Tài liệu WATER POLLUTION BY INSECTICIDES IN AN AGRICULTURAL RIVER .BASIN. I. OCCURRENCE OF INSECTICIDES IN RIVER AND TREATED MUNICIPAL WATER pdf

Water was sampled by the carbon adsorption method at a municipal water treatment plant.. Samples taken in the summer of 1960 from the stream system indicated the same order of contaminat

WATER POLLUTION BY INSECTICIDES IN AN AGRICULTURAL

H Page Nicholson, Alfred R Grxenda, Gerald J Lauer,

William S Cox, and John I Teasle y

U.S Department of Health, Education, and Welfare, Public Health Scrvicc

Division of Water Supply and Pollution Control, Region IV, Atlanta, Georgia

ABSTRACT

Insecticide contamination was studied in a stream from the summer of 1959 through the winter of 1962-63 The stream drains a 400-square-milt (1,036-km’) watershed in northern Alabama, in which approximately 15,000 acres (6,070 ha) of cotton are culti- vntcd annually Estimates of insccticicle usage on cotton varied from 58,000 lb technical (26.3 metric tons) in 1960 to 139,000 lb (63.0 metric tons) in 1962 Toxaphcne, DDT, and BHC (benzcnc hcxachloride) comprised over 90% of the insccticidcs used

Water was sampled by the carbon adsorption method at a municipal water treatment plant ‘The data are reported as concentrations of insecticide recovered

DDT was ncvcr detcctcd, but toxaphcne and BHC wcrc detected in all samples Toxa- phcnc ranged from 7 to 410 parts per trillion ( ppt ), and BHC from 7 ppt to 1.0 part per billion (ppb) Samples of treated and untrcatcd water showed that the purification proc- c‘ss failed to rcmovc toxaphene and BIIC Samples taken in the summer of 1960 from the stream system indicated the same order of contamination observed at the water treatment plant, The data suggest that soil-persistent insecticides in the stream are related to levels

of usage and to the solubility of the compounds

INTRODUCTION Certain insecticides have been recov-

ered from the major waterways of the

United States Middleton and Lichten-

berg (1960) dctccted DDT in the Missis-

sippi River at Quincy, Illinois, and at New

Orleans; in the Missouri River at Kansas

City; and in Lake St Clair and the Detroit

River They detected aldrin in a sample

from the Snake River near Pullman, Wash-

ington Other evidence suggests the oc-

currence of a large variety of insecticides

in water resources It has been estimated

that during the past 3 years agricultural

poisons were second only to industrial

wastes as a cause of fish-kills in the United

States (U.S Public Health Service 1960,

1961, 1962) However, little work has

been done, particularly in the cases of

apparently sublethal contamination, to de-

termine the magnitude and seasonal peri-

odicity of pollution associated with insec-

ticide usage

This paper discusses the levels of con-

tamination over a nearly 4-year period in a

stream system and a potable water supply

located in an agricultural basin Biolog-

ical aspects of the study arc reported by Grzenda, Lauer, and Nicholson ( 1964) The study area is a watershed in north-

cm Alabama of about 400 square miles ( 1,036 km”) drained by approximately 760 miles ( 1,224 km) of streams tributary to the Tennessee River Flint Creek and the West Fork of Flint Creek are the princi- pal streams in the basin Their junction is

a few hundred meters upstream from a conventional water treatment plant draw- ing from the stream and serving 7,000 rcsi- dents of I-Iartselle and Flint, Alabama This plant receives drainage from almost the entire basin The stream receives treated sewage from Hartselle and no in- dustrial waste The land is generally roll- ing, with some slopes of 15 60% in the headwater areas Soils are loams of lime- stone or sandstone origin, and an estimated 40% of the basin is forested

The following agencies assisted in this project: the Alabama Department of Pub- lic Health and Water Improvement Com- mission; the Hartselle City Council and

Department; the Tennessee Valley Author-

310

INSECTICIDE POLLUTION IN A RIVER: THE WATER 311 ity; the Agricultural Stabilization and Con-

servation Agents and Agricultural Agents

of Cullman, Lawrence, and Morgan coun-

tics; and the U.S Forest Service Messrs

Coy Stephenson and L J Sandlin of the

Hartselle Water Treatment Plant main-

tained and operated the sampling equip-

ment Some of the work was performed

under the direction of Dr H J Webb

under the provisions of a contract between

the U.S Public Health Service and Clcm-

son College

METIIODS

Sample collection

Insecticides were recovered from water

scribed by Rosen and Middlcton ( 195s)

This procedure permits recovery of dis-

solved organic compounds from a volume

of water ( approximately 19 rn” ) large

enough to yield sufficient extract for chem-

ical analysis and bio-assay This concentra-

tion procedure was required up to the fall

of 1962 because existing chemical tech-

niqucs and available cquipmcnt were not

sensitive enough to detect tract quantities

of insecticides in whole water samples

Sampling units were installed at the mu-

nicipal water plant to sample both un-

treated and treated water The paired

samples were used to determine the cffi-

cacy of the treatment process

The contents of the carbon cylinder

(400 f 5 grams) were dried with a stream

of unheated air passed over shallow trays

containing the sample and then extracted

with chloroform for 24 hr in a Soxhlet ap-

paratus The chloroform was then cvap-

orated, the rcsiduc taken up in acetoni-

trile, and the insecticides partitioned into

petroleum ether The ether fraction was

washed with distilled water and elutriated

through a sodium sulfate column This

solution was transferred to a florisil col-

umn and the residue elutriated with mixed

ethers ( 15% diethyl ‘and 85% petroleum),

The ethers were volatilized and the resi-

due transferred to a volumetric flask with

benzene Aliquots of this solution were

used for analysis

Analytical methods Early in the study, analyses were accom- plished by fish bio-assay supplcmentcd by infrared spectroscopy and paper chroma- tography, These methods were not en- tirely satisfactory The samples generally contained too many contaminants to per- mit satisfactory use of infrared spcctros- copy Toxaphene, a mixture of related compounds, does not produce characteris- tic diagnostic peaks on the instrument at the levels encountered Toxaphenc and BHC (benzene hexachloride ) could not be differentiated by bio-assay

In late 1962, a gas chromatograph with

a microcoulometric titration detection sys- tem was obtained The gas chromato- graph assembly consisted of the following:

a Micro Tek 2530R Chromatographic unit;

combustion unit, C-100 coulometer, and T-200 microcoulometric titration cells; and

a Brown Y153X recorder cquippcd with a disc integrator This system is specific for chlorinated hydrocarbon or thiophospho- rus compounds, depending on the titration cell used A discussion of microcoulomct- ric gas chromatography is given by Cassil ( 1962) All residues remaining from ear- licr samples wcrc reanalyzed using this

mc thod

The isomers of BIIC were qualitatively identified by thin-layer chromatography using the method given by Baumler and Rippstein ( 1961) , modified as follows: the immobile phase consisted of aluminum oxide with calcium sulfate as a binder ap- plied to glass plates at a thickness of 250 p; the color-developing reagent was a mix- ture of silver nitrate in water, 2-phcnoxy- ethanol, and acetone

Survey methods Rainfall, stream turbidity, and insccti- cidc usage data were obtained to deter- mint their influence on the quantities of insecticides recovcrcd from the samples Daily rainfall records were obtained from five weather stations ( Fig 1) operated by various federal agencies These data were used to compute avcragc rainfall, by sea-

312 I-1 I? NICHOLSON, A R GRZENDA, G J LAUER, W S COX, AND J I TEASLEY

FIG 1 A map of the Flint Creek Basin show-

ing the sites where carbon adsorption samples

(CAS) were taken ancl the location of biological

stations

sons, using the Thiessen method (Foster

1949) Cotton acreage data were obtained

from three County Agricultural Stabiliza-

tion and Conservation offices of the U.S

Department of Agriculture These data

are determined annually by photogram-

metry and site inspection Each fall, from

1959 to 1962, a stratified random sample

was drawn representing from 10 to 17% of

the total cotton acreages Stratification was

based on the proportional allocation of

cotton acreage found in the part of each

county occurring within the basin Needed

information was obtained by personal in-

terview Confidence limits (0.05) were set

on the estimates by using the sample

ureas treated in the Flint Creek Basin DDI’ usage in the Flint Creek Basin

Year plant& Arca

in

cotton

Hectares Area treated*

Insecticide use in metric tons (technical) j

1959 6,(36,2 3,750 ( -t- 154 ) 38.8( Z!I 2.3)

1960 6,062 2,057(-t- 96) 26.5(-1 1.8)

1961 6,541 4,029( -c 108) 45.1( f 2.1)

1962 6,630 5,575( -t- 81) 63.3( IL 1.8)

* The 95% confidence limit on the estimate is given in

parentheses

t The fraction of the total contributed by BHC is

standard error, the Student “t,” and a fi- nitc population correction term

Agricultural statistics Diversified agriculture is practiced in the area, but cotton is the only crop re- quiring significant quantities of insecti- cides The cotton farms are small, and only a few exceed 100 acres (40 ha) The surveys indicate that from 1959 to 1962, there were from 14,980 to 16,382 acres ( 6,062-6,630 ha) of cotton cultivated in the basin Average cotton acreage per farm increased from 10.0 acres (4.0 ha) in

1959 to 12.7 acres (5.1 ha) in 1962

Most of the insecticides are applied as dusts by ground equipment, generally from June through the middle of Septem- ber However, about 60% (by weight) is used from mid-July to mid-August Dur- ing the study period, farmers using insec- ticide made, on the average, from 3.2 to 4.8 applications per growing season Estimates of the total weight of insccti- tides (technical grades) used and the area treated are given in Table 1 Each year,

over 90% of the total In 1959, 1960, and

1961, the approximate relative usage was

as follows : toxaphene, 65% of the total tonnage estimate; DDT, 27%; and gamma BHC, 4.5% In 1962, DDT comprised 35%

of the total tonnage estimate, toxaphcne

Other pesticides used during one or more

Usage estimate in metric tons (technical)* Year

Toxaphene DDT Gamma BHCP

1959 25.6( + 2.1) 10.4( z!z 0.7) 1.7( ?I 0.2)

1960 17.2( + 1.7) 7.6( 2 0.8) 1.2( L 0.1)

1961 29.3( ?I 1.5) 11.9(-t- 0.9) 2.1( + 0.3)

1962 32.7( f 1.4) 22.0( k 0.7) 3.2( * 0.2)

* The 95% confidence limit on the estimate is given in parentheses

t The ratio of gamma isomer to other isomers differs among manufacturers Therefore, depending on the brands used by the farmers, the total quality of BHC used in the basin could vary approximately from two to

INSECTICIDE POLLU’l1ON IN A RIVER: THE WATER 313

TABLE 3 Mean concentration of toxaphene recouered from Flint Creek, Alabama, at the Hartselle

Water Treatment Plant by the carbon adsorption method

Period

Summer

1959

1960

1961

1962

Fall

1959

1960

1961

1962

Winter

1959430

1960-61

1961-62

1962-63

Spring

1960

1961

1962

Number of snmplcs

5

3

12

1:

3

No samples

No samples

Parts per trillion

of the years were aldrin, endrin, dieldrin,

Sevin, Malathion, parathion, methyl para-

thion, Trithion, Methyl Trithion, Guthion,

and sulfur The use of insecticides as

emulsifiable concentrates was estimated to

increase from 5,844 lb technical (2.6 met-

ric tons ) in 1959 ( 6.8% of the total) to

26,294 lb (11.9 metric tons), or 18.8% of

the total in 1962 This increase was inde-

pendent of the total quantity used in the

basin

Forestry practices

About 24 square miles ( 62 km2) of the

basin are in the William B Bankhead Na-

tional Forest This area is on the slopes of

the divide in the southwest portion of the

watershed In 1961, 1,210 acres (490 ha)

of the forest within the Flint Creek Basin

were treated, by mistblowing, with 3,600

lb or 1.6 metric tons ( acid equivalent) of

the isoctyl ester of 2,4,5-T (2,4,5-trichloro-

phenoxyacetic acid) for the control of

hardwoods.1 In 1962, the hardwood stems

1 Personal communication, U.S Forest Service,

Haleyville, Alabama

growing on 860 acres ( 348 ha) were treated individually, by hand, with an un- determined quantity of the same herbi- cide

Insecticide recovery

Microcoulometric gas chromatography was used to estimate the recovery of toxa- phene and gamma BHC from carbon ad- sorption samples Toxaphene and gamma BHC were added to 200 liters of tap wa- ter so that each was present in the same sample at a concentration of 0,5 or 1.0 part per billion ( ppb ) The solutions were passed through standard 8- x 46-cm carbon columns at 2 liters/min Three rep- licates were made at each concentration The mean recoveries and ranges for the 0.5-ppb level were toxaphene, 42% (39 47% ), gamma BHC, 43% (3846%); and, for the LO-ppb solutions, toxaphene, 48% (42-51% ), gamma BHC, 48% ( 43-54%) Analysis of water samples collected from the storage vessels and the column effluent indicated that the toxaphene and BHC were not lost prior to passage through the carbon column and were totally removed

314 H I? NICHOLSON, A R GRZENDA, G J LAUER, W S COX, AND J I TEASLEY

TABLE 4 Mean concentration of BHC recovered from Flint Creek, Alabama, ut the Hartselle Water

Treatment Plant by the carbon adsorption method

Parts per trillion Period

Summer

1959

1960

1961

1982

Fall

1959

1960

1961

1962

Winter

1959-60

1960-61

1961-62

1962-63

Spring

1960

1961

1962

Number of samples Untreated Treated water water

4

No samples

No samples

Mean concentration Range Untrcatcd Treatrcl Untreated Trcatcd water water water wntcr

from water by the carbon The field data

were not corrected for the losses indicated

by the above laboratory tests They are

calculated concentrations based on the

sample volume and the quantity of insec-

ticide recovered

Toxaphene was present in all samples

reanalyzed by gas chromatography (Ta-

blc 3) Although there were differences

between the chemical analyses and the

bio-assays for toxaphene, the range of the

results was of the same magnitude, 5-410

parts per trillion (ppt) and l-433 ppt, re-

spectively

springs of 1961 and 1962 were exhausted

before the gas chromatograph was avail-

able Although paper chromatography and

infrared spectroscopy indicated that BHC

or other chlorinated hydrocarbon insccti-

tides were not present in concentrations

toxic to fish, the residues did kill fish and

produced a syndrome described by Hen-

derson, Pickering, and Tarzwell (1959) as

being associated with chlorinated hydro-

carbon poisoning Since infrared spcctros-

copy and paper chromatography do not give a definitive indication of toxaphenc, the above is presumptive evidence that toxaphene was present in these samples,

In summary, microcoulometric gas chro- matography indicated toxaphene in all samples collected from the summer of

1959 to the winter of 1962-63 except those not available for reanalysis (spring 1961 and 1962 ) Presumptive evidence indi- cated that toxaphene was present in those samples

Infrared spectroscopy indicated the presence of gamma BIIC in only six paired samples of raw and treated water taken during the summer of 1959, but gas chro- matography showed BHC to be as ubiqui- tous as toxaphcne ( Table 4) The time required for isomeric separation with the gas chromatograph was prohibitive, so 12 pairs of raw and treated water samples were selected for analysis by thin-layer chromatography The samples were col- lected in May, October, November, and December 1960; June, July, September, and October 1961; and January, July, Sep-

INSECTICIDE POLLUTLON IN A RIVER: TIIE WATER 315 tcmber, and October 1962 The alpha,

beta, and gamma isomers of BHC were

present in all samples Several showed a

spot near that of the epsilon isomer, but

the evidence was not conclusive

Treatment at the Hartselle water plant

apparently did not reduce the toxaphenc

and BHC content of the water (Tables 3

and 4) Prior to the spring of 1962, the

treatment consisted of lime-alum coagula-

tion, sedimentation, rapid sand filtration,

chlorination, and pH adjustment, Later,

activated carbon treatment was added, but

this, at the carbon levels used ( unknown),

also failed to remove the trace amounts of

insccticidcs Cohen et al ( 1960), simu-

lating municipal water treatment in the

laboratory, found that alum coagulation

and chlorination did not reduce the toxa-

phene content of water Reductions rang-

ing from 40 to 95,% were obtained in wa-

ter containing 0.3 ppm toxaphene when

treated with from 1 to 9 ppm of carbon,

but this concentration of toxaphene greatly

exceeded those found at IIartselle

Eight sites on the two principal streams

and their tributaries (Fig 1) were sam-

pled sequentially with two portable car-

bon adsorption units from 11 August to

29 September 1960 One sample was col-

lected at each site, and gas chromatogra-

phy indicated that all samples contained

toxaphene and BHC The mean recovery

for toxaphene was 154 ppt, and the range

was from 30 to 304 ppt; for BHC they

were 84 ppt and 13-210 ppt, respectively

This was the magnitude of contamination

noted at the water treatment plant There-

fore, the insecticides probably originated

from the watershed in general, rather than

from a few cotton fields or subwatersheds

Factors influencing insecticicle

contuminution

The mean seasonal recoveries for toxa-

phenc appear to be unrelated to usage in

the basin (Tables 2 and 3) In contrast,

mean summer recoveries of BHC appear

to be related to agricultural use The

means for the summers of 1961 and 1962,

both years of high usage, are greater than

those in years of low to moderate usage (1959 and 1960) DDT, which was used

in greater quantities than BIIC, was ab- sent from all samples The apparent ab- sence of DDT and other insecticides used

in the basin is not believed to be an error associated with the carbon adsorption method The authors ( unpublished data ) have noted the prescncc of DDT, endrin, chlordane, dieldrin, diazinon, and para- thion in carbon adsorption samples taken from other water sources

The persistence of insecticides in soils will vary, in part, with the characteristics

of the compound, the soil, and the rate of application (Lichtenstein and Schulz 19598) Generally, BHC is much less stable in soils than DDT or toxaphene (Chisholm and Kublitsky 1959; Foster et al 1956) En- drin, aldrin, and dieldrin, which arc used

in small quantities in the basin, are as per- sistent as BIIC or more so (Foster et al 1956) Collectively, the data presented by these authors indicate that all the insecti- cides mentioned would bc cxpcctcd to persist in soil at least a year IIcnce, on the basis of persistence alone, all could bc cxpccted to contribute to the year-round contamination of the stream Conversely, from the standpoint of total use in the basin, toxaphcnc, DDT, and BHC, in that order, would be the most probable con- taminants This lcads to a consideration of the water solubility of these three insecti- cides

Bowman, Acrce, and Corbett ( 1960)) working with carbon 14 labeled DDT, re- ported its solubility as 1.2 ppb or less at 25C Sladc (1945) noted the following solubilities in water at 20C for the various isomers of BHC: 10 ppm for the alpha, gamma, and delta isomers, and 5 ppm for the beta isomer The solubility of toxa- phene in water at 25C is 0.4 ppm (Cohen

et al 1960) The quantities of toxaphene and BHC recovered from the stream, even when corrected for losses associated with the sampling method, are well within the range of solubility The absence of dctect- able quantities of DDT (more than 10 ppt) may be related to its low solubility,

316 I-1, P NICHOLSON, A R GRZENDA, G J LAUER, W S COX, AND J I TEASLEY

Since toxaphcne, DDT, and BHC were

used in the greatest quantities, and since

DDT was absent in the samples, it appears

that the sustained presence of soil-persist-

ent insecticides in the stream is related to

both the total usage and the solubility of

the compound

Seasonal mean recoveries for toxaphene

and BHC (Tables 3 and 4) were not cor-

related with the mean seasonal rainfalls

Similarly, there was no correlation among

individual recoveries and the mean rainfall

during the l-Z-week period required to

obtain an adsorption sample There was

also no correlation between mean stream

turbidity and insecticide recovery

were made of nonagricultural factors that

might contribute to insecticide contamina-

tion Coverage was best from May to No-

vember 1960, when an investigator resided

in the basin Two events involving the di-

rect application of toxaphene to the West

Fork of Flint Creek were noted that year

On 18 July, a farmer accidentally spilled 1

gal (3.8 liters) of emulsifiable concentrate

into the river while filling a spray drum

About a week and a half later, fish poach-

ers introduced 100 lb (45.3 kg) of dust

(probably 20% active ingredient) into the

stream Both incidents occurred 15-20

miles (24-32 km) upstream from the water

treatment plant Each resulted in a local-

ized fish-kill (Grzenda, Lauer, and Nich-

olson 1964 )

Herbicide recovery

Microcoulometric gas chromatography

indicated that the 2,4,5-T used in the Na-

tional Forest did not reach the Hartselle

water treatment plant in detectable quan-

tities (more than 25 ppt ) This also ap-

plies to 2,4,5trichlorophenol, which is sus-

pected to be a decomposition product and

an impurity in formulated 2,4,5-T

Along the course of the West Fork of

Flint Creek, the principal stream draining

the forest, there are approximately 36

miles (58 km) between the forest bound-

ary and the water intake of the Hartselle

nlant Manv of the small tributaries

within the forest flow only during periods

of high discharge Hence, it seems that if any residue had leached from the soil, it would have been diluted to a nondetcct- able level before reaching the water treat- ment plant

SUMMARY

and BHC was observed at a municipal water plant in treated and untreated river water Samples taken from the stream sys- tem indicated the same order of contam- ination observed at the municipal plant DDT, which was used in substantial quantities in the basin, was absent from all samples

The recovery of toxaphcne appeared to

be independent of its agricultural usage However, the data suggested that the re- covery of BHC ,during the dusting season was related to use in the basin

REFERENCES

schichtchromatographischer Nachweiss von Insektizidcn Helv Chim Acta, 44: 1162-

1164

BOWMAN, M C., F ACREE, JR., AND M K COR- HETT 1960 Solubility of carbon-14 in water 1 Aar Food Chcm., 8: 406-408 CASSIL, C 6 i962 Pesticide residue analysis

by microcoulomctric gas chromatography Residue Reviews (Academic, New York), 1: 37-65

CHISIIOLM, R D., ANI) L KUBLITSICY 19,59 Accumulation and dissipation of pesticide rcsiduc in soil Trans N Am Wildlife Conf., 24th, 118-123

COHEN, J M., L J KAMPJIAKE, A E LEMKE, C HENDERSON, AND R L WOODWAIZD 1960 Effect of fish poisons on water supplies I Removal of toxic materials J Am Water Works Assoc., 52: 1551-1566

Oceanog., 9: 318-323

FOSTER, A C., V R BOSWELL, R D CHISHOLM,

R II CARTTX, G L GILPIN, B B PEPPEI~,

AND M GEIGER 1956 Some effects of in- sccticide spray accumulations in soil on crop plants U.S Dep Agr Tech Bull 1149

36 p

FOSTER, E E 1949 Rainfall and runoff Macmillan, New York 487 p

GRZENDA, A R., G J LAUER, AND H P NICHOL- SON 1964 Water pollution by insecticides

in an agricultural river basin II The zoo- plankton, bottom fauna, and fish Limnol

INSECTICIDE POLLUTION IN A RIVER: TIIE WATER 317 HENDERSON, C., Q H PICKERING, AND C M Chem., 52 ( 6) : 99A-102A

TARZWELL 1959 Relative toxicity of ten ROSEN, A A., AND F M MIDDLETON 1959 chlorinated hydrocarbon insecticides to four

spccics of fish Trans Am Fishcries Sot., Chlorinated insecticides in surface waters

LICEITENSTEIN, E P., AND K R SCHULZ 1959 SLADE, R E 1945 The gamma isomer of hex- Pcrsistcnce of some chlorinated hydrocarbon achlorocyclohexane ( gammexane ) , an inscc- insccticidcs as influenced by soil types, rate ticidc with outstanding properties Chem

of application, and temperature J Econ Ind., 40: 314-319

MIDDLETON, F M., AND J J LICIITENDERG lution-caused fish kills U.S Public Health

1960 Measurement of organic contami- Service Publ No 847 1960 cd 21 p., 1961 nants in the nation’s rivers Ind Eng cd 20 p,, 1962 cd 21 p