Simulation of Herbicide Persistence in Soil

Pestic. 9ci.1976, 7, 50-58p.trsiSimulation of Herbicide Persistence in SoilII. Simazine and Linuron in Long-term ExperimentsAllan WalkerNational Vegetable Research Station, llellesbourne, Warx,ick CV35 gEF(Manuscripr received l9 September 1975)Qsktethe. Six sat5,2t6%weigticate fl)oThe rates of degradation of simazine and linuron were measured in soil from plotsnot treated previously with these herbicides. Degradation of both compoundsfollowed first-order kinetics and soil temperature and soil moisture content had amarked effect on the rate of loss. With linuron, half-lives increased from 36 to 106days with a reduction in temperature from 30' to 5"C at 4lsoil moisture, and from29 to 83 days at 12\ soil moisture. Similar temperature changes increased the halflifeof simazine from 29 to 209 days and from l6 to 125 days at soil moisture contentsof 4 and l2)( respectively. A computer program which has been developed forsimulation of herbicide persistence was used in conjunction with the laboratory dataand the relevant meteorological records for the years 1964 to 1968 in order to testthe model against previously published field persistence data for the two herbicides.The results with simazine showed a close correspondence between observed andpredicted residue levels but those for Iinuron, particularly in uncropped plots. weresatisfactory for limited periods only.I. lntroductionExperirnents were begun in 1963 at the Agricultural Research Council Weed Research Organization,Oxford to examine possible long-term effects following repeated application of the sanreherbicides to the same soil. Crop yield and herbicide persistence results for the years 1963 to 1968were reported by Fryer and Kirkland.l The most extensive persistence data were presented for theherbicides simazine and linuron for the period May 1964 to November 1968. Simazine was appliedpre-emergence to maize at L68 kg/ha once each year, and was also applied at 3.36 kg/ha at approximately6 month intervals to plots which were maintained vegetation-free. Linuron was applied tocarrots, usually at 0.84 kg/ha pre-emergence, and on three occasions also post-emergence at thesame rate about 8 weeks aiter the first treatment. As with simazine, linuron was also sprayed onbare plots. The rate used was usually 1.68 kg/ha and this was applied three times each year. Theherbicide persistence data from these experiments provided an opportunity to test the simulationmodel for herbicide persistence described by Walker2 against data for 5 years in both cropped anduncropped situations. In order to do this, degradation rates of the two herbicides in the Oxford soilwere measured at different temperatures and soil moisture contents.2. ExperimentalThe soil used was from the unsprayed control plot paired with the vegetation-free simazine treatmentin the long-term experiment at Oxford. The herbicides used were a commercial wettablepowder formulation of simazine (50% a.i.) and samples of [carbonyl-1rC]-linuron(specificacrivity.6.85 mCi/e) and unlabelled linuron.2.1. Simazine treatmentsA fresh sample of soil taken from the surface 5 cm of the plot was passed through a 2 mm meshsieve and allowed to air-dry for 24h. After this period, simazine was mixed into a subsample ofthe surfaair-dto give atreated soiflasks IDuplicate saide wash, The extrat5 ml diox;scintillatlon cur'The extracts5 ml and store:chromatograPproducts. A thi5 x 20 cm glassl0 cm above thAfter developsive 0.5 cm baning on a TracerA further sarflask, and 10 mair, saturated w0aCl-Carbon damine in metheof 70 days, the50 ml with mescintillation co100 ml redistillin the soil extratography to de3.1. LaboratorThe effects of tI and on linurand the seriesunder these coThe degradzcharacterised i

Long-term persistence of simazine and linuron in soilthe soil (2.5 kg) by the method previously described3 to give a final concentration of 4.0 pglg drysoil. Separate 25 g amounts of the treated soil were weighed into wide-mouthed 100 ml conicalflasks and the appropriate amount of water was added to each to give the required soil moistureccntent. Six samples were prepared for incubation at both 4 and 121 soll moisture and temperaturesof 5, 10. 15,20,25 and 30'C. Further samples were incubated at 25'C with soil moisture contentsof 8 and 16)1. The initial weights of all the flasks were recorded and they were returned to theseoriginal weights by adding water at 7 day intervals.Duplicate flasks from each treatment were removed for analysis on three occasions during thesubsequent 56 days. The analytical method used for simazine has been described previously.3lotsrndsrda106romralfentsfordatatestides.andwereh Organizarfthe same963 to I 968nted for thewas appliedat approxisapplied totence at thesprayed onh year. The: simulation:ropl rndOxford soil2,2. Linur on treatmentsA rnethanol solution containing50 pg IaC]-linuron/ml and 750 pg unlabelled linuronimlwas pipettedover the surface of a sample of the air-dry soil (2.5 kg) to give a final concentration of 4.0 pglinuron/g air-dry soil. The solvent was allowed to evaporate and sufficient water was added to thesoil to give a flnal soil moisture content of approximately 4'%. Mixing was achieved by passingthe treated soil several times through a 2 mm mesh sieve. Treatments were prepared in 100 mlconical flasks for incubation under the same conditions as those used for simazine (section 2.1).Duplicate samples were removed for analysis at intervals during the subsequent 56 days. Theherbicide was extracted by shaking with 50 ml redistilled methanol on a wrist-action shaker forI h. The extracts were filtered and duplicate 2 ml subsamples were transferred to counting vialswith 5 ml dioxan-based scintillator.a Activities were determined using a Tracerlab Corumatic 200liquid scintillation counter. Quench corrections were determined using a sample channels ratiocalibration curve.The extracts remaining after preparation of samples for counting were concentrated to about5 ml and stored in glass vials at 4'C. Several of these extracts were examined by the thin-layerchromatographic method described by Katzs to separate linuron from its possible breakdownproducts. A thin streak of extract was applied to freshly prepared 250 pm layers of silica gel G on5 x 20 cm glass plates which had been activated for 30 min at I10"C. The plates were developed forl0 cnr above the baseline in chloroform-methanol-pyridine (100:5: I by volume).After development the plates were dried and the peaks of activity determined by scraping successive0.5 cnr bands from the plates, dispersing in methanol on 5 cnr aluminium planchets and countingon a Tracerlab, low-background, end-window Geiger counter.A further sample (100 g) of the original linuron-treated soil was weighed into a 250 ml conicalflask, and l0ml water was added to give a moisture content of about 141. Carbon

52A.Woo4% Soit moislu/ec:3l?'h Soil moisturejFlgure 2. Effr'; O, 20"; l, 25".:'rT'.T l,:f::,'!:t,r"+:erature40Trme {doys)on simazine degradation at 4 and r2lsoitmoisture. A,5.; a, 10"; o, 15.;OU20cTable l' The effects of temperature and soil moisture content on simazine and linuron degradationLinuronSoil moisturecontent(%)Simazine+812l68l2loI0683209125I{alf-life (days) at temperature (oC)l0 l5 20 25 30;o,384 67 5745 36:50Il=63I5343I293414080show that the effects of soil moisture content on the rate of degradation aretemperaturessimilar atfortheanydiflerentone herbicide. with tinuron a reduction from 12\75)(ofsoil moisturefieldcapacitv)tolT.soil(aboutmoisture (25/,otfieldcapacity;increaseothehalflifebyabout30"1 and a similar change with simazine increasld the half-life by about g0f(.similar to Thesethose reported increases areby Usoroh and Hance,? and walker3 forlinuron and simazine respectively.925360344028162329l6toc?5t?.3Figure 3. Effe,l, 25'; n, 30'

A. WalkerLong-term persistence of simazine and linuron in soil 53oo:-csot.o4.047" Sorl morstureiai2% Soil morslure40Tlme (doys )Figure 2. Effects of temperature on linuron degradation at 4 and lzlO, 20'; l' 25"; C, 30'C.soil moisture.' l' 5'; a, lo";S rmoz ineruron degradation\_\*--.-----t-.--*--.-.2530E4542JI343629a-.--\---*----.-a4028262329l6milar at the different;oil moisture (abouthe halfJife by about. These increases areimazine respectively.Z.f, 5 tO ?.O 2.5Sorl moislu.e content {o/");Flgurer' Effectsofsoilmoisturecontentonsimazineandtinurondegradation.A,5.; a,r0"; o,r5"; o,20";l, 25'; n, rO.c.

A. Walker Wry-term Persistence ofA feature of the results in Table I is the greater effect of both soil temperature and soil moisturecontent on the rate of degradation of simazine compared with linuron. The half-life of linuron insoil at 30"C andl2l soil moisture is almost twice that of simazine, whereas at4lsoil moistureand 5"C, the half-life of simazine is almost twice that of linuron. The relationships between halflivesand soil moisture content are shown for both simazine and linuron in Figure 3. The straightlines drawn through the data at 25"C show an approximate fit to the empirical equation relatingthe half-life (H) to soil moisture content (M):H:AM B (l)in which A and B are constants. This equation has previously been shown to fit similar data withpropyzamidee, napropamide,2 simazine3 and prometryne.3 The fact that straight lines with similarslope to that at 25"C can be drawn through the two points at the other temperatures is a furtherillustration that the relative effects of moisture content on degradation rates are similar at differenttemperatures.The relationship between the rates of degradation of both herbicides and temperature can beexpressed in terms of the Arrhenius equation:' Ht lF'l l\be'n,:#i(n-Z I tztin which Ht and 11: are the half-lives at temperatures Ir and Tz and -lE is the activation energy.The relationships between the half-lives of simazine and linuron and temperature at 4 and 12i(soil moisture are shown in Figure 4, in which the half-life on a logarithmic scale is plotted against;= oro.oo35| /Absolute temperolure ( K -')Figure 4, Test for the fit of the Arrhenius equation to data for simazine and linuron degradation. O,4%Soilmoisture ; O, l2',r" soil moisture.the inverse of the absolute temperature. From the slopes of the lines, the activation energy forlinuron degradation is calculated as 30.2 kJ/mol and for simazine as 57.4kJ/mol.The data for linuron degradation in Figure 2 and Table I were based on the amounts of radioactivityextracted from the soil expressed as concentrations of linuron. The thin-layer chromatogramsof the extracts from the samples incubated at 25"C and all the extracts at 56 days showedthat over 957" of the activity on the plates was located at the Rr,-value of the parent herbicide(between 0.90 and 0.95), therefore demonstrating that in experiments with [carbonyl-1aC]-linuronchanges in extractability of radioactivity give a valid measure of linuron degradation. This con-clusion is further suPlin Figure 5. The amotloss of [raC]-carbon don a log scale againsexpressed as linuron,spondence with first-'srooQa\t z,)between the calculategraphic analysis of ttof the extractable raCthose of Bcirner,lo wcarbonyl group and n3.2. Simulation of peThe meteorological rrand l0 cm soil temp(the computer prograrsoil temperature ancNational Vegetable IWeed Research Orgaweather records arepattern, on a day-tosimulation program'simulation program €simazine plots in Figlinuron plots are sho,with the simulation nof measured residueThe results from theactual amounts repeidicted and to give thThe results with sshape to those obserrthe uncropped plots,the determined resid0.56 to l.g9 kg/ha wPlots, and the model

A. WalkerIrng-term persistcnce of simazinc and linuron in soil55d soil moistureie of linuron in( soil moistures between half-3. The straightuation relatingclusion is further supported by the results from collection of ILlc]-carbon dioxide which are shownin Figure 5. The amount of linuron remaining in the soil (calculated by subtraction of the cumulativeloss of [1aC]-carbon dioxide from the radioactivity present initially, expressed as linuron), is plottedon a log scale against time of incubation. Also shown is the value for extractable activity, againexpressed as linuron, on termination of the experiment at 70 days. The data show a close correspor.ldencewith first-order kinetics giving a half-life of about 30 days. There was good agreement(l)nilar data withres with sinrilarres is a furtherilar at differentt coFs50= :_..erature can be(2)ivati< .tergy.at 4 and 12\plotted against:n. C, 4i{ SoilIn energy forrnts of radio-3r chromato-Cays showed:nt herbicide.raClJinuronn. This con-6(d 4t fbTime {doys)Figure 5, Loss of linuron from soil calculated fromcumulative amounts of liCOz collected ( o) and amountof linuron extracted from the soil after 70 days (O).between the calculated amount remaining after 70 days and that determined. Thin-layer chromatographicanalysis ol the methanol extract at the end of the experiment again showed that over 957"of the extractable radioactivity was in the form of linuron. These data are therefore consistent withthose of Bcirner,lo which showed that loss of linuron from soil is closely related with loss of theerbonyl group and not associated with buildup of the demethylated or demethoxylate derivatives.3.2. Simulation of persistence in the fieldTh,) nleteorological records of rainfall (mm/day), evaporation from an open water surface (mm/day)and l0cm soi[ temperature ("C) for the period January 1964to December 1968 were combined inthe computer programz with the respective constants relating half-lives of simazine or linuron withsoil temperature and soil moisture content. The weather data used were those recorded at theNational Vegetable Research Station since evaporation measurements were not available from theWeed Research Organization. The two stations are approximately 35 miles apart, and the averageweather records are very similar. There would have been differences, particularly in the rainfallpattern, on a day-to-day basis, but such differences were probably of little importance when thesimulation program was run for long periods. The disappearance curves obtained from use of thesitnuiation program are compared with the published results of Fryer and Kirklandl flor the croppedsimazine plots in Figtrre 6 and for the uncropped plots in Figure 7. Similar results for the croppedlinuron plots are shown in Table2,and for the uncropped plots in Table 3. In previous experimentswith the simulation model, the determined degradation curves have been drawn following correctionof measured residue levels for initial recoveries2,3'e but in the present study this was not done.The results from the model, shown in Figures 6 and 7 and Tables 2 and 3 were calculated using theactual amounts repeatedly applied, in order to determine whether residue build-up would be predictedand to give the model a strict test against field results.The results with simazine are encouraging: the simulated disappearance curves are of similarshape to those observed and the predicted residues are often very similar to those determined. Withthe uncropped plots, the predicted residue at the end of the fifth year is 0.58 kg/ha compared withthe determined residue of 0.63 kg/ha. Fryer and Kirklandl conclude that a residue varying from0'56 to 1.89 kg/ha was present in the soil at the time of the next application of simazine to theseand the model predicts a minimum residue of 0.49 kg/ha following spring application and a

56 A. Walker persistence o,.oI?t1964/3 r965/6ooo\.\oEoi z.oct966/ 76080t967/8tro0t20r40r50r80200Time ( weeks )Figure 7. PersistencFigure 6. Persistence of simazine in cropped plots. O, Observed; o, simulated; time 0:27 May 1964.WeeksafterDate Dose applisprayed(kg/ha) cation12.6.64 t.t2 085.8.64 0.56 07l42725.5.65 0. 84 0822.7 .65 0. 84 06l32l2835Table 2. Simulation of linuron persistence in the cropped plotsResidue (ke/ha)Observed0. 700.420. 980. 700. 70.,:0.l40. 63Simulated.:0. 14 0. 380. 98 (0. 98)' l . 220.49 (0.57) 0.710.42 (0.35) 0.430.21 (0.24) 0.290.21 (0.18) 0.22

A. Walkerpersistence of simazineL,ong-termand linuron in soilI.,olL',--c'o^-

.\. WalkerTable 3. Simulation of linuron persistence in uncropped plotsDatesprayedResidue (kgi ha)Dose Weeks after(kg/ha) application Observed Simulated12.6.645.8.&28.tO.6425.5.65zt.t.6J2.24t.t2l 68I .68I .680807l206t2l82329080bl21.820. 9l2.52| .47t.262.66l .541 .260. 98r .050.77l .401 .332.943.36.3.992.24r.042.16l.200. 902. 58l .931.55| .250. 990. 64) ?fI .042.72L59l.0la From this ooint onwards the observedata become unreliable.and linuron to this particular soil will not lead to a buildup of residues of either herbicide. Theyconfirm the conclusions of Fryer and Kirklandr that repeated applications of simazine for 5 yearsdid not alter the soil's capacity to degrade the herbicide, but the discrepancies between observedand predicted linuron residues following three applications of relatively high doses in each of fivesuccessive years suggest the possibility that some change did occur.AcknowledgementsThanks are expressed to Mr H. A. Roberts for his encouragement during this work and to MissP. A. Brown for technical assistance.Referencesl. Fryer, J. D.; Kirkland, K. ll'eed Res. I970, 10, 133.2. Walker, A. J. Environ. Quality 1974,3,396.3. Walker, A. Pestic. Sci. 1976,7, 41.4. Bray, G. A. Analvt. Biochem. 1960, 1,279.5. Katz, S. "/. l.rs. of. anolyt. Chem. 1967,50,912.6. Hance, R. J. Pesrlc. Sci. 1973, 4. 817.7. Usoroh, N. J.; Hance, R. J. l4eed Res. 1974, 14, 19.8. Obien, S. R.; Creen, R. E. lleed Sci. 1969, 17, 509.9. Walker, A. In Proceedings European lYeed Research Council Synposiunt, Herbicidts and the Soil Paris, 1973,p.240.10" Borner, H. Z. P.flnnzenkrankh Pfanzensthutz 1965,7?-,516.Ycgetable Rcsrccelved 19Thc effects ofof propyzamidHalf-lives in srfrom 63 to I Icapacity was fiwith the relevaprogxam, predplots were inresidues of Prtments concerthose observecomputer progJto predict theiments.z The mrare made undermoisture contenlmodel are in goobe specific to thtThe present experpropertisoil type, and to drapply to oth(in lettuce,thharvest was also (2.1. MaterialsSoils from flve fieldproperties are showrconcrete slabs sunkiof about 25 cm withraked level, andfrom the field uThe herbicide use