Abstract
The 2010 4-yearly national survey of pesticides in groundwater included 162 wells. The aims of the survey were to update the national overview of pesticides in New Zealand's groundwater systems, to investigate temporal variation in pesticide concentrations between surveys, and to identify environmental factors associated with pesticide contamination. Pesticides were detected in 38 wells (23%), with two or more pesticides detected in 15 wells (9%). Pesticides were measured in wells in nine of the 14 regions sampled. One well showed a dieldrin concentration greater than the maximum acceptable value for drinking water. A total of 22 different pesticides were detected. Herbicides were the most common pesticide group detected. Levels of only three of the 66 pesticide detections exceeded 1 mg/m3. Comparisons with earlier surveys indicate that a similar percentage of wells had detectable pesticides in the last four surveys once correction for variable detection limits was made.
Introduction
Pesticides used on land can contaminate groundwater through leaching, spillage and preferential flow through soils (Close et al. Citation2001). In several regions of New Zealand, groundwater is an important source of drinking water. Ten per cent of the volume of groundwater abstracted is used for public water supplies (Ministry for the Environment Citation2006) and approximately half of the community drinking water supplies and many rural households rely on groundwater as a source of drinking water (Close et al. Citation2001; Davies Citation2001). Groundwater is also used extensively in primary production as a source of water for irrigation and stock, and nationally the volume of groundwater abstracted is increasing (Ministry for the Environment Citation2006). Regional and national surveys of groundwater have reported pesticide contamination of groundwater and in particular contamination of shallow and unconfined systems (Close Citation1993, Citation1996; Taranaki Regional Council Citation1995; Canterbury Regional Council Citation1997; Hadfield & Smith Citation1999; Close & Rosen Citation2001; Close & Flintoft Citation2004; Gaw et al. Citation2008). Although pesticides have generally been detected at low concentrations in groundwater, occasional exceedances of the corresponding maximum acceptable values (MAVs) for pesticides in drinking water have been reported (Close et al. Citation2001) making ongoing monitoring necessary for regional councils. Regular monitoring is also necessary to assess and demonstrate whether measures to minimise and prevent pesticide contamination of groundwater have been successful.
National surveys of pesticides in New Zealand groundwater have been undertaken with the assistance of regional and unitary authorities, approximately every 4 years since 1990 (Close Citation1993, Citation1996; Close & Rosen Citation2001; Close & Flintoft Citation2004; Gaw et al. Citation2008). The Sixth National Survey of Pesticides in Groundwater was undertaken in late 2010 and the results are reported here. The aims of this survey were to update the national overview of pesticides in New Zealand's groundwater systems, to investigate temporal variation in pesticide concentrations, and to identify environmental factors associated with pesticide contamination of groundwater.
Methods
Well selection
Fourteen regional councils and unitary authorities with responsibility for groundwater management participated in the 2010 survey (). To enable comparison with previous surveys of pesticides in groundwater, similar well selection criteria were applied, including the regional importance of the aquifer; the known use of pesticides in the area; and the vulnerability of the aquifer to contamination. Where possible, wells sampled in previous surveys were included to provide a direct temporal comparison. For each well sampled the following information was recorded: well location, water level, depth of the well screen, the type of aquifer (confined or unconfined) and the surrounding land-use. Most of the wells sampled (75% of the wells for which confinement status was known) were from unconfined aquifers. The selection included wells in most types of New Zealand groundwater (alluvial gravel, sand, shell bed and fractured volcanic aquifers), with the exception of very deep aquifers that would not be expected to show pesticide contamination.
Figure 1 Location map showing regional council boundaries within New Zealand. The West Coast region was not included in this survey.
Sampling and analysis
Sampling was undertaken from September to December 2010. Single samples were collected into solvent-washed 1-l glass bottles from selected wells by regional council and unitary authority staff and analysed by AsureQuality, Wellington. Samples were collected by either directly from down-hole pumps or with in situ pumps sampling as close to the borehead as possible. Where possible, wells were purged for three well volumes before samples were collected. Electrical conductivity, dissolved oxygen and pH were measured in the field at the time of sampling where possible. Some nitrate concentrations were measured from samples taken in conjunction with the pesticide samples. Where values for these parameters were not measured at the time of sampling, median values were taken of previously collected data.
A total of 162 wells from 14 regions were sampled and the groundwater analysed. The samples were analysed for acid and hydrophilic herbicides and a suite of organochlorine, organophosphorus and organonitrogen pesticides (OC/OP/ON) (Appendix) using gas chromatography with mass spectrometry detection (GC-MS). The acid herbicide analysis involved solid phase extraction and derivatisation of the extract with diazomethane followed by GC-MS analysis using single ion monitoring (SIM) and was based on method 6640 (APHA Citation2005). The OC/ON/OP pesticide analysis involved extraction with dichloromethane and a pre-concentration step followed by GC-MS analysis with quantification using the SIM mode (method 8270, USEPA Citation1989). Field blanks were not collected but samples from eight wells (5%) were collected in duplicate as blind duplicate samples and analysed for quality control purposes. Laboratory blanks and spike recoveries were analysed with each batch of samples by the laboratory as part of the laboratory QC procedure. There were no pesticides detected in the laboratory blanks. The spike recoveries were required to be in the range of 75–130%. Six of the samples from Environment Southland were analysed using similar methods by Hill Laboratories in Hamilton.
Data analysis
Wells were categorised based on the presence or absence of pesticides and the total concentration of pesticides present in each well was calculated. Results below the detection limit were assigned a value of zero to avoid overestimating the total concentration of pesticides present in each well. T-tests were carried out using SYSTAT to explore the association of well parameters and environmental factors, namely well depth, well diameter, temperature, pH, nitrate-N concentration, conductivity and dissolved oxygen, with the pesticide presence/absence data. These well parameters and environmental factors are jointly referred to as groundwater parameters in the remainder of the paper. The F statistic was used to determine whether the variances should be pooled or kept separate (Rothery Citation2000). Pearson's correlation coefficient was used to determine correlations between the total pesticide concentration and groundwater factors for wells with pesticides detected. The wells were also categorised based on the type of aquifer (unconfined, semi-confined or confined) to test the association of aquifer type with detection of pesticides in groundwater. For wells that had been sampled in four or more surveys, an assessment for temporal trend was carried out. Pearson's correlation coefficient was used to determine significant correlations between the total pesticide concentration and the year of the survey.
Results
Blind duplicate samples from eight wells (5%) were analysed as a quality control measure. There was good agreement for all duplicate analyses (). Pesticides were not detected in either of the blind duplicate samples from seven wells. Terbuthylazine was detected in both duplicate samples from well 207244 with very good agreement.
Table 1 Comparison of blind duplicate samples.
Overall survey results
Pesticides were detected in nine of the 14 participating regions (the West Coast was the only region that did not participate in the survey). Pesticides were not detected in wells from the Bay of Plenty, Taranaki, Hawke's Bay, Marlborough and Canterbury regions. Of the 162 groundwater wells that were sampled, 38 (23%) tested positive for pesticides and in 15 of these wells two or more pesticides were detected (). The maximum number of pesticides detected in one well was five and there were four pesticides detected in four wells.
Table 2 Summary of pesticide concentrations measured in the 2010 groundwater survey; regions are arranged north to south.
Twenty-two different pesticides and pesticide metabolites were detected in the wells sampled (). Herbicides were the pesticide group most commonly detected with 17 different herbicides found, followed by insecticides (three), and fungicides (two). There was a total of 66 pesticide detections and of these, 60 (91%) were herbicides with 40 detections of triazine herbicides. Terbuthylazine (17 wells) and simazine (10 wells) were the two most frequently found pesticides (). Three insecticides (chlorpyrifos, pp-DDT and dieldrin) were detected in this survey in three different wells. Along with the detection of each insecticide there were multiple herbicides detected in the same well.
Table 3 Characteristics of detected pesticides for the 2010 groundwater survey.
Concentrations of only three of the 66 pesticide detections exceeded 1 mg/m3, and only one of the positive pesticide detections, dieldrin, exceeded its corresponding MAV of 0.04 mg/m3 for drinking water (). Dieldrin was detected in one well at a concentration of 0.13 mg/m3 in November 2010. The well was resampled 7 months later in May 2011 and the concentration of dieldrin in the second sample was 0.14 mg/m3, confirming that the well contained high levels of dieldrin. Terbuthylazine was detected in 17 wells from six regions. The highest level detected was 5.8 mg/m3, which was 73% of the MAV (). Alachlor was detected in four wells with the highest concentration being 12 mg/m3, which is 60% of the MAV. Most of the pesticides detected were at concentrations of less than 1% of the MAV.
The mobility and degradation characteristics, groundwater ubiquity scores (GUS) and FAO class (FAO Citation1996) for each pesticide are also given in . A review and collation of mobility and degradation values for pesticides has been carried out by the United States National Pesticide Information Center (US NPIC Citation2011) and the mobility and degradation values are from this source unless otherwise noted. The selected value listed in this database, plus the range of values in the literature, are given in . The mobility is represented by the soil organic carbon sorption coefficient (K oc). K oc is calculated by measuring the ratio, K d, of sorbed to solution pesticide concentrations after equilibration of a pesticide in a water/soil slurry and then dividing by the weight fraction of organic carbon present in the soil. This assumes that the pesticides are sorbed to the organic matter and not to the clay or mineral content in the soil. High K oc values indicate compounds with high adsorption to soils and low mobility. The soil half-life is the time it would take for half the amount of pesticide to degrade in soil, assuming a first-order degradation process. The GUS scores are a simplified assessment of whether a pesticide is likely to leach (Gustafson Citation1989) and are calculated as:
Gustafson (Citation1989) used GUS values greater than 2.8 to indicate that the compound would leach relatively readily and a GUS score of less than 1.8 to indicate a ‘non-leacher’. There was a transitional zone between 1.8 and 2.8 where pesticides could leach under favourable conditions. Primi et al. (Citation1994) suggested a wider transitional zone with GUS criteria of 1.5 and 3.0 to differentiate leachers and non-leachers, and these criteria were used in this study.
Two wells were sampled in all six national surveys of pesticides in groundwater and a further 61 wells were sampled in at least four of the surveys. Of the 63 wells sampled in four or more of the surveys, 30 (48%) had no pesticides detected in any of the surveys. Five (8%) wells showed significant (P<0.5) decreasing total pesticide concentrations and two (3%) wells showed significant (P<0.5) increasing total pesticide concentrations. There were a further two wells that showed increasing levels, and a further five wells that showed decreasing levels, of total pesticides concentrations at lower significance levels (0.05 < P<0.20). The remaining 19 wells showed no trend. Wells with no detections also belong to the group of no trends.
Effects of groundwater parameters and aquifer confinement status
When there were sufficient data, the relationships between groundwater factors and the presence/absence of pesticides were investigated using t-tests (). also summarises the means and standard deviations for the groundwater variables. In contrast to previous surveys, there was only a significant difference for nitrate concentrations between wells with and wells without detected pesticides. Wells with pesticide detections had higher nitrate concentrations. There was a significant correlation for wells with pesticides detected between the total pesticide concentration and the nitrate concentration (; r=0.71, n=23, P<0.001). The aquifer confinement status was known for 101 of 162 wells. All pesticide detections were from unconfined aquifers (23 wells) or from aquifers with unknown status (15 wells). No pesticides were detected in wells from semi-confined or confined aquifers. If only the wells with known aquifer confinement status were analysed then statistically more pesticide detections (χ2=15.2, P<0.001, likelihood ratio chi-square test) occurred in the unconfined aquifers than would be expected on the basis of an even distribution, compared with the semi-confined and confined aquifers.
Figure 2 Comparison of nitrate with total pesticide concentrations.
Table 4 Summary of t-test results between groundwater parameters and the presence/absence of pesticides.
Discussion
Herbicides comprised 60 of the 66 detections (91%) for pesticides. This higher number of detections for herbicides compared with insecticides and fungicides is consistent with estimates that herbicides comprise at least 60% of the total amount of pesticides sold annually in New Zealand (Manktelow et al. Citation2005). In addition, although mobility and degradation properties of herbicides vary widely according to their chemical classification (Weber Citation1994), they are often more polar and water soluble than insecticides and fungicides, making it more likely that they will leach to groundwater.
Consistent with previous surveys of pesticides in groundwater (Close et al. Citation2001), the triazine group of herbicides was the most frequently detected, comprising 61% of the detections, consistent with the percentage for triazine detections (50–76%) in the previous four surveys (). One triazine metabolite, desethyl atrazine, was detected in one well in the 2010 survey. No triazine metabolites were detected in the 2006 survey but six metabolites were detected in the 2002 survey (Close & Flintoft Citation2004) and 12 metabolites were detected in the 1998 survey (Close & Rosen Citation2001). The higher frequency of detection of herbicides in New Zealand groundwater is comparable with results of groundwater surveys undertaken in Norway (Haarstad & Ludvigsen Citation2007), Spain (Hernandez et al. Citation2008) and the USA (Kolpin et al. Citation1998).
Table 5 Summary statistics for the five national surveys of pesticides in groundwater in New Zealand.
The infrequent detection of insecticides is consistent with the previous national surveys of pesticides in groundwater (Close & Rosen Citation2001; Close & Flintoft Citation2004; Gaw et al. Citation2008). Two of the three insecticides detected, chlorpyrifos and pp-DDT, have been detected in a previous survey and the third insecticide, dieldrin, has been detected in groundwater in the Waikato region (Hadfield & Smith Citation1999). The detected insecticides all have very high K oc values corresponding to low mobility () but have very long half-lives, and DDT and dieldrin were used extensively in the past all over New Zealand (Boul Citation1995; Close et al. Citation2001). The pesticide half-lives listed in are for the topsoil and persistence of pesticides can be ten times those values in the vadose and groundwater environments (Levy & Chesters Citation1995).
Procymidone (fungicide) has been detected in one or two wells in five of the six national surveys, including this survey, and has been detected in some regional investigations of pesticides in groundwater (Hadfield & Smith Citation1999). Procymidone, which had a K oc value of 1500 ml/g and a soil half-life of 15 days (USDA pesticide database www.arsusda.gov/acsl/ppdb.html as at March 2005) and therefore not expected to leach to groundwater, was included in a series of field trials (Close et al. 2008) after detection in the 1990 survey. The field trials from three New Zealand soils gave a median K oc value of 352 ml/g and a median soil half-life of 850 days (Close et al. Citation2008). Thus procymidone is much more likely to leach from New Zealand soils than would have been expected from previous literature values. The GUS score calculated using the new leaching parameters () indicates that procymidone is likely to leach in New Zealand conditions.
The detection limits for the surveys undertaken since 1998 have been lower than the limits for the 1994 and 1990 national surveys by a factor of between 5 and 10, making direct comparison among the six surveys difficult. If the detection limits for the 1990 and 1994 surveys are applied to the 2010 survey, then pesticides would have been detected in 12 out of the 162 wells sampled (7%). In comparison, 8% of the 163 wells in 2006 would have contained detectable pesticides, 9% of the 133 wells in 2002, 11% of the 95 wells in 1998, 14% of the 118 wells in 1994 and 7% of the 82 wells in 1990, when results are adjusted for higher detection limits (). These values indicate that a similar percentage of wells had detectable pesticides in each survey once the results were corrected for variable detection limits and that there has been a slight decrease in wells with detectable pesticides since the 1994 survey.
The MAVs for the detected pesticides in drinking water range from 0.04–700 mg/m3 (). Similar to the earlier surveys (Close Citation1993, Citation1996; Close & Rosen Citation2001; Close & Flintoft Citation2004), most pesticides were detected at concentrations below 1% of the MAV. Dieldrin was detected in one well at a concentration of 0.13 mg/m3 (the MAV for dieldrin equals 0.04 mg/m3 and is actually for the sum of aldrin and dieldrin; Ministry of Health Citation2008), with a repeat sample 7 months later containing a concentration of 0.14 mg/m3. This well was not used for drinking purposes but was used for irrigation and stock water purposes. An old sheep dip site was identified at the site located approximately 20–30 m up-gradient from the well. Dieldrin was used in New Zealand primarily for the government-required control of ectoparasities on sheep in the 1960s. Most livestock farms in New Zealand would probably have had a sheep or cattle dip site. Even though dieldrin has not been used for this purpose since the mid-1960s, its long persistence means that it can be detected in the soil where the dip site wastewater was disposed of and occasionally in the underlying groundwater. Hadfield & Smith (Citation1999) carried out an investigation into dieldrin in groundwater in the Waikato region. Their results indicated that dieldrin contamination could be widespread and that concentrations in shallow groundwater (about 5 m below ground level) could be expected to increase, even though usage had ceased 30–40 years previously. Dieldrin has a low MAV (0.04 mg/m3), which means that even low concentrations in groundwater can easily exceed the MAV for drinking water.
Terbuthylazine was the most commonly detected pesticide, being found in 17 wells at levels ranging from 0.014–5.8 mg/m3 (). The maximum value detected of 5.8 mg/m3 is 73% of the MAV. This well had been sampled previously in 2002 with no pesticides detected. All other detections of terbuthylazine in the 2010 survey were <0.7 mg/m3.
Some of the wells with high levels of pesticides detected in the 2010 survey also had these pesticides detected in the 2006 survey. For example, well 372034 (Horizons) had high (12 mg/m3) levels of alachlor, plus trace levels of metalaxyl, metribuzin and pendimethalin detected in 2010 (). In 2006 this well had even higher levels of alachlor (34 and 18 mg/m3), plus trace levels of metalaxyl and metribuzin. This indicates either continuing usage of these pesticides in the area or persistence of these pesticides for a long time in the subsurface environment. The half-life for alachlor in topsoil is around 1 month so this would imply a much longer persistence below the root zone and in the groundwater, consistent with the lower levels of organic carbon and microbial activity in these layers. Pesticide persistence in the groundwater can be much longer than in the active root zone (orders of magnitude) as shown by Pang & Close (Citation1999) and Levy & Chesters (Citation1995). Alachlor is a herbicide used to control annual grasses and broad leaf weeds in crops (The Pesticide Manual Citation1994). The well is used as a source of stock water and the surrounding land-use comprises market gardening and grazing pasture. Three wells in the Auckland region had low levels of bentazone in both the 2006 and 2010 surveys and three wells from Southland had low levels of terbuthylazine, simazine and propazine in both surveys.
Of the 22 pesticides and metabolites detected, GUS values indicated that 15 were leachers, three were transitional and four were non-leachers (). This was a reasonably high proportion of non-leacher pesticides. Detection of non-leacher pesticides can be an indication that normal leaching processes are not responsible for their presence in the groundwater and that other pathways, such as spillages or preferential flow, are taking place. As discussed earlier for dieldrin, it may also be the result of widespread use of long-lived pesticides in previous decades that has resulted in high soil concentrations.
The small number of wells showing significant trends in total pesticide concentrations with time is explained, at least partially, by the small number of surveys even though these have taken place over a 20-year period. There are more wells showing decreases in total pesticide concentrations compared with those showing increases, but the majority of wells show no change in total pesticides concentrations with time. About half of all wells that had been sampled in at least four surveys had no pesticides detected in any survey. There was a slight decrease in the number of wells with pesticide detections with time for the last five surveys (adjusted for common detection limits; ). Our data indicate that there has been a slight decrease in pesticide concentrations in groundwater over the past 16 years for the assayed pesticides. This decrease in groundwater pesticide concentrations is consistent with reported reductions in pesticide use on horticultural crops following the introduction of integrated pest management schemes and changes to weed management techniques in orchards (Manktelow et al. Citation2005). However, if the un-adjusted detection limits are used for the last four surveys, then there has been a large decrease in detections from high rate (35%) observed in the 1998 survey to the recent surveys, with the last three surveys having similar (19–24%) rates of detection.
A range of groundwater factors including well design and depth, land-use, aquifer type, and climate can make groundwater more or less vulnerable to pesticide contamination (Close et al. Citation2001; Worrall & Kolpin Citation2004). For example, as most pesticide contamination results from leaching from the land surface, shallow groundwaters will usually be more contaminated and wells at greater depth will tend to be much less contaminated. There were significantly more pesticide detections in the unconfined aquifers than would be expected on the basis of an even distribution of wells in these aquifers, compared with the semi-confined and confined aquifers. Unconfined aquifers by definition are more likely to be become contaminated by pesticides because they do not have an overlying impermeable layer that prevents contamination from infiltrating into the aquifer, which results in less protection.
Pesticides and nitrate are often used concurrently on horticultural and agricultural land. Nitrate has been proposed as a possible indicator of likely pesticide contamination (Close et al. Citation2001). Mean nitrate concentrations were significantly higher for wells with pesticide detections than for wells without pesticide detections. This pattern was also observed in the 1998 and 2002 surveys (Close & Rosen Citation2001; Close & Flintoft Citation2004). A significant correlation existed between nitrate and total pesticide concentration for wells where pesticides were found (; r 2=0.50). Similarly high concentrations of nitrate were also observed in wells that tested positive for pesticides in a Waikato region survey (Hadfield & Smith Citation1999).
A limitation of these types of pesticide surveys are the budgetary constraints and the availability of analytical tests. It is not feasible to analyse for every pesticide that is registered for use in New Zealand and likely to enter groundwater. For example, the non-selective herbicide glyphosate is widely used in New Zealand (Manktelow et al. 2004), but was not included in the current survey. In addition, our study included only a limited number of pesticide transformation products. Transformation products have been detected in groundwater at higher concentrations and more frequently than the parent compounds in some studies (Hernandez et al. Citation2008; Steele et al. Citation2008).
Acknowledgements
This study was funded by the regional and unitary authorities that manage groundwater resources in New Zealand. We are grateful for the comments by Liping Pang and Chris Nokes (ESR) and two anonymous journal reviewers who improved the quality of this paper.
References
- APHA 2005 . Standard methods for the examination of water and wastewater. , 21st edition . Washington , DC , American Public Health Association, American Water Works Association and Water Pollution Control Federation .
- Australian Government 2004 . Australian Drinking Water Guidelines . Australia : National Health and Medical Research Council and Natural Resource Management Ministerial Council .
- Boul , HL . 1995 . DDT residues in the environment—a review with a New Zealand perspective . New Zealand Journal of Agricultural Research , 38 : 257 – 277 . doi: 10.1080/00288233.1995.9513126
- Canterbury Regional Council 1997 . Canterbury Regional Council Regional Environment Report (1995/96). CRC Report 97(10) . Christchurch , Canterbury Regional Council .
- Ciba-Geigy Corporation 1993 . Summary of toxilogical data on atrazine and its chlorotriazine metabolites . Attachment 12, 56 FR 3526 .
- Close , ME . 1993 . Assessment of pesticide contamination of groundwater in New Zealand. II. Results of groundwater sampling . New Zealand Journal of Marine and Freshwater Research , 27 : 267 – 273 . doi: 10.1080/00288330.1993.9516566
- Close , ME . 1996 . Survey of pesticides in New Zealand groundwaters 1994 . New Zealand Journal of Marine and Freshwater Research , 30 : 455 – 461 . doi: 10.1080/00288330.1996.9516734
- Close , M and Flintoft , M . 2004 . National Survey of Pesticides in Groundwater 2002 . New Zealand Journal of Marine and Freshwater Research , 38 : 289 – 299 . doi: 10.1080/00288330.2004.9517238
- Close , ME and Rosen , MR . 2001 . 1998/99 National survey of pesticides in groundwater using GCMS and ELISA . New Zealand Journal of Marine and Freshwater Research , 35 : 205 – 219 . doi: 10.1080/00288330.2001.9516992
- Close , ME , Rosen , MR and Smith , VR . 2001 . “ Fate and transport of nitrates and pesticides in New Zealand's aquifers ” . In Groundwaters of New Zealand , Edited by: Rosen , MR and White , PA . 185 – 220 . Wellington : New Zealand Hydrological Society Inc .
- Close , ME , Lee , R , Sarmah , AK , Pang , L , Dann , RL , Magesan , GN , Watt , JPC and Vincent , KW . 2008 . Pesticide sorption and degradation characteristics in New Zealand soils—a synthesis from seven field trials . New Zealand Journal of Crop & Horticultural Science , 36 : 9 – 30 . doi: 10.1080/01140670809510217
- Davies , H . 2001 . “ Groundwater and Health ” . In Groundwaters of New Zealand , Edited by: Rosen , MR and White , PA . 253 – 272 . Wellington : New Zealand Hydrological Society Inc .
- Gaw , S , Close , ME and Flintoft , MJ . 2008 . Fifth national survey of pesticides in groundwater in New Zealand . New Zealand Journal of Marine and Freshwater Research , 42 : 397 – 407 . doi: 10.1080/00288330809509968
- Gustafson , DI . 1989 . Groundwater ubiquity score: a simple method for assessing pesticide leachability . Environmental Toxicology and Chemistry , 8 : 339 – 357 . doi: 10.1002/etc.5620080411
- FAO 1996 . Disaggregated data sets for pesticides. FAO Pesticides Database . Rome , UN Food and Agricultural Organisation .
- Hadfield J , Smith D 1999 . Pesticide contamination of groundwater in the Waikato region . Environment Waikato Technical Report 1999/9 .
- Haarstad , K and Ludvigsen , GH . 2007 . Ten years of pesticide monitoring in Norwegian Groundwater . Ground Water Monitoring and Remediation , 27 : 75 – 89 . doi: 10.1111/j.1745-6592.2007.00153.x
- Hernandez , F , Marin , JM , Pozo , OJ , Sancho , JV , Lopez , FJ and Morell , I . 2008 . Pesticide residues and transformation products in groundwater from a Spanish agricultural region on the Mediterranean Coast . International Journal of Environmental Analytical Chemistry , 88 : 409 – 424 . doi: 10.1080/03067310701724772
- Kolpin , DW , Barbash , JE and Gilliom , RJ . 1998 . Occurrence of pesticides in shallow ground water of the United States: initial results from the National Water-Quality assessment programme . Environmental Science and Technology , 32 : 558 – 566 . doi: 10.1021/es970412g
- Levy , J and Chesters , G . 1995 . Simulation of atrazine and metabolite transport and fate in a sandy-till aquifer . Journal of Contaminant Hydrology , 20 : 67 – 88 . doi: 10.1016/0169-7722(95)00061-Y
- Manktelow D , Stevens P , Walker J , Gurnsey S , Park N , Zabkiewicz J , Teulon D , Rahman A 2005 . Trends in pesticide use in New Zealand: 2004 . Report to the Ministry for the Environment, Project SMF4193. HortResearch Client Report no. 17962 .
- Ministry for the Environment 2006 . Snapshot of water allocation in New Zealand . Ministry for the Environment Report ME 782 .
- Ministry of Health 2008 . Drinking-Water Standards for New Zealand 2005 (Revised 2008) . Wellington , Ministry of Health .
- NATA 2012 . Guidelines for the validation and verification of quantitative and qualitative test methods . NATA Technical Note 17. August 2004, amended & reissued March 2012. Silverwater, NSW, National Association of Testing Authorities .
- Pang , L and Close , ME . 1999 . Attenuation and transport of atrazine and picloram in an alluvial gravel aquifer: a tracer test and batch study . New Zealand Journal of Marine and Freshwater Research , 33 : 279 – 291 . doi: 10.1080/00288330.1999.9516877
- The Pesticide Manual 1994 . The pesticide manual, incorporating the agrochemicals handbook. , 10th edition . Tomlin C . The British Crop Protection Council and The Royal Society of Chemistry .
- Primi , P , Surgan , MH and Urban , T . 1994 . Leaching potential of turf care pesticides: A case study of Long Island golf courses . Ground Water Monitoring and Remediation Summer , 1994 : 129 – 138 . doi: 10.1111/j.1745-6592.1994.tb00474.x
- Rothery , P . 2000 . “ Basic Concepts ” . In Statistics in ecotoxicology , Edited by: Sparks , T . 1 – 32 . Chichester : John Wiley & Sons Ltd .
- Steele , GV , Johnson , HM , Sandstrom , MW , Capel , PD and Barbash , JE . 2008 . Occurrence and fate of pesticides in four contrasting agricultural settings in the United States . Journal of Environmental Quality , 37 : 1116 – 1132 . doi: 10.2134/jeq2007.0166
- Taranaki Regional Council 1995 . Report on an investigation of pesticides in shallow groundwater in Taranaki . Technical Report 95–14 .
- USEPA 1989 . Semivolatile organic compounds by gas chromatography/mass spectrometry (GC/MS) . EPA 600/S 4-89/010. United States Environmental Protection Agency .
- US NPIC 2011 . United States National Pesticide Information Center pesticide properties databases . http://npic.orst.edu (accessed in April 2011) .
- Weber , JB . 1994 . “ Properties and behaviour of pesticides in soil ” . In Mechanisms of pesticide movement into ground water , Edited by: Honeycutt , RC and Schabacker , DJ . 15 – 41 . Lewis Publishers .
- Worrall , F and Kolpin , DW . 2004 . Aquifer vulnerability to pesticide pollution—combining soil, land-use and aquifer properties with molecular descriptors . Journal of Hydrology , 293 : 191 – 204 . doi: 10.1016/j.jhydrol.2004.01.013