Modeling the fate of paddy field pesticide in surface water and environmental risk assessment

Journal of Environmental Sciences Vol. 12,No. 3.pp.337- -343 ,20001S581001070Articke ID: 100 0742000)03-0337.07 CLC number: X820.4Docement code: AModeling the fate of paddy field pesticide in surface water andenvironmental risk assessmentLI Shi-yu' , Tohru Morioka?(1. Institute of Environmental Science, Zhongshan Universty, Guagzhou 510275. China. E- mail: eslsy@ 2su. edu. cn; 2.Department of Environmental Engineering, Osaka University, Yamada-oka 2-1, Suita, Osaka 565, Japan. E - mail; morioka@ env.eng. osaka u. ac. ip)Abstract: The risk of drinking water is greatly concemed because of the large armount of pesticide applied to paddy field and thecontamination of drinking water sources due to the runoff. A mathermatical model is developed, based om the mass balance, to predictthe fate of paddy field pesticides from application, runoff and mixing in 8 river, taking account of the physical-chermical propetties andprocesses of vlailization, degradation, adsorption and desorption. The model is aplied to a tiver basin in Japan to estimnate thes drinking water induced by each pesticideconcerned is estimated and evaluated by comparing with the acceptable daily intake values( ADI) and with that induced bytrihalomnethanes. An index to evaluate the total risk of all pesticides appearing in water is proposed. The methods for risk managementare also disussed.Key words: modeling; pesticide; paddy feld; environmental risk 8sessment; drinking waterIntroductionMicro-organic compounds have been identified as one of the main risk agent groups in drinkingwater. In surface water, agricultural chemicals account for a large part of the organic compoundcontents. In East and Southeast Asia, including China and Japan, cultivating of rice, the mainfood in the region, depends more and more on the use of pesticides. As a result, the applicationamount of these chemicals has been increasing. In Japan, about forty percent of agriculturalpesticides are applied in paddy fields. During the rice-farming scason, especially the period of riceseedling transplant when various kinds of agricultural chemical are intensively used, the residualpesticides are discharged from the paddy fields effluent and runoff with rainwater and then enterthe receiving waters. High contaminant levels of pesticides have been monitored in Japan inirrigation canals for paddy field and some rivers that serve as drinking water sources. The healthrisk induced from the chemicals has become a public concern.On the other hands, tihalomethane(THM) is also a main risk agent in drinking water. Unikepesticides, THM is a by-product formed in the chlorinating process of water purification, in which chloninereacts to the organic substances in the raw water , mainly humic substances. THM could be controlled orreduced within the water purification processes or by changing the traditional way of disinfection bychlorine. Comparing to this, the pollution control of pesticides in a river is dificult.In this paper a mathematical model is developed to trace the applied pesticides and to examinethe impact of the runoff chemicals on river waters. The model is applied to a typical river basin topredict the concentration of popularly used pesticides in the river water. The risk in drinking wateris estimated and evaluated based on drinking water standards and compared with the risk of THM.1 Model description1.1 Pesticide model中国煤化工,rom the fieldResidual pesticides in paddy field enter the wateassociated with irrigation water and/or rainwater anc:MHCNMHGawaterintakepoint. To model the fate in the environment , the behavior and movement ot pesticides between thewater, soil and atmosphere and the degradation in water and soil must be taken into account. Thepaddy field model was established based on the mass balance of a pesticide in the water and soil ofone unit area of paddy field(Li, 1992b) as shown in Fig.1..Paddy field water:338LI Shi-yu et al:Vol. 12PesucideVolatilization Aird(hgC.)=p-qCw- K.hwCwapplicationdtDesonptionWaterDegnadationKwghwCw + Kwh,r,C, .- KwrhwCw，(1)RunofTPaddy field soil:AdsorptionSoilDegradationd(h,r,C,)=- K,h,r.C,- Kwhsx.C.dFig. 1 Concept of pesticide fate in paddy field+ KwshwCw.2)where Cw andC。are the concentration ofpesticide in paddy field water and soil(mg/m3 ，mg/t), respectively; hw is the depth of paddy fieldwater(m); h, is the effective soil depth of adsorption(m); t is time(d); P is the rate of pesticideapplication (mg/(m2.d)) ; q is the direct runoff of water from one unit area of paddy field( m/d);r。is the specific gravity of soil(t/m' ); Kw is the first order degradation rate of pesticide in water(1/d); K。is the first order degradation rate of pesticide in soil(1/d); K w is the volatilization rateof pesticide from water to atmosphere(1/d); K ws is the adsorption rate of pesticide from water tosoil(1/d), and Ksw is the desorption rate of pesticide from soil to water(1/d). At present lttleinformation about dynamics of adsorption and desorption of pesticides is available. A third equationhas to be introduced to express the exchange of chemicals between the water and soil. Suppose thatthe adsorption of pesticides in the water onto the soil and the desportion can reach an equilibriumstate in a short period of time, the Freundlich' s adsorption isotherm is linear，and the adsorbent isthe organic carbon of the soil, the following relationship is given:C, = KoM,Cw，(3)where Ko is the adsorption equilibrium constant(m'/t); and M, is the organic carbon content ofsoil. By substituting Equation(2) and (3)into Equation(1), the pesticide model is obtained:(hw + h,y.KoM,)= p-qCw-Cw”- KwhwCw- KwnhwC.- Kghsr,KxM,Cw.(4)1.2 Water managementThe water management for rice cultivation depends 5toa great extent on the breed of rice, the texture of the客1es, -soil and the weather conditions. Farmers usually decide 日SMAAthe way of irigation based on these factors as well as李'20 400 80 100their personal experience. Fig. 2 shows a typical pattern1. deys after transplantof irigation practices in Japan， In order to keep theFig. 2 Typiceal irigatin patternwater depth in the field during various stages of ricegrowth, the consumed water is to be supplermented. The water consumed per day is the waterrequirement expressed in water depth per day, which reflcts the evapotranspiration and the lossesdue to percolation. Most of the paddy field has a water requirement ranging from 1.5 to2. 0 cm/d.The volume of runoff water from one unit o[ paddy field can be estimated by:|f(d - 8),if R≤dq:if(R(5)where d is the water requirement (m/d); 8中国煤化工); R is the rainfll .(m/d); and f is the runoff coefficient of padd:YHCNMHG2 Model application2.1 Model verificationThe pesticide model was verified (Li, 1992b) before application. The experimental site usedin the model testing is a 2760 hm2 watershed located at the upstream basin of Shibuya River inNo.3Modeling the fete of paddy field pesticide in surface water and environmental risk asessment339Kanagawa Prefecture of Japan. The watershed includes 700 hm2 of paddy field. Three herbicides.chlornitrogen( CNP), thiobencatb and simetryn, were popularly used in the rice transplant periodand their concentrations in the river were monitored for seven weeks with a sampling interval of oneweek( lizuka, 1982). It was found by comparing the calculated results produced by the model withthe observed data that the model simulates the runoff profiles quite well.2.2 River basinThe pesticide model is applied to the Yodo River Basin， Japan， to predict the pesticideconcentrations in the river. The Yodo River is merged by the Kizu River, Uji River and KatsuraRiver at one point and serves as drinking water sources of over ten million people in the Kansaiarea. Many water intakes are located along the riverbanks. Two main water intakes, Isojima andKunijima，are located at the left bank, 6. 4 km downstream of the junction, and at the right bank,23. 8 km downstream of the junction, respectively. They have capacities of 200000 and 1500000m23 /d, respectively, and serve a total population of about 700000. Due to the difference in waterquality of the three confluent tributaries, transverse mixing occurs in the mainstream. Li et al.(Li, 1987) introduced a factor called mixing ratio, which is the volumetric share of watercontributed by a tributary within one unit of water sampled at an intake, to represent the mixinglevel at the intake. Therefore, the concentration at the intakes could be expressed as:C=之i aijC,(i = 1,2)，(6)where i represents Isojima(i= 1) and Kunjima(i= 2) intake, respectively; j represents the KizuRiver(j= 1), U]ji River (j=2) and Katsura River(j = 3)，respectively; C; is the concentration atthe ithintake;G is theTable 1 Data of the Yodo River Basinconcentration of the jth tributary;and aju is the mixing ratio of theRiverFlow rate. Rasin area, Paddy are8,Mixing ratio, %m'/skm2kmn2Isojima Kunijimaith tributary at the ith intake.The values of mixing ratio could beKizu R.201643164.723.12 12.74detrmined by simulation with aUji R.044334626.068.86 66.24two-dimensional diffusion equationKatsura R.33117281.88.0221.02(Li, 1992a). The river' s hydraulicdata and areas are listed in Table 1.2.3 Pesticide applicationOf the pesticides used for paddy field, six are chosen for simulation, including two insecticides(ienitrothion (MEP) and diazinon), two fungicides (kitazin-p( IBP) and isoprothiolane) and twoherbicides ( CNP and thiobencarb). The principles toBPdiazinonMEPchoose these pesticides are: (a) they are popularly used inisoprothiolanesoprothiolaneIBPthe basin; ( b) their physico-chemical properties and. CNPenvironmental dynamic data are known; and(c) there arthibencarbdrinking water quality standards for them.Fig.3 shows a typical schedule of application MtyJuneJuly Aug Septrecommended by the pest and blight prevntion and conrol。: sinetaletieidrerpieaien .guidelines proposed by the local government. The situ中国煤化工of paddy field pesticide application in the Yodo River:TYHCNMHG340LI Shi-yu et al.Vol. 12Table2_ Pesticide application in Yodo River Basin(1991)Use rate.NumberArea of spplication, hm2Pesticideskg/hm2of timeKizu R.Uji R.Katsura R.MEP0.70218875271864I 052520Diazinon1341860IBP6.80S16460Isoprothiolane832106NP2.8039214719279Thiobencarb16107662706_Generally, the season of rice seedling transplant may continue for several weeks. This resultsin different starting time of irigation as well as pesticide dusting. In case where the study area islarge, it is possible to consider these human activities as random events occurring within a certainlength of time. In this study, it is assumed that these events obey a standard normal distribution.The daily use of pesticides is estimated based on this assumption.2.4 Environmental dynamic dataSeveral environmental factors in the model have to be determined. However, it is difficult toobtain the environmental data of all the pesticides from experimental results because the number ofpesticides in use is huge. Many researches were focused on the relationships between the physico-chemical properties of organic compounds and their environmental dynamics ( Kenaga，1979;Kanazawa，1987). By using these relationships, the environmental dynamic data of the sixpesticides are estimated and shown in Table 3. Other parameters of the model are shown in Table4. The volatilization rate is estimated by using the following formula(Liss, 1974; Mackay, 1975):Table 3 Physico-cbenical properties and eovironmental dyoamies of pesticidesMVP(20C)，K..K,PesticideMWlogPmKu，1/377.2 142.046.1x10一曲10223.16x10-26.32x10-2304.3 413.145743.72x10-27. 44x10-2288.3 4002. 082. 3x10-6621.13x10-12. 26x 10-1290.4 482.811.4X 10~5194.47x10-38.94x 10-8CNP318.5 0. 252.671.7x 10。93572.51 x 10-35.02x 10-3257.8 303.421.5x10-5728.52x10-31.70x10-2Note: Po is the octanol/water parition coefficientKw= 1/L.[1/KL + 1/(HK,)]-',(7)KL = 4. 752(44/MW)47,(8)Kg = 720(18/MW)1/2,(9)H = 16.04VP. MW/(WS●T),(10)where Lw is the depth of liquid (m); H thedimensionless Henry' s Law constant; T the2$，d，δabsolute temperature(" K); MW the molecularmc/m’m/d m/dweight; WS the water solubility ( mg/L); and1.50.015 0.005 0.6VP the vapor pressure( mmHg) .2.5 ResultsThe pesticide concentrations in the threepesticide model under the above mentioned中国煤化工in the basin of eachtributary was used as input data. For the1YHC N M H Ghe anally averagedconcentration is given and shown in Table 5, rnvugul Lue uuCI wwu givc dynamic output. Theresults of simulation reveal that pesticides runoff from paddy field is related to the amount of use,but it is dominated mainly by the physico- chemical properties. Pesticides with high watersolubility, such as IBP , runoff relatively easily with irigation and rainfall water and as a result, itappears in the river in high concentration after the application. As contrast, despite its large“No. 3Modeling the fate of paddy field pesticide in surface water and environmental risk ssessment41amount of use, CNP ,Table 5 Pesticide concentrntiom in the river by model simulatlonwhich has low waterAverage concentration, ug/LMaximum concentration +xg/Lsolubility and strongPesticideKicuR.__ UjiR. KasuraR._ KiruR. UiR__ Katsura R.adsorptivity, appears in MEP0.200.1060.095.43.82.5disazinon0.040.01).5the river in lowIBP0.640.0020.63..4concentration and last isoprothiolane0. 640.360.1:9.0.050.020.5.2for a longer time.thiobencarb0.680.590.1810. 111.94. g3 Risk assessment3.1 Comparison with the drinking water standardComparing the risk agent concentration with the drinking water standard is one of the mainmethods of risk assessment for water sources. In Japan, the environmental standard and drinkingwater standard for 15 items of pesticides were set in 1993， including the six items in this study.These chemicals had been used for more than one decade by then without regulation for water. InChina，there is no environment or drinking water standard for any pesticide in current use.Concentrations of paddy fieldTable 6 Concealntion ot risk gents nt water intakepesticides at Isojima and KunjjimaConcentration al intake, ug/L Drinking waterintake are shown in Table 6, associatedChemicalIsojimuaKunjimastandard, g/Lwith the drinking water standard levels.-The concentrations were calculated by MEP0. 12(4.00%) 0. 1(3.67%)Equation (6) with annually averaged diarion0.02(0. 40%) 0.02(0. 40%)concentrations in the three tributaries of0.22(2.75%) 0. 16(2.00%)the Kizu River, Uji River and KatsuraRiver as shown in Table 5, and mixingisoprothiolane0.40(1.00%) 0. 3S(0.88%)0.21(4. 20%) 0. 20(4.00%)ratios in' Table 1. Because thesepesticidesnon-carcinogenic thiobencatb0.58(2.90%) 0.51(2. 55%)chemicals，the standard levels are THMPP47. 5(47.5%) 59.2(59.2%)100*determined based on the acceptable daily Notes: the data in parentheses Teree percentage of drinking warerintake values for human being ( ADI ).standard;。the standard is set for THM.The drinking water standards are the ADI value in drinking water (ADId), which is 3% of theADI for MEP and 10% for the others.The results show that at present the pesticides appear in low concentration in the water, about0.36% to4. 18% of the ADId' s values. Attention must be paid to the tributaries in which some ofthe maximum concentrations of the pesticides are higher than the standards: MEP in Kizu River,IBP in Kizu River and CNP in Uji River. Therefore, there still exists the possibility that thepesticide concentrations at the intakes exceed the standards.3.2 Comparison with the THMThe load of THM formation potential (THMFP) in a river includes the contribution from bothwastewater discharge and runoff water swept from farm and forest areas. By plotting the surveyedTHMFP load data of the Kizu River, Uji River and Kat:oainst the flow rates, Li et al.(Li，1996) found that the THMFP loads increase whe中国煤化工and a powerfunction could be employed to express the relationship:YHCNMHG.Lτ= aQ'(11)where Lr is the THMFP load in a river (kg/d); Q is the river flow rate(m'/s); and a and b areparameters. Regression analysis was conducted with the surveyed data to determine the values of aand b and the correlation coefficient r. The correlation coefficients of the three rivers range from0. 90,to0.98, which indicate high correlation between Lr and Q.万万数据342LI Shi-yu et al.Vol. 12THMFP loads in the three rivers are predicted by Equation(11), and the concentrations at theintakes are calculated by Equation(6), respectively. The results are shown in Table 6. Assumingthat all THMFP transforms into THM after chlorination in the water purification processes, theTHM content is about 50% of the standard value. The predicted peak concentrations of THMFP inIsojima and Kunjima would be 109 and 91 ug/L，higher than or close to the standard level.Obviously, the risk of pesticide pollution is much lower than that of THM in the water sources ofthe Yodo River. At present, the priority of health risk management for drinking water could begiven to THM control. However, this does not mean that the contamination of pesticides could beignored.3.3 Tbe problem of standard for pesticidesThe drinking water standards are set with targets for individual pesticides. It has clearmeaning when only one pesticide in water is considered and it is easy to judge if the water issuitable for drinking based on information of water quality monitoring. For carcinogenicsubstances, the standard values are set with a carcinogenic rate in the lifetime of 100000 peoplewho are exposed to the chemicals. For non- carcinogenic substances, they are set with an ADIdvalue that mainly reflects the chronic effects on human. If the concentration of a non-carcinogenicpesticide is lower than the standard value, however, it is difficult to employ the ADId value toestimate the potential risk because little information is provided by the standard. When multiplepesticides appear in water simultaneously and each of their concentrations is lower than thecorresponding ADId value, it may even have problem to judge if the water is safe for drinking withthe standard. From the viewpoint of risk assessment , since these organic compounds are toxic andmay cause adverse health effects on human， the sum of them should be more dangerous thansingle one. Unfortunately, no well-developed method is available at present to evaluate this kind ofrisk of multiple non- carcinogenic chemicals whose concentrations are lower than the standards.Nakanishi ( Nakanishi，1994) proposed a provisional index by the sum of ratios of eachconcentration to its standard value:Rr=Zc/ADld，(12)where Rr is the total risk index, n is the number of risk agents, and Ci and ADld, are theconcentration and drinking water standard of the ith risk agent (um/L), respecively. If Rr>l,then the water is considered as unsafe.The Rr values are calculated for the six pesticides in the water at Isoima and Kunjima withEquation(12) and the results are 0. 1516 and 0. 1330，respetively. Although the values are stillsmaller than 1, they are much bigger than that of a single substance.It must be noticed that pesticides regulated by the drinking water standard are only a smallpart of the total number of agricultural chemicals in use. In Japan, over 100 items of pesticideshave been registered for use. Most o[ them are toxic to human, animal and the ecosystems andshould be regulated as risk agents. The total risk of pesticides would be much higher when all ofthem are taken into account.4 Risk management中国煤化工4.1 Pesticide use controlControling the use of pesticides is essenMYHCNMHGdbythechemicalsindrinking water. Pesticides should be refrained from use and in fact this could be done in manycases. When it is a necessity, they should be applied in proper manner, in terms of the use rate,the dusting method and timing. Low toxic pesticides are recommended.4.2 Irrigation managementAfter the pplication of pesticides, irigation is the most important human activity afectingNo. 3Modeling the fate of paddy field pesticide in surface water and environmental risk asesssentthe runoff of the chemicals from paddy fields. Simulation of CNP and thicbencarb was done withthe model for two cases of irrigation management. Case 1 is the traditional irrigation pattern whilecase2 has a three days stop of irrigation water after the application of the pesticides. The resultsshowed that runoff rates of pesticides could be reduced by 12% to 17% in case 2. Other methods ofwater management such as repeatedly using the irrigation water can also be utilized to protect therunoff of pesticides from paddy field by extending the retention time and cutting down the volumeof runoff water.5 ConclusionResidual pesticide in surface water is one of the main sources of environmental risk in drinkingwater. A mathematical model has been developed to predict the runoff of pesticide from paddy fieldin which various kinds of pesticide are intensively applied during the rice-farming season. Thesimulation results provide daily variation of pesticide concentration in a river and the amount ofrunoff. The model is applied to the Yodo River Basin. The health risk caused by the pesticide inthe water sources are estimated and evaluated by comparing the concentration with the drinkingwater standard and by comparing with the risk caused by THM. It is indicated that controlling theuse of pesticide and employing appropriate irrigation pattern are important ways to manage therisk.References:KonazawaJ, 1987. Environment Monitoring and Assenent[J], 9(1):57- -70.lizuka H. Iwanade S, 1982. Water and w astewaterl[J].24(6):13- -19.Kenagsa E E, Goring A 1, 1979. Aquatic toxicology[ M]. ASTM sTP.707 :667.LiS, 1992. Water Science and Technology[J], 26(7-8) :1823- -1830.Lis, MigiaJ, 1992. Water Science and Techanology[J], 25(11):69- 76.Li s, Morioka T.1996. Acta Scientiarum Naturalium Universitstis Sunyaseni[J], 35()111-11511Lis, Yagi S, Sueishi T, 1987. Proceedings of the 31st Japanese conference on hydraulics[M]. Tokyo: Japan Society of CivilEninering. 311-316.LissPS. Slater PG, 1974. Nature[J], 247:181-184.Mackay D, Leinonen PJ, 1975. Environmenal Science and Technology[J], 9(13);1178- -1180.Nakanishi J, 1994. The environmental surategy for water[ M]. Tokyo: Iwanami Press. 149- 163 .(Recived for review December 18, 1998. Accepted March 8, 1999)中国煤化工MYHCNMHG