JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Meyer, M. T.
Right arrow Articles by Goolsby, D. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Meyer, M. T.
Right arrow Articles by Goolsby, D. A.
Agricola
Right arrow Articles by Meyer, M. T.
Right arrow Articles by Goolsby, D. A.
Related Collections
Right arrow Water Quality
Right arrow Agricultural Pesticides
Right arrow Surface Water Quality
Right arrow Water Pollution
Journal of Environmental Quality 30:1836-1843 (2001)
© 2001 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORT
Surface Water Quality

Differentiating Nonpoint Sources of Deisopropylatrazine in Surface Water Using Discrimination Diagrams

M. T. Meyer*,a, E. M. Thurmanb and D. A. Goolsbyc

a U.S. Geological Survey, 4500 SW 40th Ave., Ocala, FL 34474
b U.S. Geological Survey, 4821 Quail Crest Place, Lawrence, KS 66049
c U.S. Geological Survey, Box 25046, Denver Federal Center, MS 406, Denver, CO 80225

* Corresponding author (mmeyer{at}usgs.gov)

Received for publication July 5, 2000.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Pesticide degradates account for a significant portion of the pesticide load in surface water. Because pesticides with similar structures may degrade to the same degradate, it is important to distinguish between different sources of parent compounds that have different regulatory and environmental implications. A discrimination diagram, which is a sample plot of chemical data that differentiates between different parent compounds, was used for the first time to distinguish whether sources other than atrazine (6-chloro-N2-ethyl-N4-isopropyl-1,3,5-triazine-2,4-diamine) contributed the chlorinated degradate, deisopropylatrazine (DIA; 6-chloro-N-ethyl-1,3,5-triazine-2,4-diamine) to the Iroquois and Delaware Rivers. The concentration ratio of deisopropylatrazine to deethylatrazine [6-chloro-N-(1-methylethyl)-1,3,5-triazine-2,4-diamine], called the D2R, was used to discriminate atrazine as a source of DIA from other parent sources, such as cyanazine (2-[[4-chloro-6-(ethylamino)-1,3,5-triazin-2-yl]amino]-2-methylpropionitrile) and simazine (6-chloro-N,N'-diethyl-1,3,5-triazine-2,4-diamine). The ratio of atrazine to cyanazine (ACR) used in conjunction with the D2R showed that after atrazine, cyanazine was the main contributor of DIA in surface water. The D2R also showed that cyanazine, and to a much lesser extent simazine, contributed a considerable amount (~40%) of the DIA that was transported during the flood of the Mississippi River in 1993. The D2R may continue to be a useful discriminator in determining changes in the nonpoint sources of DIA in surface water as cyanazine is currently being removed from the market.

Abbreviations: ACR, concentration ratio of atrazine to cyanazine • DIA, deisopropylatrazine • DEA, deethylatrazine • D2R, concentration ratio of deisopropylatrazine to deethylatrazine • GC–MS, gas chromatography–mass spectrometry


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
OVER THE LAST decade pesticide degradates have accounted for a significant portion of the total pesticide load transported in surface water (Thurman et al., 1991, 1992) and in ground water (Kolpin et al., 1996, 1998). Pesticides with similar structures may share similar degradation pathways and degrade to the same compound. To more fully understand the fate and geochemical transport processes of pesticide degradates with multiple parent sources, it is important to differentiate the contributions of each parent to a common degradate. Furthermore, in order to do accurate risk assessment for pesticide exposure to humans, it is important to know the original source of the degradates.

The concept of discrimination diagrams has been widely applied in geology (e.g., Pearce et al., 1984) and similar concepts have also been used in organic geochemistry (e.g., Hitoshi and Yoshinari, 1986; Sicre et al., 1987). For example, discrimination diagrams have been developed to distinguish between sources of hydrocarbons (Sicre et al., 1987) and humic material (Hitoshi and Yoshinari, 1986) using chain lengths of aliphatic hydrocarbons, elemental ratios, and carbon and oxygen isotopic data as discriminators. The search for dependable biomarkers to discriminate between biogeochemical sources and processes is also an important area of research in organic geochemistry. Discrimination diagrams are a means to describe and compare data sets in a more uniform way and also to more easily grasp complex concepts and processes within and among data sets. To our knowledge, the concept of the discrimination diagram has not been applied to environmental chemistry. This paper provides a first application for the construction and use of a discrimination diagram with nonpoint-source water-quality data for herbicide occurrence in surface water.

The discrimination diagram is applied to the degradation product, deisopropylatrazine (DIA), which is a degradate that may originate from atrazine (Gunther and Gunther, 1970; Thurman et al., 1994), cyanazine (Sirons et al., 1973; Meyer, 1994; Thurman et al., 1994), and simazine. The literature has shown that all three of these parents give rise to DIA (Fig. 1). This result has been discussed for atrazine and simazine (Gunther and Gunther, 1970; Mills and Thurman, 1994) and for cyanazine (Benyon et al., 1972a, b; Sirons et al., 1973; Muir and Baker, 1978; Thurman et al., 1994) but has not yet been exploited as a tool for examining nonpoint-source water-quality data. The concentration ratio of DIA to deethylatrazine (DEA), called the D2R (Mills and Thurman, 1994; Thurman et al., 1994) was developed as a discriminator and was compared with the concentration ratio of atrazine to cyanazine (ACR) to determine if nonpoint contributions of DIA not attributable to atrazine were primarily from cyanazine. The discrimination diagram is used as a tool in this paper to better understand the nonpoint-source contributions of DIA and DEA to the Mississippi River at Baton Rouge. The paper was developed in three stages. The first stage included field plot studies of the disappearance of atrazine and cyanazine and the appearance of the two degradates, DEA and DIA, in surface runoff. Second, we applied the D2R to two river basins that had high application rates of atrazine and high and low application rates of cyanazine. Little simazine was applied in either basin. Finally, the use of the D2R discrimination diagram was applied to water-quality data from the Mississippi River collected during the flood of 1993 as an indicator of DIA contributions from cyanazine and simazine.



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 1. Structure diagram showing common dealkylation pathways of atrazine, cyanazine, simazine, and propazine for the formation of deethylatrazine (DEA) and deisopropylatrazine (DIA).

 

   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Field Studies
The field study was conducted at the Kansas State University, Kansas River Valley Experimental Farm, near Topeka, KS. Three adjacent field plots (4.6 by 30.5 m) with slopes of less than 1% were used for the study. Previous dissipation studies of atrazine, but also simazine and propazine [6-chloro-N,N'-bis(1-methylethyl)-1,3,5-triazine-2,4-diamine], had been conducted at these sites in late 1989 and 1990 using manufacturer-prescribed application rates. Cyanazine was applied as Bladex (Dupont, Wilmington, DE) to two field plots and atrazine was applied as AAatrex (Ciba Geigy, Greensboro, NC) to one field plot on 20 May 1992. Both herbicides were applied with a boom sprayer at a rate of 2.2 kg/ha (approximately 32 g of active ingredient per plot). The herbicides then were incorporated to a depth of 10 cm with a disk after which corn (Zea mays L.) was planted. The plots were redisked on 18 June 1992 for weed control. The soil consisted of Eudora silt-loam (coarse-silty, mixed, superactive, mesic Fluventic Hapludoll) with a particle-size distribution of 44 to 62% silt, 26 to 50% sand, and 5 to 21% clay and an organic carbon content of 1.0% in the top 15 cm (Adams and Thurman, 1991; Meyer, 1994).

Each field plot was instrumented with a surface water collection system, and porous suction-cup lysimeters were installed in duplicate at depths of 30, 60, 90, 120, 150, and 300 cm. Prior to application soil pore water was collected from the lysimeters and 350-cm-deep soil cores were obtained from each field plot and analyzed for herbicides in 15-cm intervals. In the pore water at shallow depths only trace levels of DEA were detected and in the shallow soil cores low concentrations of DEA and atrazine. Runoff was initiated by rainfall and sprinkler irrigation. During each event surface water was continuously collected and pumped from 20-L buckets installed at the foot of each field plot into 1200-L galvanized troughs. The cumulative rainfall during the course of the study was 420 mm and the cumulative precipitation from three sprinkler irrigation events was 150 mm (approximately 50 mm per event). Sprinkler irrigation was used once in May, June, and July. The volume of surface runoff captured in the troughs from each of the three field plots during each precipitation event varied from 25 to more than 1200 L. The 4-L sediment–water samples then were collected from each trough after a runoff event. The runoff samples then were filtered through a 2-L Buchner funnel lined with a preweighed, no. 2 Whatman (Maidstone, UK) filter to separate the water from the suspended sediment. Three, 125-mL aliquots of the water samples collected from each field plot were stored in glass bottles and refrigerated until analysis. More detail on this study is given in Meyer (1994) and Thurman et al. (1994).

The Iroquois and Delaware Rivers were part of a regional storm event study of nine rivers across a five-state region of the Midwest and details of sampling are thoroughly described by Scribner et al. (1994). Briefly, successive storm events from the spring through the mid-summer of 1990 were sampled from spring 1991 through early spring of 1992. The stream sites were at U.S. Geological Survey gauging stations. The basin sizes ranged from 100 km2 to 4700 km2 (Delaware River = 1116 km2 and Iroquois River = 5416 km2). Automated water samplers were installed at each sampling site in late March and early April before the onset of herbicide application. The samplers were programmed to collect samples at a frequency sufficient to define a hydrologic event, usually every two to four hours over storm runoff events and every one to two days during nonrunoff periods. All of the samples were screened by immunoassay for atrazine and a subset of these samples then was chosen for analysis by gas chromatography–mass spectrometry.

The study of the Mississippi River, which was sampled at Baton Rouge, was one of several main-stem and tributary sites of the Mississippi River monitored in 1993. Water samples were collected at the Baton Rouge site from July 1993 to March 1994 using depth-integrated techniques at three to five locations across the width of the sampling site. For all studies water samples were collected in glass or Teflon containers, and filtered through 1-µm-pore-diameter glass-fiber filters into baked, 125-mL glass bottles. The water samples then were shipped to the laboratory on ice within 3 d after being collected, where they were refrigerated until analysis.

Analytical Methods
Atrazine, cyanazine, simazine, propazine, deethylatrazine, and deisopropylatrazine and seven other herbicides were extracted from water samples and analyzed using solid-phase extraction (SPE) and gas chromatography–mass spectrometry (GC–MS) methods of Thurman et al. (1990) and Meyer et al. (1993). Briefly, water samples were extracted in sets of 14, which included two blanks, and two check standards. A surrogate standard, terbuthylazine [2-(tert-butylamino)-4-chloro-6-(ethylamino)-s-triazine], was added to all samples. The samples were automatically extracted using a Waters Millilab Workstation (Millipore Corp., Bedford, MA). The SPE C18 cartridges were sequentially rinsed with 2 mL methanol, 6 mL ethyl acetate, 2 mL methanol, and 2 mL distilled water and the samples pumped through the cartridges at 10 mL/min. The cartridges then were eluted with 3 mL ethyl acetate and an internal standard (phenanthrene–d10) then was added followed by a transfer step to remove the ethyl acetate (top layer) from the residual water (bottom layer) in the eluate. The ethyl acetate then was evaporated to approximately 80 µL for GC–MS analysis. The recovery of all the analytes exceed 90% except for DIA, which was 60%.

Samples were analyzed using a Hewlett–Packard (Palo Alto, CA) Model 5890 GC and a 5970B mass selective detector with a direct capillary interface at 280°C; ionization voltage, 70 eV; ion-source temperature, 280°C; electron multiplier, 400 V above autotune. Samples were injected in the splitless mode into the GC. Injector temperature was 210°C. Herbicides were separated on a 12-m-long x 0.2-mm-i.d., DB-5-like column. Twenty-nine ions divided into four acquisition groups were monitored during each sample analysis. The area of the base-peak ion for each compound was divided by the single, 188 ion peak of the phenanthrene–d10 and the 214 ion peak of terbuthylazine for quantification. Compound confirmation was based upon the presence of the molecular ion, one to two confirming ions (with area counts ±20%), and a retention-time match of ±0.2% relative to phenanthrene–d10. The quantitation limit for the measured compounds was 0.05 µg/L.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
Developing the Deisopropylatrazine to Deethylatrazine Ratio
Atrazine and cyanazine were applied to separate field plots and the concentration of dealklyated degradates were measured in the runoff from each of the field plots (Fig. 2A). Mass calculations of each herbicide and DIA and DEA showed that more than 99% of the herbicide and metabolite transport was in the dissolved phase (Meyer, 1994). The two cyanazine field plots had concentrations of DIA that ranged from 1 to 3.5 µg/L during the first 35 d after application. After this time, the concentration of DIA decreased to less than 0.2 µg/L. Both replicate cyanazine field plots had much greater concentrations of DIA than DEA, about 2 to 20 times greater than DEA. Although cyanazine does not degrade to DEA, this result was expected as analysis of soil cores and shallow pore-water samples from porous suction-cup lysimeters from each field plot prior to herbicide application showed the presence of some residual DEA. This result was unavoidable, because of the previous application of atrazine over the last decade to this agricultural test farm.



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 2. (A) Temporal variation of the concentration of deethylatrazine (DEA) and deisopropylatrazine (DIA) in surface runoff. (B) Scatter plots showing the relationship of DIA to DEA for surface runoff from the cyanazine and atrazine field plots. (C) Temporal variation of the concentration ratio of DIA to DEA (D2R) for samples from the cyanazine and atrazine field plots.

 
On the atrazine field plot, however, the concentration of DEA (from 2 to 10 µg/L) was about two times greater than the concentration of DIA (from 1 to 4 µg/L) during the first 35 d after application. The concentration of DEA and DIA decreased to trace levels after approximately 60 d. The ratio of DIA to DEA was predictable, as shown by the high correlation in Fig. 2B with a regressed slope of 0.4. There was no correlation of DIA and DEA, as expected, for the two cyanazine plots (Fig. 2B). These results were consistent with previous field-plot studies that have shown that cyanazine degrades to DIA (Benyon et al., 1972a, b; Sirons et al., 1973; Muir and Baker, 1978; Blumhorst and Weber, 1992) and that atrazine degrades to both DIA and DEA (Gunther and Gunther, 1970; Skipper and Volk, 1972; Muir and Baker, 1978; Frank et al., 1991; Adams and Thurman, 1991; Mills and Thurman, 1994). Additionally, Thurman et al. (1994) obtained similar linear relationships between DIA and DEA in surface runoff from five atrazine field-plot studies with mean regressed slopes of approximately 0.40.

The concentration ratio of DIA to DEA (D2R) provides a more detailed look at the relationship between DIA and DEA over time (Fig. 2C). The D2R ranged between 0.30 and 0.70 in runoff samples from the atrazine field plot (Fig. 2C). In the early season, when the concentrations of DIA and DEA were highest, the D2R ranged only from 0.30 to 0.50; only one sample had a D2R that exceeded 0.60 during the study. A grand mean D2R of 0.38 ± 0.08 (1 SD), was obtained using data from five atrazine field-runoff studies reported by Thurman et al. (1994). These data show that the D2R is narrowly restricted when atrazine is the only herbicide responsible for the DIA and DEA transported to surface water.

Field and laboratory studies have shown that deethylation proceeds more rapidly than deisopropylation (Sirons et al., 1973; Kruger et al., 1993a, b; Mills and Thurman, 1994). For example, radiolabled soil column studies of atrazine with an Iowa silt loam by Kruger et al. (1993a)( HREF="#BIB14">b) found that the production of DIA was one-third to one-half that of DEA. Atrazine radiolabeled studies on Oregon soils by Skipper and Volk (1972) showed that the ethyl group was removed faster than the isopropyl group. A residue study of atrazine in the top 5 cm of a Perth clay-loam soil by Sirons et al. (1973) showed that the concentration of DIA was about one-third that of DEA two to three months after application. Mills and Thurman (1994) studied the dealkykation of simazine and propazine in Kansas silt- and clay-loam soils and found that deethylation occurred two to three times faster than deisopropylation. The results from this field study and previous degradation studies of atrazine studies conducted over the years indicate that the production ratio of DIA to DEA is fairly well constrained between 0.3 and 0.5. Thus, the mean D2R of 0.38 obtained from the atrazine field dissipation studies is well within estimates reported for the dealkylation reactions of atrazine over a wide range of conditions.

For the cyanazine field plots the D2R was 0 before application then rapidly increased to 20 within 20 d after application (Fig. 2C). The mean D2R of the surface-runoff samples collected from the two cyanazine field plots was about 7. Thus, mean D2R was approximately 18 times higher for the cyanazine field plots than for those obtained from this and other atrazine field plot studies. The data from the atrazine and cyanazine field plots also suggest that the D2R should increase from values of about 0.4 as the amount of cyanazine increases relative to atrazine, so that when the concentration of cyanazine reaches that of atrazine the ratio could double to 0.8, or more, as the amount of DIA increases in runoff. From these experiments and considerations, D2R ranges were established from 0.3 to 0.5 (low cyanazine application), 0.5 to 0.7 (moderate cyanazine application), and >0.7 (high cyanazine application). Next, we tested the discrimination diagram on actual river systems to see if the assumptions of low and high applications of cyanazine were correct.

The Discrimination Diagram
The Delaware River, Kansas, and the Iroquois River, Illinois, are examples of two basins with much different atrazine to cyanazine usage. Approximately 15 times more atrazine was applied in the Delaware River basin and about three times more atrazine was applied in the Iroquois Basin than cyanazine (calculated using the data of Gianessi and Puffer, 1991; previously published data in Scribner et al., 1994). The concentrations of atrazine, cyanazine, simazine, and propazine, potential sources of DEA and DIA (Fig. 1), along with stream flow for the two rivers, are shown in Fig. 3. The concentrations of simazine and propazine were generally less than 0.2 µg/L in both rivers. The concentrations of these two herbicides were much lower than atrazine in both rivers and were much lower than cyanazine in the Iroquois River and slightly lower than cyanazine in the Delaware River. The concentration data for simazine and propazine are consistent with concentrations detected in regional stream water studies conducted in the late 1980s and early 1990s (Scribner et al., 1993, 1994; Thurman et al., 1991, 1992) and national usage data (Gianessi and Puffer, 1988, 1991; Gianessi and Anderson, 1995). Based on the concentration and usage data, atrazine was considered the only significant potential source of DEA and atrazine and cyanazine of DIA.



View larger version (41K):
[in this window]
[in a new window]
  		Fig. 3. Temporal variation of the concentration of atrazine, cyanazine, simazine, propazine, and stream flow for the Delaware River, KS and Iroquois River, IL, 1990.

 
To visually examine whether the amount of atrazine relative to cyanazine affects the amount of DIA flushed into surface water and thus, test the validity of the D2R as a discriminator, another ratio was required. This was the concentration ratio of atrazine to cyanazine (ACR). This ratio will typically be a value greater than 1.0 because considerably more atrazine is applied in the Midwest on row crops (Gianessi and Puffer, 1988, 1991; Thurman et al., 1991, 1992; Scribner et al., 1993; Gianessi and Anderson, 1995).

To test the D2R discrimination diagram, the ratio of the degradates was plotted against the ratio of parent compounds for each sample and arbitrary ranges or windows were established for the predicting variable, which in this case was the ACR. For example, samples (i.e., ACR values) from both study sites were divided into four, color-coded, arbitrary groups: samples with an ACR less than 2.0; 2.0 to 5.0; 5.0 to 10; and greater than 10. The variation of the ACR for the two rivers is shown in Fig. 4A and B.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 4. (A, B) Shows the temporal variation of the concentration ratio of atrazine to cyanazine (ACR) among ACR categories for the Delaware River, KS and the Iroquois River, IL. (C, D) Discrimination diagram showing the influence of the ACR on the temporal variation of the concentration ratio of deethylatrazine to deisopropylatrazine (D2R) among discrimination fields for the Delaware River, KS and the Iroquois River, IL, 1990.

 
The ACR varied between 9.0 and 200 in May and June in the Delaware River (Fig. 4A). There was pronounced peak in the ACR as it rose from about 20 to 100 and then declined to 10 in early May. The ACR fluctuated between 20 and 200 thereafter. The ACR exceeded 10, the highest ACR category, for all but one sample. Note the dark blue points versus one light blue point. The mean ACR was 54 ± 37 for samples from the Delaware River. The high ACRs demonstrate that significantly more atrazine than cyanazine was applied in this basin.

Comparison of Fig. 4A and B shows that the temporal variation of the ACR in the Iroquois River is much lower than in the Delaware River. This marked difference between these two rivers is represented by the distribution of the colored points in Fig. 4A and B. The ACR of the Iroquois River samples varied only between 0.50 and 11. The ACR was less than 10 for 93% of the samples, less than 5.0 for 87% of the samples, and less than 2.0 for 48% of the samples (ntotal = 46; Fig. 4B). The mean ACR of the Iroquois River samples (3.2) was more that 15 times lower than the mean ACR obtained from the Delaware River samples. A Mann–Whitney U test showed a significant difference (p < 0.0001) between the ACR data sets of these two rivers.

Only atrazine was detected in samples collected from the Iroquois River before April application of herbicides. Note the absence of ACR symbols in the preapplication period (Fig. 4B). Thurman et al. (1991)( 1992) found that while cyanazine is rarely detected in stream water prior to application, atrazine is commonly detected prior to application because of carryover of herbicide from the past year and ground water contributions to surface water. Thus, the ACR in late April, marked by the first, solid vertical line in Fig. 4B, probably represents the first flush of herbicides from the fields during the onset of application. With the first large runoff event (Fig. 3) period the concentration of atrazine and cyanazine increased through early May, with atrazine rising more rapidly than cyanazine. The ACR increased from approximately 2.0 to 10.0 from late April through early May. The peak concentration of atrazine occurred with the highest ACR, marked by the vertical dashed line in Fig. 4B. From early to mid-May, during the next large storm event, the concentration of cyanazine in the Iroquois River increased, and also increased relative to the concentration of atrazine (Fig. 3), indicated by the decrease in the ACR to less than 1.0. The peak concentration of cyanazine occurred with the lowest ACR (less than 1.0), marked by the second solid vertical line (Fig. 4B). The rapid decrease and increase in the ACR during the first two large storm events in the Iroquois River may indicate differential spatial or temporal differences in the application of atrazine and cyanazine throughout the basin. During the postapplication period the ACR increased linearly from less than 1.0 to more than 5.0 at a rate of 0.038 ± 0.014 per day (95% confidence limits; Fig. 4B). The systematic increase in the ACR with time is consistent with literature findings that have demonstrated that cyanazine degrades more rapidly than atrazine (Muir and Baker, 1978; Adams and Thurman, 1991; Meyer, 1994; Mills and Thurman, 1994).

The D2R discrimination diagram was constructed with cutoffs of 0.3 to 0.5, 0.5 to 0.7, and >0.7 based on field plot studies as previously discussed (Fig. 4C,D). When the ACR in the Delaware River was greater than 10 (dark blue squares in Fig. 4C) the D2R was generally less that 0.50. For example, the D2R of 81% of the samples was less than 0.50 and for 94% of the samples was less than 0.55. Only one sample had a D2R greater than 0.60. The mean D2R of 0.44 ± 0.08 obtained from the Delaware River samples was just slightly higher than the mean D2R obtained from the atrazine field-plot studies (D2R of 0.38). Thus, the distribution of the D2R from the Delaware River was similar to that predicted from the analysis of the atrazine field-plot studies and the previously discussed studies of atrazine degradation.

Figure 4C also shows that the D2R did not fall below 0.30, indicating that in basins where atrazine is the primary triazine herbicide applied, approximately a minimum of 3 molecules of DIA is flushed from the fields for every 10 molecules of DEA (mass ratio was used rather than mole ratio but the molecular weights are within 10% of each other). The temporal variation of the D2R in the Delaware River exhibits some subtle fluctuations associated with large changes in the ACR that occurred in early May and early June (Fig. 4C). However, the overall distribution of the data set indicates that the relationship between DIA and DEA was mostly tied to normal fluctuations in the production and degradation of DIA and DEA resulting from the degradation of atrazine.

Comparison of the D2R discrimination diagram (Fig. 4C,D) shows that the variation of D2R from the Iroquois River samples is much higher than the Delaware River samples. The D2R of 44% of the samples was greater than 0.70, and 77% of the samples was greater than 0.50 (n = 43). The color-coded scheme shows the influence of the ACR on the D2R (Fig. 4) and confirmed the hypotheses developed from the field dissipation studies that as the amount of cyanazine relative to atrazine increases, the D2R increases.

For example, for samples with an ACR greater than 10 the D2R was less than 0.5. As the ACR decreased to less than 5, the D2R generally increased to greater than 0.5. Also as the ACR decreased to less than 2 the D2R generally increased to greater than 0.7 and always exceeded 0.5. This result is shown by comparing the ACR and D2R color codes for the high-ACR dark blue dots with the low-ACR red dots. A Mann–Whitney U test also showed a significant difference (p < 0.0001) between the D2R data sets of the Delaware and Iroquois Rivers. The mean D2R of the Iroquois River was 0.65. A simple estimate of the percentage of DIA transport that was due to cyanazine in the Iroquois River was made using the following calculation:

where the mean atrazine D2R is 0.38. Based on this calculation the amount of DIA transported in the Iroquois River was due to cyanazine. This calculation shows that cyanazine accounted for about 40% of the DIA transported in the Iroquois River during this study, whereas in the Delaware River atrazine was the overwhelming contributor of DIA. These data indicate that the D2R is an effective discriminator of atrazine sources of DIA from other parent sources and that the D2R discrimination diagram can be extrapolated from field-scale to watershed-scale studies. Finally, we tested the discrimination diagram on a regional scale using data from the Mississippi River at Baton Rouge, LA.

Discrimination Diagram for the Mississippi River
The usefulness of the discrimination diagram was lastly used to determine sources of DIA transported to the Mississippi River during the 1993 floods and to test the validity of the D2R discrimination diagram on a regional-scale water-quality problem. The Mississippi River at Baton Rouge, LA drains two-thirds of the Midwest. The ACR in the Mississippi River at Baton Rouge rose from approximately 2.0 to 8.0 between July and October of 1993 (Fig. 5A). The flat ACR signature (1.9 to 3.0) during July indicates that the Mississippi River was transporting water flushed from areas where atrazine and cyanazine had been actively applied or reapplied between late June and mid-July because of flooding in the Missouri River Basin. Two significant linear increases in the ACR occurred in August. In the first, the ACR increased from approximately 2.0 to 4.0 at a rate of 0.079 ± 0.028 per day (95% confidence limits, r2 = 0.91, n = 7). After an abrupt decrease, the ACR then increased from approximately 2.0 to 8.0 at a rate of 0.122 ± 0.017 per day (95% confidence limits, r2 = 0.96, n = 12) from mid-August into September. The two increasing ACR events indicate that water was flushed from two different low-ACR sources that were undergoing postapplication degradation, which is similar to what occurred in the Iroquois River (Fig. 4B).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 5. (A) Temporal variation of the concentration ratio of atrazine to cyanazine (ACR) among ACR categories. (B) Discrimination diagram showing the influence of the ACR on the temporal variation of the concentration ratio of deethylatrazine and deisopropylatrazine (D2R) among discrimination fields for the Mississippi River at Baton Rouge, LA, 1993.

 
The ACR fluctuated between 6.0 and 8.0 from mid-September into December, indicating that atrazine was the predominant parent triazine herbicide being transported in the Mississippi River. The flat late-season ACR signature indicates that water transported in the Mississippi River was probably from ponded water (e.g., reservoirs) and not from water actively flushed from the fields by flooding. The ACR should have continued to increase in water flushed from fields where active postapplication degradation would occur because of the faster degradation rate of cyanazine than atrazine.

The D2R discrimination diagram (Fig. 5B) shows that the D2R varied between 0.6 and 0.8 in early July and steadily increased from about 0.6 to 0.8 through early August. The D2R then decreased to about 0.6 through mid-September. The D2R signature then flattened at around 0.6 into November, corresponding to the flattening of the ACR. The discrimination diagram shows that cyanazine was contributing a substantial amount of the DIA transported in the Mississippi River through the spring and summer floods and well into the fall. Even as the ACR increased to more than 5, atrazine was not the sole source of DIA in the Mississippi River. The mean D2R was 0.68 and 0.59 for samples with an ACR less than 5.0 and greater than 5.0, respectively, indicating that the ACR was affecting the DIA contributions from cyanazine relative to atrazine. A Mann–Whitney U test did show a significant difference (P = 0.0002) between the distribution of the D2R for samples associated with these two ACR groups confirming this observation. Thus, as with the field dissipation studies and the watershed scale studies of the Delaware and Iroquois Rivers, the D2R discrimination diagram effectively discriminated sources of DIA in the Mississippi River.

It is estimated from the D2R that cyanazine contributed approximately 40% of the DIA transported in the Mississippi River from July through early September, and approximately 30% of the DIA transported in the Mississippi River from September to November. Thus, the ACR diagram and D2R discrimination diagram can provide evidence of past atrazine to cyanazine application practices long after the growing season. The calculations of DIA contributed by cyanazine are reasonable based on the usage of cyanazine in the basin relative to atrazine (approximately twice as much atrazine to cyanazine, 60 million pounds versus 30 million pounds; Gianessi and Puffer, 1988, 1991; Gianessi and Anderson, 1995).

This study shows that the D2R is an effective discriminator of parent sources of DIA in surface water and also demonstrates that discrimination diagrams can be useful in assessing contaminant transport processes from field-plot, basin, and regional-scale studies. The D2R discrimination diagram may also be useful in determining regulatory compliance. For example, in Wisconsin, regulatory compliance for the USEPA maximum contaminant level of atrazine (3.0 µg/L) is done by summing the concentrations of atrazine and its degradates DEA and DIA (Wisconsin Ground Water Act 410 [1983], Rule under the Law, Enforcement Standard, Chapter NR 140, Wisconsin ADM CODE [1991]). Thus, the discrimination diagram may be used to ascribe the relative amount of DIA that is due to atrazine from that of cyanazine and other parent sources of DIA. In addition, because of the proposed phase-out of cyanazine usage by the year 2000, the D2R discrimination diagram may also be useful in determining the length of residual inputs of DIA from past cyanazine. Finally, D2R may be used in conjunction with the ACR to determine whether cyanazine usage relative to atrazine usage is decreasing during the cyanazine phase-out period.


   ACKNOWLEDGMENTS
 
The authors wish to thank Gail Mallard, formerly of the U.S. Geological Survey's Toxic Substances Hydrology Program, for funding of this work and Phil Barnes of the Kansas State University Silver Lake Experimental Farm for support of the field-plot studies.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
The use of trade names is for identification purposes only and does not imply endorsement by the U.S. Geological Survey.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
R. F. Spalding, D. G. Watts, D. D. Snow, D. A. Cassada, M. E. Exner, and J. S. Schepers
Herbicide Loading to Shallow Ground Water beneath Nebraska's Management Systems Evaluation Area
J. Environ. Qual., January 1, 2003; 32(1): 84 - 91.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Meyer, M. T.
Right arrow Articles by Goolsby, D. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Meyer, M. T.
Right arrow Articles by Goolsby, D. A.
Agricola
Right arrow Articles by Meyer, M. T.
Right arrow Articles by Goolsby, D. A.
Related Collections
Right arrow Water Quality
Right arrow Agricultural Pesticides
Right arrow Surface Water Quality
Right arrow Water Pollution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome