PEST MANAGEMENT: WEEDS

Environmental Implications of Pesticides in Rice Production J.D. Mattice, B.W. Skulman, and R.J. Norman ABSTRACT For the past three years we have collected and analyzed water from four sites each on the L’Anguille and St. Francis rivers from near Jonesboro in the north to near Marianna in the south. This past year we included four sites on Lagrue Bayou from just below Peckerwood Lake north of Stuttgart to near the mouth southeast of DeWitt. Over this period the most frequently detected compounds were molinate (Ordram), quinclorac, (Facet), and clomazone (Command). Most (71-78%) of the detections that were over 2 ppb were less than 5 ppb, but each year there were between one and three detections of a compound in the 30 to 50 ppb range. The L’Anguille River consistently has the most detections with most of those coming at the two most upstream sites. There is no trend for the overall frequency of detections over 2 ppb (9.2 % in 2000, 12.0% in 2001, and 5.2% in 2002). INTRODUCTION Some rice pesticides have been found to persist in surface waters in California. This project is to determine if there is a persistence problem with rice pesticides or if they are being found more frequently in Arkansas waters. Monitoring for pesticides in water may allow us to detect a potential problem and address it before it becomes a major problem. In the past, we have analyzed water collected from the Mississippi River both as it enters and leaves the state, in addition to water from one site each on the L’Anguille, St. Francis, Arkansas, and White rivers. Most of the detections occurred in the L’Anguille and St. Francis rivers. The higher frequency of detections in these rivers is

135

AAES Research Series 504

likely due, at least in part, to the smaller volumes of water they have, resulting in less dilution of any compounds that are present. Also, a large percentage of the water in the other rivers comes from areas outside the rice-growing region which might not contain pesticides used in rice production. For these reasons, these two smaller rivers would be more sensitive indicators of changes occurring in either pesticide concentrations or frequency of detections. Therefore, beginning in the year 2000 we sampled only these two rivers and we sampled them more intensively by collecting water from four different sites on each river from near Jonesboro in the north to near Marianna in the south. This past year, in response to a request from the Rice Research and Promotion Board to analyze water from further south, we added four sites on Lagrue Bayou from just below Peckerwood Lake north of Stuttgart to near the mouth southeast of DeWitt. PROCEDURE Sampling Sites Surface water samples were collected at eight locations during 2000 and 2001 and at twelve locations in 2002. Four samples were taken from the L’Anguille River: where it crosses highways US 79 near Marianna, US 64 near Wynne, State 14 near Harrisburg, and near Claypool reservoir north of Harrisburg. Four samples were taken from the St Francis River: where it crosses US 79 near Marianna, US 64 near Parkin, State 75 near Marked Tree, and State 18 east of Jonesboro. In 2002 an additional four samples were taken on Lagrue Bayou: at a county road approximately a quarter mile below Peckerwood Lake, the second bridge on highway 146 west of the highway 33 junction, near the town of Lagrue at highway 33 before the junction with highway 153, and where the Lagrue crosses highway 1 outside of DeWitt (Fig. 1). Sampling Procedure Water samples were collected and extracted onto C18 Speedisks using a mobile field extractor which allows us to extract the samples immediately after collecting them while we are driving to the next site. At one site on each river four replicate subsamples were collected. For quality control, two subsamples were fortified with known amounts of the compounds and two were left unfortified. Analysis of these samples allowed us to verify recovery and reproducibility. Sampling was performed at 2-week intervals during the rice production season from May through August. The extracted samples were transported to the lab on ice where they were eluted for analysis. Analysis Procedure Pesticides selected for monitoring were Bolero (thiobencarb), Facet (quinclorac), Garlon (triclopyr), methyl parathion, Ordram (molinate), Sevin (carbaryl), Command (clomazone), Blazer (acifluorfen), Pursuit (imazethapyr) and 2,4-D. A 500 mL aliquot of

136

B.R. Wells Rice Research Studies 2002

each sample was extracted onto C18 disks in the field with the mobile extractor using conventional C18 disk technology. The disks were stored on ice and eluted on return to the lab. Samples were then analyzed by gas chromatography mass spectrometry (GCMS) and high performance liquid chromatography (HPLC). RESULTS AND DISCUSSION In order to make comparisons from year to year, the original cut-off point of 2 ppb is used, although for all compounds we can detect lower levels. Our rationale is that it would not be surprising to find low levels of compounds in runoff water adjacent to fields where the compounds are used, especially with the sensitive analytical equipment that is now available. Trying to find meaningful trends in frequency of detection when looking at changes in small fractions of a part per billion concentration in water would be difficult. There will be variability, but not necessarily meaningful variability in the sense of identifying a developing problem. Our objectives are to see if over years there is an increase 1) in the frequency of detections above what could be considered to be background in these samples at these locations at these times, and 2) in the levels of the highest concentrations. Since these are river water samples from small rivers surrounded by rice fields, not drinking water, the 2 ppb level would be reasonable for making comparisons. All the detections from 2002 are listed in Table 1. Table 2 lists the frequency of detections for each year from 2000 to 2002. There are more possible detections in 2002 because Lagrue Bayou was added to the list of rivers to sample. There is no trend in the percent of samples that have at least one compound in them. The most frequently detected compounds each year were molinate (Ordram) in 39% of the samples in 2000, quinclorac (Facet) in 36% of the samples in 2001, and quinclorac in 28% of the samples in 2002. The three compounds detected most often each year include molinate, quinclorac, and clomazone (Command). The distribution of concentrations is given in Table 3. The concentration distribution has been quite consistent over the 3-year period with between 71 and 78% of the samples that contained a pesticide having concentrations between 2 ppb and 5 ppb. Each year there were 1 to 3 detections of a compound between 30 to 50 ppb. Table 4 shows the distribution of detections by site and by river. The L’Anguille consistently has the most detections and the number of detections has been fairly constant over the 3- year period. The L’Anguille, especially the upper portion, is completely surrounded by rice fields, so virtually all the water is coming from areas under commercial rice production. The number of detections in the St. Francis has been quite variable over the 3-year period. Site E had 11 detections in 2001 which tied with two other sites for the largest number of detections in the first 2 years; however, in 2002 there were no detections over 2 ppb at this site. Site F at Marked Tree has consistently had the lowest or tied for the lowest number of detections. Site F had only two detections in 2000 and four detections in 2001 and none in 2002. A likely reason for this is that the St. Francis Sunken Lands Wildlife Management Area is approximately 1 mile up-

137

AAES Research Series 504

stream from the site and extends up to the Missouri border. This not only provides a buffer zone where there is no agriculture close to the river that could provide immediate runoff, but also because of the slow, meandering nature of the river here, there is ample opportunity for any pesticides that come in upstream to degrade before reaching the sampling site. Both years all detections at this site were in the lowest concentration range (2 to 5 ppb). Table 5 shows the number of samples with more than one compound. The number of samples with more than one compound increased in 2001, but decreased in 2002 to results very similar to what was seen in 2000. Detection of the same compound at the same site on consecutive sampling periods could indicate that the compound is being continually introduced into the river, as opposed to a limited, intermittent introduction. Table 6 shows when and where there were consecutive detections of a compound. Most of the detections of the same compound on consecutive sampling times were on the upper L’Anguille. SIGNIFICANCE OF FINDINGS It is not surprising to find some pesticides in surface water in an agricultural area during the growing season. Most of the detections have been of low level and sporadic. Exceptions for being sporadic would be for clomazone (Command) in the first part of the sampling season and for quinclorac (Facet) in the middle part of the season (Table 6). These compounds were detected fairly frequently, but usually at low levels. In 2000, the most frequently detected compound was molinate; it was the third most frequently detected compound in 2001 and 2002. Also, in 2000 the 10 highest concentrations were for molinate. In 2001, only two of the ten highest concentrations were for molinate, and there were five compounds represented in the ten highest concentrations. In 2002, there were three compounds represented in the ten highest concentrations (molinate with three, clomazone with three, and quinclorac with four). EPA does not have guidelines on acceptable levels for most of these compounds in either the National Recommended Water Quality Criteria - Corrected (1999) or the 2002 Edition of the Drinking Water Standards and Health Advisories (2002). There was a listing of 70 ppb for the Maximum Contaminant Level (MCL) for 2,4-D in their drinking water standards. The highest level we found in river water was 2.7 ppb. The California Department of Pesticide Regulation has a performance goal of 10 ppb for molinate. The performance goal is a guide that is not enforceable, but is a level at which there can be toxic affects to some test species. In a personal call to the project leader in California she likened it to a canary in a mine situation - a reason to be watchful. In the past three years we have had 11 detections of molinate above 10 ppb (2 in 2000, 8 in 2001, and 1 in 2002); all but one were on the L’Anguille. Our results are similar to those reported by the California Department of Pesticide Regulation (DPR) in their Rice Pesticides Program Monitoring Data August 20, 2002-Final Update (2002). They have had periodic detections of molinate above the performance goal, and the concentrations were similar to what we have found.

138

B.R. Wells Rice Research Studies 2002

Since there are no specific guidelines for tolerances for most of these compounds, and since we are not aware of any environmental problems that are occurring in these rivers, we have no reason to say there is a problem. However, based on the performance goals used in California, there is a reason to be watchful. The concentrations of the compounds appear to be similar each year over the 3-year period, but the frequency of detections can fluctuate. ACKNOWLEDGMENTS We would like to acknowledge the Arkansas Rice Research and Promotion Board for funding this project. LITERATURE National Recommended Water Quality Control Criteria - Correction, USEPA, Office of Water, 4304, Washington DC, April 1999. 2002 Edition of the Drinking Water Standards and Health Advisories, EPA 822-R-02038, Summer 2002, USEPA, Office of Water, Washington DC. Rice Pesticides Program Monitoring Data, August 20, 2002 - Final Update, KayLynn Newhart, California Department of Pesticide Regulation, Environmental Monitoring Branch, 1001 I Street, Sacramento, CA 95812

Fig. 1. Sampling sites for the 2002 water monitoring program.

139

Table 1. Results for the year 2002 water samples that contain at least one detection of a pesticide. Limit of quantitation of 2 ppb. Pesticides detectedy Date

River

Site

z

molin

clomaz

quinc

triclopyr

2,4-D

aciflu

------------ (ppb corrected for percent recovery) ---------15 May

28 May 11 June

25 June

9 July

23 July

6 Aug 20 Aug z

y

x

L’Anguille L’Anguille L’Anguille St. Francis Lagrue Lagrue L’Anguille L’Anguille Lagrue L’Anguille L’Anguille L’Anguille St. Francis Lagrue Lagrue Lagrue L’Anguille L’Anguille L’Anguille L’Anguille St. Francis Lagrue Lagrue L’Anguille L’Anguille L’Anguille L’Anguille St. Francis Lagrue L’Anguille L’Anguille L’Anguille St. Francis L’Anguille

A B C G L M A B N A B C G L M N A B C D H M N A B C D H N A B C H B

5.0 -- x -5.2 -5.3 ---3.1 ---3.8 -2.6 ---38.3 -3.2 -2.2 --------6.4 13.4 -3.9 ----------33.0 -------2.9 ---3.3 -7.3 ---2.8 -----------------------------

--------2.9 ---2.8 ---5.8 ---2.6 -----------2.9 ---6.6 -------2.7 -----2.4 -2.3 ---2.9 ---3.0 ---3.5 -2.7 5.0 6.3 ---2.5 ---2.6 ---2.3 ---2.0 ---3.4 ---5.4 -2.2 -3.3 ---5.3 ---2.0 ---2.4 ---2.9 ----2.3 --2.4 ---2.4 ---3.1 ---2.0 ---no detections ---------------------------

A-D - L’Anguille upstream to downstream; E-H - St. Francis upstream to downstream; K-M Lagrue upstream to downstream, see text for details. molin = molinate (Ordram); clomaz = clomazone (Command); quinc = quinclorac (Facet); triclopyr (Garlon); 2,4-D, aciflu = acifluorfen (Blazer). ‘--’ = not detected. Table 2. Frequency of detections over 2 ppb of pesticides in water by year.

Year

2000

2001

2002

Possible detections Detections Percent

576 53 9.2

565 68 12.0

958 49 5.1

140

Table 3. Concentration distribution of pesticides in water by year. Distributionz Range

2000

2001

2002

--------------------------------- (ppb) --------------------------------2-5 5-10 10-20 20-30 30-40 40-50 z

38 7 3 2 1 2

(72%) (13%) (6%) (4%) (2%) (4%)

48 13 4 2 0 1

(71%) (19%) (6%) (3%) (0%) (2%)

38 8 1 0 2 0

(78%) (16%) (2%) (0%) (4%) (0%)

Percents may not total to 100 due to rounding to nearest percent.

Table 4. Detection frequency of pesticides in water over 2 ppb by site. L’Anguille

St. Francis

Lagrue

Site

2000

2001

2002

Site

2000

2001

A B C D Total

8 8 9 7 32

9 11 7 8 35

14 9 7 2 32

E F G H Total

5 2 7 7 21

11 4 7 11 33

2002 0 0 2 5 7

Site

2002

K L M N Total

0 3 3 4 10

Table 5. Multiple detections of pesticides in water over 2 ppb per sample. Compound in sample

2000

2001

2002

28 11 1 0 0

18 7 6 2 2

23 9 1 0 1

1 2 3 4 5

Table 6. Consecutive detections of a given pesticide by site. Date 15 May 28 May 11 June 25 June 9 July 23 July 6 Aug z

molinate

A A A

clomazone Az A A A

B C B C B B

H H

quinclorac N A C MN A B C D HMN A B C D H N B C H B

2,4-D

A A

A-D = L’Anguille; E-H = St. Francis; and K-N = Lagrue.

141