Field and laboratory study of pesticide leaching in a Motupiko silt loam (Nelson) and in a Waikiwi silt loam (Southland).
The majority of groundwaters in New Zealand appear free from pesticide contamination (Ministry for the Environment 1997). However, recent New Zealand-wide targeted surveys in 1998 and 2004 have detected pesticides in 35 and 21% of the wells sampled, respectively, albeit mostly at concentrations below 0.1 mg/[m.sup.3] (Close and Rosen 2001; Close and Flintoft 2004). Other localised surveys by Regional Authorities (e.g. Smith 1993; Hadfield and Smith 2000) have also found a number of pesticides in vulnerable groundwater systems, i.e. shallow groundwaters underlying free-draining soils with high pesticide use. These surveys have indicated that there is a need for better information on the processes involved in pesticide fate and transport, particularly under New Zealand conditions. One possible management tool for the prevention of pesticide contamination of groundwaters is the use of pesticide leaching models to assess the possibility of leaching in vulnerable areas and to limit the application of certain pesticides in those areas if necessary. It is important that such decisions are based on good scientific data.
This study is part of a nationwide approach to assess pesticide behaviour in key New Zealand soils under different climatic conditions, and the suitability of simulation models to predict transport to depth, where there could be impact on groundwater quality. Research has been carried out previously at 5 field sites in the North Island, and the 2 South Island field sites reported here complete this research program. The 5 North Island sites comprise 2 trials in the relatively dry Hawkes Bay region on sandy/gravelly alluvial soils, derived predominantly from greywacke (Close et al. 1999), an allophanic soil near Hamilton in the Waikato, derived from alluvial volcanic material (Close et al. 2003), a bouldery clay loam with some allophane content, derived from basalt, in Northland, and a coastal sandy soil in Manawatu (Close et al. 2005). Relevant background to the rationale for the studies is given in (Close et al. 2003).
The Nelson and Southland trials reported here add a northern and southern South Island aspect to the studies reported previously. Salient features are given in Table 1 and compared with those from the previous studies. The Nelson soil, a Motupiko silt loam, is a weathered fluvial recent soil derived from greywacke, whereas the Southland soil, a Waikiwi silt loam, is a deeper soil derived from loess from tuffaceous greywacke and schist. The Nelson site is much warmer and drier than the Southland site, which is the coldest of all the field sites in the program (Table 1).
The objectives of this study were to (i) monitor the movement of selected pesticides and tracers through 2 contrasting free-draining soils in different climatic conditions, (ii) determine their mobility and persistence characteristics, and (iii) evaluate the performance of several leaching models of different complexity. This paper describes the study and presents results of tracer and pesticide monitoring with depth in the soil profiles. A companion paper (Sarmah et al. 2006) describes the simulations of these results using 4 pesticide leaching models and compares the estimates of mobility and degradation with the available laboratory and literature values.
Materials and methods
Soils and sites
The site was located at Wakefield about 25 km south-west of Nelson on a Motupiko silt loam (Fig. 1). Following the New Zealand Soil Classification (Hewitt 1992) the soil is classified as a Weathered Fluvial Recent Soil. It is a shallow recent soil with sandy alluvial gravels occurring from about 0.40 m depth. Limited soil profile details, together with specific soil chemical and physical properties for measured horizons, are given in Table 2. Land use at the site prior to the study was grazing for dairy cows. There had been no application of any of the chemicals used in the study in the previous 5 years. The site had no vegetation during the period of study because of the combined effect of the applied herbicides, and there was some sparse weed re-growth by the end of the study. Rainfall and temperature gauges were installed at the site from November 2002, when the pesticides were applied. Prior to this time, rainfall and temperature data were obtained from the Appleby climate station (G13305), about 11 km to the north-east. Radiation data were obtained from the Appleby climate station for the whole study period. The rainfall normal values were taken from a closed rainfall station (G13211) located at Appleby about 13 km from the study site.
[FIGURE 1 OMITTED]
The site was located on the AgResearch Woodlands research station about 20km west of Invercargill (Fig. 1), on a Waikiwi silt loam. Following the New Zealand Soil Classification (Hewitt 1992) the soil is classified as a Typic Firm Brown Soil. Limited soil profile details, together with specific soil chemical and physical properties for measured horizons, are given in Table 2. Land use at the site prior to the study was grazing for sheep. As with the Nelson site, there had been no application of any of the chemicals applied in the study in the previous 5 years, and the site was clear of vegetation throughout the study, with only sparse weed re-growth at the end of the study period. Rainfall data were obtained from the nearby Environment Southland rainfall station at Woodlands about 2 km from the site. Radiation and temperature data were obtained from the Invercargill climate station (168435), approximately 22 km away. The rainfall normal values were taken from a closed rainfall station (I68433) located at Invercargill Airport about I km from site 168435.
Both trials were set up in a similar way to the previous studies (Close et al. 2003). The Southland site covered an area 15 m by 15 m, similar to previous studies, but the Nelson site was slightly larger (18 m by 18 m) to enable soil sampling of the stony profile using soil pits. There were pairs of suction cups at 4 different depths for sampling of soil water at each site. The suction cups were located randomly throughout the plot (excluding the outer I m of the plot), and were covered with plastic water meter boxes. The suction cups sloped downwards at angles ranging from 15 [degrees] to 52 [degrees] so that the selected sampling depths were achieved and that water would collect in the bottom of the cups on application of suction. The angle of the suction cups also ensured that the collected leachate had percolated through an undisturbed soil profile. The gap between the soil and the suction cups was sealed with bentonite to prevent leakage of water down the sides of the suction cups. Duplicate suction cup samplers were installed at depths of 0.20, 0.30, 0.62, and 0.92 m at the Nelson site. The lower 2 sets of suction cups penetrated the gravel profile and were installed vertically using a steel spike and then sealed around the top with bentonite to prevent water movement down the tube. There were only 3 depths for suction cups at the Southland site, with 4 cups at 0.40 m, and 2 cups each at 0.70 and 1.30 m. The pesticide application at the Southland site took place in very wet conditions and the bentonite had not had sufficient time to seal around the suction cups. It was obvious at the time of application that 3 of the suction cups had been affected by some of the applied pesticides moving down the sampling tube. These cups were removed, leached with HCI and water to remove any sorbed pesticide and re-installed in different locations 3 weeks later. After the first pesticide analytical results were received it was apparent that the other suction cups had also been affected and so the remainder of the suction cups were removed, leached with HCI and water, and re-installed in different locations in April 2003. An additional sample was collected from the suction cups a week after the re-installation.
Atrazine, bromacil, diazinon, hexazinone, and terbuthylazine were applied to both sites in November 2002, with procymidone also being applied to the Nelson site, and trifluralin being applied to the Southland site. These pesticides were selected because they had been detected in groundwaters in New Zealand, there was a reasonable range in leaching properties (required for model evaluation), and all the pesticides applied to each site could be analysed in a single method. Bromide was applied at both sites as a tracer. A summary of properties and application rates is given in Table 3. The pesticide application rates were designed to be 10 kg/ha, which were approximately 1-8 times normal label application rates depending on the crop and soil type, to ensure measurable quantities leached to depth for modelling purposes. However, some of the actual application rates were higher than this, particularly for hexazinone, and the rate for atrazine was much lower than anticipated. The pesticides and bromide tracer were applied using a pressurised hand sprayer at a rate of around 12 mm/h to minimise any preferential flow which might result from the application. No surface ponding was observed during application at the Nelson site but there was heavy rainfall occurring at the Southland site at the time of application and the soil profile became saturated. Pesticide application was stopped at that stage and the remainder of the application was carried out the next day. There was a significant possibility of preferential flow at the Southland site at the time of pesticide application.
Sampling and analysis
The suction cups were sampled by applying a vacuum of 60 kPa for 10-20h at the Southland site and 35-40h at the much drier Nelson site. Soil samples (composites of 4-5 cores) were taken to a maximum depth of 1 m at the Southland site. At the Nelson site, 2 soil pits were dug on each sampling occasion and samples taken from opposite sides of each pit. Both suction cup and soil samples were taken at monthly intervals for the first 3 months after application and thereafter at 3-monthly intervals. The last sampling period was in February 2004, approximately 15 months after the pesticide application. An additional round of suction cup samples was taken in April 2003 at the Southland site after 5 of the suction cups had been re-installed. One of the suction cups at 0.92 m at the Nelson site only gave 1 sample during the whole study, and one cup at 0.3 m was found to have a large natural cavity in the soil next to it when the cup was excavated at the end of the study, so the results from these suction cups were omitted from the study. Soil moisture was measured gravimetrically on the soil samples. During site closure, all of the suction cups were excavated and the depth for each suction cup was measured accurately.
The water samples were analysed for pesticides using gas chromatography with mass spectrometry (GCMS) with a detection limit of 0.1 mg/[m.sup.3] and bromide was analysed by ion selective electrode (detection limit was 0.1 g/[m.sup.3]). Two metabolites, desethyl atrazine and desisopropyl atrazine, were also analysed in the soil water samples using GCMS. The samples from the duplicate suction cups were combined prior to analysis. The soil samples were analysed for the suite of pesticides by extracting the soil with ethyl acetate and anhydrous sodium sulfate. The extracts were cleaned using gel permeation chromatography and analysed using high resolution gas chromatography. The detection limits were 0.01 mg/kg for the soil samples. The soil samples from both sites were analysed for bromide following extraction of 20 g soil combined with 10 g acid-washed sand with distilled water. The bromide was analysed using an ion selective electrode and the detection limit for bromide in soil was approximately 0.5 mg/kg. For the Nelson site in the gravels below 0.40m, the analyses were carried out on the fine fraction, and after adjustment for percentage stones, the detection limit for bromide was approximately 0.2 mg/kg.
When the soil samples were collected, the roots and vegetation were removed from the sample before analysis. This mainly applied to the 0-0).10m sample. The mass recoveries for bromide indicated that a significant amount was sorbed to the root material in the first month for the Nelson site and for the first 2 months for the Southland site. The distribution of bromide through the profile indicated that little, if any, leaching had occurred below the sampling depth. Hexazinone was also very conservative for the first 2 months, so a factor was applied to adjust the concentrations of bromide and hexazinone in the 0-0.10 m sample to give approximately 100% mass recovery for the profile. This factor was applied uniformly to all pesticide concentrations in the 0-0.10m interval, although it is acknowledged that the sorption will be non-aniform for different pesticides. The factor was 1.5 for the Nelson site and 2.5 and 2.2 for the first and second months, respectively, for the Southland site. Lauren et al. (1990) analysed the sorption of 4 insecticides ([K.sub.oc] ranging from 46 to 1355 mL/g) to both the soil and the turf in the top 0.10m of soil, and found that that > 50% of the insecticide was associated with the turf rather than the soil. In the first month after application this was 60-80% for the moderately sorbed insecticides ([K.sub.oc] 46-300mL/g). Thus, the correction applied in this study is reasonable and is possibly an under-estimate.
[FIGURE 2 OMITTED]
The Nelson profile was very stony below 0.37m (Table 2) and was assessed as having 60% stones (M. McLeod, Landcare Research, pets. comm.). As the soil analysis was carried out on the fine fraction, the pesticide and bromide concentrations, organic carbon (OC) content, and bulk density were adjusted for the % stones before data analysis. This adjustment also affects the detection limits for these layers.
Laboratory batch sorption experiments
Batch sorption isotherms were performed for each of the pesticides in the topsoil and subsoil at each site. The topsoil sample was taken from 0 to 0.10m in the A horizon for both sites and the subsoil sample was taken from 0.40 to 0.50m (BC1 horizon) at the Nelson site and from 0.55 to 0.70m in the Cl horizon at the Southland site. The batch experiment was carried out on the fine fraction for the BC1 horizon at the Nelson site. The concentrations used for the batch sorption tests were 10, 50, 100, 500, 1000, 5000, and 10 000 p.g/L. The soil : solution ratio was 25 g soil in 50 mL solution, which was untreated Christchurch groundwater (electrical conductivity 16 mS/m). Samples were shaken for 1 h, then left overnight (approximately 20 h) for further equilibration and settling before being centrifuged. A glass syringe was used to sample the clear supernatant, which was analysed using GCMS as described previously. Blanks were run at 50, 500, and 5000[micro]g/L to correct for any sorption onto the containers and were used as the standard concentrations. Soil blanks were analysed for each soil and subsoil and were below detection for all pesticides. One sample was analysed in duplicate for each soil layer and indicated good reproducibility. The sorption coefficient values ([K.sub.d]) were converted into organic carbon sorption coefficients ([K.sub.oc]) by dividing by the organic carbon content. The data were fitted to both Freundlich and linear isotherms. The Freundlich isotherm is given below and becomes a linear isotherm when n=1:
S = [KC.sup.n]
where S is the amount sorbed ([micro]g/g), C is the equilibrium concentration in solution ([micro]g/mL), K is the sorption coefficient (mL/g), and n is the exponent.
[FIGURE 3 OMITTED]
Results and discussion
The total rainfall at the Nelson site between November 2002 and February 2004 was 1254mm, which was similar to the expected or normal rainfall over this period (1267mm). There was much more rainfall than normal in June 2003 (2.4 times the normal amount) and there were 2-3 months with much lower than normal rainfall, which balanced out over the whole study period (Fig. 2a). The total rainfall at the Southland site between November 2002 and February 2004 was 1307mm, which was about 90% of the expected or normal rainfall over this period (1447 mm). Heavy rainfall occurred at the Southland site at the time of application and the soil profile became saturated. This meant that there was a significant possibility of preferential flow at the Southland site at the time of pesticide application, and this can be observed in some of the leaching patterns.
Pesticides and bromide tracer
The soil water concentrations of bromide, bromacil, and hexazinone at the Nelson site (Fig. 3) indicate fairly rapid movement of the bromide tracer down the profile. Bromacil and hexazinone were retarded with respect to bromide, and all compounds showed a reduction in peak concentrations as they moved down the soil profile. The first 2 observations at 0.3 m depth are higher and are an exception to the general leaching pattern. This may reflect spatial variability of the soil, or micro-topography resulting in the application amount being higher around that suction cup. The soil water concentrations of bromide, bromacil, hexazinone, and terbuthylazine at the Southland site (Fig. 4) show movement of these compounds down the profile but the rising limb of the concentration breakthrough curves (BTC) were often at a similar time. This indicates that preferential flow was probably taking place at this site, which is consistent with the rainfall pattern around the time of application. There were no data for the suction cups at 1.3 m before April 2003 (154 days after application) as both suction cups were replaced at that time. However, Fig. 4 indicates that the peak and BTC for both the tracer and pesticides were still observed at that depth and arrived after 180 days.
[FIGURE 4 OMITTED]
The soil concentrations for bromide, hexazinone, and procymidone with time at the Nelson site (Fig. 5) show that traces of the bromide tracer had moved down the profile within the first month but most of the tracer did not leach from the topsoil until after 3 months. The mass recoveries (Table 4) indicate that 36% of the tracer was still in the profile at 6 months but this was reduced to 3% at 9 months, consistent with the rainfall inputs in June 2003 (Fig. 2). It should be noted that a log scale is used for the soil concentration figures so that both bromide, which was applied at higher concentrations, and the pesticide data could be displayed. The log scale tends to accentuate the lower concentrations. The different detection limits for bromide and the pesticides should also be noted.
Hexazinone showed a similar leaching pattern to bromide, but the procymidone remained in the topsoil and the concentrations were quite variable. Table 4 indicates that mass recoveries were 23-34% for the first 6 months, and then increased to around 60% at the 9- and 12-month sample round. It is not clear what caused this increase. While some of the variation may be due to analytical error and spatial variability of the soil concentrations, it is likely that there was also some sorption onto dead plant material. This material was not extracted and analysed, and desorption of the pesticide as the plant material decomposed could result in these increases in concentration in the top 0.10m.
The soil concentrations for bromacil, diazinon, and terbuthylazine with time at the Nelson site (Fig. 6) show that diazinon was only detected at low concentrations and only persisted in the profile for the first 3 months. Terbuthylazine and bromacil had similar leaching patterns, with bromacil being more mobile and persistent. The mass recoveries (Table 4) indicated that bromacil decreased to around 70% after 2 months and then stayed between 59 and 73% until after the 9-month sample. The bromide soil concentrations were fairly constant below 0.40 m on most sampling occasions (Fig. 5), and bromide was observed in both the 0.62 m and 0.92 m suction cups at the same time, indicating that leaching was rapid once the water and tracers moved below 0.40 m in the gravel media.
The soil concentrations for bromide, hexazinone, and trifluralin with time at the Southland site (Fig. 7) show that some of the bromide tracer had moved down the profile within the first month with nearly half (43%) of the tracer having leached from the topsoil by 3 months. The mass recoveries (Table 4) indicate that 38% of the tracer was in the profile at 6 months, which reduced to 2% at 12 months. Hexazinone showed slightly retarded movement compared to bromide, particularly in the topsoil as would be expected. There was a large increase in degradation of hexazinone after 3 months, as indicated by its disappearance compared to bromide. Trifluralin decreased to around 49% after 2 months and then stayed around this level, with most being located within the top 0.10 m of the profile (Fig. 7). This suggests that most of the losses of trifluralin were due to volatilisation, and that leaching and degradation of trifluralin were low. Bromacil and terbuthylazine had fairly similar leaching patterns (Table 4, Fig. 8), with a rapid drop to around 40-70% mass recovery, followed leaching through the profile over the next 12-15 months. This indicates that there were probably some initial volatilisation losses for bromacil and terbuthylazine, followed by relatively slow degradation at the Southland site. The rapid disappearance of diazinon was probably a combination of volatilisation and degradation at both sites. Evidence of preferential flow at the Southland site can be seen for trifluralin (Fig. 7), where most of the pesticide remains in the top 0.10-0.20 m, with occasional detections below 0.60 m on the second and third sampling rounds. A similar pattern is seen in Fig. 8 for terbuthylazine and, to a lesser extent, bromacil.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
One of the 5 soil cores sampled as part of the 9-month sampling round (August 2003; 273 days after application) was taken near the area where the pesticide solution had been mixed and dispensed. There had been some spillage and concentrations of the less mobile, more persistent pesticides, such as terbuthylazine and trifluralin, were elevated in the top 0.30 m of the profile for that sampling occasion compared to the previous (6 month) sampling round.
As the sampling for the soil cores and the suction cups took place at the same time, and each soil core was analysed for each 0.10m interval collected, there were more soil sample results available for analysis compared to suction cup data. This was accentuated by the absence of some suction cup data from the Southland soil at the start of the study because of the saturated profile at the start of the study and the contamination of the suction cups, as discussed earlier. This ratio of soil sampling to suction cup data is in contrast to the first 3 studies where the majority of the data collected was from suction cups. Both the soil sampling and suction cup data was used to derive field-based leaching parameters by inverse modelling using 3 different pesticide leaching models and is reported in the companion paper by (Sarmah et al. 2006).
Laboratory batch sorption experiments
The parameters for both the Freundlich and linear isotherms are given in Tables 5 and 6 for the Motupiko and Waikiwi silt loams, respectively. The sorption data with both isotherms for the topsoil samples plotted are given in Fig. 9 for atrazine, bromacil, diazinon, and terbuthylazine. There were generally good fits, as indicated by the [R.sup.2] values, for both the linear and Freundlich isotherms. The [K.sub.oc] values obtained for the Motupiko silt loam were generally higher than the 'best available' values from the literature (Table 3), with the exception being diazinon, which was lower than the 'best available' values. The subsoil [K.sub.oc] values were very high and indicate that sorption on mineral surfaces was significant. Green and Karickhoff (1990) state that there is little error with respect to OC content in [K.sub.oc] estimates for OC levels > 1%, but that the error can range from 2 to 10 times for OC levels <0.2%, as other sorption surfaces are important at these low OC levels. They indicate that the [K.sub.oc] concept should be valid if the ratio of clay mineral fraction to OC content is <40. The Freundlich exponent, n, was quite variable, with most values less than or close to 1, except for hexazinone (1.34) and terbuthylazine (1.24).
The [K.sub.oc] values for the Waikiwi silt loam topsoil (Table 6) were similar to the 'best available' values from the literature and within the range of values reported in the literature (Table 3). The exception, as for the Motupiko silt loam, was diazinon, which was about 15 times lower than the 'best available' value. The isotherm is non-linear and the [K.sub.oc] value is significantly influenced by the 2 highest concentrations. If these were deleted then the [K.sub.oc] derived for a linear isotherm would be 310, which is still about 5 times lower than the 'best available' value. This means that diazinon will be much more mobile than expected in these soils. This is consistent with the detection of diazinon at very low levels (0.014).03 mg/[m.sup.3]) in 2 wells in both the 1998 and 2002 national surveys of pesticides in groundwater in New Zealand (Close and Rosen 2001; Close and Flintoft 2004). [K.sub.oc] values for diazinon measured previously at the Northland and Manawatu sites were 107 and 460 mL/g (Close et al. 2005), and were also much lower than literature values (Table 3). The Freundlich exponent, n, was fairly consistent for the Waikiwi topsoil (Table 6), with most values close to 0.8. There was relatively little sorption for the Waikiwi subsoil but the higher [K.sub.oc] values than for the topsoil indicate that there was some sorption to mineral at the lower organic carbon levels (Table 6). Figure 9 indicates that there is good agreement between the linear and Freundlich isotherms for most pesticides at low concentrations, as would be expected. However, there is some divergence between the isotherms at higher concentrations. The [K.sub.oc] values from the batch isotherms were consistent with the movement observed at the Southland (Waikiwi silt loam) field site, with hexazinone and bromacil having the lowest [K.sub.oc] values and being transported most rapidly down the profile.
[FIGURE 9 OMITTED]
The bromide tracer had moved through the profile after 9-12 months at both sites. Hexazinone was the most mobile of the pesticides, followed by bromacil at both sites. Diazinon and atrazine dissipated rapidly at both sites, while the trifluralin and, to a lesser extent, terbuthylazine and bromacil data indicated some volatilisation losses. The diazinon removal was probably a combination of degradation and volatilisation. There was evidence of some preferential flow at the Southland site, as rainfall was very heavy around the time of pesticide application. The [K.sub.oc] values from the sorption isotherms were generally similar to literature values for the Waikiwi soil but were much higher for the Motupiko soil. The exception was diazinon, where the [K.sub.oc] values were much lower than the literature values for both sites. The [K.sub.oc] values from the batch isotherms were consistent with the movement observed at the Southland site (Waikiwi silt loam), with hexazinone and bromacil having the lowest [K.sub.oc] values and being transported the most rapidly in the field. The results from the Nelson indicate that there is a low likelihood of terbuthylazine, diazinon, or procymidone reaching shallow groundwater when applied in normal agricultural practice. However, where heavy rainfall occurs around the time of pesticide application, as occurred for the Southland site, there is a higher likelihood of pesticides moving through the soil profile and into the shallow groundwater.
The authors thank Evan Baigent, Wakefield, Nelson, and Kevin Knowler, Farm Manager, AgResearch, Woodlands, Southland, for allowing the use of their land for the research trials. We thank Danny Thornburrow and Janine Ryburn (Landcare Research), Gordon Curnow and Tom Kennedy (Tasman District Council), and Jim Risk (Environment Southland) for assistance with the field work. The research was funded by contracts CO3X0303 (ESR) and CO9X0017 (Landcare Research) from the Foundation for Science, Research and Technology (New Zealand).
Manuscript received 12 October 2005, accepted 31 May 2006
Close ME, Flintoft MJ (2004) National Survey of Pesticides in Groundwater in New Zealand--2002. New Zealand Journal of Marine and Freshwater Research 38, 289-299.
Close ME, Lee R, Magesan GN, Stewart M, Skuse G, Bekesi G (2005) Field study of pesticide leaching in a Himatangi sand (Manawatu) and in a Kiripaka bouldery clay loam (Northland). 1. Results. Australian Journal of Soil Research 43, 457-469.
Close ME, Magesan GN, Lee R, Stewart MK, Hadfield JC (2003) Field study of pesticide leaching in an allophanic soil in New Zealand. 1: Experimental results. Australian Journal of Soil Research 41, 809-824. doi: 10.1071/SR02080
Close ME, 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.
Close ME, Watt JPC, Vincent KW (1999) Simulation of picloram, atrazine, and simazine transport through two New Zealand soils using LEACHM. Australian Journal of Soil Research 37, 53-74. doi: 10.1071/S97080
Green RE, Karickhoff SW (1990) Sorption estimates for modeling. In 'Pesticides in the soil environment'. Soil Science Society of America Book Series No. 2. pp. 79-101. (Soil Science Society of America: Madison, WI)
Hadfield J, Smith D (2000) Pesticide contamination of groundwater supplies in the Waikato region. In 'Water 2000'. Auckland, New Zealand. p. 10. (New Zealand Water & Wastes Association)
Hewitt AE (1992) New Zealand soil classification. DSIR Land Resources, No. 19.
Lauren DR, Henzell RF, Wrenn NR (1990) Control of grass grub (Costelytra zealandica) adults with soil insecticides. New Zealand Journal of Agricultural Research 33, 165-171.
Ministry for the Environment (1997) 'The State of New Zealand's Environment 1997.' (Ministry for the Environment, NZ)
New Zealand Meteorological Service (1983) 'Summaries of climatological observations to 1980.' NZ Meteorological Service Miscellaneous Publication 177. p. 172. (NZ Meteorological Service)
Sarmah AK, Close ME, Dann R, Pang L, Green SR (2006) Parameter estimation through inverse modelling and comparison of four leaching models using data from two contrasting pesticide field trials in New Zealand. Australian Journal of Soil Research 44, 581-597.
Smith VR (1993) Groundwater contamination by triazine pesticides, Levels Plain, South Canterbury. Canterbury Regional Council Report 93(36).
The Pesticide Manual (1994) 'The pesticide manual.' (Ed. C Tomlin) p. 1341. (The British Crop Protection Council and The Royal Society of Chemistry: Cambridge, UK)
M. E. Close (A,E) A. K. Sarmah (B), M. J. Flintoft (A) J. Thomas (C), and B. Hughes (D)
(A) Institute of Environmental Science and Research, PO Box 29-181, Christchurch, New Zealand.
(B) Landcare Research NZ Ltd, Private Bag 3127, Hamilton, New Zealand.
(C) Tasman District Council, Private Bag 4, Richmond, New Zealand.
(D) Environment Southland, Private Bag 90116, Invercargill, New Zealand; Present address: Sinclair Knight Merz, PO Box 8298, Christchurch, New Zealand.
(E) Corresponding author. Email: email@example.com
Table 1. Characteristics of current and previous pesticide leaching field sites Climate data taken from New Zealand Meteorological Service (1983) Sites Soil Parent Mean ann. material rainfall (mm) Hawkes Bay Silt loam and Greywacke 798 (2 sites) sandy loam alluvium over gravels Waikato Silt loam over Volcanic 1214 alluvial alluvium sands/silts Northland Clay loam, Basalt 1514 compact subsoil Manawatu Sand Dune sand 1120 Nelson Silt loam over Greywacke 970 gravels alluvium Southland Silt loam Loess 1079 Mean daily temp. ([degrees]C) Sites Max. Min. Mean Hawkes Bay 18.8 7.0 12.9 (2 sites) Waikato 19.2 7.6 13.4 Northland 19.9 10.9 15.4 Manawatu 17.1 8.9 13.0 Nelson 17.4 7.8 12.6 Southland 14.4 5.2 9.8 Sites Mean RH [K.sub.sat] (B) (%) (A) (mm/h) Hawkes Bay 82 15-25 (2 sites) Waikato 84 8 Northland 82 75 Manawatu 79 204 Nelson 80 >60 Southland 84 4 (A) Relative humidity. (B) [K.sub.sat] is saturated hydraulic conductivity of the topsoil. Table 2. Soil chemical and physical characteristics for experimental site OC, Organic carbon; BD, bulk density Horizon Depth(m) Description Nelson site Motupiko silt loam A 0-0.15 Silt loam Bw 0.15-0.37 Silt loam BC1 0.37-0.65 Very gravelly sand BC2 >0.65 Slightly gravelly sand Southland site-Waikiwi silt loam A 0-0.20 Silt loam AB 0.20-0.26 Silt loam B2 0.26-0.39 Silt loam BC 0.39-0.55 Silty clay loam C1 0.55-0.92 Silty clay loam C2(IuB) 0.92-1.20+ Silty clay loam Horizon pH Clay Silt Nelson site Motupiko silt loam A 19 57 Bw 26 59 BC1 BC2 Southland site-Waikiwi silt loam A 5.7 24 69 AB 6.0 18 76 B2 6.1 12 82 BC 5.9 18 74 C1 21 75 C2(IuB) 28 68 Horizon OC Stones (%) Nelson site Motupiko silt loam A 2.38 0 Bw 1.23 0 BC1 0.26 (B) 60 (A) BC2 60 (A) Southland site-Waikiwi silt loam A 4.1 0 AB 2.7 0 B2 1.6 0 BC 0.8 0 C1 0.9 0 C2(IuB) 0 Horizon BD Total (g/[cm.sup.3]) porosity (%, v/v) Nelson site Motupiko silt loam A 1.42 46 Bw 1.38 50 BC1 2.1 (A) BC2 2.1 (A) Southland site-Waikiwi silt loam A 1.13 57 AB 1.29 54 B2 1.29 53 BC 1.46 47 C1 1.42 48 C2(IuB) 1.34 51 (A) The % stones was estimated visually and the BD adjusted for % stones. (B) OC% for fine (<2 mm) soil fraction. Table 3. Selected information on chemicals applied on each site Selected 'best available' values from www.arsusda.gov/acsl/ppdb.html as at December 2004 with range in parentheses are given for mobility and persistence Mobility ([K.sub.oc] from Chemical Primary use literature, mL/g) Bromide Water tracer Atrazine (C) (herbicide) Arable cropping 147 (38-288) Bromacil (herbicide) Arable cropping 14 (2-33) Diazinon (insecticide) Wide range of uses 1520 (1007-1842) Hexazinone (herbicide) Forestry 40 (34-74) Procymidone (B) (fungicide) Vegetables, 1500 (1500-1945) strawberries and stone-fruit Terbuthylazine (A) Arable cropping 220 (162-278) (herbicide) Trifluralin (B) (herbicide) Arable cropping 7200 (1200-13 700) Chemical Persistence Application rate (kg/ha) (field half-life in days, from literature) Nelson Southland Bromide No degradation 150 195 Atrazine (C) (herbicide) 173 (13-402) 0.7 0.6 Bromacil (herbicide) 207 (61-349) 10.0 11.4 Diazinon (insecticide) 7 (7-49) 27.6 22.9 Hexazinone (herbicide) 79 (30-180) 73.0 38.2 Procymidone (B) (fungicide) 15 (7-120) 14.4 -- Terbuthylazine (A) 60 (30-60) 13.8 7.4 (herbicide) Trifluralin (B) (herbicide) 81 (15-149) -- 4.6 (A) Data for terbuthylazine come from The Pesticide Manual (1994). (B) Trifluralin was only applied at the Southland site and procymidone was only applied at the Nelson site. (C) The application rates for atrazine were much lower than planned. Table 4. Mass recoveries (%)for tracers and pesticides for each plot For the purposes of calculating mass recoveries, values <DL were set equal to zero. The detection limit is equivalent to mass recoveries ranging from 1.2 to 1.5% for the different pesticides, and from 2 to 4% for Br, depending on the soil moisture status and the input mass. Note: One of the soil cores for the Southland site sampled in August 2003 was taken from the area next to where there had been some leakage while filling the pesticide spray containers; this resulted in increases in concentrations for some pesticides, particularly terbuthylazine and trifluralin Sampling Days since Depth of Br date application sampling (m) Nelson site 2 Dec. 02 22 0.80 102 14 Jan. 03 65 0.80 70 4 Feb. 03 86 0.80 61 7 May 03 178 0.80 36 6 Aug. 03 269 0.80 3 1 Nov. 03 356 0.80 1 19 Feb. 04 466 0.80 0 Southland site 2 Dec. 02 26 1.00 95 14 Jan. 03 69 1.00 63 4 Feb. 03 90 1.00 57 7 May 03 182 1.00 38 6 Aug. 03 273 1.00 20 5 Nov. 03 364 1.00 2 16 Feb. 04 469 1.00 0 Sampling Atrazine Bromacil Diazinon date Nelson site 2 Dec. 02 10 86 4 14 Jan. 03 3 73 2 4 Feb. 03 2 68 0 7 May 03 0 59 0 6 Aug. 03 0 73 0 1 Nov. 03 0 29 0 19 Feb. 04 0 10 0 Southland site 2 Dec. 02 4 41 1 14 Jan. 03 3 46 0 4 Feb. 03 2 52 0 7 May 03 0 59 0 6 Aug. 03 0 45 0 5 Nov. 03 0 37 0 16 Feb. 04 0 25 0 Sampling Hexazinone Terbuthylazine Procymidone/ date trifluralin (A) Nelson site 2 Dec. 02 102 51 25 14 Jan. 03 91 46 34 4 Feb. 03 54 24 24 7 May 03 5 10 23 6 Aug. 03 5 18 59 1 Nov. 03 1 12 61 19 Feb. 04 0 1 49 Southland site 2 Dec. 02 104 89 71 14 Jan. 03 68 80 49 4 Feb. 03 71 70 44 7 May 03 4 63 42 6 Aug. 03 2 113 69 5 Nov. 03 2 51 45 16 Feb. 04 1 10 24 (A) Procymidone at Nelson, trifluralin at Southland. Table 5. Summary of batch isotherm results for Motupiko silt loam Freundlich isotherm [K.sub.f] ([micro] [g.sup.1-n] Pesticide n [mL.sup.n]/g) Topsoil 0-0.10m (OC % 2.38) Atrazine 1.03 4.82 Bromacil 1.14 3.42 Diazinon 0.87 15.24 Hexazinone 1.34 1.35 Terbuthylazine 1.24 14.96 Subsoi1 0.40-0.50m (OC% 0.26) (A) Atrazine 0.92 2.60 Bromacil 0.93 3.33 Diazinon 0.96 3.13 Hexazinone 1.08 1.88 Terbuthylazine 1.00 4.91 Freundlich isotherm Pesticide [K.sub.foc] [R.sup.2] Topsoil 0-0.10m (OC % 2.38) Atrazine 202 0.97 Bromacil 144 0.90 Diazinon 640 0.99 Hexazinone 57 0.93 Terbuthylazine 629 0.89 Subsoi1 0.40-0.50m (OC% 0.26) (A) Atrazine 999 0.95 Bromacil 1282 0.99 Diazinon 1202 0.99 Hexazinone 722 0.92 Terbuthylazine 1889 0.98 Linear [K.sub.d] regression [K.sub.d] [K.sub.oc] Pesticide (mL/g) (mL/g) [R.sup.2] Topsoil 0-0.10m (OC % 2.38) Atrazine 4.47 188 0.99 Bromacil 4.88 205 0.96 Diazinon 13.62 572 0.92 Hexazinone 2.29 97 0.99 Terbuthylazine 13.68 575 0.99 Subsoi1 0.40-0.50m (OC% 0.26) (A) Atrazine 3.32 1276 0.68 Bromacil 3.49 1343 0.97 Diazinon 2.62 1008 0.85 Hexazinone 3.71 1427 0.82 Terbuthylazine 6.90 2656 0.99 (A) Batch isotherms and OC% were analysed on the fine (<2 mm) soil fraction. Table 6. Summary of batch isotherm results for Waikiwi silt loam Freundlich isotherm [K.sub.f] ([micro] [g.sup.1-n] Pesticide n [mL.sup.n]/g Topsoil 0-0.10m (OC% 3.89) Atrazine 0.76 3.07 Bromacil 0.79 0.60 Diazinon 0.82 5.51 Hexazinone 0.79 1.02 Terbuthylazine 0.83 6.48 Subsoil 0.55-0.70m (OC% 0.87) Atrazine 0.87 0.86 Hexazinone 1.03 1.54 Terbuthylazine 1.24 0.95 Freundlich isotherm Pesticide [K.sub.foc] [R.sup.2] Topsoil 0-0.10m (OC% 3.89) Atrazine 79 0.99 Bromacil 16 0.97 Diazinon 142 0.98 Hexazinone 26 0.91 Terbuthylazine 167 0.99 Subsoil 0.55-0.70m (OC% 0.87) Atrazine 98 0.90 Hexazinone 177 0.94 Terbuthylazine 109 0.83 Linear [K.sub.d] regression [K.sub.d] [K.sub.oc] Pesticide (mL/g) (mL/g) [R.sup.2] Topsoil 0-0.10m (OC% 3.89) Atrazine 2.92 75 0.96 Bromacil 0.73 19 0.98 Diazinon 3.98 102 0.92 Hexazinone 1.75 45 0.81 Terbuthylazine 6.87 177 0.95 Subsoil 0.55-0.70m (OC% 0.87) Atrazine 1.71 197 0.99 Hexazinone 3.12 359 0.99 Terbuthylazine 3.85 442 0.96