WATER QUALITY MODELING AND POLLUTION CONTROL FOR THE EASTERN ROUTE OF SOUTH TO NORTH WATER TRANSFER

253Available online at www.sciencedirect.comsCIENCEdInecr.Journal of HydrodynamicsSer.B, 2006, 18(3): 253-261 .sdlj.chinajournal.net.cnWATER QUALITY MODELING AND POLLUTION CONTROL FOR THEEASTERN ROUTE OF SOUTH TO NORTH WATER TRANSFER PROJECTIN CHINA*WANG Chao, WANG Yan-ying,W ANG Pei-fangKey Laboratory of Integrated Regulation and Resource Development on Shallow Lakes， College ofEnvironmental Science and Engineering, Hohai University, Nanjing 210098, China,E-mail: cwang @ mailsvr .hhu.edu.cn(Received Oct. 14, 2005)ABSTRACT: South to North Water Transfer Project in Chinas designed to solve water shortage problem andis the largest project over centuries to solve the water shortagesupport sustainable social and economic developmentproblem in vast areas of northern China. It comprises of threein northern China. Eastern Route Project (ERP) is oneroutes: the eastern, central and western route and this studymainly focused on the eastern route. As water quality is the keyof the three components of the water transfer project.factor for the eastern routpaper examined the mainPumping water from the Y angtze River in Y angzhou,factors infuencing water quality of the main route south of theJiangsu Province, ERP utilizes the Grand Canal andYellow River, by investigating the point source, non-pointits parallel rivers to transfer water from south to north.source (diffusive source) and internal source pollutions alongIt also employs Hongze Lake, Huoma Lake, Nansithe main eastern route and in its drainage basins, and assessingLake and Dongping Lake for flood regulation. Therethe current water quality in the waterways. According to theare two water transfer ways after Dongping Lake. Onecomplicated and combined systerms of rivers and lakes in thisroute, one-dimensional water quantity and quality model forway is northward: water goes through the Yellowrivers and two-dimensional model for lakes were developed toRiver in tunnels near Weishan Hill, enters Nan Canalsimulate the hydrodynamic and pollutant transport processes.and finally flows to Tianjing City. The total length ofThe numerical method and model algorithm were described. .this route is 1156 km, of which 646 km aresouth ofThe values of model parameters were estimated by usingYellow River, 17 km through the Yellow River andfield. monioring data along the Irouteand the iverse .modeling technique. Established models were employed to493 km north of Yellow River. The other way ispredict the degradations of CODyn and NH4*-N in the maineastward: water is transferred to Yantai City and .stream, under the conditions of current pollution loads andWeihai City through 704 km long rivers in thedifferent hydrologic conditions. Schemes were presentJiaodong area. Thirteen levels of pumping stations arecontrolling total quantities of pollutants from point source andset from Y angtze River to Dongping Lake with a totalnon-point source along the main route to secure water qualitydelivery head of 65 m. The main route for ERP is .for the eastern route.shown in Fig. 1.KEY WORDS: South to North Water Transfer, Eastern RouteThe design flow rate for ERP is 600 m'/s. As .Project, water quality, pollution loads, pollution controlSouth to North Water Transfer Project spans 23municipalities and 105 counties, it is excessive of thetotal quantity of pollutants from point sources in these1. INTRODUCTIONareas，the non-point sources in the basins, and theSouth to North Water Transfer Project in Chinainternal sources from watercrafts, resulting in very* Project supported by the National Basic Research Program of China (Grant No: 2002CB412303) and the National NatrualScience Foundation of China (Grant No: 50379012).Biography: WANG Chao (1958- ), Male, Ph.D., Professor中国煤化工YHCNMH G.254Water transport to Tianjing CityWater transpon to Yantai CilyYellow RiverDongping Lake+Nansi Lake1Luoma Lake10Hongze LakeFloodgate reach2 Sanyang River reach3 Nanyunxi Floodgate to Huaian reach4 Huaian to Hongze Lake reach5 Nanyunxi Floodgate to HongzeSanjiangying6 Huaian to Huaiying reachpumping station on7 Hongze Lake to Huaian reachYangize River8 Huaiying to Luoma Lake reach9 Hongze Lake to Fangting River reach10 Luoma Lake 10 lower lake reach12 Upper lake to Dongping Lake reachFig. 1 The main route for water transfer in eRPbad water quality in main rivers. So water pollutionriverbed mud and so on.control is vital for the success of the ERP. AccordingThe pollution loads of urban point sources,to investigations and analysis on pollution loads of thenon-point sources， and shipping internal sourcesmain waterways for water transfer, one- dimensionaland two-dimensional water quantity and qualityby corresponding methods， of which non-pointmodels were developed to simulate water quality insources assessment referred to Jeffrey 5. It indicatesERP. And the allowable quantities of pollutantsthat the amount of COD from urban point sources isdischarged into the rivers were analyzed andabout 78.5% of the total COD. So point sources arecalculated based on water quality targets designed foroverwhelmingly and should be the priority for controlwater transfer. Then pollution control and pollutantand abatement. Non-point pollution sources shouldabatement scheme were established along the mainalso be paid adequate attention, because the NH4 *-Nroute. These results provided scientific support fofrom these sources accounts for 34.5% of the totalpollution control and water environment protection inNH4*-N, and for COD, it accounts for 20.9%.ERP.MODELS OF WATER QUANTITY AND2. THE POLLUTION LOADS ALONG THEQUALITY SIMULATIONMAIN ROUTE OF WATER TRANSFER3.1 Mathematical formulation for calculating waterAccording to the characteristics of their spacialquantity and quality in riversdistribution, the pollution sources are classified aFlow dynamics and changes of the chemicalpoint sources, non-point sources and internal sourcescomposition of water in rivers and their tributaries candischarged into rivers and lakes in ERP山2. Pointbe adequately described by a system of. nonlinearpollution sources mainly include industrial andpartial dfferential equations as listed below 14.51.domestic wastewater from urban areas. Non-pointpollution sources comprise of the pollutants frondh. dQsurface runoff from agriculture lands， town lands,B9=G(1)0t dswoodlands, meadows and so on. Internal sources are中国煤化工commonly caused by shipping and released fromMHCNMH G.255n2Q|Q|. a1-i=0 (2)aζ. dhu, dhv=0(5)4/3dt dx dygA| 0t as( AA2 Rd:aC0C_d(du.. Ju.. Jr子+v( d'u. 子u)tOs)s(E。AAs-kC+r=0 (3)ddx2+ dy"where h is the cross-section average of the riverTgn^depth (h= A/B), Q is flow rate, C is solute .fv+-su(u2 +2)"(6)hp。 Iconcentration，A and B are the cross-section areaand the average width of the river, respectively, d is0vdvaζ(子2v，d2vthe rate of lateral inflow， g is the magnitude of+u=-g+Vgravitational acceleration, R is the hydraulic radius)ydx2 dyn is the Manning coefficient, i is river bed slope,E。is the coefficient of longitudinal dispersion 0),fu+ - gnv(u2+“2(7)k is self-purification coefficient, r is the substancehp。，43input from external sources, s is coordinate alongthe channel, t is time. .Two independent conditions are required at thedC. dC dCClocation of confluence and diffluence for a riverdidy 00xnetwork system , they are:(a) The condition of continuity:-| E-kC+ Sc8)Loatal = Qbranch1 + Qbranch2 + ." + Qbranchndy(b) The condition of equal water level:where u and V are depth-averaged horizontalcomponents of velocity vector of water movementZrotal= Zbranch1 = Zbranch2= "" = Zbranchnalong axes x and y，C is depth-averaged soluteFor the floodgate boundary, continuity conditionconcentration，5 is the elevation of free waterof flux and the relationship between the upstream andsurface of the water body，f is Coriolis' parameter,the downstream of the floodgate water levels shouldn is the Manning coefficient, h is the depth of thebe considered in the calculation. Taking thwater body, T and T，are the components of thesubmerged hydraulic gate as an example,wind friction stress, P。andV, are the average densityQ= ubeV2g(z,-z)2(4nd turbulent viscosity, respectively, E、and E,,are the eddy diffusivity in x and y directions, k iswhere b is the net width of water gate，μ is the fluxcoefficient influenced by submergence， e is theself-purification coefficient， Sc is the substanceunlocking height of floodgate, Z ; is the water levelinput and leak from external sources. Numericalsolutions of equations refer to W ang'upstream of the floodgate, Z;1 is the water leveldownstream of the floodgate.MODELCONDITIONAND3.2 Mathematical formulations for calculating waterDETERMINATION OF PARAMETERSquantity cand quality in lakesA two-dimensional (horizontal) model was4.1 Model gridsAccording to the hydrological characteristics ofdeveloped to describe the dynamics and quality oflakes and rivers along the main route south of Y ellowwater in lakes and large rivers. The governingRiver, one and two-dimensional water quantity andequations are 8-10:中国煤化工YHCNMH G.256quality models were adopted for simulation andrespectively in this paper. Another important modelcalculation . The main route for water transferparameter is the dispersion coefficient in the equationscomprises in total of 12 reaches (Fig. 1).of pollutant concentration."3 longitudinalOne- dimensional model was used for the reachcoefficient E、 in one-dimensional equations ofsimulation. The grid size, commonly between 50 mconcentrations in rivers is usually expressedto 200 m, was determined by actual conditions ofasE =aU H，where a is a constant coefficient ofeach reach. The grids were denser near urban areashaving more sewage outlets. Two-dimensionallongitudinal dispersion, U" is shear velocity， H ismodels were employed to simulate the four lakes inwater depth. In the two-dimensional equations ofthe route, Hongze Lake, Luoma Lake, Nansi Lakeconcentrations for lakes,he eddy diffusivityand Dongping Lake. According to the lake shapes,E,=a,U*H，E,=a,U*H ,here 0, and a，areorthogonal grids system was adopted for the partitioncoefficients of dispersionof calculation grids. The grid sizes with4X =0Ydirections, respectively. The estimated values of thevaried between 100 m and 300 m. Grid sizes wereparameters are shown in Table 1.gradually changed in the reaches between rivers andlakes to keep the numerical solution stable.4.2 Determination of parametersOpposite ling(1) Determination of pollutant degradationcoefficient15%Based on water quality monitoring in ERP, themain water pollutants are CODMn and NH4*-N in ERP.And their degradation coefficients were estimated inthe main waterways.Water qualities of reaches along the main routefor water transfer were monitored specially in Sept.,2000 by Water Resources Protection Agency of HuaiRiver. According to the monitoring data, thCalculated Q/m's'(a) Comparison of river fluxesparameters of models for reaches and lakes weredetermined using the method of inverse modeling'0.6 l Opposite line(Table 1). Parameters were also calculated using theHSY calculation method with the Monte Carlomethod and based on routine monitoring data of water0.4quality from 1997 to 1999 by the Environmental18%Science and Engineering Department of TsinghuaUniversity. Table 1 shows that the average values of0.2the parameters calculated by both two methods arealmost the same. However, the pollutant degradationcoefficients are largely different between dredged0.2 0.4 0.6Calculated h(-3.0)/mriver reaches and natural ones. And pollutant(b) Comparison of water levelsdegradation coefficients of large and newly dredgedof rivers and lakesreaches are lower than those of natural reaches. This ismainly because of the different ecological!7 pcharacteristics and geometrical dimensions betweenOpposite linegthe two kinds of reaches. In addition, the coefficients当13- 22%of lakes are smaller than those of rivers due to lowerflow rates.(2) Determination of roughness coefficients anddispersion coefficients of pollutants兰22%The roughness coefficients of rivers and lakes, inrelation with the coarseness degree of river beds, arevery important parameters in the hydrodynamicmodels.Generally speaking,roughnessCalculated C/mgL*1(C) Comparison of CODMncoefficients of newly dredged or dug rivers areconcentrationsmaller than those of natural rivers. As the rivers inERP consist of both dredged and natural reaches, theFig. 2 The co中国煤化工values andmeasureroughness coefficient of each reach was determinedMYHCNM HG.257Table 1 Determined parameters in water quantity and quality modelsRoughness coefficientsCoefficients of pollutant degradation (1/d)and dispersion coefficientsReachesMethodsof pollutantTypes .KcDMnKNH4-NnInverse modelingMean value0.095500.0834000.035 95Sanjiang Ying-0.0831680.075424Nanyunxi FloodgateHSY methodMean variance0.0498210.0140580.025800.0667020.037 98Sanyang River0.0139760.0748520.0058290.0142210.102500.1123300.038 100NanyunxiFloodgate The inlet of0.0619100.075327Hongze Lake0.0154490.0141220.0675000.0855000.040 86The outlet of Hongze0.0491790.109113Lake _Fangting RiverMean V ariance0.0078580.022050The outlet of Luoma0.0445000.0515000.028 101Lake-The inlet of0.0105870.022157lower lake0.0061830.0129620.0456000.0387000.041 85Hanzhuang Cannal-The inlet of0.0096280.021018Dongping Lake0.0028470.0030110.0258000.0195000.022585.2Huoma Lake0.0212000.0205000.02565 5.7Nansi Lake,0.0316000.02755 5.0中国煤化工YHCNMH G.2585. PREDICTION OF WATER QUANTITY ANDQUALITY, AND POLLUTION CONTROLThe Grand CanalLaoyiRiver油Yi RiverSCHEMES.1 Results of simulated water quantityThe calculation of water quantities for the mainroute of water transfer in ERP is very complex. This isHuangdun|i↓i "Luoma Lakebecause the water quantity imported from Y angtzeFlashboardPeihongXinyiRiver, imported or exported from branches all need toRiverKIVErbe considered./ZhangsanFigure 3 shows the velocity field of Luoma LakeYanghetanunder different influent and effluent conditions. AsFlashboard '}FlasboardZhongseen from Fig. 3(a), when the transferred water entersLiutang Riverinto the lakes from the Zaohe Flashboard and leavesfrom the lakes from the Huangdun Flashboard, there islittle mixing between the transferred water and theSuqiunCity ◎间water bodies of Luoma Lake. A clear mainstreamregion is formed in Luoma Lake between the inlet anda)the outlet. The velocity distribution of Luoma Lakeshows that the water in the southern and the easternThe Grd Canaloyi River用Yi Riverparts of the lake are not affected by water transfer. Butfor the northern lake, water coming from Yi River andLaoyi River is affected, and the flow velocitydistribution is modified as shown in Fig 3(a). So thequality of transferred water will be affected if YiRiver or Laoyi River is polluted. Figure 3(b) presentsPeihong RiverXinyi Riverthe velocity field when the Grand Canal flows into thesouthern part of the Luoma Lake. Seen from the figure,Zhangsan Flashboardmost of the mixing happens between the transferredwater and the lake water while the transferred waterenters the lake from the south and leaves from thewest. Obviously, Luoma Lake will be beneficial inperforwing quantity adjustment for water transfer. Amainstream region is also formed between the inletSuqian Cityand the outlet, though the velocity there is lower thanthat in the Fig. 3(a). Meanwhile, the velocity field isalso affected by the conflux of Yi River and LaoyiFig.3 Velocity field of Luoma Lake under different influentRiver in the north, so the water quality of transferredand effluent conditionswater is also affected by the two rivers.5.2 Results of simulated water quality4.3 Model validationBased on the results of water quantity andDegradation coefficients were employed in waterpollutant quantity discharged into the water bodies,quantity and quality simulationpollutant concentrations were calculated in the maincorresponding river reaches and lakes. The monitoringwaterways. The following results are obtained: thedata from Dec.8th to 12th, 2003, which was 10 timesconcentrations of pollutants increases gradually alongas many as before, was used for model validation.with current pollution loads andFigure 2 is the comparison between the calculateddifferent water quantity， the concentrations at thevalues and measured ones. It shows that error rangesinlets of the four lakes are all higher than those at the)f river flow rate， water level and pollutantoutlets，which demonstrates that the lakes areconcentration, which are within 15%, 18%, and 22%advantageous for water quality improvement in ERP .respectively, all meet the precision requirement for theThe calculation results also indicate that: the waterprediction of water quality.qualities degrade distinctively when the transferred中国煤化工MHCNMH G.259Xuhong RiverEr RiverSui RiverXinbian RiverSubei rmrigation Channel10.05.0、12.0Hui River" 14.011.0Huai RiverSan River14.02.0 4.0 6.0 10.0 15.0 (mg/L)Chi RiverThe legcnt of CODIn concentrationFig. 4 Distribution of CODMn concentration in Hongze Lakewater passes through urban areas. In particular whenmain pollutants of CODMn and NH4*-N werethe water passes through the reaches of Pizhou andcalculated using water quality models (Table 2). WithXuzhou City， the concentrations of CODMn andthe abatement in effect, the concentrations of CODMnNH4+-N drastically exceed the level of water qualityand NH4 -N will decrease to the range between 1 andtarget set by ERP. Therefore， it is necessary t6 mg/L， which meets the requirements on waterenhance pollution control in urban areas along thequality in ERP.main route to secure water quality. Figure 4 shows thedistribution of CODMn concentration in Hongze Lake.The concentration of CODMn in most areas of Hongze6. CONCLUSIONSLake is less than 8.0 mg/L as required by ERP. But inAccording to the investigation and analysis onthe southwestern water bodies where about 20% areapollution sources along the main route for waterof the lake is influenced by the pollutant inflow fromtransfer and its adjacent area, the main sources thatHuai River, the concentration of CODMn increases.influence the water quality of waterways wereConsequently, the measure of pollution control inidentified in this paper. The urban point pollutionHuai River drainage basin should be enhanced tcsources are the principal sources， the non-pointsecure the water quality of Hongze Lake for ERP.pollution sources take the second place， and the5.3 Schemes for pollution control .internal pollution sources are less significant. SoIn order to meet the requirement on water qualitywasterwater treatment plants in cities and towns alongfor ERP, stringent measures of controlling pollutionthe main route should be built as soon as possible toloads into the rivers north of Pizhou Bridge should beensure the water quality in ERP.implemented. According to the current quantity ofBased on the complicated combination of riverspollutants discharged from urban and rural areas alongand lakes in中国煤化工laracteritics,the main route, the allowable quantity dischargedone-dimensionality modelsintothe water bodies and the quantity abatement 0for rivers and tYHcNMHGlakeswere.260Table 2 Quantities of current, allowable and abatement of CODMn and NH:+-N discharged into waterwaysQuantities of CODMnQuantities of NH4+-N dischargedControl sectionsdischarged into waterways (Va)into waterways (Va)CurrentAllowableAbatementNew Tongyang Cannal745837294385830Beichengzi River6382.48095573.4383.861222.8Huai River1962.88801082.865.43431.4L aobian River2060.86021458.8184.272112.2Huaian Canal14455.91090.8Xusha River1620.31105555.3153.552Fangting River2278756001718767.7651.7Bulao River29684.501164.7Peiyan River3560.13213347.1693.576517.5Fuxin and Yi River30436.11276917667.13009.5105 11958.5Picang Floodway757623595217401.179332.1Hangzhuang Canal1688.727.5Chengguo River15075.62522.8Dongyu River2958.56482310.551.2843.2Old Wanfu River898.6522376.695.5283.5Baima River3754.76243130.71798.225973.2Si River3999.818902109.8283.20776.2Zhuzhao New River124882054104342365.8532214.8Guangfu River51306.21302338283.233111831Liangji Canal2238.67001538.6I1590.8Zhaowang River1616670946112.550.51095.4223.4Dawen River23365.5744415921.53540.13214.1Dongping Lake4445.79933452.7437.2303.2 .Total2052955.2596341993321.222993.4517617817.4 .established to simulate and calculate the flow rate,hydrodynamic condition of the main route for waterwater level, CODMn and NH4*-N concentrations intransfer is very complex. Especially in theERP. The calculated results agree with thconvergence of rivers and the lakes, the flow state offield-measured values, which demonstrates that it isthe water changes greatly. And transferred water doespracticable to use these models to simulate andnot mix completely with the lake water before flowingcalculate the water quality of the complicatedout of it, but it mixes locally with the lake water andcombination of rivers and lakes in ERP.then flows ou中国煤化工the shortestFlow rate prediction indicate that under the effectdistance.C NMH G,of transferred water and branches influx， theWater qua.y't.. ..... .hat under the.261conditions of present pollution loads and differentPEREKAL'SKII V. M. Modeling of pollution spreadingin the Northern Dvina[J]. Water Resources, 2000, 27(5):hydrologic design criteria, the water quality of the524-528.main rivers is drastically different from that of the7] LIANG Bin, WANG Chao, WANG Pei-fang. Thelakes, the water quality of the lakes is better than thatanalysis on investigation of pollutants discharged intoof rivers, the quality of the water flowing out of therivers in Complicated River network Area[J]. Waterlakes is better than that entering the lakes. TheResource Protection, 2003, 71(3): 26-30. (in Chinese)pollutants discharged into the main route of water8] ZANNETTIInvironmental modeling Vtransfer should be strictly managed, because they haveComputer methods and software for simulatingenvironmental pollution and its adverse effects[M].great impact on the water quality. On the whole, theLondon，New York: Elsevier Applied Science, 1993,water quality upstream Luoma Lake outlet in the mainroute is better than downstream one.9] SHI Jiu-xin, LE Ken-tang, YIN Bao-shu et al. Numericalsimulation of the tidal motion and suspended sedimentREFERENCEStransport in the radial sandbank area of the SouthernY ellow Sea[J]. Journal of Hydrodynamics, Ser. B,1] XU Gui-quan, CHU Jun-da. The discussion of water2004, 16(5): 55-563import used to improve water environment in Shanghai[10] ZHU Dang-sheng, WANG Chao, CHENG Xiao-bing.City[J]. Water IResource Protection, 2001, 65(3): 26-30.he theory and technique of water resourceinChinese)(in Chinese)protection and layout[ M]. Beijing; The Press of China[2] XIONG Wan-yong. The Practice and cognition ofHydraulic and Hydroelectric, 2002, 88- 100. (in Chinese)importing Vater to improve water quality of urban rivers[11] WANG Chao. Modeling of water dynamics andin Fuzhou City[J]. China Water and Waste Water,pollutant spreading in the Luomahu reservoir for Water2000, 116(3):26-28. (in Chinese)Transfer from South to North of China[J]. Journal of[3] JEFFREY A. T., WALTER R. and MARJORIE M. H.Hydrodynamics, Ser. B, 2001, 13 (4): 14-23.Assessment and control of nonpoint source pollution[12] XU Zu-xin, YIN Hai-long. Development of coupled)f aquatic ecosystems[M]. USA: The Parthenon1D-2D mathematical models for tidal river[J]. Journal ofPublishing Group Inc., 1999, 296-299.Hydrodynamics, Ser. B, 2004, 16 (6): 767-776.[4] DATSENKO Yu S.， IVANENKO S. A. an[13] WANG Chao, GU Bin-jie. The optimize calculation ofKORYAVOV P. P. Mathematical model of waterdynamics and pollutant spreading in the Ivan 'kovotransforming model[J]. Advances in Water Science,reservoir[J]. Water Resources, 2000 27(3): 259-270.2002, 13(2): 184- 190. (in Chinese)rces,Model of calculating5] WANG Chao, WANG Pei-fang.[14] WANG Dao-zeng, ZHAO Yu-feng, LIN Wei-qing.water quantity needed to dilute and purity pollutants inInverse problem method for water enviroment totalriver network and its applications[J]. Journal ofcapacity control of a tidal river[J]. Journal ofHydrodynamics, Ser. B, 2005, 17 (4): 418-428.Hydrodynamics, Ser. B, 2005, 17 (3): 323-329.[6] BREKHOVSKIKH v. F，BYLINYAK Yu A. and中国煤化工MHCNMH G.