3 ,m 1, \ „ • ._ :,SYN1 POSIUM

H E , , " ] • : , L . ,. E C • D. -, -1:3 B m 7." Fr., 7-c3rA: iu 'TR

February 21-22, 1978

Crystal City, Marriott HotelArlington, Virginia

United States Department of AgricultureUnited States Environmental Protection Agency •

COORDINATORDavid E, Ketcham

MODERATORSJan B. Wine, EPA

Barry Flamm, USDA

An addendum of papers presented after the Symposium is being distributed to recipients of these, proceedings_

DEPARTMENT OF AGRICULTUREOFFICE OF THE SECRETARYWASHINGTON, D.C. 20250

for sate by tne Superintendent of Documents, ■J.S. Government Printing Ottec.eWashington. D.C. 20402

Stock Number 00 1-000-0384 2-7

t

t.I NM

11.=

Vt,infigelie*•

POST-SYMPOSIUM RESPONSES TO QUESTIONSSUBMITTED TO LOGAN NORRIS

QUESTION: Many thousands of people 60 miles...est of your home breathe for many days smoke from.;nits triggered with 2,4,5 .-T and then burned_ Are youi...are of this? Have you monitored the air?

ANSWER: I know that some units treated with1,5-T are later burned and those in the area are likely

*3 breathe some of the smoke. I have not personallymonitored the air from this type of operation to deter--line the level of either 2,4,5-T or TODD.

QUESTION: The data of Stehl and Lamparski1977) predict that "spray and burn" with 2,4,5-T might

pe equivalent to spraying with 2,4,5-T with a dioxin con-sent of 1.6 ppm, almost three times the median concen-: •ation of 2,4,5-T used in Vietnam. How can you justifyspray and burn?

ANSWER: The previous two questions do nots pecifically so state, but I believe the concern relates to

l e possibility that combustion of 2,4,5-T and 2,4,5-T--e.ated materials may result in the production of signifi--ant quantities of 2,3,7,8-tetrachlorodibenzo-p-dioxin. The'mparison between "spray and burn" as done in fores-

' ry and the dioxin content of 2,4,5-T used in Vietnam issreposterous.

In laboratory experiments it is possible under.-;ecial conditions to produce TODD on heating or burn-

2,4,5-T-treated materials. The probability of signifi-3 nt production of TODD in the field from burning,

however, is vanishingiy small. The amount of TODD pro-duced is dependent on the amount of 2,4,5-T which ispresent, and current research shows more than 90percent of the 2,4,5-T applied to the forest is gone in 30days. Therefore, unless burning occurs within the first30 days alter application, no increase in TODD levelsover the initial level after application is expected to occur.The probability of fire occurring within 30 days of treat-,merit in areas treated with 2,4,5-1 is remote. Based onthe detailed analysis below, I do not believe the possi-bility of thermal production of TODD on burning areassprayed with 2,4,5-1 is sufficiently high to warrant seriousconcern.

There are several questions involved when con-sidering the probability and consequences of TODD pro-duction from combustion of 2,4,5-1. The first questiondeals with the formation of TODD from the combustionof 2,4,5-T. The second question deals with the quantitiesof TODD which might be produced from burning areastreated with 2,4,5-T. A third question considers thepossible environmental implications of this TODD pro-duction.

1. is it possible to produce TODD on heatingor burning of 2,4,5-T or 2,4,5-T-treated materials?

ANSWER: Yes, in laboratory tests. The conditionsof combustion and herbicide concentration are crucial.The vests reported by Baughman and others show TCDDformation when 2,4,5-T is heated in a closed containerunder alkaline condition such that the sodium salt oftrichiorophenol is a significant degradation product. Theamount of herbicide employed in these tests was veryhigh. Langer et at., (1973) showed control of the decom-position reaction to produce trichlorophenol was neces-sary since heating above the decomposition point (300°C)produced no TODD. Concentration of herbicide is veryimportant because the formation of TODD is apparentlya bimolecular reaction; that is, it requires the joiningtogether of two molecules of sodium 2,4,5-trichloro-phenate. If conditions of heat and alkalinity are con-ducive to the condensation of the phenol to form TODD,then the extent of condensation varies with the number ofmolecules available to interact with one another. This ismost easily explained by analogy. Take two 50-gallonbarrels and place 10 small lead pellets in one and 1000lead pellets in the other. Close the lids and shake eachbarrel for 5 minutes and count the number of times thattwo pellets collide in each barrel. Obviously, there willbe more collisions in the barrel with 1000 pellets thanthere will be in the barrel with only 10. Formation ofTCDD results from the interaction (collision) between twomolecules of sodium 2,4,5-trichtorophenate.

95

Experiments like those of Baughman and othersare useful only to show that thermal production of TODDis chemically possible. Experiments which use closedsystems and high concentrations of 2,4,5--T drasticallyoverestimate the levels of TODD which might be pro-duced in burning situations in the field because (a) theconcentrations of herbicide are several times greaterthan the levels of 2,4,5-T which occur in the field and(b) heating is prolonged and uniform, but combustiondoes not actually occur. Temperatures at which thermaldecomposition of TCDD occurs (800°C) are not attainedin these test situations. Actual burning will result intemperatures near those used in laboratory tests onlybriefly. As temperatures approach 800°C, thermal de-composition of TCDD will also occur. When combustionscan take place with a free exchange of air, temperaturesabove 1200°C are common. Under these conditions weexpect complete oxidation of all carbon compounds in-cluding 2,4,5-1, trichlorophenol, and TCDD.

For these various reasons I conclude the labora-tory studies of the thermal production of TODD from2,4,5-T maximize TCDD production and minimize theopportunities for its dissipation. Laboratory studies, then,will vastly overestimate the level of TODD productionwhich might occur in the natural environment.

2. How much TODD is produced when 2,4,5-Tis burned?

ANSWER: There are only limited experimentaldata. Watts and Storher (1973) noted burning and heatingof such 2,4,5-T-treated products as vegetation, meat, andfat did not produce detectable tetrachiorodioxin. Thesensitivity of their analysis was not adequate, however,to detect environmentally important quantities of TCDD.Present methodology with sensitivities that approach 10parts per trillion is sufficient.

The most pertinent data comes from a laboratoryexperiment in which grass treated with 2,4,5-T at 12pounds per acre was burned under conditions margin-ally resembling those which might occur in the field(Stehl and Lamparski, 1977). Their study showed anapproximate 0.00016 percent conversion of 2,4,5-1 toTODD. This involved a semi-closed system, however.Thus, any TCDD which might normally have been lost tothe air as vapor or adsorbed on smoke particles in forestburning was captured and retained in this system.

The amount of TODD produced is dependent onthe concentration of 2,4,5-1 in the vegetation. Studies byNorris, et al., (1977) of the persistence of 2,4,5-1 in Ore-gon forests shows levels of herbicide and calculatedlevels of TODD (Table 1) which might be produced by

burning, assuming the conversion ratio reported by Stehland Lamparski holds in this case.

TABLE 1-2,4,5-T residues on vegetation(measured) and TCDD (caiuciated) that might

be produced on burning vegetation'

Possible TODDlevel if

burning occursMonths at time

after indicatedApplication (PPt)

0 95. 152.9.1 14.

3 0.10 0.196 0.07 0.11

12 0.01 0.02

' Assumes 0.00016 percent percent conversion of 2,4,5-T toicon (Ste ti; end Lamparsk, 1977j.

'From Norris et ar. 1977.

Clearly the amount of TCDD produced depends toa major degree on when burning occurs; after treatment_Burning which takes place from 1 to 3 months after theapplication may result irr TCDD levels of 14 and 0.2 partsper trillion, respectively. In some brush types, burning isdelayed for 12 months or more. Immediat:-.4y after appli-cation the level of TCDD present on the vegetation isapproximately 10 parts per trillion, assuming the 2,4,5-Tcontained 0.1 parts per trillion TCDD. Research of Getzan--darter and Hummel (1975) and Crosby and Wong (1971)indicates the TODD originally applied will be largely gonewithin 1 month of the application_ Therefore, the levelsof TODD which might be produced by burning are riotexpected to substantially exceed TCDD levels present asa result of the original application of herbicide.

A forest fire could conceivably occur at any time.However, they are relost likely to occur during the driestparts of the year. These do not usually coincide Withperiods when 2,4,5-T is sprayed for vegetation control.Unless a forest fire occurs within 1 month of herbicideapplication, the possible TCDD production from burningwould be les, than that calculated for the case where2,4,5-T is used as a preburn desiccant.

3. What are the implications of possible TCDDproduction on the burning of Z4,5 -I treated vegetation?

96

ANSWER: Toxic hazard from TCDD requires thatthe organism receive exposure to toxicologically signifi-cant quantities of the chemical. A substantial amount oftoxicology has been done on TODD. In a recent 13-weekfeeding study (Kociba et al., 1975), 0.01 micrograms ofTCDD per kilogram of body weight per day did not affectrats. Assuming these animals consume 10 percent oftheir body weight per day' in food, the no-effect dosagelevel in this experiment is equivalent to 100 parfs pertrillion TCDD in the diet. The calculated possible levelsof TODD in forest vegetation immediately after treatmentwith 2,4,5-1 (about 10 parts per trillion) is only one-tenthof the established no-effect level in this study. If burningoccurs 1 month after spraying, the concentration of TCDDwhich might be produced is still approximately one-tenthof the no-effect level. However, there are several factorswhich operate to further reduce the probable exposurelevels. Areas which are burned do not make favorablehabitat for animals. Therefore, animals are not likely tobe present in such areas for any significant period oftime. The burning required to produce TCDD destroysthe vegetation that would have to be consumed if inges-tion of the combustion products is to occur. Burningproduces charcoal which is a highly effective adsorbentfor TCDD and is likely to prevent dermal absorption andreduce or eliminate absorption in the intestine.

TCDD uptake by ingestion and dermal absorptionas a result of burning 2,4,5-T-treated vegetation is un-likely (even if substantial quantities of TCDD are pro-duced on burning). Inhalation is the only other probablemeans of exposure. The production of TCDD by burningis associated with the rapid heating, expanding, andrising of air mass. TODD in that air mass will be dispersedand greatly diluted, thereby minimizing the exposure forindividual organisms.

REFERENCES

Crosby, D. G., and A. S. Wong_1977. Environmental degradation of 2,3,7,8-tetra-chiorodibenzo-p-dioxin. Science 195:1337•1338.

Getsandaner, M. E., and R. A. Hummel'.1975. Disappearance cf TODD from grass followingfield treatment with Esteron 245 herbicide. TheDow Chemical Company Internal Report GHC 792.February 18, 1975.

Kociba, R. J., P. A. Keeler, C. N. Park, and P. J. Gehring.1975. 2,3,7,8-tetrachlordibenzo-p-dioxin (TODD): re-sults of a 13-week oral toxicity study in rats.Toxicol. App. Pharmacol. 35(3):553-574.

Langer, H. G., T. P. Brady, and P. R. Briggs.1973. Formation of dibenzodioxins and other con-densation products from chlorinated phenols andderivatives. Environ. Health Perspect. Exper. IssueNo. 5, September 1973. p. 3-7.

Norris, Logan A., Marvin L. Montgomery, and Eugene R.Johnson.

1977. The persistence of 2,4,5-1 in a Pacific North-west forest. Weed Sci. 25:417-422.

Stehl, R. H., and L L. Lamparski.1977. Combustion of 2,4,5-trichlorophenoxyaceticacid and derivatives: formation of 2,3,7,8-tetra-chlorodibenzo-p-dioxin. Science 197:1008.

Watts, R. R., and R. Storherr.1973. Negative finding of 2,3,7,8-tetrachtorodibenzo-p-dioxin in cooked fat containing actual and forti-fied residues of ronnel and/or 2,4,5-trichlorophenol.J. Assoc. Off. Anal. Cheat. 56(4)1026.

QUESTION: Can you comment on the incident atthe Gnat Creek Fish Hatchery in Oregon, which Dr. Cutlerreferred to in his speech this morning. Were herbicidesto blame? Are herbicides such as 2,4,5-1 likely to befound in streams in biologicaliy significant quantitiesfollowing forest applications?

QUESTION: What level of herbicide was found insamples taken at the time of the death of the fish fromthe Gant Creek Fish Hatchery? Are data available forexposure of similar fish at OSU?

ANSWER: Both of these questions refer to anincident of fish mortality which occurred at the Stateof Oregon, Department of Fish and Wildlife hatchery atGnat Creek in northwest Oregon. This hatchery raiseswinter steelbead trout and had a population of about475,000 fingerlings (about 750 fish per pound) at thetime of the incident. Herbicide spraying which took piece:a relatively short distance upstream from the fish hatcheryis alleged to have been the cause of the fish mortality.

97

2,000 Ft.

----- Spray Unit BoundaryStreamWater IntakeWater LineHatchery Site'Water Sampling Site

FIGURF 1. SPRAY UNIT, STREAMS AND(IN t'.-1 CREEK FISH HATCHERY.

•

As is true for most instances of this kind, somevisual observations and a few pieces of hard data areavailable, but they do not provide a direct answer to thequestion "Did the chemical brush control project causefish mortality?" The hard data, the literature, and myexperience are the bases upon which I have evaluatedthe incident and reached my conclusions. The followingis a summary of the observations and hard data.

May 11-143 acres (Figure 1) were treated with 2 poundsper acre each of 2,4,-D and 2,4,5-T in 10 gallons of spraycarrier per acre. The specific formulation was Dow'sEsteron Brush Killer. An orientation flight was made witha representative from the State Department of Forestryprior to the application of herbicide. They noted no watervisible from the air in Unit VI.

May 15-0.6 inches rain in 24. hours ending at 9:00 a.m.Unusual behavior and mortaility noted on May 15. Onehundred fifty dead fish were counted at 5:00 p.m.

May 15-0.2 inches rain in 24 hours ending at 9:00 a rn.Fish tanks were vacuumed on this day and 1500 fishcollected. The cleaning operation collects both deadfish which float (floaters) and those which sink (sinkers).The 1500 fish collected in the cleaning operation on May16 therefore reflect fish which died on May 15 as well asMay 16.

May 17-0.4 inches rain in 24 hours ending at 9:00 a.mThree hundred dead fish counted (floaters) at 5:00 p.mA water sample containing 0.4 ppb/2,4-D' was collectedat the outfall from one of the six tanks in which mortailitywas occurring. A sample, taken at about the same timefrom a small tributary within the spray unit, contained 1.8ppb 2,4-D and 7.2 ppb 2,4,5-T. A water sample takenfrom Gnat Creek at the water intake site showed nodetectable levels of 2,4-D or 2,4,5-T.

May 18—No rainfall in the previous 24 hours. Fish tankswere cleaned again. Fish mortality was 5000. The 5000fish represent a large number of sinkers which mayhave died any time between 5:000 p.m. on the 16th and5:00 p.m. on the 18th.

A pathologist examined the dead fish and reportedno obvious signs of disease or parasitism in the deadfish. The Oregon State Department of Agriculture analysisof dead fish from the Gnat Fish Hatchery did not finddetectable residues of 2,4-0 or 2,4,5-T. The minimum

98

'.•-•,1.1.•&20.,T•WwwW,_

TankCleaned3

WaterSamplesCollected2

TankCleaned3

1700 17001700 1700 17001FISH MORTALITY 150 15001 310 5000

RAIN

Midnight Midnight Midnight Midnight Midnight MidnightTIME

5/14 5/15 5/16 5/17

5/18

1 Includes floaters (dead fish that float) for one day and sinkers for 2 days.

2 Water Sample Location Date TimeHerbicide, ppb2, 4-0 2, 4, 5-1

A In Spray Unit 5/17 1600 Hrs. 1.8 7.2

B Gnat Creek at 5/17 1600 Hrs. 0 0Water Intake (Approx.)

C in Hatchery 5/17 1600 Hrs. 0.4 0.05

3 Dead fish did not contain detectable herbicide residues (<0.01 ppm).

FIGURE 2. Schematic of events at Gnat Creek Fish Hatchery during May 1977 fish mortalityincident.

0900, 09000900

0900OA" 09°

limit of detection was 0.01 ppm. Figure 2 shows the pat-terns of precipitation, fish mortaility, and water sampling.

' The Research Division of the Oregon State De-partment of Fish and Wildlife conducted a bioassay ofwinter steelhead fry from the Bi g Creek hatchery (nearGnat Creek) and the same herbicide formulation whichwas applied to the area upstream from the Gnat CreekHatchery. Following a 48-hour acclimatization to the testtanks, fish were exposed to the toxicant on May 31, 1977.The percent survival of various exposed steelhead fry isin Table 1.

One steelhead fry died at nominal concentrationof Esteron of less than 800 pob during the 96-hourexposure. At the two high concentrat ions, no mortalities

occurred during this 24 hours, but 50 percent of the groupexposed to 1200 ppb died between 24-48 hours of ex-posure. The 96-hour L.C 50 for steelhead trout fry on thebasis of this test would he about 1 ppm. .

Some of the fish exposed to herbicide were ana-lyzed for whole body residues (Table 2). Data in Table 2is conclusive. Fish exposed to sufficient 2.4-D and 2,4,5-Tin water to cause death will have measurable residues ofthese herbicides in their bodies.

After inspecting this site, studying the data, and .consulting with the fish toxicologist, I reached the follow-ing conclusions. I -CONCLUSION: The herbicide cannot be discounted as contributing to the fish mortality, but the

A

99

4

probability of its involvement is extremely low. The FACTORS WHICH INDICATE HERBICIDE WAS INVOLVEDevidence for and against herbicide involved is listed A. Location and time of herbicide applicationbelow. made it physically possible for herbicide to

enter the hatchery system.B. The fish mortality occurred coincidentally with

heavy rains. The rain could have mobilizedTABLE 1—Survival of Big Creek winter herbicide in and around the ephemeral stream

steelhead fry exposed to Esteron (Fairplay channels in the spray unit, resulting in herbi-Laboratory, OSU, Corvallis) May 31-June 4, cide in the water supply for the hatchery.

1977 C. Herbicide residues were (0.4 ppb 2,4-D) in thehatchery water during the time fish were dy-ing. Although 0.4 ppb herbicide is not a toxicconcentration, the levels could have beenhigher earlier.

D. Disease, parasites, temperature, and turbidity.do not appear to be factors in the reortiatity.

FACTORS INDICATING HERBICIDES WERE NOTINVOLVED

A. Herbicide residue levels measured in thehatchery are not toxic to fish. Water sampleB collected at the water intake contained nodetectable herbicide; however, stream ditu-Lion would have reduced the concentrationof herbicides from sample Point A, so it wouldnot be detectable at B. The residence time of

°Test tanks set up June 2 in a.m.; exposure initiated in p.m. hatchery is about f hours, so residues insample C reflect conditions in Gnat Creekat least 6 hours earlier. Presence of residuesin sample C, but not in B, indicates herbicideresidues in the water were transitory.

TABLE 2--Herbicide residues in winter B. Fish and Wildlife Service research with ansteelhead fry exposed to Esteron Brushkiller Esteron 2,4-D formulation shows a no-e..fiect

in water. level at 40 ppb in very young cutthroat troutexposed coritirtuously for GO days_ Dilutionpotential in moving from the spray unit intoGnat Creek is at least 40 to 1 (20 cfs in GnatCreek vs. art estimated 0.5 cfs in the streamin the unit). Dilution with downstream move-merit in Gnat Creek is at least 40 to 1. There- -fore, if the small tributary in the spray unit(sample point A) had at one point an herbicideconcentration of 1000 ppb (higher than meas-tired in Oregon in 15 years), the concentrationin Gnat Creak would be 25 ppbwhich is lessthan the no-effect level for cutthroat trout, Ifan additional 40 to 1 dilution occurred with

' Fish placed in clean water for 48 hours before being sacrificed downstream` movement, the eoncerrration atfor chemical analysis. the hatchery would be 0.6 ppb.

.r.

NominalConcentration

(ug/liter)

Exposuretime(h)

Percentsurvival(range)

Control (60)' 96 100.0

1 (60) 96 100.0

7 (60) 96 100.0 •

25 (60) 96 98.3 (100-96.7)

75 (60) 96 100.0

150 (60) 96 100.0

450 (60) 96 100.0

800 (60) 48b 88.3 (100-76.7)

1200 (63) 484 50.3 ( 80-24.2)

'Total number of fish exposed in replicated tanks under staticwater in the line from Gnat Creek to theconditions.

Exposure (nominal)Water Timeppb hours

FishCondition 2,4-D

Residue2,4,5-T

PPm

75 96 Live 0.9 1.9

150 96 Live 41 44

800 48 Live 20 26

800 48 Dead 64 60

1200 96 Live & clearance 8 8

1200 48 Dead 165 100

100

The lack of measurable herbicide residues indead fish argues strongly against herbicideas a causal agent of death.Fish mortality occurs periodically in the courseof hatchery operations_ Gnat Creek Hatcherypersonnel related other incidences of fishmortality resulting from hatchery operationspaint, zinc pipes) which were more extensivethan the mortality which occurred in this in-cident. Failure to find other causative agentsbesides temperature, turbidity, parasites, ordisease does not automatically assure thatherbicide was the cause of death.

To put this incidence in perspective, the total ofapproximate ly 7500 fish which died in this incident repre-sented less than 2 percent of the hatchery populationand weighed a total of 10 pounds. Hatchery personnelfelt that the magnitude of this incidence would have ab-solutely no impact on productivity of the hatchery.

* * *

A AC e..•~1.••■■-■•••; *1 L•1/0"

101