Fate and transport of chlormequat in subsurface environments
- 349 Downloads
Background, aim and scope
Chlormequat (Cq) is a plant growth regulator used throughout the world. Despite indications of possible effects of Cq on mammalian health and fertility, little is known about its fate and transport in subsurface environments. The aim of this study was to determine the fate of Cq in three Danish subsurface environments, in particular with respect to retardation of Cq in the A and B horizons and the risk of leaching to the aquatic environment. The study combines laboratory fate studies of Cq sorption and dissipation with field scale monitoring of the concentration of Cq in the subsurface environment, including artificial drains.
Materials and methods
For the laboratory studies, soil was sampled from the A and B horizons at three Danish field research stations—two clayey till sites and one coarse sandy site. Adsorption and desorption were described by means of the distribution coefficient (Kd) and the Freundlich adsorption coefficient (KF,ads). The dissipation rate was estimated using soil sampled from the A horizon at the three sites. Half life (DT50) was calculated by approximation to first-order kinetics. A total of 282 water samples were collected at the sites under the field monitoring study— groundwater from shallow monitoring screens located 1.5–4.5 m b.g.s. at all three sites as well as drainage water from the two clayey sites and porewater from suction cups at the sandy site, in both cases from 1 m b.g.s. The samples were analysed using LC-MS/MS. The field monitoring study was supported by hydrological modelling, which provided an overall water balance and a description of soil water dynamics in the vadose zone.
The DT50 of Cq from the A horizon ranged from 21 to 61 days. The Cq concentration-dependant distribution coefficient (Kd) ranged from 2 to 566 cm3/g (median 18 cm3/g), and was lowest in the sandy soil (both the A and B horizons). The KF,ads ranged from 3 to 23 (µg1 − 1/n (cm3)1/n g−1) with the exponent (1/n) ranging from 0.44 to 0.87, and was lowest in the soil from the sandy site. Desorption of Cq was very low for the soil types investigated (<10%w). Cq in concentrations exceeding the detection limit (0.01 µg/L) was only found in two of the 282 water samples, the highest concentration being 0.017 µg/L.
That sorption was highest in the clayey till soils is attributable to the composition of the soil, the soil clay and iron content being the main determinant of Cq sorption in both the A and B horizons of the subsurface environment. Cq was not detected in concentrations exceeding the detection limit in either the groundwater or the porewater at the sandy site. The only two samples in which Cq was detected were drainage water samples from the two clayey till sites. The presence of Cq here was probably attributable to the hydrogeological setting as water flow at the two clayey till sites is dominated by macropore flow and less by the flow in the low permeability matrix. In contrast, water flow at the sandy site is dominated by matrix flow in the high permeability matrix, with negligible macropore flow. Given the characteristics of these field sites, Cq adsorption and desorption can be expected to be controlled by the clay composition and content and the iron content. Combining these observations with the findings of the sorption and dissipation studies indicates that the key determinant of Cq retardation and fate in the soil is sorption characteristics and bioavailability.
The leaching risk of Cq was negligible at the clayey till and sandy sites investigated. The adsorption and desorption experiments indicated that absorption of Cq was high at all three sites, in particular at the clayey till sites, and that desorption was generally very limited. The study indicates that leaching of Cq to the groundwater is hindered by sorption and dissipation. The detection of Cq in drainage water at the clayey till sites and the evidence for rapid transport through macropores indicate that heavy precipitation events may cause pulses of Cq.
Recommendations and perspectives
The present study is the first to indicate that the risk of Cq leaching to the groundwater and surface water is low. Prior to any generalisation of the present results, the fate of Cq needs to be studied in other soil types, application regimes and climatic conditions to determine the Cq retardation capacity of the soils. The study identifies bioavailability and heavy precipitation events as important factors when assessing the risk of Cq contamination of the aquatic environment. The possible effects of future climate change need to be considered when assessing whether or not Cq poses an environmental risk.
KeywordsCCC Chlormequat Chlorocholine chloride Cycocel Freundlich isotherm Groundwater Pesticide Quaternary ammonium herbicides Soil Sorption
1 Background, aim and scope
Chlormequat (hereafter Cq) is a plant growth regulator used throughout the world. In six out of 18 European countries, it is among the five compounds used most frequently to treat wheat crops (Anonymous 2007a). Concern for the potential effects of Cq on mammalian health and fertility has been raised and discussed (Kramers et al. 1975; Sussmuth and Lingens 1976; Olson and Hinsdill 1984; Fairbrother et al. 1986; Yamano and Morita 1993; Langhammer et al. 1999; Torner et al. 1999; Andersen et al. 2002; Reynolds et al. 2004; Leffers et al. 2006; Sorensen and Danielsen 2006; Sorensen et al. 2009). Some in vitro experiments with mice spermatozoa and oocytes indicate that Cq intake may reduce the functional competence of the spermatozoa resulting in significantly diminished oocyte fertilisation and cleavage (Torner et al. 1999; Sorensen and Danielsen 2006). Given the potential effects of Cq on the environment and human fertility, surprisingly little is known about the fate and transport of Cq in the subsurface environment, particularly in relation to fate and transport mechanisms in soil and the possible risk of groundwater contamination.
Cq is an onium compound acting through the inhibition of the gibberellin biosynthesis in the target organism (Rademacher 2000). The active compound Cq is generally applied as the chloride salt, i.e. chlormequat-chloride (2-chloroethyltrimethylammonium chloride, CAS No. 999-81-5, also known as chlorocholine chloride and CCC). Regulatory issues relating to the use of Cq have been reviewed by the US EPA (Anonymous 2007b). Much is known about its mode of action (Hughes 1967; Tolbert 1960), distribution in treated plants (Dekhuijz and Vonk 1974), effect on crop yield (Alexopoulos et al. 2006; Berry et al. 2004; Browne et al. 2006; Espindula et al. 2009; Gencsoylu 2009; Leitch and Kurt 1999) and residue concentration in food and agricultural products (Granby and Vahl 2001; Juhler and Vahl 1999; Zhao et al. 2000; Ueberschar et al. 2003). No studies have yet been published on the environmental fate of Cq and the contamination risk it may pose to the aquatic environment, however.
The present study is the first to examine the subsurface fate and transport of Cq to assess the risk of leaching through the vadose zone to the groundwater. The study combines laboratory fate studies of Cq degradation, adsorption and desorption processes in both clayey till and sandy soils with monitoring studies of the transport of Cq under field conditions. As several studies of Cq metabolism indicate that all the metabolites of Cq are compounds commonly present in the environment (Fig. 1), the present study focused on the parent compound, Cq.
2 Materials and methods
2.1 Field site descriptions
Physical and chemical properties of the soil in the A and B horizons of the three Danish sites (Lindhardt et al. 2001) together with dissipation and sorption characteristics of Cq in the soils
2.2 Field monitoring study
Leaching of Cq under field conditions was evaluated at the three field sites. Each of the sites was cultivated in line with conventional agricultural practice in the region. Cq was applied in the maximum permitted dose in accordance with the regulations. At the Estrup and Silstrup sites, 698 g Cq/ha (1.2 L Cycocel 750/ha) was applied to winter wheat on 11 April and 13 April 2007, respectively. At the Jyndevad site, 581 g Cq/ha (1 L Cycocel 750/ha) was applied to triticale on 13 April 2007.
2.2.1 Water sampling
Water samples were collected from groundwater monitoring screens (no. of samples: Estrup 74, Silstrup 101 and Jyndevad 13), tile drains (no. of samples: Estrup 45 and Silstrup 21) and suction cups (no. of samples: Jyndevad 28) over a 12-month period following application of Cq. The groundwater samples were collected from vertical and horizontal monitoring wells at depths ranging from 1.5 to 4.5 m, monthly at the two clayey till sites and half-yearly at the sandy site. The drainage water samples were collected by means of flow-proportional sampling at the clayey till sites, with subsamples being collected for every 3,000 L of drainage runoff using a refrigerated ISCO (Teledyne Isco, Inc., Lincoln). The chemical analyses were performed weekly. Porewater samples were collected monthly from groups of Teflon suction cups (Prenart, Frederiksberg) installed at 1 m b.g.s. at two locations at the sandy site.
To avoid influencing pesticide leaching at the field sites due to the installation and presence of sampling equipment in the ground, all installations and soil sampling deeper than 20 cm b.g.s. were restricted to a buffer zone surrounding the treated area of each site. Detailed information about the instrumentation and sampling methods are provided elsewhere (Lindhardt et al. 2001; Kjaer et al. 2005).
2.2.2 Water balance
The monitoring data were supported by hydrological modelling, which provided an overall water balance and a description of soil water dynamics in the vadose zone. The MACRO model version 5.1 (Larsbo et al. 2005) was applied to each site covering the soil profile to a depth of 5 m, always including the water table. The model was parameterised, calibrated and validated using the considerable amount of data available for each of the field research stations. A detailed description of data acquisition, model set-up and model performance is provided elsewhere (Barlebo et al. 2007).
2.3 Laboratory fate studies
2.3.1 Soil sampling
The soil samples for the sorption and dissipation studies were collected from the topsoil (0–20 cm b.g.s.) and subsoil (40–60 cm b.g.s.) at the Estrup, Silstrup and Jyndevad sites using a 2-cm i.d. soil sampling auger as described in Henriksen et al. (2009). After mixing, a 1-kg composite sample was transferred to a rilsan bag (Rotek, Sdr. Felding) and stored at −18°C until analysis. The physical and chemical properties of the A and B horizon soil at each of the sites are summarised in Table 1. Sampling was performed in March 2007 prior to the application of Cq. According to available data dating back to 1993 for the Estrup site, 1983 for the Silstrup site and 1995 for the Jyndevad site, Cq had only been applied once previously, namely at the Silstrup site in 1985, i.e. 22 years previously.
2.3.2 Batch sorption and desorption experiments
The Freundlich adsorption coefficient (KF, ads, µg1 − 1/n (cm3)1/n g−1) was estimated in batch experiments. Air-dried samples from the field sites were sieved with mesh size 2 mm before initiation of the sorption experiments. For each assay a 2-g homogenised soil sample was placed in a flask with Milli-Q water to yield a soil (dry weight):water ratio of 1:6 for the topsoil samples and 1:4 for the subsoil samples.
2.3.3 Batch dissipation study
The procedure for estimating dissipation time (the time for a 50% decline of the initial pesticide concentration, DT50) is described in detail elsewhere (Juhler et al. 2008). Soil samples collected in the field were homogenised, sieved (mesh size 2 mm) and stored at −20°C until needed for the dissipation experiment. Soil corresponding to 30 g fresh weight was spiked with 1.0 mL of an aqueous Cq solution to a concentration of 183 µg Cq/kg soil and incubated at 10°C in darkness. For each soil, 15 replicates were prepared and incubated, with batches of three replicates being analysed after 1, 7, 14, 28 and 63 days of incubation. The soil was extracted using pressurised liquid extraction and analysed using LC-MS/MS, as described elsewhere (Henriksen et al. 2009).
2.4 Chemicals, analysis and quantification
Cq (2-chloroethyltrimethylammonium, CAS No. 7003-89-6) was purchased as its chloride salt chlormequat-chloride (2-chloroethyltrimethylammonium chloride, CAS No. 999-81-5) from Ehrenstorfer. For field application a commercial formulation, Cycocel, was used (BASF, Ludwigshafen, Germany). The water samples and soil extracts were analysed by LC-MS/MS (Henriksen et al. 2009). The equipment was a Waters Alliance 2695 HPLC system (Milford, USA) connected to a Quattro Ultima triple quadrupole mass spectrometer from Micromass (Manchester, UK). The chromatography was based on ion-pair chromatography using a 150 × 2 mm Zorbax C8 column (5-µm particle size) from Agilent (Santa Clara, CA, USA) at 30°C with a flow rate of 0.15 mL min−1. The solvent components were HPLC grade methanol (Fisher Scientific, Loughborough, UK) and heptafluorobutyric acid (HFBA, 97.4%, CAS RN 375-22-4; Sigma–Aldrich). A 60% aqueous solution of 3-chloro-2-hydroxypropyltrimethylammonium chloride (CHTC; Sigma-Aldrich, Schnelldorf, Germany) was used as the internal standard. The mobile phase consisted of methanol/50 mM heptafluorobutyric acid (25:75, v/v). Electrospray ionisation was performed in positive ionisation mode with MS/MS detection of Cq at the MS/MS ion trace: [M]+m/z 122 → m/z 58.
The limit of detection (LOD) was 0.01 µg/L. Every 4 months, external control samples containing 0.039 µg Cq/L were analysed along with the monitoring water samples from the three sites. The control samples were prepared in the field by quantitatively transferring a standard solution to a 1-L measuring flask. The standard solution was diluted and adjusted to the mark with groundwater from an upstream well. After thorough mixing, the control sample was transferred to a sample bottle and transported to the laboratory together with the regular samples, the identity of the control samples being unknown to the laboratory. Recovery of the spiked samples was generally good, the average recovery ranging from 54% to 95%. No Cq was detected in blank samples, thus indicating that no contamination of the samples occurred.
The fate and transport of Cq were examined by means of a 1-year monitoring study (April 2007–April 2008) encompassing all three field sites. In support of the field monitoring, soils were sampled from the A and B horizon at each site for determination of Cq sorption, desorption and dissipation.
3.1 Field monitoring study
Annual water balances for the monitoring period (1 April 2007–31 March 2008) and quality control of monitoring programme for the Estrup, Silstrup and Jyndevad field research stations: ‘1st month precipitation’ and ‘1st month percolation’ refer to accumulated precipitation and percolation within the first month following application of chlormequat
Water balances (mm/year)
Measured tile-drainage water
Quality control of monitoring programme
Number of samples analysedb
Number of samples containing chlormequat
Maximum concentration (µg/L)
1st month precipitation (mm)
1st month percolation (mm)a
3.2 Laboratory fate studies
The general absence of Cq in the groundwater and drainage water samples could be due to retention of Cq in the soil matrix or transformation or other processes causing loss of the active compound. Sorption and dissipation of Cq in soil from the A and B horizons of each of the three sites are summarised in Table 1. Sorption and desorption were evaluated at four Cq concentrations ranging from 0.5 to 10 µg Cq/mL water. The adsorption isotherms for Cq sorption on soil were generally non-linear. The Freundlich regression constant ranged from 0.42 to 0.87 (n−1, Table 1), indicating a concentration-dependant process and a heterogeneous soil matrix with binding sites displaying a variety of adsorption energies associated with the binding sites. The Freundlich regression constant (n−1) was significantly higher at the sandy Jyndevad site than at the clayey Estrup site (one-way ANOVA between group design with Tukey grouping, F(2.5); p < 0.04). The Freundlich adsorption coefficient (KF,ads) for the two horizons at the three sites ranged from 3 to 23 (µg1 − 1/n (cm3)1/n g−1) and was significantly lower at the sandy Jyndevad site than at the clayey Estrup and Silstrup sites (one-way ANOVA between group design with Tukey grouping, F(2.5); p < 0.02).
Following sorption, less than 10% of the sorbed Cq could be desorbed from the solid phase when pure water was added to the system. Due to the low desorption of Cq, it was not possible to determine desorption isotherms for the six soils. The desorption experiments clearly indicated very low reversibility of Cq sorption to the soil matrix, however. The results are in agreement with a previous study on montmorillonite and Cq in which approximately 10% of the adsorbed Cq could be desorbed by the addition of water (Maza et al. 1989). The authors assigned this pool of Cq to a fraction weakly retained on the external surface of the mineral by weak Van der Waals forces.
In addition to the sorption characteristics of a pesticide residue, dissipation is also a central parameter in risk assessment. Dissipation is commonly described by DT50, the time required for a 50% dissipation of the initial pesticide concentration (Beulke and Brown 2001). Dissipation of the pesticide can be attributable to a multitude of processes, including transformation, metabolisation and sorption. In DT50 experiments, harsh extraction conditions are used, i.e. more of the bound residue can be recovered using the DT50 protocol than with simple water addition in a batch desorption experiment. The DT50 of soil sampled from the A horizon at the three field sites ranged from 21 to 61 days (Table 1), in agreement with an average DT50 of 32 days reported for four soils incubated at 10°C (Anonymous 1999a).
Sorption in both the A and B horizons was highest in the soil from the two clayey till sites, a finding apparently attributable to differences in the composition of the soils. The two sites differ from the sandy Jyndevad site in that the soil has a higher clay and iron content. These constituents could be important sorbents for Cq, an observation that is in agreement with studies of Cq on a standard clay mineral montmorillonite (Maqueda and Morillo 2001; Maza et al. 1989). It appeared that Cq was retained on the mineral predominantly by interlamellar cationic exchange, supplemented by a small amount of Cq adsorption to the external surface of the mineral. Retention of the Cq was reversible and related to formation of well-defined interlayer complexes that were stable in air. Copper and other ions also affected sorption in the study of Maqueda and Morillo (2001). Thus, adsorption decreased when Cu and electrolytes were present in the aqueous medium. The presence of a ‘two-region competitive sorption process’ has been suggested, i.e. high-affinity sites for metal binding at the edge of the minerals and interlamellar binding sites for Cq and Cu. The findings of Maqueda and Morillo (2001) and Maza et al. (1989) together with the present findings indicate that the clay composition and content as well as the iron ion content could be key determinants of Cq sorption and desorption in the soil environment.
In a previous study on Cq sorption to the clay mineral montmorillonite, the adsorption isotherm was found to be S-shaped (Maqueda and Morillo 2001). The authors interpreted this as indicating an increase in affinity for Cq after a few molecules have been adsorbed, possibly due to opening of the silicate layers when some Cq molecules have adsorbed to the interlamellar binding sites on the mineral (Maqueda and Morillo 2001; Maza et al. 1989; Giles et al. 1960). In the present study, the isotherms were not S-shaped, but instead tended to be L-shaped, in agreement with the study of Maza et al. (1989). L-shaped sorption has also been observed for the cationic herbicides diquat and paraquat in soils from Western Australia (Kookana and Aylmore 1993). Sorption characteristics of this type would indicate that as sites are filled, it becomes increasingly difficult for Cq to adsorb to the soil components. As the present study was made using disturbed heterogeneous soils rather than standardised minerals, the adsorption characteristics are likely to reflect the variability of real-world systems and thus the sum of several sorption processes. As the results reflect a mixture of sorption mechanisms, the curve shape observed in the present study should not be interpreted too rigorously in relation to a single sorption mechanism. Rather, the present results indicate that in complex soil samples, the sorption of Cq is highly effective at low pesticide concentrations. This may be an important characteristic in that it could hinder the leaching of Cq through natural soils.
Compared to literature values (Maqueda and Morillo 2001; Maza et al. 1989; Giles et al. 1960), the Kd values estimated for the Estrup and Silstrup sites were one to two orders of magnitude higher. This is not unexpected, as distribution constants would be expected to be highly dependent on the soil matrix properties, as well as on the pesticide concerned. In the laboratory fate studies presented here, disturbed heterogeneous soil samples were used to determine sorption, desorption and dissipation. These structureless soil samples represent a simplified and relatively homogenous model system reflecting a subset of the soil matrixes existing under field conditions, but do not take into account important characteristics such as soil aggregates, micropores and macropores that are present under actual field conditions (Smith et al. 2003). The differences in sorption and desorption reactions in steady-state and dynamic systems have been discussed in several papers, moreover, and it has been argued that the use of simple steady-state Kd studies in modelling is problematic (Reardon 1981; Bethke and Brady 2000; Smith et al. 2003). Despite these drawbacks, the laboratory methods are generally accepted as an approach for setting a range for the parameters needed for modelling. In the present study, the laboratory sorption and desorption studies were supplemented with field experiments where the Cq was exposed to the variability in the geological, geochemical and hydrological setting at the three sites.
Statistical analysis revealed that Kd was significantly lower at the sandy Jyndevad site than at the clayey Estrup site (ANOVA, p < 0.05). In addition to having the lowest Kd, the soils from the sandy Jyndevad site exhibited the fastest dissipation of Cq (Table 1 and Fig. 3). From dissipation studies on other pesticides, it is known that bioavailability of the compound in question can be an important determinant of its dissipation (Gevao et al. 2000; Barbash 2007; Zhang et al. 1998). Moreover, model studies have identified desorption resistance as an important controlling parameter for bioavailability of organic chemicals in porous media (Liu et al. 2007). Thus, the observed rapid dissipation of Cq in soils with low Kd (i.e. high concentration in the aqueous phase, where bioavailability can be expected to be relatively high) and the slower dissipation in soils with relatively higher Kd (i.e. high concentration in the solid soil phase, where bioavailability can be expected to be relatively low) indicates the importance of bioavailability in the transformation of Cq.
Given the results of the laboratory fate studies and the precipitation pattern, the Jyndevad site would be expected to be the site most prone to Cq leaching. Compared to the clayey till soils (Silstrup and Estrup), precipitation and percolation following Cq application were highest at the sandy Jyndevad site, while soil retardation characteristics (Kd, KF and DT50) were lowest (Table 1). Nevertheless, Cq was detected in drainage water from the two clayey till sites but not in water from the sandy Jyndevad site. This apparent contradiction is probably explained by the hydrogeological and geochemical setting at the sites. Thus the hydrogeological setting of the two clayey till sites is dominated by macropore flow and less by the flow in the low permeability matrix. In contrast, the sandy Jyndevad site consists of a high permeability matrix where macropore flow is less dominant (Kjaer et al. 2005). Positively charged free iron oxides are present in the A horizon at all three sites. Given that both the A horizon matrix and Cq are positively charged under natural soil conditions, Cq could be affected by cation exclusion from the A horizon as suggested by Maqueda and Morillo (2001). This process is likely to be more prevalent at the clayey till sites (Estrup and Silstrup) as the iron content is much higher there than at the sandy site (Jyndevad). Moreover, as macropore flow dominates at the clayey till sites, Cq could more readily sorb to the clay and organic matter lining the macropores (Rosenbom et al. 2008). Heavy precipitation events will then entail a non-equilibrium situation in the macropore resulting in rapid vertical flow and transport to the drainage system. It is known that both organic and inorganic colloids may transport pesticides towards the groundwater, and such carrier processes may also be of importance to the Cq transport in macroporous soils (Pateiro-Moure et al. 2009; Rytwo et al. 2002; Rytwo and Tropp 2001). The quantitative impact of such transport was not evaluated in the present study. In general the sorption capacity and property of the coating present in the macropores will determine the release of Cq to the liquid phase in the macropore or in colloid-bound form (de Jonge et al. 2004). In both cases in which Cq was detected in a drainage water sample, heavy precipitation events occurred approximately 1.5 weeks beforehand. This could have resulted in Cq transport from the macropores to the drains. Clear evidence of such macropore transport was found at the Estrup site. While piston flow through the low permeability soil matrix would involve a transit time to the drainage system of about 98 days, the Cq was detected in drainage water 56 days after application. The heavy precipitation event (34 mm/day) there on 29 May 2007 apparently seems to have induced rapid transport of Cq though the macropores. This finding is thus consistent with previous transport studies of another strongly sorbing pesticide, glyphosate, at the Estrup and Silstrup sites (Kjaer et al. 2005), as well as other field studies demonstrating rapid macropore-mediated transport of pesticides (Flury 1996; Jarvis 2007).
The present study combines laboratory fate studies on naturally remoulded soil samples with field scale monitoring to assess the risk of Cq contamination of the groundwater. The sorption and desorption experiments indicated that Cq was highly adsorbed to the soils investigated. Moreover, once Cq was adsorbed, the fraction that could be desorbed was highly limited. In the field monitoring experiments, Cq was only detected in two of the 282 water samples, in both cases at a concentration of less than 0.02 µg/L. Thus even though Cq is highly soluble in water, properties related to sorption hinder it from leaching to the groundwater. The detection of Cq in two of the drainage water samples indicates that Cq can be rapidly transported through macropores in structured loamy soil.
The present study indicates that the leaching risk posed to the aquatic environment by Cq is negligible at the three field sites investigated. The study also demonstrates the usefulness of combining field and laboratory experiments. The data are still too limited to allow extensive generalisation, however.
6 Recommendations and perspectives
In view of the widespread use of Cq and the ongoing discussion on possible effects on man and the environment, it is recommended that the present study be repeated in other parts of the world. It appears from our study that the sorption characteristics of Cq are such that the majority of the pesticide is retained in the soil and that only a small fraction can be desorbed. The combined fast dissipation and low sorption in the soils from the sandy Jyndevad site suggest that bioavailability may be an important determinant of the fate of Cq in the subsurface environment. This finding should be investigated further. No evidence was found of extensive Cq leaching, although rapid transport through macroporous structured soils may occur after heavy precipitation events.
The project was supported financially by Copenhagen Energy (Københavns Energi A/S). We wish to thank P. Stockmarr (GEUS) for skilled technical assistance in connection with method development and sample analysis, P. Olsen (Aarhus University), L. Gudmundsson (GEUS), C. Andersen (Jyndevad Research Station), P. Boesen (Estrup Research Station) and J. Molbo (Silstrup Research Station) for assistance with the field leaching experiments, and D. I. Barry for linguistic assistance.
- Alder L, Startin JR (2005) Determination of chlormequat and mepiquat in foods by liquid chromatography/mass spectrometry or liquid chromatography/tandem mass spectrometry: interlaboratory study. J AOAC Int 88:1762–1776Google Scholar
- Alexopoulos AA, Akoumianakis KA, Passam HC (2006) Effect of plant growth regulators on the tuberisation and physiological age of potato (Solanum tuberosum L.) tubers grown from true potato seed. Can J Plant Sci 86:1217–1225Google Scholar
- Anonymous (1999a) The pesticide manual, 11th edn. British Crop Protection Council, SurreyGoogle Scholar
- Anonymous (1999b) Soil taxonomy—a basic system of soil classification for making and interpreting soil surveys, 2nd edn. USDA-NRCS Agricultural Handbook, vol 436. United States Department of Agriculture, New YorkGoogle Scholar
- Anonymous (2000) Adsorption-desorption using a batch equilibrium method. Section 1: physical chemical properties, 106, OECD guideline for the testing of chemicals. OECD, ParisGoogle Scholar
- Anonymous (2006) Reregistration eligibility decision for aliphatic alkyl quaternaries (DDAC). EPA-HQ-OPP-2006-0338/EPA 739-R-06-008. US Environmental Protection Agency, Prevention, Pesticides and Toxic Substances, Washington, D.C.Google Scholar
- Anonymous (2007a) The use of plant protection products in the European Union, data 1992-2003. Nadin, Pierre. http://epp.eurostat.ec.europa.eu. ISBN 92-79-03890-7. Eurostat, European Communities, Luxembourg
- Anonymous (2007b) Reregistration eligibility decision for chlormequat chloride (RED). EPA-HQ-OPP-2007-0968/EPA 738-R-07-014. US Environmental Protection Agency, Prevention, Pesticides and Toxic Substances, Washington, D.C.Google Scholar
- Barlebo HC, Rosenbom AE, and Kjaer J (2007) Evaluation of pesticide scenarios for the registration procedure. Environmental Project No. 1178, Danish Environmental Protection Agency, CopenhagenGoogle Scholar
- Berry PM, Sterling M, Spink JH, Baker CJ, Sylvester-Bradley R, Mooney SJ, Tams AR, Ennos AR (2004) Understanding and reducing lodging in cereals. Adv Agron Vol 84(84):217–271Google Scholar
- Hughes RG (1967) Chlormequat to control lodging in wheat. Agriculture 74:133Google Scholar
- Juhler RK, Vahl M (1999) Residues of chlormequat and mepiquat in grain—results from the Danish National Pesticide Survey. J AOAC Int 82:331–336Google Scholar
- Kramers PGN, Knaap AGAC, Voogd CE (1975) Lack of mutagenicity of chlormequat chloride in Drosophila and in bacteria. Mutat Res 31:65–68Google Scholar
- Langhammer M, Kuhla S, Schneider F, Renne U, Spitschak M, Hagemeister H (1999) Effects of chlorocholine chloride-treated wheat on special fertility traits in laboratory mice. J Anim Physiol An N 81:190–202Google Scholar
- Lesueur C, Gartner M, Mentler A, Fuerhacker M (2007) Qualitative and quantitative analysis of polar pesticide multiresidues in leaf samples with a liquid chromatography-ion-trap mass-selective detector. Int J Environ Anal Chem 87:1013–1032Google Scholar
- Lindhardt B, Abiltrup C, Vosgerau H, Olsen P, Torp S, Iversen BV, Jørgensen OJ, Plauborg F, Rasmussen P, Gravesen P (2001) The Danish Pesticide Leaching Assessment Programme, Site characterization and monitoring design. Geological Survey of Denmark and Greenland, GEUS, CopenhagenGoogle Scholar
- Rademacher W (2000) Growth retardants: effects on gibberellin biosynthesis and other metabolic pathways. Annu Rev Plant Phys 51:501–531Google Scholar
- Rosenbom AE, Ernstsen V, Fluhler H, Jensen KH, Refsgaard JC, Wydler H (2008) Fluorescence imaging applied to tracer distributions in variably saturated fractured clayey till. J Environ Qual 37:448–458Google Scholar
- Ueberschar KH, Danicke S, Wolff J (2003) Gaschromatic determination of chlormequat in silo cereals and bran. Landbauforsch Volk 53:37–42Google Scholar
- Wang JH, Lu XY, Huang M, Wu Z, Ma GP, Xu CY (2007) Rapid screening and confirmation of chlormequat and mepiquat residues in tomato and its products using liquid chromatography with electrospray tandem mass spectrometry. Chin J Anal Chem 35:1509–1512Google Scholar