2005. [248] H. Vereecken, "Mobility and leaching of glyphosate: a review," Pest Management Science, vol. 61, pp. 1139-1151, 2005.

239

References

[249] A. L. Gimsing, O. K. Borggaard, and M. Bang, "Influence of soil composition on adsorption of glyphosate and phosphate by contrasting Danish surface soils," European Journal of Soil Science, vol. 55, pp. 183–191, 2004. [250] W. H. Wischmeier and D. D. Smith, Predicting rainfall-erosion losses from cropland east of the rocky mountains. Washington, D.C., 1978. [251] A. L. Gimsing, C. Szilas, and O. K. Borggaard, "Sorption of glyphosate and phosphate by variable-charge tropical soils from Tanzania," Geoderma, vol. 138 pp. 127–132, 2007. [252] A. L. Gimsing and O. K. Borggaard, "Competitive adsorption and desorption of glyphosate and phosphate on clay silicates and oxides," Clay Minerals vol. 37, pp. 509–515, 2002. [253] K. Dideriksen and S. L. S. Stipp, "The adsorption of glyphosate and phosphate to goethite: a molecular-scale atomic force microscopy study," Geochimica et Cosmochimica Acta, vol. 67, pp. 3313–3327, 2003. [254] S. V. V, "Nitrogen gains and losses in sugarcane (Saccharum sp.) agroecosystems on the coast of Peru," Plant and Soil, vol. 67, pp. 147-156, 1982. [255] H. V. Olphen and J. J. Fripiat, Data handbook for clay materials and other non-metallic minerals. Oxford, UK: Pergamon Press, 1979. [256] J. S. McConnell and L. R. Hossner, "pH-Dependent adsorption isotherms of glyphosate," Journal of Agricultural and Food Chemistry, vol. 33, pp. 1075- 1078, 1985 [257] E. Shannon, A. McDougall, K. Kelsey, and B. Hussey, "Watercheck - a co- ordinated extension program for improving irrigation efficiency on Australian cane farms," presented at the 18th Australian Society of Sugar Cane Technologists, 1996. [258] H. d. Jonge, L. W. d. Jonge, O. H. Jacobsen, T. Yamaguchi, and P. Moldrup, "Glyphosate sorption in soils of different pH and phosphorous content," Soil Science, vol. 166, pp. 230-238, 2001. [259] H. M. Dion, J. B. Harsh, and H. H. Hill, "Competitive sorption between glyphosate and inorganic phosphate on clay minerals and low organic matter soils," Journal of Radioanalytical and Nuclear Chemistry, vol. 249, pp. 385- 390, 2001. [260] R. J. Hance, "Adsorption of glyphosate by soils," Pesticide Science, vol. 7, pp. 363-366, 1976. [261] H. d. Jonge and L. W. d. Jonge, "Influence of pH and solution composition on the sorption of glyphosate and prochloraz to a sandy loam soil," Chemosphere, vol. 39, pp. 753-763, 1999. [262] E. J. Schmidt, "The role of irrigation in the South African sugar industry," in South African Sugar Technologists’ Association, 1989, pp. 108-113. [263] E. J. Schmidt, "Impacts of sugarcane production on water resources," in South African Sugar Technologists’ Association, 1997, pp. 73-75. [264] M. J. Robertson, N. G. Inman-Bamber, and R. C. Muchow, "Opportunities for improving the use of limited water by the sugarcane crop," in Intensive Sugarcane Production: Meeting the Challenges beyond 2000, B. A. Keating and J. R. Wilson, Eds., ed Wallingford, UK: CAB International, 1997, pp. 287– 304.

240

References

[265] O. K. Borggaard and A. L. Gimsing, "Fate of glyphosate in soil and the possibility of leaching to ground and surface waters: a review," Pest Management Science, vol. 64, pp. 441-456, 2008. [266] O. K. Borggaard and B. Elberling, Pedological biogeochemistry. Denmark: Københavns Universitet, 2004. [267] N. C. Brady and R. R. Weil, The nature and properties of soils, 12th ed.: Prentice-Hall, 1999. [268] A. L. Gimsing, O. K. Borggaard, O. S. Jacobsen, J. Aamand, and J. Sørensen, "Chemical and microbiological soil characteristics controlling glyphosate mineralisation in Danish surface soils," Applied Soil Ecology, vol. 27, pp. 233– 242, 2004. [269] H. d. Jonge, L. W. d. Jonge, and O. H. Jacobsen, "[14C]Glyphosate transport in undisturbed topsoil columns," Pest Management Science, vol. 56, pp. 909- 915, 2000. [270] I. S. Fomsgaard, N. H. Spliid, and G. Felding, "Leaching of pesticides through normal-tillage and low-tillage soil-a lysimeter study. II.," Journal of Environmental Science and Health, vol. B38, pp. 19-35, 2003. [271] P. Laitinen, K. Siimes, L. Eronen, S. Rämö, L. Welling, S. Oinonen, et al., "Fate of the herbicides glyphosate, glufosinate-ammonium, phenmedipham, ethofumesate and metamitron in two Finnish arable soils," Pest Management Science, vol. 62, pp. 473-491, 2006. [272] M. Flury, "Experimental evidence of transport of pesticides through field soils - a review," Journal of Environmental Quality, vol. 25, pp. 25-45, 1996. [273] V. Contardo-Jara, E. Klingelmann, and C. Wiegand, "Bioaccumulation of glyphosate and its formulation Roundup Ultra in Lumbriculus variegatus and its effects on biotransformation and antioxidant enzymes," Environmental Pollution, vol. 157, pp. 57-63, 2009. [274] A. Paganelli, V. Gnazzo, H. Acosta, S. L. López, and A. E. Carrasco, "Glyphosate-based herbicides produce teratogenic effects on vertebrates by Impairing retinoic acid signaling," Chemical Research in Toxicology, vol. 23, pp. 1586-1595, 2010. [275] P. Duchnowicz, M. Koter, and W. Duda, "Damage of erythrocyte by phenoxyacetic herbicides and their metabolites," Pesticide Biochemistry and Physiology, vol. 74, pp. 1-7, 2002. [276] B. Szatkowska, M. Kwiatkowska, J. Michałowicz, P. Sicińska, B. Huras, and B. Bukowska, "Impact of chlorfenvinphos, an organophosphate insecticide on human blood mononuclear cells (in vitro)," Pesticide Biochemistry and Physiology, vol. 102, pp. 175-181, 2012. [277] B. Bukowska, K. Goszczyńska, and W. Duda, "Effect of 4-chloro-2- methylphenoxyacetic acid and 2,4-dimethylphenol on human erythrocytes," Pesticide Biochemistry and Physiology, vol. 77, 2003. [278] B. Bukowska, J. Michałowicz, A. Krokosz, and P. Sicińska, "Comparison of the effect of phenol and its derivatives on protein and free radical formation in human erythrocytes (in vitro)," Blood Cells, Molecules, and Diseases, vol. 39, pp. 238-244, 2007. [279] C. N. Martini, M. Gabrielli, and M. d. C. Vila, "A commercial formulation of glyphosate inhibits proliferation and differentiation to adipocytes and induces

241

References

apoptosis in 3T3-L1 fibroblasts," Toxicology in Vitro, vol. 26, pp. 1007-1013, 2012. [280] "Data sheets on pesticides: glyphosate; international programme on chemical safety," World Health Organization, Food and Agriculture Organization, Geneva, Switzerland 1996. [281] G. M. Williams, R. Kroes, and I. C. Munro, "Safety evaluation and risk assessment of the herbicide Roundup and its active ingredient, glyphosate, for humans," Regulatory Toxicology and Pharmacology, vol. 31, pp. 117-165, 200. [282] D. W. Brewster, J. Warren, and W. E. Hopkins, "Metabolism of glyphosate in Sprague-Dawley rats: Tissue distribution, identification, and quantification of glyphosate-derived materials following a single oral dose," Fundamental and Applied Toxicology, vol. 17, pp. 43-51, 1991. [283] M. Birch, "Toxicological investigation of CP 67573-3," U. S. Environmental Protection Agency, Washington DC.1993. [284] N. J. Rattray, "Glyphosate acid: 4-hour acute inhalation toxicity study in rats," Zeneca Agrochemicals, Central Toxicology Laboratory, Cheshire, England1996. [285] É. Claira, R. Mesnage, C. Travert, and G.-É. Séralini, "A glyphosate-based herbicide induces necrosis and apoptosis in mature rat testicular cells in vitro, and testosterone decrease at lower levels," Toxicology in Vitro, vol. 26, pp. 269-279, 2012. [286] D. Blaszcak, "Primary dermal irritation study in rabbits for glyphosate technical (wetcake)," U.S. Environmental Protection Agency, Washington DC.1993. [287] D. Blaszcak, "Eye irritation study in rabbits for glyphosate technical (wetcake)," U.S. Environmental Protection Agency, Washington DC.1993. [288] "The metabolism of glyphosate in sprague dawley rats- Part I. Excretion and tissue distribution of glyphosate and its metabolites following intravenous and oral administration," Monsanto Corporation, Report no. MSL-7215, 1988. [289] R. M. Bodden, "Metabolism study of synthetic 13C/14C-labeled glyphosate and aminomethylphoshonic acid in lactating goats," 1988. [290] R. M. Bodden, "Metabolism study of synthetic 13C/14C-labeled glyphosate and aminomethylphosphonic acid in laying hens," 1988. [291] R. Fink and J. Beavers, "One-generation reproduction study in mallard ducks: glyphosate technical," U.S. Environmental Protection Agency, Washington DC.1993. [292] "Reregistration Eligibility Decision (RED): Glyphosate; EPA-738-R-93-014," U.S Environmental Protection Agency, Office of Prevention, Pesticides, and Toxic Substances, Office of Pesticide Programs, U.S. Government Printing Office, Washington DC.1993. [293] S. Castle, L. Ruzo, and S. Katheryn, "Degradation study: photodegradation of carbon 14 glyphosate in a buffered aqueous solution at pH 5, 7, and 9 by natural sunlight," Pharmacology and Toxicology Research Laboratory, Inc., U.S. Environmental Protections Agency, Washington DC.1993. [294] B. Brightwell and J. Malik, "Solubility, volatility, adsorption and partition coefficients, leaching, and aquatic metabolism of MON 0573 and MON 010, Report No: MSL-0207 (00108192)," 1978.

242

References

[295] W. Ridley and K. Mirly, "The metabolism of glyphosate in Sprague-Dawley rats. Part I. Excretion and tissue distribution of glyphosate and its metabolites following intravenous and oral administration," U.S. Environmental Protection Agency, Washington DC.1993. [296] L. C. Folmar, H. O. Sanders, and A. M. Julin, "Toxicity of the herbicide glyphosate and several of its formulations to fish and aquatic invertebrates.," Arch. Environm. Contain. Toxicol., vol. 8, pp. 269-278, 1979. [297] C. M. Howe, M. Berrill, B. D. Pauli, C. C. Helbing, K. Werry, and N. Veldhoen, "Toxicity of glyphosate-based pesticides to four North American frog species," Environmental Toxicology and Chemistry, vol. 23, pp. 1928- 1938, 2004. [298] W. McAllister, M. McKee, and M. Schofield, "Chronic toxicity of glyphosate (AB-82-036) to Daphnia magna under flow-through test conditions. Chronic toxicity final report ABC 28742," U.S. Environmental Protection Agency, Washington DC.1993. [299] R. C. Wester, J. Melendres, R. Sarason, J. McMaster, and H. I. Maibach, "Glyphosate skin binding, absorption, residual tissue distribution, and skin decontamination," Fundam Appl Toxicol, vol. 16, pp. 725-732, 1991. [300] H. I. Maibach, "Irritation, sensitization, photoirritation and photosensitization assays with a glyphosate herbicide," Contact Dermatitis, vol. 15, pp. 152-156, 1986. [301] Y. Hori, M. Fujisawa, K. Shimada, and Y. Hirose, "Determination of the herbicide glyphosate and its metabolite in biological specimens by gas chromatography-mass spectrometry. A case of poisoning by roundup herbicide," Journal of Anlytical Toxicology, vol. 27, pp. 162-166, 2003. [302] R. Howe, R. Chott, and R. McClanahan, "Metabolism of glyphosate in Sprague-Dawley rats. Part II: Identification, characterization, and quantitation of glyphosate and its metabolites after intravenous and oral administration," U.S. Environmental Protection Agency, Washington DC.1993. [303] A. Talbot, T. S. Ku, C. L. Chen, G. C. Li, and H. P. Li, "Glyphosate levels in acute Roundup herbicide poisoning," presented at the 1994 Toxicology World Congress Abstracts, 1995. [304] D. Pieniążek, B. Bukowska, and W. Duda, "Comparison of the effect of Roundup Ultra 360 SL pesticide and its active compound glyphosate on human erythrocytes," Pesticide Biochemistry and Physiology, vol. 79, pp. 58-63, 2004. [305] B. Bukowska, D. Pieniążek, and W. Duda, "Hemolysis and lipid peroxidation in human erythrocytes incubated with Roundup," Current Topics in Biophysics vol. 26, pp. 245-49, 2002. [306] C. Gasnier, C. Dumont, N. Benachour, E. Clair, M.-C. Chagnon, and G.-E. Séralini, "Glyphosate-based herbicides are toxic and endocrine disruptors in human cell lines," Toxicology vol. 262, pp. 184-191, 2009. [307] N. Benachour and G.-E. Séralini, "Glyphosate formulations induce apoptosis and necrosis in human umbilical, embryonic, and placental cells," Chemical Research in Toxicology, vol. 22, pp. 97-105, 2008. [308] A. J. D. Roos, A. Blair, J. A. Rusiecki, J. A. Hoppin, M. Svec, M. Dosemeci, et al., "Cancer incidence among glyphosate-exposed pesticide applicators in

243

References

the agricultural health study," Environ Health Perspect., vol. 113, pp. 49-54, 2005. [309] R. Jasper, G. O. Locatelli, C. Pilati, and C. Locatelli, "Evaluation of biochemical, hematological and oxidative parameters in mice exposed to the herbicide glyphosate-Roundup®," Interdisciplinary Toxicology, vol. 5, pp. 133-140, 2013. [310] Y. x. Gui, X. n. Fan, H. m. Wang, G. Wang, and S. d. Chen, "Glyphosate induced cell death through apoptotic and autophagic mechanisms," Neurotoxicology and Teratology, vol. 34, pp. 344-349, 2012. [311] W. G. Siems, O. Sommerburg, and T. Grune, "Erythrocyte free radical and energy metabolism," Clinical Nephrology, vol. 53, pp. S 9–17, 2000. [312] M. Sihtmäe, I. Blinova, K. Künnis-Beres, L. Kanarbik, M. Heinlaan, and A. Kahru, "Ecotoxicological effects of different glyphosate formulations," Applied Soil Ecology, vol. 72, pp. 215-224, 2013. [313] S. Guilherme, M. A. S. Santos, C. Barroso, I. Gaivão, and M. Pacheco, "Differential genotoxicity of roundup formulation and its constituents in blood cells of fish (Anguilla anguilla): considerations on chemical interactions and DNA damaging mechanisms," Ecotoxicology, vol. 21, pp. 1381-1390, 2012. [314] V. J. Koller, M. Fürhacker, A. Nersesyan, M. Mišík, M. Eisenbauer, and S. Knasmueller, "Cytotoxic and DNA-damaging properties of glyphosate and Roundup in human-derived buccal epithelial cells," Archives of Toxicology, vol. 86, pp. 805-813, 2012. [315] L. J. Moore, L. Fuentes, J. H. R. Jr., W. W. Bowerman, G. K. Yarrow, W. Y. Chao, et al., "Relative toxicity of the components of the original formulation of roundup® to five north american anurans," Ecotoxicology and Environmental Safety, vol. 78, pp. 128-133, 2012. [316] W. A. Battaglin, K. C. Rice, M. J. Focazio, S. Salmons, and R. X. Barry, "The occurrence of glyphosate, atrazine, and other pesticides in vernal pools and adjacent streams in Washington, DC, Maryland, Iowa, and Wyoming, 2005– 2006," Environmental Monitoring and Assessment, vol. 155, pp. 281-307, 2009. [317] EC, "Directive 2000/60/EC of the European Parliament and of the council of 23 October 2000 establishing a framework for community action in the field of water policy," Official Journal of the European Union L 327, 2000. [318] V. K. Nandula, Glyphosate resistance in crops and weeds: history, development, and management. Hoboken, New Jersey: John Wiley & Sons, Inc., Publication, 2010. [319] C. M. Benbrook, "Glyphosate tolerant crops in the EU. Prepared for Greenpeace International.," 2012. [320] F. Sanchez-Bayo, R. V. Hyne, and K. L. Desseille, "An amperometric method for the detection of amitrole, glyphosate and its aminomethyl-phosphonic acid metabolite in environmental waters using passive samplers," Analytica Chimica Acta, vol. 675, pp. 125–131, 2010. [321] B. Sahasrabuddhey, A. Jain, and K. K. Verma, "Determination of ammonia and aliphatic amines in environmental aqueous samples utilizing pre-column derivatization to their phenylthioureas and high performance liquid chromatography," Analyst, vol. 124, pp. 1017-1021, 1999.

244

References

[322] P. L. Alferness and Y. Iwata, "Determination of glyphosate and (aminomethyl) phosphonic acid in soil, plant and animal matrices, and water by capillary gas chromatography with mass-selective detection," Journal of Agricultural and Food Chemistry, vol. 42, pp. 2751-2759, 1994. [323] M. Piriyapittaya, S. Jayanta, S. Mitra, and N. Leepipatpiboon, "Micro-scale membrane extraction of glyphosate and aminomethylphosphonic acid in water followed by high-performance liquid chromatography and post-column derivatization with fluorescence detector," Journal of Chromatography A, vol. 1189, pp. 483-492, 2008. [324] C. C. Hsu and C. W. Whang, "Microscale solid phase extraction of glyphosate and aminomethylphosphonic acid in water and guava fruit extract using alumina-coated iron oxide nanoparticles followed by capillary electrophoresis and electrochemiluminescence detection," Journal of Chromatography A, vol. 1216, pp. 8575-8580, 2009. [325] L. Goodwin, J. R. Startin, B. J. Keely, and D. M. Goodall, "Analysis of glyphosate and glufosinate by capillary electrophoresis–mass spectrometry utilising a sheathless microelectrospray interface," Journal of Chromatography A, vol. 1004, pp. 107-119, 2003. [326] Y. Zhu, F. Zhang, C. Tong, and W. Liu, "Determination of glyphosate by ion chromatography," Journal of Chromatography A, vol. 850, pp. 297-301, 1999. [327] Z. X. Guo, Q. Cai, and Z. Yang, "Determination of glyphosate and phosphate in water by ion chromatography - inductively coupled plasma mass spectrometry detection," Journal of Chromatography A, vol. 1100, pp. 160- 167, 2005. [328] K. Sato, J. Y. Jin, T. Takeuchi, T. Miwa, K. Suenami, Y. Takekoshi, et al., "Integrated pulsed amperometric detection of glufosinate, bialaphos and glyphosate at gold electrodes in anion - exchange chromatography," Journal of Chromatography A, vol. 919, pp. 313-320, 2001. [329] M. A. González-Martínez, E. M. Brun, R. Puchades, A. Maquieira, K. Ramsey, and F. Rubio, "Glyphosate immunosensor. Application for water and soil analysis," Analytical Chemistry, vol. 77, pp. 4219-4227, 2005. [330] R. L. Glass, "Colorimetric determination of glyphosate in water after oxidation to orthophosphate," Analytical Chemistry, vol. 53, pp. 921-923, 1981. [331] M. R. Jan, J. Shah, M. Muhammad, and B. Ara, "Glyphosate herbicide residue determination in samples of environmental importance using spectrophotometric method," Journal of Hazardous Materials, vol. 169, pp. 742-745, 2009. [332] F. Rubio, L. J. Veldhuis, B. S. Clegg, J. R. Fleeker, and J. C. Hall, "Comparison of a direct ELISA and an HPLC method for glyphosate determinations in water," Journal of Agricultural and Food Chemistry, vol. 51, pp. 691-696, 2003. [333] S. Watanabe, "Simultaneous analysis of glyphosate and glufosinate in vegetables and fruits by GC-FPD," Journal of the Food Hygienic Society of Japan, vol. 45, pp. 38-43, 2004. [334] S. H. Tseng, Y. W. Lo, P. C. Chang, S. S. Chou, and H. M. Chang, "Simultaneous quantification of glyphosate, glufosinate, and their major metabolites in rice and soybean sprouts by gas chromatography with pulsed

245

References

flame photometric detector," Journal of Agricultural & Food Chemistry vol. 52, pp. 4057–4063, 2004. [335] J. L. Tadeo, C. Sánchez-Brunete, R. A. Pérez, and M. D. Fernández, "Analysis of herbicide residues in cereals, fruits and vegetables," Journal of Chromatography A, vol. 882, pp. 175–191, 2000. [336] J. V. Sancho, F. Hernández, F. J. López, E. A. Hogendoorn, and E. Dijkman, "Rapid determination of glufosinate, glyphosate and aminomethylphosphonic acid in environmental water samples using precolumn fluorogenic labeling and coupled-column liquid chromatography," Journal of Chromatography A, vol. 737, pp. 75-83, 1996. [337] E. A. Hogendoorn, F. M. Ossendrijver, E. Dijkman, and R. A. Baumann, "Rapid determination of glyphosate in cereal samples by means of pre-column derivatisation with 9-fluorenylmethyl chloroformate and coupled-column liquid chromatography with fluorescence detection," Journal of Chromatography A, vol. 833, pp. 67-73, 1999. [338] L. Zhou, Z. Luo, S. Wang, Y. Hui, Z. Hu, and X. Chen, "In-capillary derivatization and laser-induced fluorescence detection for the analysis of organophosphorus pesticides by micellar electrokinetic chromatography," Journal of Chromatography A, vol. 1149, pp. 377-384, 2007. [339] E. Alexa, M. Hafner, M. Negrea, and A. Lazureanu, "HPLC and GC determination of glyphosate and aminomethylphosphonic acid (AMPA) in water samples " presented at the 43rd Croatian and 3rd International Symposium on Agriculture, 2008. [340] R. J. Vreeken, P. Speksnijder, I. Bobeldijk-Pastorova, and T. H. M. Noij, "Selective analysis of the herbicides glyphosate and aminomethylphosphonic acid in water by on-line solid-phase extraction–high-performan liquid chromatography–electrospray ionization mass spectrometry," Journal of Chromatography A, vol. 794, pp. 187–199, 1998. [341] S. Y. Chang and C.-H. Liao, "Analysis of glyphosate, glufosinate and aminomethylphosphonic acid by capillary electrophoresis with indirect fluorescence detection," Journal of Chromatography A, vol. 959, pp. 309-315, 2002. [342] M. G. Cikalo, D. M. Goodall, and W. Matthews, "Analysis of glyphosate using capillary electrophoresis with indirect detection," Journal of Chromatography A vol. 745, pp. 189-200, 1996. [343] H. Miyano, T. Toyo'oka, K. Imai, and T. Nakajima, "Influences of metal ions on the reaction of amino and imino acids with fluorogenic reagents," Analytical Biochemistry, vol. 150, pp. 125-130, 1985. [344] R. L. Glass, "Liquid chromatographic determination of glyphosate in fortified soil and water samples," Journal of Agricultural and Food Chemistry, vol. 31, pp. 280-282, 1983. [345] C. Druart, O. Delhomme, A. d. Vaufleury, and E. N. M. Millet, "Optimization of extraction procedure and chromatographic separation of glyphosate, glufosinate and aminomethylphosphonic acid in soil," Anal Bioanal Chem, vol. 399, pp. 1725–1732, 2011. [346] G. A. Shabir, "Validation of high-performance liquid chromatography methods for pharmaceutical analysis: Understanding the differences and similarities between validation requirements of the US Food and Drug Administration, the

246

References

US Pharmacopeia and the International Conference on Harmonization.," Journal of Chromatography A, vol. 987, 2003. [347] I. S. Krull and M. Swartz, Pharmaceutical Regulatory Guidance Book, 2006. [348] "Validation of analytical methods: Definitions and terminology," presented at the International Conference on Hermonization (ICH), 1995. [349] "Validation of analytical methods: Methodology, Q2B," presented at the International Conference on Harmonization, 1995. [350] "The fitness for purpose of analytical methods: A laboratory guide to method validation and related topics " Eurachem1998. [351] "Quality control procedures for pesticide residues analysis," European Commission, Health and Consumer Protection Directorate-General, Brussels2006. [352] "Reviewer Guidance: Validation of chromatographic methods," United States Food and Drug Administration (US-FDA)1994. [353] "Guidance for Industry: Analytical procedures and methods validation," United States Food and Drug Administration (US-FDA)2000. [354] "Methods of analysis and sampling," Codex Alimentarius Commission, Budapest1995. [355] M. Thompson, S. L. R. Ellison, and R.Wood, "Harmonized guidelines for single- Laboratory validation of methods of analysis," Pure and Applied Chemistry, vol. 74, pp. 835-855, 2002. [356] "General requirements for the competence of testing and calibration laboratories," International Organization for Standardization,, Geneva1999. [357] M. Ribani, C. H. Collins, and C. B. G. Bottoli, "Validation of chromatographic methods: Evaluation of detection and quantification limits in the determination of impurities in omeprazole," Journal of Chromatography A, vol. 1156, pp. 201-205, 2006. [358] F. M. Rubio, J. A. Itak, A. M. Scutellaro, M. Y. Selisker, and D. P. Herzog, "Performance characteristics of a novel magnetic particlebased enzyme‐linked immunosorbent assay for the quantitative analysis of atrazine and related triazines in water samples," Food and Agricultural Immunology, vol. 3, 1991. [359] J. C. Hall, T. D. V. Deynze, J. Struger, and C. H. Chan, "Enzyme immunoassay based survey of precipitation and surface water for the presence of atrazine, metolachlor and 2,4-D," Journal of Environmental Science and Health Part B, vol. 28, pp. 577-598, 1993. [360] N. Graziano, M. J. McGuire, A. Roberson, C. Adams, H. Jiang, and N. Blute, "National atrazine occurrence monitoring program using the abraxis ELISA method.," Environmental Science and Technology, vol. 40, 2004. [361] T. L. Fare, J. A. Itak, T. S. Lawruk, F. M. Rubio, and D. P. Herzog, "Cross- reactivity analysis using a four-parameter model applied to environmental immunoassays.," Bulletin of Environmental Contamination and Toxicology, vol. 57, pp. 367-74, 1996. [362] B. D. Johnson and J. C. Hall, "Fluroxypyr- and triclopyr-specific enzyme- linked immunosorbent assays: Development and quantitation in soil and Water," Journal of Agricultural & Food Chemistry, vol. 44, pp. 488-496, 1996. [363] J. S. Parnell and J. C. Hall, "Development of an enzyme-linked immunosorbent assay for the detection of metosulam," Journal of Agricultural & Food Chemistry, vol. 46, pp. 152-156, 1998.

247

References

[364] "Professional scientist remuneration survey summary report," Professional Australia ISSN 1444-0253, 2014. [365] "Labour costs, Australian Bureau of Statistics, Australia," 63480DO001_201011, 2012. [366] "Costs of setting up an office in sydney," New Zealand Trade and Enterprise 2005. [367] F. Graef, U. Stachow, A. Werner, and G. Schütte, "Agricultural practice changes with cultivating genetically modified herbicide-tolerant oilseed rape," Agricultural Systems, vol. 94, pp. 111-118, 2007. [368] G. R. Stringam, V. L. Ripley, H. K. Love, and A. Mitchell, "Transgenic herbicide tolerant canola—The Canadian experience," Crop Science, vol. 43, pp. 1590-1593, 2003. [369] S. Gilroy, "Canola in Canada: A grower’s perspective.," Australian Oilseeds Federation Forum, Sydney2007. [370] E. Mallat and D. Barceló, "Analysis and degradation study of glyphosate and of aminomethylphosphonic acid in natural waters by means of polymeric and ion-exchange solid-phase extraction columns followed by ion chromatography–post-column derivatization with ﬂuorescence detection," Journal of Chromatography A, vol. 823, pp. 129-136, 1998. [371] D. Humphries, G. Byrtus, and A.-M. Anderson, "Glyphosate residues in Alberta’s atmospheric deposition, soils and surface waters, Pub No. T/806," 2005. [372] I. W. G. WRB, "World reference base for soil resources," World Soil Resources, Rome 103, 2007. [373] S. S. Staff, A basic system of soil classification for making and interpreting soil surveys, 2nd ed.: Govt. Printer: Washington,DC, 1999. [374] R. F. Isbell, "A brief history of national soil classification in Australia since the 1920s," Australian Journal of Soil Research, vol. 30, pp. 825-842, 1992. [375] A. W. Moore, "The USDA Soil Taxonomy in relation to some soils of eastern Queensland.," Brisbane, 1986. [376] R. F. Isbell, The Australian Soil Classification, Revised edn ed. Melbourne: CSIRO, 2002. [377] D. T. Morand, "The world reference base for soils (WRB) and soil taxonomy: an appraisal of their application to the soils of the northern rivers of New South Wales," Soil Research, vol. 51, pp. 167-181, 2013. [378] P. Buurman, B. V. Lagen, and E. J. Velthorst, Manual for soil and water analysis., 1996. [379] "Guide to the sampling and investigation of potentially contaminated soil part 1: Non volatile and semi-volatile compounds," Australian Standard AS4482.12005. [380] "Guideline of laboratory analysis of potentially contaminated soil," National Environment Protection (Assessment of Site Contamination) Measure, NEPM 1999. [381] Keith, "Environmental sampling and analysis: A practical guide," Lewis1991. [382] J. E. Brodie and K. Fabricius, "Terrestrial runoff to the Great Barrier Reef and theimplications for its long term ecological status," 2008.

248

References

[383] J. Brodie, C. Christie, M. Devlin, D. Haynes, S. Morris, M. Ramsay, et al., "Catchment management and the Great Barrier Reef," Wat. Sci. Tech., vol. 43, pp. 203-211, 2001a. [384] J. E. .Brodie, J. Binney, K. Fabricius, I. Gordon, O. Hoegh-Gouldberg, H. Hunter, et al., "Scientific consensus statement on water quality in the Great Barrier Reef," Queensland Department of the Premier and Cabinet, Brisbane2008a. [385] J. Brodie, J. Binney, K. Fabricius, I. Gordon, O. Hoegh-Guldberg, H. Hunter, et al., "Synthesis of evidence to support the scientific consensus statement on water quality in the Great Barrier Reef," Queensland Department of the Premier and Cabinet, Brisbane2008b. [386] J. E. Brodie, M. Devlin, D. Haynes, and J. Waterhouse, "Assessment of the eutrophication status of the Great Barrier Reef lagoon (Australia)," Biogeochemistry vol. 106, pp. 281-302, 2011a. [387] G. De'ath and K. Fabricius, "Water quality as a regional driver of coral biodiversity and macroalgae on the Great Barrier Reef," Ecological Applications, vol. 20, pp. 840-850, 2010. [388] M. Furnas, "Catchments and corals: terrestrial run-off to the Great Barrier Reef," Australian Institute of Marine Science, Townsville2003. [389] J. King, F. Alexander, and J. Brodie, "Regulation of pesticides in Australia: The Great Barrier Reef as a case study for evaluating effectiveness," Agriculture, Ecosystems and Environment, vol. 180, pp. 54-67, 2013. [390] B. Sajjakulnukit, "Bio-fuels in Thailand," in Proceedings of ASEAN-U.S. enhanced partnership biofuels and the automotive industry seminar, Bangkok, Thailand, 2007. [391] L. Mamy, E. Barriuso, and B. Gabrielle, "Environmental fate of herbicides trifluralin, metazachlor, metamitron and sulcotrione compared with that of glyphosate, a substitute broad spectrum herbicide for different glyphosate- resistant crops," Pest Management Science, vol. 61, pp. 905–916, 2005. [392] "Energy independence and security Act of 2007," presented at the 110th United States Congress, Washington DC., 2007. [393] "2007 Census of agriculture: United States summary and state data, Vol. 1, Geographic area series, Part 51," United States Department of Agriculture 2009. [394] C. Williams, M. Agassi, J. Letey, W. Farmer, S. Nelson, and M. Ben-Hur, "Facilitated transport of napropamide by dissolved organic matter through soil columns," Soil Science Society of America Journal, vol. 64, pp. 590-594, 2000. [395] "A basic system of soil classification for making and interpreting soil surveys," in Soil Use and Management, Agricultural Handbook 436. vol. 17, 2nd ed, 2001, pp. 57-60. [396] M. McBride and K.-H. Kung, "Complexation of glyphosate and related ligands with iron (III)," Soil Science Society of America Journal, vol. 53, pp. 1668- 1673, 1989. [397] R. Vanantwerpen and J. H. Meyer, "Soil degradation under sugarcane cultivation in northern kwazulu-natal," in South African Sugar Technologists' Association, 1996, pp. 29-33. [398] "Managing soil quality: challenges in modern agriculture," P. Schjønning, S. Elmholt, and B. T. Christensen, Eds., ed Wallingford, UK: CABI, 2004.

249

References

[399] T. S. Steenhuis, C. J. Ritsema, L. W. Dekker, and J. Y. Parlange, "Fast and early appearance of solutes in groundwater by rapid and far-reaching flow," presented at the 15th International Congress on Soil Science, Mexico, 1994. [400] D. J. Sullivan and M. M. Sallaway, "Development of soil conservation specifications in the coastal Burnett district," in Australian Society of Sugarcane Technologists, 1994, pp. 178-186. [401] S. Shoval and S. Yariv, "The interaction between roundup (glyphosate) and montmorillonite. Part 1. Infrared study of the sorption of glyphosate by montmorillonite," Clays and Clay Minerals, vol. 27, pp. 19-28, 1979. [402] G. Mckay and S. J. Allen, "Surface mass transfer processes using peat as an adsorbent for dyestuffs," The Canadian Journal of Chemical Engineering, vol. 58, pp. 521-526, 1980. [403] T. K. Srivastava, G. B. Singh, and S. N. Srivastava, "Efficacy of weed control measures and soil persistence of atrazine in sugarcane (Saccharum officinarum) as influenced by irrigation and nitrogen " Indian Journal of Agronomy, vol. 44, pp. 835-840, 1999. [404] R. L. Yadav, S. R. Prasad, R. Singh, and V. K. Srivastava, "Recycling sugarcane trash to conserve soil organic carbon for sustaining yields of successive ratoon crops in sugarcane," Bioresource Technology, vol. 49, pp. 231–235, 1994. [405] A. Piccolo, G. Celano, and P. Conte, "Adsorption of glyphosate by humic substances," Journal of Agricultural and Food Chemestry, vol. 44 pp. 2442– 2446, 1996. [406] K. H. Tan, Principles of soil chemistry, 3rd ed.: CRC Press, 1998. [407] K. Verburg, B. A. Keating, M. E. Probert, K. L. Bristow, and N. I. Huth, "Nitrate leaching under sugarcane: interactions between crop yield, soil type and management strategies," presented at the 9th Australian Agronomy Conference, Wagga Wagga, NSW, 1998. [408] D. Ghosh and K. G. Bhattacharyya, "Adsorption of methylene blue on kaolinite," Applied Clay Science, vol. 20, pp. 295–300, 2002. [409] R. E. Treybal, Mass-transfer operations, 3rd ed.: McGraw-Hill, 1981. [410] O. Hamdaoui and E. Naffrechoux, "Modeling of adsorption isotherms of phenol and chlorophenols onto granular activated carbon: Part I. Two- parameter models and equations allowing determination of thermodynamic parameters," Journal of Hazardous Materials, vol. 147, pp. 381-394, 2007. [411] L. C. Salazar and A. P. Appleby, "Herbicidal Activity of Glyphosate in Soil," Weed Science vol. 30, pp. 463-466, 1982. [412] C. R. Worthing and S. B. Walker, The Pesticide Manual. Suffolk, England: Lavenham Press, 1987. [413] P. H. Nicholls and A. A. Evans, "Sorption of lonisable organic compounds by field soils. Part 2: Cations, bases and zwitterions," Pesticide Science, vol. 33, pp. 331-345, 1991. [414] U. B. Cheah, R. C. Kirkwood, and K. Y. Lum, "Adsorption, desorption and mobility of four commonly used pesticides in malaysian agricultural soils," Pesticide Science, vol. 50, pp. 53-63, 1997. [415] T. Kiely, D. Donaldson, and A. Grube, "Pesticides industry sales and usage," U.S. Environmental Protection Agency, Office of Pesticide Programs, Washington DC.2004.

250

References

[416] Genetically modified crop fact sheet, Department of Agriculture, Fisheries and Forestry. Available: www.daff.gov.au [417] "Fate, transport and transformation test guidelines," United States Environmental Protection Agency2008. [418] N. Toride, F. J. Leij, and M. T. v. Genuchten, "The CXTFIT code for estimating transport parameters from laboratory or field tracer experiments," U. S. DEPARTMENT OF AGRICULTURE, Riverside, California1995. [419] S. Grundmann, U. Dörfler, B. Ruth, C. Loos, T. Wagner, H. Karl, et al., "Mineralization and transfer processes of 14C-labeled pesticides in outdoor lysimeters," Water Air and Soil Pollution, vol. 8, pp. 177-185, 2008. [420] N. S. Nomura and H. W. Hilton, "The adsorption and degradation of glyphosate in five Hawaiian sugarcane soils," Weed Research, vol. 17, pp. 113-121, 2006. [421] F. X. M. Casey, H. Hakk, J. Šimůnek, and G. L. Larsen, "Fate and transport of testosterone in agricultural soils," Environmental Science & Technology, vol. 38, pp. 790-798, 2004. [422] C. S. Buergisser, M. Cernik, M. Borkovec, and H. Sticher, "Determination of nonlinear adsorption isotherms from column experiments: an alternative to batch studies," Environmental Science & Technology, vol. 27, pp. 943-948, 1993. [423] P. D. Capel, W. Giger, P. Reichert, and O. Wanner, "Accidental input of pesticides into the Rhine river.," Environmental Science and Technology, vol. 22, pp. 992-997, 1988. [424] D. L. Corwin, "An assessment of the significant physicochemical interactions involved in pesticide diffusion within a pesticide-sediment-water system," Chemosphere, vol. 13, pp. 1295-1317, 1984. [425] V. Petit, R. Cabtidenc, R. P. J. Swannell, and R. S. Sokhi, "Review of Strategies for modelling the environmental fate of pesticides discharged into riverine systems," Environment International, vol. 21, pp. 167-176, 1995. [426] Fate of pesticides and chemicals in the environment, 1991.

251

Appendix A

Quotes from Commercial Laboratory

252

Page 1 of 3

QT-tba

NMI QUOTATION – GT12/0830CA_UNINxx (Amended)

Date of quote: 30 August 2012 Amended 31 August 2012 105 Delhi Road Quote valid to: 31 December 2012 North Ryde NSW 2113 Contact: Kamrun Nahar PO Box 138 Organisation: University of NSW @ Australian Defence Force Academy North Ryde NSW 1670 Address: Desk 14, Rm 381/21, School of Engineeering & Information Technology Phone: +61 2 9449 0111 Northcott Drive, Canberra ACT 2600 Fax: +61 2 9449 0296 Telephone: 02 6268 6057 Mobile: 0425 910 200 www.measurement.gov.au e-mail address: [email protected] ABN: 74 599 608 295

Project Reference: Analysis of Glyphosate in water and soil.

Dear Kamrun, Thank you for the opportunity to quote for the analysis of glyphosate in water and soil as detailed in your email. Please see the details and notes below and on the following pages. Amended to add additional soil tests.

No. Limit Price per Sample Sample type of samples Test of Reporting EXCL GST $ Water 20 ** Glyphosate (and AMPA *) 10 μg/L 150 Glyphosate (and AMPA *) 0.05 mg/kg 200 Available Phosphorus (Bray or Colwell) 0.1/5 mg/kg 36 Soil 20 ** Cation Exchange Capacity (CEC) 0.08 meq/100g 54 Iron (Fe) 0.5 mg/kg 26 Aluminium (Al) 0.5 mg/kg for both Total Solids in soil 0.1% 10 * Amino Methyl Phosphonic Acid, the main breakdown product of Glyphosate. ** Can be any combination of 20 or more samples, eg. 10 water and 10 soil or 15 water and 6 soil. Preferred sample sizes: Water: at least 500 mL collected in amber glass bottles. Soil: at least 200 grams collected in a glass jar.

Goods and Services Tax The prices in this Quotation are EXCLUSIVE of Goods and Services Tax (GST). The current GST rate is 10%. At the completion of the job NMI will issue valid tax invoices and if required, adjustment notes, as per the requirements of the GST legislation. Handling Fee: A standard Handling Fee of $33.00 inclusive of GST component $3.00, applies per invoice. Minimum invoice amount NMI has a minimum invoice amount of $275 including handling fee & GST. Legal proceedings samples If results of the requested analyses may be used in legal proceedings please advise NMI as procedures and pricing may vary Handling and Storage Samples to be kept cold during storage and transit to the Laboratory. Samples may be delivered to the laboratory between Requirements: 0830 hrs and 1730hrs Monday to Friday. Recommended Holding NMI will use its best efforts to commence preparation and analysis of samples within recommended holding time (RHT) Time (RHT) provided that samples are delivered to the laboratory in an appropriate timeframe. NMI is not responsible for any breach of RHT. Click the link on the left for details of RHT. Hazardous Material Advice See ‘Treatment of Samples’ section of our Service Terms and Conditions. Limits of Reporting LORs are targets only and may be revised if your samples are found to contain substances which cause interferences, or if matrix effects become apparent, during analysis. While every effort will be made to minimise such effects, NMI has no control over the sample constituents, and hence cannot guarantee to reach the LORs quoted when persistent interferences occur. Agreed Turnaround Time: 7 working days from receipt at laboratory depending on batch sizes. Premiums apply for shorter TATs which must be requested, and agreed to by NMI in advance of sample despatch. Moisture Content Results (mg/kg) for soil, sediment & other environmental solids are reported on a dry weight basis unless otherwise requested. The following GST inclusive amount per sample will be charged for determination of moisture or total solids content: $11.00

______gt0830ca_uninxx_water_soil National Measurement Institute Page 2 of 3

Please ensure the above NMI Quotation Number GT12/0830CA_UNINxx appears on all correspondence, and sample submission or chain of custody forms. This quotation and your acceptance thereof are subject to the Terms and Conditions on the following page.

I trust this quotation is suitable to your requirements. If you have any questions or require any further information, please call me on 02 9449 0111, fax 02 9449 0296, or email to

QUOTED BY:

CUSTOMER SERVICE MANAGER Brian Woodward 31 August 2012 Signature Date

Delivery address for samples: NMI, Attn: Sample Receipt, 105 Delhi Road, North Ryde, NSW 2113

______gt0830ca_uninxx_water_soil National Measurement Institute NMI SERVICE TERMS & CONDITIONS Page 3 of 3

The lodgement of an order or receipt of samples for Terms of payment are strictly 30 days from date of FORCE MAJEUR NMI services constitutes an acceptance of the invoice. A late fee of up to $11.00 inclusive of GST NMI shall not be responsible or liable for any delay to following terms and conditions. may apply if the client does not pay in full by the due perform any of its obligations when such delay or date. failure to perform any of its obligations is caused by Unless otherwise agreed in writing, the following terms unforeseen circumstances beyond its reasonable and conditions apply to services conducted by NMI, TREATMENT OF SAMPLES control and without its fault or negligence, including, resulting from engagement of NMI either by accepting Unless NMI has otherwise agreed in writing, the client without limitation, Acts of God, fire, explosion, riot, a quotation and/or submission of samples to NMI. The is responsible for collecting samples and for delivering sabotage, strike or other labour dispute, shortage of client agrees to be bound by and comply with these samples for testing to the address nominated in the materials, transportation difficulties or compliance with terms and conditions. Any terms and conditions you quotation. any order, action, governmental officer, department, notify to NMI, will apply only if and to the extent that agency, authority or committee thereof that renders NMI agrees to them in writing. When providing samples to NMI, the client must give performance impracticable or impossible for NMI. written notice of all known safety or health hazards and SERVICES special procedures relevant to the handling, testing, EXCLUSION OF WARRANTY NMI reserves the right to review prices at any time if: storage, transport and disposal of samples. NMI To the full extent permitted by law NMI and its (a) significant changes to our costs are incurred reserves the right to refuse to conduct any test where proprietor exclude all warranties, terms, conditions or beyond our control ie. changes to legislative NMI in its absolute discretion determines such testing undertakings, ('terms'), whether express or implied, in requirements or variations in tax or excise rates; or (b) may pose a safety or health hazard. relation to services, the report or its contents. Where any of the assumptions set out in the quotation prove any legislation implies any terms which cannot be to be incorrect. Where a formal request is made, NMI will return excluded or modified then such terms shall be deemed samples to the client, at the client's expense. to be included. However, (to the full extent permitted Alterations to the scope of the quoted services by law) NMI's liability to the client is limited at NMI’s (including changes to timeframe of services, sample The client acknowledges that during conduct of the option to the re-performance of service or the refund of numbers, limits of reporting, agreed analyte suite etc), services the samples or parts of samples may be service fee. prior to commencement of the services, may require a altered, damaged, lost or destroyed. NMI shall not be review of the quotation. liable to the client or any third party for any samples Without limiting the generality of this clause, it is that are altered, damaged, lost or destroyed during agreed that, to the full extent permitted by any Alterations to client requirements requested after conduct of the services. applicable Commonwealth, State or Territory laws commencement of the testing process will incur an having jurisdiction, NMI and its proprietor will not be administration fee of $22 inclusive of GST plus charges The client is responsible for ensuring that samples liable to the client or any other person for any loss of for extra service delivery costs incurred by NMI, if any. supplied for testing are representative of the product or profits or business whether directly or indirectly material to be analysed and for retaining any duplicate incurred or any special, indirect or consequential Records will be kept for the required minimum period or control samples. damages arising from the client's use of NMI's services unless otherwise requested and agreed to by NMI (eg or reports. NATA technical accreditation requires records are kept Unless NMI has otherwise agreed in writing, NMI shall for a minimum of three years). not be obliged to return samples to the client and may CLIENT'S RELEASE AND INDEMNITY in its discretion store, experiment on, destroy or The client hereby releases and indemnifies and shall BUSINESS HOURS dispose of samples. continue to release and indemnify NMI, its proprietor, Services will be provided by NMI during normal its officers, employees and agents from and against all business hours Monday to Friday (excluding public NMI reserves the right for samples deemed hazardous actions, claims, proceedings or demands (including holidays), unless otherwise agreed. by NMI to be returned to the client, at the client's any costs and expenses in defending or servicing expense. same) which may be brought against it or them, in Any services conducted outside NMI premises will be respect of any loss, death, injury, illness or damage to performed Monday to Friday (excluding public Where possible a representative sample will be kept persons or property, and whether direct or indirect and holidays) between 9am and 5pm, unless otherwise for a period of one (1) month from the date of final in respect of any infringement of any industrial or agreed. report. NMI will charge for costs incurred for longer intellectual property rights, howsoever arising out of the term storage, or for disposal of noxious samples. use of the report or the services of NMI. TURNAROUND TIMES Any samples received after 1630 hrs Monday to Friday OWNERSHIP CLIENT'S ACKNOWLEDGMENT or on Public Holidays are deemed to have been NMI will own the final report until such time as full The client acknowledges that: received the following working day. payment for the services is received, beyond which The client at its own risk uses the report and its time the client will own the final report. contents and any advice, opinions or information It is the client’s responsibility to ensure that NMI has supplied by NMI, its proprietor, its officers, employees access to all information and premises necessary to All intellectual property rights associated with sample or agents concerning the service; commence the services as agreed. analysis methods, processes and reports are vested, and shall remain vested, in NMI. No other party may The service is performed on the understanding that the The due date of the services may be delayed where replicate or appropriate the method or any part thereof client will not hold NMI, its proprietor, its officers, such information or access is not provided, or is judged for any use, be it commercial or otherwise, without the employees or agents liable for any loss or damage by NMI to be inadequate for the services to express written consent of NMI’s General Manager or resulting from the conduct of the service or the use of commence. approved delegate. or reliance upon the report or its contents; and

ACCOUNTS & PAYMENT LEGAL OBLIGATIONS It is the responsibility of the client to make its own GST at the applicable rate (currently 10%) will be NMI, its proprietor, its officers, employees and agents assessment of the suitability for any purpose of the charged in addition to the quoted prices. NMI will issue are under no legal obligation to provide information or service, report and its contents and any information or valid tax invoices and adjustment notes as per expert witnesses as an outcome of any testing advice generated therefrom. requirements of GST legislation. undertaken at NMI. GENERAL A minimum invoice fee inclusive of GST applies (this Any requests for NMI, its proprietor, its officers, The services are governed by the laws of the State in includes a handling fee per invoice). employees and agents to provide information or expert which services have been conducted, unless witnesses will not be granted without the express Commonwealth law prevails. The establishment of a trading account is subject to the written consent of NMI’s General Manager or approved completion of an account application form. delegate. The client will not represent in any way that NMI supports or endorses the client’s business, goods or NMI reserves the right to undertake credit verification In circumstances where NMI, its proprietor, its officers, services, without NMI’s written consent. The client will of all established accounts or to request up-front employees or agents agree or are required to provide not make any press release or public statement about payment of services before services can commence. information or appear as expert witnesses as an the services or NMI without NMI’s written consent.. outcome of testing undertaken at NMI an hourly fee will be charged to the client. Current from 1 July 2004

K:\CLNTSERV\QUOTES\Quotes12\08_AUG12\GT0830CA_UNINxx_water_soil.doc

General Enquiries Ph: 1300 722 845 [email protected] www.measurement.gov.au ABN: 74 599 608 295 NATIONAL MEASUREMENT INSTITUTE QUOTATION 105 Delhi Rd, North Ryde NSW 2113 1/153 Bertie St, Port Melbourne VIC 3207 26 Dick Perry Ave, Kensington WA 6151

Quotation Number: UNIN28A-KC1509N Date of Issue: September 2, 2015 Valid Until: December 31, 2015 LIMS Reference (NMI use only): QT-02018

Contact Name: Robert Niven Company: University of NSW ABN: Street Address: Northcott Drive, Canberra ACT 2600 Postal Address: Telephone/Mobile: 02 6268 6057 email address: [email protected] Customer Reference: As per your email to NMI on 2/9/15 Job / Project Reference: Glyphosate in soil and water

Dear: Robert Thank you for the opportunity to quote for the analysis of

LIMIT OF PRICE NO. OF MATRIX TEST REPORTING PER SAMPLE REFERENCE METHOD SAMPLES (mg/kg) (excl. GST) $ WATER - ORGANIC COMPOUNDS

WATER varies Glyphosate (plus AMPA, glufosinate) 10ug/L $ 280.00 USGS, Report 01-454, 2002

SOLIDS - ORGANIC COMPOUNDS USGS, Report 01-454, SOLIDS varies Glyphosate (plus AMPA, glufosinate) 0.5mg/kg $ 280.00 2002

HANDLING FEE: A standard Handling Fee of $33.00 inclusive of GST (GST component $3.00) applies per invoice. $33.00 MINIMUM INVOICE FEE: A minimum Invoice of $275 inclusive of GST applies. (Includes the $33.00 handling fee). $275.00 COMMENTS & SPECIAL CONDITIONS

LORs are targets only and may be revised if samples are found to contain substances which cause interferences, or if matrix effects become apparent during analysis

RECOMMENDED HOLDING TIME (RHT) NMI will use their best efforts to commence preparation and analysis of samples within recommended holding time (RHT) provided that samples are delivered to the laboratory in an appropriate timeframe. NMI is not responsible for any breach of RHT. SAMPLE DELIVERY Please deliver samples to 105 Delhi Rd, North Ryde NSW 2113 with completed sample submission (chain of custody) form (attached) Samples to be preserved in accordance with RHT requirements and securely packed in eskies during storage and transit to the Laboratory. Please do not submit samples to the laboratory until you have established a credit account

AGREED TURNAROUND TIME Typically 5-7 working days from receipt at laboratory. Premiums will apply for fast TATs if available. Fast TATs must be agreed to before dispatch of samples and requested in writing on the chain of custody / sample submission form. Premiums applied to fast TATs: 24 hrs - 100% Premium, 48 hrs - 50% Premium, 3-4 days - 25% Premium

QUOTED BY: Katie Chambers - Customer Service Officer DATE: 2/09/2015

When providing samples to NMI, the client must give written notice of all known safety or health hazards and special procedures relevant to the handling, testing, storage, transport and disposal of samples. NMI reserves the right to refuse to conduct any test where NMI in its absolute discretion determines such testing may pose a safety or health hazard. NMI CHAIN OF CUSTODY (SAMPLE SUBMISSION) FORM ENVIRONMENTAL SAMPLES to be submitted to: NMI: 105 Delhi Rd, North Ryde NSW 2113 Ph: 1300 722 845 email: [email protected] SENT FROM: Internal use only Company Name: University of NSW NMI Quote Number: UNIN28A-KC1509N Valid until: December 31, 2015

Address: Northcott Drive, Canberra ACT 2600 0 LIMS Reference: QT-02018 0 Contact: 02 6268 6057 Additional email(s) for report / invoice (if required): TURN AROUND TIME REQUESTED (Working days):

Phone: 0 24 hrs 48 hrs 3-4 5-7 10 20 other *Fast TATs are not available for all tests (please and MUST be agreed to prior to sample ABN: 0 100% 50% 25% Standard dioxins specify) submission Contact email: [email protected] If a PO number is required on your invoice, it must be provided at sample submission. PO's received after sample submisison will not appear on final invoice Purchase order required: Y / N PO Number:______

TESTS REQUIRED (Please list all tests required here and tick required tests against samples)

NMI LRN SAMPLE REFERENCE SAMPLE MATRIX (NMI USE ONLY - please do not (Sample ID / Description / DATE & TIME SAMPLED COMMENTS (water / soil / biota) write in this column) Number)

NOT

FILL

NMI

USE

ONLY

Relinquished by: Received at NMI laboratory by: Print Name: Print Name: PAGE No: of PAGES Date & Time: / / : hrs Date & Time: / / : hrs Signature: Signature: If multiple pages, ensure ALL pages are stapled together NMI SERVICE TERMS & CONDITIONS

The lodgment of an order or receipt of samples for Terms of payment are strictly 30 days from date of FORCE MAJEUR NMI services constitutes an acceptance of the invoice. A late fee of up to $11.00 inclusive of GST NMI shall not be responsible or liable for any delay following terms and conditions. may apply if the client does not pay in full by the due to perform any of its obligations when such delay or date. failure to perform any of its obligations is caused by Unless otherwise agreed in writing, the following unforeseen circumstances beyond its reasonable terms and conditions apply to services conducted by TREATMENT OF SAMPLES control and without its fault or negligence, including, NMI, resulting from engagement of NMI either by Unless NMI has otherwise agreed in writing, the without limitation, Acts of God, fire, explosion, riot, accepting a quotation and/or submission of samples client is responsible for collecting samples and for sabotage, strike or other labor dispute, shortage of to NMI. The client agrees to be bound by and delivering samples for testing to the address materials, transportation difficulties or compliance comply with these terms and conditions. Any terms nominated in the quotation. with any order, action, governmental officer, and conditions you notify to NMI, will apply only if department, agency, authority or committee thereof and to the extent that NMI agrees to them in writing. When providing samples to NMI, the client must give that renders performance impracticable or written notice of all known safety or health hazards impossible for NMI. SERVICES and special procedures relevant to the handling, NMI reserves the right to review prices at any time if: testing, storage, transport and disposal of samples. EXCLUSION OF WARRANTY (a) significant changes to our costs are incurred NMI reserves the right to refuse to conduct any test To the full extent permitted by law NMI and its beyond our control ie. changes to legislative where NMI in its absolute discretion determines proprietor exclude all warranties, terms, conditions requirements or variations in tax or excise rates; or such testing may pose a safety or health hazard. or undertakings, ('terms'), whether express or (b) any of the assumptions set out in the quotation implied, in relation to services, the report or its prove to be incorrect. Where a formal request is made, NMI will return contents. Where any legislation implies any terms samples to the client, at the client's expense. which cannot be excluded or modified then such Alterations to the scope of the quoted services terms shall be deemed to be included. However, (to (including changes to timeframe of services, sample The client acknowledges that during conduct of the the full extent permitted by law) NMI's liability to the numbers, limits of reporting, agreed analyte suite services the samples or parts of samples may be client is limited at NMI’s option to the re- etc), prior to commencement of the services, may altered, damaged, lost or destroyed. NMI shall not performance of service or the refund of service fee. require a review of the quotation. be liable to the client or any third party for any samples that are altered, damaged, lost or Without limiting the generality of this clause, it is Alterations to client requirements requested after destroyed during conduct of the services. agreed that, to the full extent permitted by any commencement of the testing process will incur an applicable Commonwealth, State or Territory laws administration fee of $22 inclusive of GST plus The client is responsible for ensuring that samples having jurisdiction, NMI and its proprietor will not be charges for extra service delivery costs incurred by supplied for testing are representative of the product liable to the client or any other person for any loss of NMI, if any. or material to be analysed and for retaining any profits or business whether directly or indirectly duplicate or control samples. incurred or any special, indirect or consequential Records will be kept for the required minimum damages arising from the client's use of NMI's period unless otherwise requested and agreed to by Unless NMI has otherwise agreed in writing, NMI services or reports. NMI (eg NATA technical accreditation requires shall not be obliged to return samples to the client records are kept for a minimum of three years). and may in its discretion store, experiment on, CLIENT'S RELEASE AND INDEMNITY destroy or dispose of samples. The client hereby releases and indemnifies and shall Estimations of measurement uncertainty (MU) are continue to release and indemnify NMI, its available upon request. NMI reserves the right for samples deemed proprietor, its officers, employees and agents from hazardous by NMI to be returned, to the client, at and against all actions, claims, proceedings or BUSINESS HOURS the client's expense. demands (including any costs and expenses in Services will be provided by NMI during normal defending or servicing same) which may be brought business hours Monday to Friday (excluding public Where possible a representative sample will be kept against it or them, in respect of any loss, death, holidays), unless otherwise agreed. for a period of one (1) month from the date of final injury, illness or damage to persons or property, and report. NMI will charge for costs incurred for longer whether direct or indirect and in respect of any Any services conducted outside NMI premises will term storage, or for disposal of noxious samples. infringement of any industrial or intellectual property be performed Monday to Friday (excluding public rights, howsoever arising out of the use of the report holidays) between 9am and 5pm, unless otherwise OWNERSHIP or the services of NMI. agreed. NMI will own the final report until such time as full payment for the services is received, beyond which CLIENT'S ACKNOWLEDGMENT TURNAROUND TIMES time the client will own the final report. The client acknowledges that: Any samples received after 1630 hrs Monday to The client at its own risk uses the report and its Friday or on Public Holidays are deemed to have All intellectual property rights associated with contents and any advice, opinions or information been received the following working day. sample analysis methods, processes and reports supplied by NMI, its proprietor, its officers, are vested, and shall remain vested, in NMI. No employees or agents concerning the service; It is the client’s responsibility to ensure that NMI has other party may replicate or appropriate the method access to all information and premises necessary to or any part thereof for any use, be it commercial or The service is performed on the understanding that commence the services as agreed. otherwise, without the express written consent of the client will not hold NMI, its proprietor, its officers, NMI’s General Manager or approved delegate. employees or agents liable for any loss or damage The due date of the services may be delayed where resulting from the conduct of the service or the use such information or access is not provided, or is LEGAL OBLIGATIONS of or reliance upon the report or its contents; and judged by NMI to be inadequate for the services to NMI, its proprietor, its officers, employees and commence. agents are under no legal obligation to provide It is the responsibility of the client to make its own information or expert witnesses as an outcome of assessment of the suitability for any purpose of the ACCOUNTS & PAYMENT any testing undertaken at NMI. service, report and its contents and any information GST at the applicable rate (currently 10%) will be or advice generated therefrom. charged in addition to the quoted prices. NMI will Any requests for NMI, its proprietor, its officers, issue valid tax invoices and adjustment notes as per employees and agents to provide information or GENERAL requirements of GST legislation. expert witnesses will not be granted without the The services are governed by the laws of the State express written consent of NMI’s General Manager in which services have been conducted, unless A minimum invoice fee inclusive of GST applies (this or approved delegate. Commonwealth law prevails. includes a handling fee per invoice). In circumstances where NMI, its proprietor, its The client will not represent in any way that NMI The establishment of a trading account is subject to officers, employees or agents agree or are required supports or endorses the client’s business, goods or the completion of an account application form. to provide information or appear as expert witnesses services, without NMI’s written consent. The client as an outcome of testing undertaken at NMI an will not make any press release or public statement NMI reserves the right to undertake credit hourly fee will be charged to the client. about the services or NMI without NMI’s written verification of all established accounts or to request consent. up-front payment of services before services can commence. Current as of 1 January 2011