SIWAPRO DSS: A TOOL FOR COMPUTER AIDED FORECASTS OF LEACHATE CONCENTRATIONS

Transcription

1 SIWAPRO DSS: A TOOL FOR COMPUTER AIDED FORECASTS OF LEACHATE CONCENTRATIONS Oliver Keesies KP Ingenieurgesellschaft für Wasser und Boden bh Bahnhofstr. 7, D Gunzenhausen, Gerany Oliver.Keesies@ibwabo.de René Blankenburg Departent of waste anageent and containated sites Technical University of Dresden Pratzschwitzer Str. 15, D Pirna, Gerany Rene.Blankenburg@ailbox.tu-dresden.de KEYWORDS Model, leachate concentration, unsaturated zone. ABSTRACT Leachate forecasts are claied to evaluate the hazard to the groundwater caused by containations in the subsurface. It is advisable to eploy a nuerical odel to siulate the coplex nature of hydraulics and solute transport in the unsaturated zone. However any paraeters are needed to forulate an appropriate siulation odel. To iniize the efforts and the costs of exploration, knowledge of the ain processes is necessary. Scenario and sensitivity analyses were done to evaluate the influence of each paraeter on the results. The analyses showed, that there is no need for transient siulation if the solute is not degradable. On the contrary, if there are organic copounds, the solute concentration is very high subjected to tie and also to degradation. INTRODUCTION Leachate forecasts are claied to evaluate the hazard to the groundwater caused by containations in the subsurface. The source is described by a concentration either in the leachate or in the soil vs. tie. The hazardous aterial is subjected to retardation and in soe s also to degradation or decay during its transport as solute through the unsaturated zone to the saturated or groundwater zone. the fluctuating groundwater table. The assessent of these paraeters is expensive and prone to errors, the uncertainty of the values is very high. Therefore we need to know, which paraeters are really iportant to the result and need to be assessed ost accurately and which paraeters do not have such an influence on the calculated concentration in the leachate. In a first step, we took a real for our scenario- /sensitivity analysises of boundary conditions and paraeters. At the chosen site a chroiu plating factory was established. The factory started at the end of the sixties and worked until the beginning of the nineties. A assive groundwater containation with chroiu and chroate was detected during the investigations in the nineties. The arising questions were 1. what will be the highest chroate concentration in the leachate 2. how long does it take until the ax. concentration reaches the groundwater table. which chroiu ass will enter the groundwater per year u. GOK (19,15 NN) 0,0 1,0 2,0,0 4,0 5,0 6,0 7,0 19,15 191,15 190,45 188,15 186,65 2,00 A, x, u', h' 2,70 S, fs, gs' Grundwasserspiegel (, 8)5,00 S, fs, u' 6,50 S, fs, u', gs', fg', g' B2 1,0 2,0,0 4,0 5,0 6,0 7,0 Feststoff: Chro VI in g/kg , , 0,12 1,0 2,0,0 4,0 5,0 6,0 Eluat: Chro VI in g/l ,9 15,4 2,4 0,59 river runoff evapotranspiration rainfall groundwater Figure 1: Illustration Figure Because of the coplex nature of soils and their transport properties it is recoended to use coputer odels for the leachate forecast. But quite a lot of paraeters are needed to describe the properties of the source, of the transport through a heterogeneous unsaturated zone and also of the cliatical boundary conditions and Figure 2: Soil and Concentration Profile at Location B2 MODELS The scenario and sensitivity analysis were done using the coputer progra SiWaPro DSS. The nae is the geran synony for Sickerwasserprognose (leachate forecast) Decision Support Syste. The progra is based on the coonly used siulation code SWMS_2D (Šiunek et al. 1994). Flow Model The flow odel describing unsaturated one diensional vertical water flow in the unsaturated zone is given by RICHARD s-equation (1)

2 h k z z and θ hp = C( hc ) t t θ t p ( θ) + 1 = w0 (1a) (1b) where the independent variables are tie t and spatial coordinate z. The dependent variables of equation (1) are the water pressure head h p =p w /ρ w g (h c = -h p ) and the water content θ. w 0 is the sink/source ter. The capillary capacity function C(h c ) is the first derivative of the hysteretic soil water retention curve drawn in Figure. The unsaturated hydraulic conductivity k(θ) depends on the water content in the soil. The hysteretic paraetric odel of soil water retention curve is given after van Genuchten (1980) and Luckner et al. (1989) by: φ A B θ = A + (2) 1 n 1 n [ 1+ ( α ) ] h c The paraeters of equation (2) are the porosity φ, the residual water content θ W,r, the residual air content θ A,r, the scaling factor α and the slope paraeter n. Figure shows a typical curve set for this hysteretic function, where PDC Priary Drainage Curve SWC Scanning Wetting Curve SDC Scanning Drainage Curve MWC Main Wetting Curve MDC Main Drainage Curve The function of unsaturated hydraulic conductivity was odeled by Muale (1976) and Luckner et al. (1989) with S λ S k( θ ) = k0 () S S 0 The paraeters of equation () are the hydraulic conductivity k 0 (θ 0 ) at a known degree of water obility S0=(θ 0 -θ W,r )/(φ-θ W,r ), the paraeter λ and the transforation paraeter. The function of unsaturated hydraulic conductivity is shown in Figure 4. The paraeters φ, k 0 and θ 0 ust be estiated in advance using lab and/or field tests. The paraeter λ in our odel ay range between 0<λ<1, but it is kept fixed at λ=0,5. water content relative pereability B =0.144 A =0.11 capillary pressure head in c porosity φ = 0.6 residual air content B =B - =θ A,r =0.10 PDC SWC SDC MWC MDC residual water content A = A - = θ W,r = capillary pressure in Pa Figure : Hysteretic Soil Water Retention Curve obility degree S 0 = 1.0 S 0 = water content Figure 4: Relative Hydraulic Conductivity Function Transport odel The well known convection-dispersion-equation (4) is used to describe the transport processes in the unsaturated zone. sfl, (u sfl,) D = r r 1424 r dispersion s { t ass storage changes convection + µ s + γ θ degradation ters q { sin ks / sources (4) where is r R spatial coordinate t s tie θ fl/ R water content D 2 R / s dispersion coefficient (D=δ v ) δ dispersivity s kg / total spezific ass (s =s fl, +s s, ) R

3 s fl, s s, kg / R kg / R spezific ass in the liquid phase spezific ass in the solid phase u R /s ean flux γ kg /( fl s) 0 order degradation coefficient µ s order degradation coefficient q kg/( R s) sinks/sources R,fl,B indexes the space, the liquid and and the soil The convection ter describes the solute transport with the water flux in the unsaturated zone. The dispersion ter in equation (4) is the su of the olecular diffusion and the hydrodynaic dispersion. Both processes are caused by concentrations gradient. The hydrodynaical dispersion is always bound to convection, but olecular diffusion is independent fro it and ay appear without any convection. The dispersivity is an epirical paraeter, it is a easure of heterogeneity of the soil and therefore depends on the scale. The reasons for dispersion are a) different velocities in the pore channels b) different pore sizes and therefore different velocities c) different flow ties because of different flow paths d) transversal spreading of particles where ρ b is the bulk density of the soil. figure 7: influence of retardation/sorption on an ipulse on top of a colun Internal reactions (decay and/or degradation) ay be described as zero or first order process as shown in figure 8. zero order ass ds = γ dt s = s γ,0 tie t first order ass ds dt s = s = µ,0 e s tie µ t Figure 5: reasons for dispersivity in pore scale (figure taken fro Luckner & Shestakov 1991) figure 8: order and description of degradation processes Figure 6 shows an exaple for the results of dispersion in a colun experient. A short input (DIRAC ipluse) was adapted on the upper boundary condition. Figure 9: influence of degradation on breakthrough curves Figure 6: influence of dispersivity on breakthrough curves The reason for retardation is sorption. The paraeter describing the linear distribution function between the solute in the liquid and at the solid phase is the distribution or HENRY-coefficient K D. The distribution and the retardation coefficient R are related to each other through equation (5). figure 7 shows the influence of retardation of the breakthrough curve at the botto of a colun experient. θ Kd = ( R 1) (5) ρb BOUNDARY CONDITIONS AND PARAMETERS Flow Model The colun has a height of.80, which corresponds to the groundwater table taken fro Figure 2. The groundwater is the lower boundary condition, it is a first kind boundary condition of eq. (1). The flux difference between precipitation and evapotranspiration is applied as second type boundary condition at the top of the colun. This is actually a transient boundary condition as shown in Figure 10. Using transient flow conditions results in a very large the coputig tie. Therefore it is a coon siplification to assue steady-state flow conditions with a ean recharge rate. But a proof is needed, if there are no differences in the results. The space discetization of the colun was held constant with dz=0.01. Each of the soil layers needs to be de-

4 scribed in its hydraulics with 5 paraeters. Their values were taken fro the field soil description and available soil data bases, like UNSODA (Schaap et al. 1999). The choosen paraeter values are listed in Table 1. water balance in /d 0,06 0,05 0,04 0,0 0,02 0,01 precipitation potential evaporation year / tie Figure 10: Precipitation and Potential Evaporation vs. Tie Table 1: Soil Hydraulic Paraeters paraeter sybol diension infilling coarse silty iddle iddle sand sand porosity φ residual water θ w,r conetent scale factor α 1/ slope factor n hydraulic k f /s conductivity bulk density ρ kg/³ dispersivity δ 0.5 GWFA legend: GWFA groundwater level below surface Transport Model The source chroate concentration at the top if colun was chosen to 0 g/l, starting fro t=0 to t=20 a. The reaining paraeter is the distribution coefficient to describe the sorption and retardation of chroate in the unsaturated zone. More than 42 drillings were done at the site. Fro these 42 drillling ore than 100 saples were taken and analyzed for their chroiu and chroate content in the soil and in the leachate. So it should not be a proble to establish a relation between the chroate content in the soil and in the liquid. But there is still a lack of ethods and understanding. Right now it is known, that the easured chroate content in the liquid is not to copare with an actual leachate concenctration in the field scale. So we had to do, what we always do in such s, we searched in the literature for siilar s and found soe hints for chroate retardation coefficients. Fro there we calculated the distribution coeffcients using eq. (5) to K D =1, 10 - ³/kg, R 6. SCENARIOS The defined s distingiush between siplification of basic equations and paraeter odels and sensitivity calculations for paraeter values. Case 1 is the basic. It is defined as a transient siulation of water flow and solute transport in the unsaturated zone for our containated site. Coonly used siplifications of equation (1) are 2: the neglection of storage ters, which results in steady-state conditions with a ean groundwater recharge rate as given flux instead of a transient flux : gravity driven flow processes, this ist actually not physically based and will not be regarded any further A further coonly used siplification of the odel of the subsurface is 4: a hoogeneous unsaturated zone The s for sensitivity analysis are defined as 5: variation of the leachate rate (=groundwater recharge rate) in the range v N =178 /a ± 20% 6: variation of the saturated hydraulic conductivity in the range /s 7: variation of the scale factor α fro eq. (2) in the range / 8: variation of the slope factor n fro eq. (2) in the range : variation of the residual water content θ r fro eq. (2) and () in the range : variation of the porosity φ fro eq. (2), () and (5) in the range 0.5 ± : variation of the sorption coefficient K D fro eq. (4) and (5) in the range ³/kg 12: variation of the source concentration c 0 in the range g/l 1: variation of the dispersivity δ fro eq. (4) in the range The 14 variation of degradation rate µ fl fro eq. (4) could not be analyzed, because chroiu/chroate is not degradable. Please note there are no s defined to investigate siplifications of the convection-dispersion-equation (2). There ust not be any siplification, because all of incorporated processes ust be taken into account to get an appropriate concentration forecast. Case 2 is the ost coonly used siplification for leachate forecasts. The results of all other s will be copared with the results of 2. The flow proble is solved if the recharge rate at the top of colun is given (= 2). Therefore one ay not expect, that the results of s 6-10 will differ significantly to the result of 2. The ean recharge rate is given with /a = /s. This value is as the functions in Figure 11 show uch ore saller than the

5 saturated and also the unsaturated hydraulic conductivity of used aterials. The water content of these aterials is always greater than the residual water content, therefore the given recharge rate can be flow through the unsaturated zone without any resistance. Please note, that we have uch ore higher recharge rates under transient conditions. They ay be greater than the saturated hydraulic conductivity. In this s runoff occurs at the surface, the soil is saturated and not able to receive ore water. The storage capacity (represented with the paraeters φ, θ w,r, α und n, s 7-10), is neglected anyway. But there ight be a sall influence of the paraeters α und n because of the incorporation of the water content in eq. (4). This ay influence the igration velocity too. This stateent also eets the paraeters φ, θ w,r, which have an influence to eq. (2) via the calculation of the water content fro eq. (1b). 4. the tie of its occurence and 5. the total chroate ass within 100a, reaching the groundwater per ². The coparisin of results fro 1 to 2, shows that there is no need for transient flow conditions, the calculated concentration, yield, ass and ties of occurebnces are always close together or identical. 1E-00 iddle sand (layer 2) hydraulic conductivity k=f(θ W ) in /s 1E-004 1E-005 1E-006 1E-007 1E-008 1E-009 1E-010 1E-011 1E-012 1E-01 θ W 0.09 θ W 0.12 infilling v N = /s 1E water content θ W iddle sand (layer ) Figure 11: hydraulic conductivity function of used aterials RESULTS The coputational results of the basic 1 are shown in Figure 12. The highest chroate concentration in the solute will be about 25 g/l at t=24 a. This results eets the observations, which were done in the 90-ies. The concentration descends after the peak went through, because the source was reoved in the beginning of the 90-ies. Within the regarded 100 years about 110 g/² chroate will reach the groundwater. The highest yield occurs after 22 years with 105 g/a. This aount is caused by a high but shortter flux rate. The results for all other s are listed in Table 2. The coluns contain 1. the highest chroate-concentration c ax, 2. the tie of its occurence,. the aount of highest yield Y ax, Figure 12: calculated chroate concentration and su of chroate ass vs. at the botto of the colun Table 2: Calculated Results concentration yield su of ass c ax year Y ax year g/l a g/a a g a b c a b a b a b a b a b a b a > b a b a b The is as entioned above an exception. It is not accepatable to reduced the flow condtions to a gravity

6 driven flow by a given first flux boundary condition at the botto. The s 4a, b and c are definded by a hoogeneous soil colun. The results show, that there is also no significant difference to the results of 2. The soil hydraulic paraeters k f ( 6), the van-genuchten- Paraeters α, n ( 7, 8), the porosity φ and residual water contant θ w,r ( 9, 10) also do not have too uch influence on the results. Significant differences are calculated with slight changes in the ean groundwater recharge rate and all transport paraeters. chroate concecntration in g/l a 4b 4c 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 10a 10 11a 11 12a 12 1a vn s 6-10: soil hydraulic paraeters δ Figure 1: calculated concentrations at the botto of the colun CONCLUSIONS The results of the scenario and sensitivity analysss showed the thesis, that it is acceptable to assue steady state flux conditions over a long ter period instead of using a transient precipitation boundary condition the soil hydraulic paraeters do not influence the calculated concentration at the botto of a colun, if their order of agnitude was quite real estiated During paraeter identification the groundwater recharge rate source concentration and distribution coefficient ust be deterined as accurately as possible. But these stateents are only valid, if the hazard is not degradable. Further investigations will be done on organic containations like a fuel depot. First calculations have shown, that the calculated concentration is highly related to the given transient flux conditions due to the degradation. A systeatization schee ust be developed. REFERENCES Luckner, L. and W.M. Shestakov. (1991). Migration Processes in the Soil and Groundwater Zone. Lewis Publishers, Inc. 485 p. Luckner, L., M. Th. van Genuchten, D. R. Nielsen. (1989). A Consistent Set of Paraetric Models for the Two-Phase Flow of Iiscible Fluids in the Subsurface, Water Resour. Res. 25, Muale, Y (1976). A New Model for Predicting the Hydraulic Conductivity of Unsaturated Porous Media. Water Resour. Res. 12, Schaap, M. G., F. J. Leij and M. Th. Van Genuchten. (1999) Bootstrap and Neuronal Network Model, In: M. Th. Van Genuchten, F. J. Leij and L. Wu (eds.): Proceedings of the International Workshop on Characterization and Measureent of the Hydraulic Properties of Unsaturated Porous Media, Part 2, University of California, Riverside, 1999 p Šiunek J., T. Vogel and M. Th. van Genuchten. (1994). The SWMS_2D code for siulating water flow and solute transport in two-diensional variably saturated edia, Version 1.1., Research Report No.12, U. S. Sa1inity Laboratory, USDA, ARS, Riverside, CA. van Genuchten, M.Th. (1980). A Closed-for Equation for Predicting the Hydraulic Conductivity of Unsaturated Soils, Soil Sci. Soc. A. J ACKNOWLEDGEMENTS This research work was granted by the Geran Departent of Education and Science (BMBF-Bundesinisteriu für Bildung und Forschung) contract # 02WP 0192, 0242, 0502 and 050. OLIVER KEMMESIES was born in Leipzig, Gerany and went to the Technical University of Dresden, where he studied civil engineering and obtained his degree in He did his Ph.D. studies with the Dresden Groundwater Research Centre and obtained his doctor degree in 1995 fro the Technical University Mining Acadey Freiberg. He worked for a couple of years in a larger consulting copany as head of the departent Nuerics in water anageent, before he founded his own consulting copany for water and soil exploration in His work is ainly subjected to odeling of unsaturated and groundwater flow as well as fate transport in the underground. His e-ail address is: Oliver.Keesies@ibwabo.de and his Web-page can be found at RENÉ BLANKENBURG was born in Soeerda, Gerany. For his studies in geodesy he went to the Technical University Dresden where he obtained the degree 'Dipl.-Ing.' in Since August 2004 René Blankenburg is eployed at the institute of Waste Manageent and Containated Site Treatent at the TU Dresden. His task is the further developent of the existing coputer software SiWaPro-DSS, a decision support syste for seepage prognosis. His e-ail address is: rene.blankenburg@ailbox.tu-dresden.de and his Web-page can be found at