RT3D version 2.5 Update Document
Contents
1. Zheng C and P Wang 1999 MT3DMS A Modular Three Dimensional Multispecies Transport Model for Simulation of Advection Dispersion and Chemical Reactions of Contaminants in Groundwater Systems Documentation and User s Guide U S Army Corps of Engineers U S Army Engineer Research and Development Center Vicksburg Mississippi SERDP 99 1 202. EPA a university consortium 17 RT3D v2 5 Update Document Dames and Moore Geosync collectively known as the Remediation Technologies Development Forum RTDF Dr Brian Hooker championed the project development efforts and Dr Clement led the modelling efforts related to the RT3D design tool development task The RTDF work was funded through a DOE research grant over a three year period from October 1995 September 1998 In 1996 Dr Yunwei Sun was hired as a post doctoral fellow at Washington State University in Prof Jim Petersen s group to support the debugging efforts and to help benchmark the code against various analytical solutions The second prototype version of RT3D which supported simple recharge and well packages was finalized in the Fall of 1996 This code ran with MT2RT a data pre processor developed by Dr Sun which facilitated the use of the RT3D within a beta version of the U S Department of Defense Groundwater Modelling System GMS 1 0 software Dr Sun also lead the code validation efforts by testing RT3D results against various analytical solutions including a new solution that he developed as part of this project effort Further Mr Christian Johnson at PNNL tested the prototype versions of RT3D and MT2RT codes and developed a series of test examples The PNNL WSU team jointly conducted a training course in November 1996 to disseminate the technology to various RTDF members In winter 1997 a subcontract was s
3. 90 code and should be compilable using any Fortran 90 complier 18 RT3D v2 5 Update Document REFERENCES AND RT3D RELATED JOURNAL PUBLICATIONS Clement T P 2001 A Generalized Analytical Method for Solving Multi Species Transport Equations Coupled with a First Order Reaction Network Water Resources Research 37 1 157 163 Clement T P C D Johnson Y Sun G M Klecka and C Bartlett 2000 Natural Attenuation of Chlorinated Solvent Compounds Model Development and Field Scale Application Journal of Contaminant Hydrology 42 2 4 113 140 Clement T P and N L Jones 1998 RT3D Tutorials for GMS Users Pacific Northwest National Laboratory Richland Washington USA PNNL 11805 Found online at http bioprocess pnl gov rt3d htm Clement T P Y Sun B S Hooker and J N Petersen 1998 Modeling Multi species Reactive Transport in Groundwater Aquifers Groundwater Monitoring amp Remediation 18 2 79 92 Clement T P B M Peyton T R Ginn and R S Skeen 1999 Modeling Bacterial Transport and Accumulation Processes in Saturated Porous Media A Review in Advances in Nuclear Science and Technology J Lewins and M Becker eds Kluwer Academic Plenum Publishers New York pp 59 78 Clement T P B M Peyton R S Skeen B S Hooker J M Petersen and D Jennings 1997 Microbial Growth and Transport in Porous Media Under Denitrification Conditions Experiment and Simulations Results Journal of Contami
4. RC NCRXNDATA which will be used in the user defined reaction module and in other pre defined reaction modules E6 ARRAY RC NCRXNDATA Format Free format one entry per line with a total of NCRXNDATA number of entries RC NCRXNDATA is the one dimensional array that stores all constant spatially invariable reaction parameters The type of parameters stored in this array will depend on the reaction model used in the simulation Input the variable reaction parameters number of arrays must be equal to NVRXNDATA the parameters will be stored in a variable VRC NCOL NROW NLAY NVRXNDATA which will be used in the user defined reaction module and in other pre defined reaction modules E7 ARRAY VRC NCOL NROW one array for each layer Then repeat VRC for NVRXNDATA number of times READER DPRARRAY a new array reader which is exactly similar to the reader RARRAY but with double precision Note that VRC NCOL NROW NLAY NVRXNDATA is a four dimensional array that stores all spatially variable reaction kinetic parameters RT3DV2 5 USER DEFINED REACTION MODULE In RT3D version 1 0 user defined reactions could be described in the rxns f file In RT3Dv2 5 an additional file named jacrxns f is also required to complement the rxns f file In the jacrxns f file the user is expected to provide the necessary information for computing an analytical Jacobian matrix for the reaction equations Both rxns f and jacrxns f subroutines will be dummy
5. Sequential Electron Acceptors 4 Rate Limited Sorption 5 Double Monod Model 6 Sequential First Order Decay up to 4 species e g PCE TCE DCE VC 7 Aerobic Anaerobic PCE TCE Dechlorination 8 module reserved for future implementation 9 module reserved for future implementation 10 User Defined Kinetics NCRXNDATA number of constant reaction parameter values NVRXNDATA number of variable reaction parameter arrays ISOLVER 0 for the instantaneous reaction modules 1 and 2 1 Automatic switching Gear stiff non stiff solver For stiff systems this option will automatically compute the Jacobian matrix using finite difference approximations 2 Automatic switching Gear stiff non stiff solver For stiff systems this option will require an external routine to compute analytical Jacobian Need to provide an external subroutine jacrxns f that specifies the Jacobian matrix for the differential reaction equations 3 Fehlberg fourth fifth order Runge Kutta method RT3D v2 5 Update Document 4 Stiff solver based on a semi implicit extrapolation method This option requires an external routine jacrxns f to compute the analytical Jacobian matrix for the differential reaction equations 5 Non stiff Runge Kutta solver IRCTOP This variable was included to maintain compatibility with the previous version Set the value to 0 or 1 preferably 1 to read sorption parameters in a layer by layer mode default mode
6. are appropriate If you have any questions or comments please contact the developers see above DISCLAIMER OF WARRANTY The RT3D computer code is provided without any warranty We make no warranties expressed or implied that the RT3D code is free of errors or whether it will meet your need for solving a particular problem You use the code at your own risk The developers disclaim all liability for direct or consequential damage resulting from your use of the code INTRODUCTION TO RT3D VERSION 2 5 RT3D Reactive Transport in 3 Dimensions is a finite difference numerical code for simulation of three dimensional multi species reactive transport in groundwater aquifers RT3D comes with 7 built in reaction kinetics modules but also has the flexibility to utilize reaction mechanisms of any complexity that are defined by the user RT3D calculates the reactive transport using the flow solution from a finite difference groundwater flow package such as MODFLOW RT3D may be applied to a wide variety of reactive flow and transport scenarios including accelerated in situ bioremediation system design natural attenuation evaluation and in situ chemical oxidation RT3D version 2 5 RT3Dv2 5 has been updated compared to version 1 0 in several aspects The base code has been updated to incorporate changes introduced into version 3 5 of the MT3DMS code see Zheng and Wang 1999 for details on these changes including a third order total variation dim
7. kck kck kCk ck kck kck kk ck ckck kck kk ck ckck kc k k kc k kk c List of calling arguments c ncomp Total number of components c nvrxndata Total number of variable reaction parameters to be input via RCT file c J I K node location used if reaction parameters are spatially variable c ml mu are for banded jacobian not used c y Concentration value of all component at the node array variable y ncomp c pd jacobian matrix ncomp x ncomp array C poros porosity of the node c reta Retardation factor ignore dummy reta values of immobile species c rhob bulk density of the node c rc Stores spatially constant reaction parameters can dimension upto 100 values c nlay nrow ncol Grid size used only for dimensioning purposes c vrc Array variable that stores spatially variable reaction parameters C End of Block LR KK KKK KKK KK k ck KK ck kc KK KK KK KK ck ko kc kc kck ck k kc k ck k ck k KK KK KK KK KK ck ck ck kc k kk C Block QolORCKCkCkCkCkckck kk KK KK Ck Ck KKK KK KK kc kck ck KK ck kc KKK k kc kckckck k ck kc kck ck k kc KKK ck k KK KK ck ck ck kk Ct Please do not modify this standard interface block IMPLICIT NONE INTEGER ncol nrow nlay INTEGER ncomp nvrxndata j i k ml mu INTEGER SAVE First time 1 DOUBLE PRECISION y pd poros rhob reta DOUBLE PRECISION rc vrc DIMENSION y ncomp pd ncomp ncomp rc 100 DIMENSION vrc ncol nrow nlay nvrxndata reta ncomp C End of block DRKEKKKK KKK
8. porosity of the node c reta Retardation factor ignore dummy reta values of immobile species c rhob bulk density of the node c rc Stores spatially constant reaction parameters can dimension upto 100 values c nlay nrow ncol Grid size used only for dimensioning purposes c vrc Array variable that stores spatially variable reaction parameters C End of Block LR KK KKK KK KKK k ck KK ck kc k ck k ck kc kckck ck k kc kc kck ck k kc k ck k k ck kc KK KK KKK KK ck ck ck kc k kk C Block Q olDORCKCkCkCkCkCkck KK KK KK KK KK kc k ck KK KK KK KK ck kc KK KK KK RK ck k KKK KK KK KKK k ck KK KK KK ck kk cx Please do not modify this standard interface block IMPLICIT NONE INTEGER ncol nrow nlay INTEGER ncomp nvrxndata j i k INTEGER SAVE First time 1 DOUBLE PRECISION y dydt poros rhob reta DOUBLE PRECISION rc vrc DIMENSION y ncomp dydt ncomp rc 100 DIMENSION vrc ncol nrow nlay nvrxndata reta ncomp C End of block DREKKK KKK KKK KK KK KK KK KK KK KK KK KK KKK KR KK KK KK KK KK KK KK kk kc KK KKK C Block Bek RR KKK KR KK KK kc kck KK kk ck kck kk kCk kc kck kk kk ck kck k ck kk ck kck k ck k ck ck ckck ck ck ck kck ck ck KK kk c Declare your problem specific new variables here C INTEGER DOUBLE PRECISION pce tce dce vc kpce ktce kdce kvc DOUBLE PRECISION ytcepce ydcetce yvcdce C End of block Bw KK KK KK RK KK KK KK KK Ck Ck KK Kk ek KK kk Ck KK Ie KK KK I ek KK K
9. procedures when one of the pre programmed reaction modules is used When the user defined reaction module is used jacrxns f will be a dummy procedure for reaction solvers 1 3 and 5 For RT3Dv2 5 the user now has the option of compiling the user defined reaction module into the main executable default for RT3Dv2 5 or into dynamic link libraries DLLs The latter option is only applicable to Microsoft Windows operating systems This section describes the rxns f and jacrxns f files with examples of each and the compilation options Example rxns f File The format for the rxns f file is unchanged from RT3D version 1 0 An example is provided here for reference only RT3D v2 5 Update Document c Reaction package for Example 2 method 1 c Please refer to the RT3Dv1 user manual for further details o SUBROUTINE Rxns ncomp nvrxndata j i k y dydt amp poros rhob reta rc nlay nrow ncol vrc C Block Lik RK KK KKK KK KK kk KK kc k ck kck kk ck ck ck kck kck kc kck kck kk kc kck kck k ck ck ckck kck kk ck ckck kc k k kc k kk c List of calling arguments c ncomp Total number of components c nvrxndata Total number of variable reaction parameters to be input via RCT file c J I K node location used if reaction parameters are spatially variable c y Concentration value of all component at the node array variable y ncomp c dydt Computed RHS of your differential equation array variable dydt ncomp C poros
10. 99 e Dr Chunmiao Zheng MT3D related technical support 1996 present e Dr James Petersen WSU project management and technical support 1995 1997 e Dr Norman Jones GMS related technical support 1996 present e Mr Christian Johnson Technical support documentation application of RT3D to waste sites publicity 1996 present e Dr Yunwei Sun Debugging development support validation and MT2RT support 1996 1998 e Mr Mike Truex Project management funding and application of RT3D to waste sites 1999 present e Dr Tirtha Gautam Code management and development support 2001 present BRIEF HISTORY OF RT3D In 1993 the U S Department of Energy U S DOE funded a field project for researchers at the Pacific Northwest National Laboratory PNNL to demonstrate pilot scale bioremediation of carbon tetrachloride at the U S DOE s Hanford site in southeastern Washington State The RT3D code development activities were initially funded from February 1994 to September 1995 as a part of this bioremediation demonstration Dr Clement originally developed a one dimensional finite difference code IDCART for simulating bioreactive transport in a denitrifying soil column reactor In winter 1995 the one dimensional model was applied to simulate Hanford bioremediation scenarios including carbon tetrachloride removal rates in a bench scale soil column reactor Later the PNNL research team including Drs Clement Hooker Skeen and P
11. Ck KK Kk Kk KK KK KK KK kk KK KK Ie kk KK KKK KK KK ko kk c Assign or compute values for new variables if required pce y 1 tce y 2 dce y 3 vc y 4 C End of block BKK KKK KK RK KK kk Ck KK Kk Kk KK Ik Kk KK Ik Kk KK Ie kk KK KKK KK KK ck kk C Block Or kK KKK KKK KK KK kk KK kk kc kck kk kk ck kck kk ck ck ck kck kk ck ckck kck kc k ck ckck kck kk ck ckck KK k kc k kk o Definition of full nx n analytical jacobian matrix pd 1 1 kpce reta 1 pd 1 2 0 0d0 pd 1 3 0 0d0 pd 1 4 0 0d0 pd 2 1 kpce ytcepce reta 2 pd 2 2 ktce reta 2 pd 2 3 0 0d0 pd 2 4 0 0d0 pd 3 1 0 0d0 pd 3 2 ktce ydcetce reta 3 pd 3 3 kdce reta 3 pd 3 4 0 0d0 pd 4 1 0 0d0 pd 4 2 0 0d0 pd 4 3 kdce yvcdce reta 4 pd 4 4 kvc reta 4 C End of block ORK KK KK KK KK Ck Ck KK KK KK KK KK KK KK KK KK KK KK KK KK KK KKK KK KK KK KK RETURN END 14 RT3D v2 5 Update Document EXAMPLE DUMMY JACRXNS F FILE An analytical Jacobian matrix will be used by RT3Dv2 5 only when solver options 2 or 4 are selected However a jacrxns f subroutine is required regardless of the solver option RT3Dv2 5 comes with a dummy jacrxns f file but for completeness the source for such a dummy file is shown below o Dummy jacrxns f file SUBROUTINE jacrxns ncomp nvrxndata j i k ml mu y pd amp poros rhob reta rc nlay nrow ncol vrc C Block Lik RK KK KKK KK KK kk KK kc k ck KK kk kc kck kck kk kc
12. D Evaluation of Jacobian Matrices If solver options 2 or 4 are invoked the user should provide the necessary information for computing an analytical Jacobian matrix for the differential reaction equations The analytical Jacobian matrix is of size NCOMP x NCOMP and the elements are the partial derivatives of the system of differential equations For a system of n number of differential equations F the Jacobian matrix is defined as in Equation 2 OF OF OR ax Ox x J x dx Ox x 2 OF OF OF x Ox x where F f x x Xn is the differential equation for species i AN EXAMPLE OF SETTING UP A JACOBIAN MATRIX This section illustrates the method for determining the analytical Jacobian matrix for a simple reaction system The test problem models first order degradation of PCE and its daughter products under anaerobic conditions For more details please refer to the Example 3 discussed in Clement et al 1998 The transport and transformation of the sequential decay chain PCE gt TCE DCE VC can be determined by solving the set of partial differential equations in Equations 3 to 6 10 RT3D v2 5 Update Document R S PCE K PCE 2 b Te APE a PCE 3 Xi ot Ox Ox Q z TCE I Y vei Peek pee PCE K ive TCE 4 Xi R T o TCE f ATCEJ Avi TCE at Oxi Ox R 6 DCE p 2E apoE 4 IDCE Yoon Kn lIl K DCE 5 t Ox ij x O
13. E CEE TETE E SEERA GORE EE ESERE E ERES 10 An Example of Setting up a Jacobian Matrix esses eene nete rennen nre 10 Example jar rns f ir 13 Example Dummy Ja rxhns f File astuce n enc th at NAE aE bei be depo eO eo ce 15 DLL VERSUS NON DLL USER DEFINED REACTION MODULE eee ee e e e ne nennen 15 How to Implement the DLL Option essent etre eerie eren ennt 16 USING MODELOW 2000 T E 16 ACKNOWLEDGMENTS sissesscsciiccscsnsadentedisutscossSeosacas 16 BRIEF HISTORY OF RT3D 5 cess scacscsusasssacossosectstcnescdasessscscsoossscosdessubescdessdnseosesas stesesasseesas ac Ue scvadsdasessesoss 17 REFERENCES AND RT3D RELATED JOURNAL PUBLICATIONS e eeeeeeeee enne netten ntn tn ntn 19 CONTACT INFORMATION There are several resources for RT3D users The official RT3D web site provides information on features manuals downloads and utilities Several of the developers of groundwater modelling packages that have RT3D interfaces e g GMS Visual Modflow Groundwater Vistas have support forums accessible thorough their respective web sites If those resources do not provide the answers you seek contact information is provided below Direct any general questions on RT3D use functionality documentation etc to Christian Johnson Complex questions wish list requests and RT3D project related inquires should be directed to Dr Clement Contact C
14. K KKK C Block As KKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKK Ck kc k ko c Initilize reaction parameters here if required IF First time EQ THEN kpce 0 005 PCE first order degradation rate ktce 0 003 TCE first order degradation rate kdce 0 002 DCE first order degradation rate kve 0 001 VC first order degradation rate ytcepce 131 36 165 8 ydcetce 96 9 131 36 yvcdce 62 45 96 9 First time 0 reset First time to skip this block later END IF C End of block LOKCKCKCKCKCKCkCKCKCKCk KK KK KK KKK KK Ck Ck KK KK KK KK KK KK KR KK KK KKK KK KK KK KKK KKK n Lp GI RT3D v2 5 Update Document C Block Br kK RK KK KK KKK KK KK KK Kk Kk KK Kk Ck KK KK KK KK kk KK KKK KK KK KK KK kk KK KKK e Assign or compute values for new variables if required pce y 1 tce y 2 dce y 3 ve y 4 C End of block BKK KKK KK KK KK KK KK KK KK KK KK KK KK KK KKK KK KK KK KK KK KK KK KK KKK C Block Q CKCKCKCk Ck kk KK k Ck ck ck kk kc kck KK kk ck kck kk kk ck ckck kk kc k ck kck kk kc kck kck KK kckck KK kk k kc k ck k kk e Differential Reaction Equations dydt 1 kpce pce reta 1 dydt 2 ktce tce kpce pce ytcepce reta 2 dydt 3 kdce dce ktce tce ydcetce reta 3 dydt 4 kvc vc kdce dce yvcdce reta 4 C End of block Q KK KKK KKK KKK KK Kk Ck KK kk Ck KK kk KK KK KK KK KK KKK KK KK KKK KK KK RETURN EN
15. KKK KKK KK KK KKK KK KK KKK KK KK KR KK KK KK KK KK KK k ko k ck kc k ko kk ok write Use solver option l or 3 write Jacobian routine is not available for this module STOP RETURN END DLL Versus Non DLL User Defined Reaction Module Numerous types of biological geochemical reactions occur in subsurface environments and it is impossible to develop a code that could model all these reactions Therefore RT3D provides a versatile user defined reaction option that allows the user to simulate reactive transport based on his her own reaction kinetics Using this option any type of kinetically limited reactive transport problem can be formulated and solved As discussed above user defined reactions in the RT3Dv2 5 environment are incorporated into the two Fortran subroutines rxns and jacrxns At the compilation stage these two routines are either compiled together with the rest of the RT3D code to build a single executable which is the standard option for RT3Dv2 5 or on 15 RT3D v2 5 Update Document Microsoft Windows operating systems they can be complied into separate dynamic link libraries DLLs Fortran make files are provided with RT3Dv2 5 for building a non DLL version of RT3D using the Compaq Visual Fortran compiler and a Fortran 90 compiler on a UNIX system How TO IMPLEMENT THE DLL OPTION To create a version of RT3Dv2 5 that uses DLL files which was the
16. NONE INTEGER ncol nrow nlay INTEGER ncomp nvrxndata j i k ml mu INTEGER SAVE First time 1 DOUBLE PRECISION y pd poros rhob reta rc vrc DIMENSION y ncomp pd ncomp ncomp rc 100 DIMENSION vrc ncol nrow nlay nvrxndata reta ncomp C End of block DRKKKK KKK KKK Ck KK KK KK KK KR KK Ck Ck KKK KK KKK Ck KK KK KK KKK KK KK KK KK ko kk C Block Bek RR KKK KK KK KK kc kck KK kk kc kck kk kk ck ckck k ck kk kc kck k ck kk ck kck k ck k ck ck KK k ck kckck ck ck ck kc KKK e Declare your problem specific new variables here DOUBLE PRECISION pce tce dce vc kpce ktce kdce kvc DOUBLE PRECISION ytcepce ydcetce yvcdce C End of block QU KCKCkCkCk KCkCk KKK KK KK KK Ck Ck KK Kk Kk KK Ik KK KK KK KK KK KKK KK KK KKK C Block L ozCKCKCKCKCKCK kk KKK KK KKK Ck KK KKK KK KKK KK KK KK Ck Ck KK KK KK KKK KK KK ck ck KK KK KK KK c Initilize reaction parameters here if required IF First time EQ THEN kpce 0 005 PCE first order degradation rate ktce 0 003 TCE first order degradation rate kdce 0 002 DCE first order degradation rate kve 0 001 VC first order degradation rate ytcepce 131 36 165 8 ydcetce 96 9 131 36 yvcdce 62 45 96 9 First time 0 reset First time to skip this block later END IF C End OE block kKkKKKKKKKKKKKKKAKKKKKKKKKAKKKKAKKKKAKKKKAKKKKAKAKAKKKAKAKKKAKAAKKKAKAKAKGK hb Ea Falk Ea C Block Br kK RK KK KK KKK KK Ck KK Ck
17. RT3D v2 5 Update Document What s New in RT3D version 2 5 January 2002 with minor update February 2003 CONTACT INFORMATION cssicsscascessssscscstescicacscsacesscastsscctsisnesedasesisess basastesssdsascosesasessaccsssoessas done sctancsssaugeassocacdszese 1 LICENSE AND COPYRIGHT INFORMATION cscsssssssssssssssscsssesssscsssesessessssssesersssssesesesessesssesssscsseesesssssoes 2 DISCLAIMER OF WARRANTY iccsissssscsossessscscoossesascesscnrscasesseuessscss seoscsstscesevsacasucssisensevsnusebassancssbieenescoetsessbsnediess 2 INTRODUCTION TO RT3D VERSION 2 5 scssssssssssssssssesssessssesscsssrscssesosessesersesssssesssesesscessesesssessessssssssseseesoss 2 RT3DV2 5 SOURCE SINK SSM PACKAGE ee eeee eene enne n tnt tnn enses snas tn sone tasse tosta sensn ses sn seen suse ta snue 3 SSM FILE STRUCTURE 5 ecciesie EEE ve AA eei esee eds sebuesctecrneeesscbicssnsstesdantenses 4 Comment on Using Artificial Injection Wells to Model Sources seen 4 RT3DV2 5 REACTION RCT PACKAGE nre ierant eo sate sort eae ee roin see eade due De dea rada serena Ro pupa AVR 5 RCT PIER STRUCTURE eeror iceride isanne aknen AEREE ARTERS EEEE ARA EAEE EE EE REAA EEE AN E EEES 5 RT3DV2 5 USER DEFINED REACTION MODULE cscsssssscsssssssesssscsssssssscessesesesessesesssessessseserscsseesoessesees 8 EXAMPLE RANS F F IDB u E E OOE ENSO RON 8 EVALUATION OF JACOBIAN MATRICES vissiin siren ent i EEn aE E Erie eR E
18. Set the value to 2 to read 3D arrays in a cell by cell mode Note NCRXNDATA defines the number of reaction parameters that are constant and NVRXNDATA defines the number of reaction parameters that are spatially variable Enter E2a if IRCTOP 1 E2a ARRAY RHOB NLAY one value for each layer READER RARRAY RHOB is the bulk density of the porous medium ML 7 Units used for defining bulk density and adsorption coefficient values should be consistent to yield a dimensionless retardation parameter Enter E2b if IRCTOP 2 E2b ARRAY RHOB NCOL NROW one array for each layer READER RARRAY RHOB is the bulk density of the porous medium ML 7 Units used for defining bulk density and adsorption coefficient values should be consistent to yield a dimensionless retardation parameter Enter E3 data if ISOTHM gt 0 Enter in E3a format if IRCTOP 1 E3a ARRAY SPI NLAY one value for each layer Then repeat SP1 array entries for each mobile component READER RARRAY SP1 is the first sorption constant Enter in E3b format if IRCTOP 2 E3b ARRAY SPI NCOL NROW one array for each layer Then repeat SP1 arrays for each mobile component READER RARRAY SP1 is the first sorption constant For linear sorption SP1 is the distribution coefficient Ka L M Other details are described in the MT3DMS manual Enter E4 data if ISOTHM gt 0 Enter in E4a format if IRCTOP 1 E4a ARRAY SP2 NLAY one value for each layer Th
19. T3D version 1 0 only required MCOMP values which includes only mobiles species This change was required because of the new point source sink types which are also applicable to immobile species Note that the source sink concentrations of immobile species are used only when ITYPE 1 or 2 For all other ITYPE values the source sink concentrations of immobile species are ignored and may be dummy values RT3D version 1 0 SSM input files must be updated to the version 2 5 format to run with RT3Dv2 5 RT3D v2 5 Update Document SSM File Structure The data structure required for the SSM input file is identical to MT3DMS with the exceptions that 1 RT3Dv2 5 does not support a mass loading source ITYPE 15 2 RT3Dv2 5 has a decaying source ITYPE 2 which MT3DMS does not support and 3 the SSM input file structure for RT3Dv2 5 includes record D8a to support the decaying source RT3D2 5 D8 _ KSS ISS JSS CSS 1 ITYPE CSS 1 CSS 2 CSS 3 CSS 4 CSS ncomp to define NCOMP number of species concentrations Note that in a change from RT3D version 1 0 a value for all species must be specified not just for mobile species RT3Dv1 0 SSM input files must be updated to the version 2 5 format if immobile species are included in the model to run simulations with RT3Dv2 5 When ITYPE 1 or 2 positive values of CSS i are used as constant decaying source concentrations and negative values are indicators that those species should no
20. arts Also discussed are some changes in the way the user defined reaction module is compiled RT3DV2 5 SOURCE SINK SSM PACKAGE RT3Dv2 5 supports two new source sink types for describing time varying constant concentration source cells The first new type ITYPE 1 is a constant concentration source type similar to setting a constant concentration boundary condition i e by setting ICBUND 1 in the Basic Transport Package The SSM constant concentration differs from setting the boundary condition in that 1 the concentration may be different in each stress period and 2 for a multi species simulation concentrations need not be held constant for all species The second new source type is a decaying source The decaying source starts at a given initial concentration and decays at a first order rate according to Equation 1 C 7 C exp 4 0 1 where C is the concentration of species i at a given time t C o is the initial concentration of species i in the source and A is the source decay constant for species i The decaying source condition is defined by setting ITYPE 2 As an example this feature could be used to model a waste pit with a depleting NAPL source RT3Dv2 5 requires that the point source sinks have NCOMP number of concentration values which includes both mobile and immobile species for CSS in Record D8 Similarly the recharge and evapotranspiration arrays of records D4 and D6 each require NCOMP arrays R
21. cial dilution of these species For example if a TCE source is created using an injection well then the concentration of TCE biotransformation byproducts DCE and VC will be assumed to be zero in the injected water Maintaining the injection flow rate at a high value would lead to artificial local dilution of the DCE and VC plumes near the injection well Maintaining low flow rates will help minimize this dilution effect RT3D v2 5 Update Document RT3DV2 5 REACTION RCT PACKAGE RCT File Structure The revised reaction package input data structure for RT3Dv2 5 is described below RT3Dv2 5 is fully backward compatible with earlier versions of RCT file formats Note that the RT3D code does not support the dual porosity and NAPL options of MT3DMS However if required dual porosity and NAPL dissolution kinetics can be defined via a user defined reaction module see the dualpore pdf and napl pdf files respectively El Record ISOTHM IREACT NCRXNDATA NVRXNDATA ISOLVER IRCTOP Format 6110 ISOTHM l Linear adsorption isotherm is simulated 2 Freundlich adsorption isotherm is simulated 3 Langmuir adsorption isotherm is simulated 0 no sorption is simulated IREACT Reaction module number 0 no reaction is simulated 1 e tracer transport 1 Two Species Instantaneous Reactions BIOPLUME II type reactions 2 module reserved for future implementation 3 Six Species First Order Rate Limited BTEX Degradation using
22. en repeat SP2 array entries for each mobile component READER RARRAY RT3D v2 5 Update Document Enter in E4b format if IRCTOP 2 E4b ARRAY SP2 NCOL NROW one array for each layer Then repeat SP2 arrays for each mobile component READER RARRAY SP2 is the second sorption constant Other details are described in the MT3D manual ES ARRAYS atol ncomp rtol ncomp Format Free format two entries per line total number of entries should be equal to ncomp If ISOLVER 1 or 2 then the atol and rtol values specified for each species are used by the solvers If ISOLVER 3 then only the first pair of values atol 1 and rtol 1 is used by the solver If ISOLVER 4 or 5 then only the first absolute error tolerance value atol 1 is used by the solvers However to maintain a general structure for the input data atol and rtol must be specified for each species regardless of the choice of reaction solver For ISOLVER 1 or ISOLVER 2 i e the Gear based solvers relative rtol and absolute atol error tolerances are needed for the solver to determine how accurate the solution for the species concentration y i must be The differential equation solver will attempt to control the species concentration local error e i such that for each species the local error has approximately a smaller magnitude than the error weight ewt 1 leG ewt i where ewt i rtol i y i atol i Here ewt i is a vec
23. ent DLL file of the same name than the one you thought was going to be used USING MODFLOW 2000 The distribution of RT3Dv2 5 comes with MODFLOW 96 for generating a flow solution If the user wishes to employ MODFLOW 2000 two things must be done First compile MODFLOW 2000 and RT3Dv2 5 using the same compiler the executable for MODFLOW 2000 distributed by the U S Geological Survey http water usgs gov software modflow 2000 html is compiled with Lahey Fortran while the RT3Dv2 5 executable is compiled with Visual Fortran resulting in incompatibilities Second in the LMT6 input file you must specify the Standard header or use a blank file because RT3Dv2 5 does not support the Extended header see U S Geological Survey Open File Report 01 82 for more details ACKNOWLEDGMENTS The RT3D software has significant applications in areas ranging from risk assessment of groundwater pollution management of large scale groundwater plumes design of natural attenuation systems and design of active bioremediation systems Over the past nine years 16 RT3D v2 5 Update Document several researchers from various organizations have participated in the RT3D development project and the key personnel and their contributions are acknowledged below e Dr Prabhakar Clement Director and principal developer 1994 present e Dr Brian Hooker Project management and technical support 1994 1996 e Dr Rodney Skeen Technical support 1994 19
24. eyton collaborated with Prof Jim Petersen s group at Washington State University and used the 1 D code to model nutrient feeding strategies The primary objective of the collaborative work was to optimize the nutrient feeding strategy to obtain better distribution of microbial growth at locations away from the nutrient injection point thereby minimizing bioclogging at the injection well Details of this preliminary modeling work and the mathematical algorithms were published in Clement et al 1995 Clement et al 1996a Clement et al 1996b Franzen et al 1997 and Clement et al 1997 In the summer of 1995 Dr Clement modified the numerical algorithms and implemented the simple aerobic reaction modules of his 1 D code within the U S Environmental Protection Agency s U S EPA MT3D modelling framework This resulted in the first prototype version of the RT3D software Reactive Transport in 3 Dimensions Simulation results based on the prototype code were presented to several prospective clients including Parson Engineering Science Based on their review comments a detailed proposal was developed to standardize RT3D and to benchmark its capabilities at a national bioremediation test facility located at Dover Air Force Base Delaware The test facility was run by a joint team consisting of industry research organizations regulatory agencies and consulting companies e g U S DOE Dow Chemical Company DuPont U S Geological Survey U S
25. hristian Johnson for discussion of how Battelle may assist your project with customized reaction modules accelerated in situ remediation design and or natural attenuation evaluations Official RT3D Web Site http bioprocess pnl gov rt3d htm RT3D v2 5 Update Document RT3D Author RT3D Solution Developer Dr Prabhakar Clement Mr Christian Johnson Auburn University Battelle Pacific Northwest Division e mail clement eng auburn edu e mail rt3d pnl gov LICENSE AND COPYRIGHT INFORMATION Like any other literary work computer programs are protected by copyright Unauthorized reproduction or distribution of this computer code or any portion of it is not permitted Users can however modify the code and use it for solving a specific problem Users are not permitted to resell or redistribute a the RT3D code or a modified version thereof or any portion of RT3D directly or via computer bulletin boards web pages or other publicly accessible archives If you have modified the code for any research application and would like to share the code with other researchers please contact the developers first and let them know the details of your work In all your publications please cite the RT3D manual PNNL 11720 and the 1998 Groundwater Monitoring and Remediation journal paper by Clement et al references are listed in the RT3Dv25 Update document and at http bioprocess pnl gov rt3d_pubs htm Also consider citing other published RT3D papers that
26. ian for the PCE degration problem c Details of this routine are discussed in the v 2 5 update manual c SUBROUTINE jrxns ncomp nvrxndata j i k ml mu y pd amp poros rhob reta rc nlay nrow ncol vrc C Block Lk RK KKK KKK KKK KK KK ck ck k KK KK Ck ck kc kc KK KK ck kc kckck k kc kc kck ck ck ck kc kck ck k KK KK KK kc kc kc k kk c List of calling arguments c ncomp Total number of components c nvrxndata Number of variable reaction parameters to be input via RCT file c J I K node location used if reaction parameters are spatially variable c ml mu are for banded jacobian not used c y Concentration of all components at the node array variable y ncomp c pd jacobian matrix ncomp x ncomp array C poros porosity of the node c reta Retardation factor ignore dummy reta values of immobile species c rhob bulk density of the node c rc Stores spatially constant reaction parameters up to 100 parameters c nlay nrow ncol Grid size used only for dimensioning purposes c vrc Array variable that stores spatially variable reaction parameters C End of Block LR KK KKK KK KKK k ck k ck ck kc KKK ck kc k ck ck KKK KKK KK KK KK k ck KK KK KK KK KK ck ck ck kc k kk 13 RT3D v2 5 Update Document C Block QolOECKCkCk KKK KKK KK KK KKK ck ck KK KK k kc kc kck ck ck kc KKK KK KR KK k kc kc kck ck KK KKK ck ck KK KK ck ck ck kk GF Please do not modify this standard interface block IMPLICIT
27. igned between Battelle and the University of Alabama to revise the prototype version to include all MODFLOW source sink packages i e river recharge drain etc and to standardize the code with a new version of MT3D Dr Clement and Dr Zheng jointly worked at the University of Alabama and developed a multi species version of the MT3D DoD 1 5 code Dr Clement later ported all reaction modules into this multi species code The first stand alone beta versions of RT3D v 1 0 and the batch reaction utility BATCHRXN were developed in May 1997 and the details of these codes were documented in Clement 1997 In 1998 the RT3D software was transferred to several industrial partners as a public domain code which facilitated integration of RT3D with commercial graphical user interfaces Dr Clement and Mr Johnson managed key collaborative research efforts between PNNL and Brigham Young University s Environmental Modeling Research Laboratory EMRL the developers of GMS to standardize the GMS interface for RT3D A detailed user document was developed to facilitate the use RT3D within the GMS 2 1 environment Clement and Jones 1998 More recently Dr Clement received a grant from the Australian Research Council with some additional industrial support from Battelle to continue the development work related to the RT3D software As a part of this project a new version of RT3D v 2 5 was released in November 2001 RT3D v 2 5 is now a standard Fortran
28. inishing advection solver and an iterative generalized conjugate gradient solver The RT3Dv2 5 code now uses standard Fortran 90 which should allow seamless use across multiple platforms RT3Dv2 5 has new source sink options of constant concentration and decaying sources Four new reaction solver options have been added including Runge Kutta solvers and solvers using an explicit Jacobian matrix for stiff problems RT3D v2 5 Update Document The input data structures of the BTN ADV and DSP files basic advection and dispersion packages respectively are identical to those described for MT3DMS version 3 5 and the reader is referred to the MT3DMS manual Zheng and Wang 1999 for those packages The input data for the source sink package SSM file is identical to the MT3DMS source sink package with the exception that the RT3Dv2 5 SSM file supports an additional boundary condition designated as the decaying constant source boundary condition The input data structure for the reaction package RCT file is similar to the data structure used in RT3D v 1 0 but the version 2 5 reaction package provides four new solver options and the sorption parameters can now be specified on a cell by cell basis in addition to the layer by layer basis This document discusses the changes made to the source sink and reaction package input file structure to incorporate the new features of RT3Dv2 5 and notes how the SSM and RCT files differ from the MT3DMS counterp
29. nant Hydrology 24 269 285 Clement T P 1997 RT3D A Modular Computer Code for Simulating Reactive Multi Species Transport in 3 Dimensional Groundwater Aquifers Pacific Northwest National Laboratory Richland WA USA PNNL 11720 Found online at http bioprocess pnl gov rt3d htm Clement T P B M Peyton R S Skeen B S Hooker J M Petersen and D Jennings 1997 Microbial Growth and Transport in Porous Media Under Denitrification Conditions Experiment and Simulations Results Journal of Contaminant Hydrology 24 269 285 Clement T P B S Hooker and R S Skeen 1996b Macroscopic Models for Predicting Changes in Saturated Porous Media Properties Cause by Microbial Growth Ground Water 34 5 934 942 Clement T P B S Hooker and R S Skeen 1996a Numerical Modeling of Biologically Reactive Transport From a Nutrient Injection Well ASCE Journal of Environmental Engineering 122 9 833 839 Clement T P B S Hooker and R S Skeen 1995 Modeling Biologically Reactive Transport in Porous Media in Proceedings of the International Conference on Mathematics and Computations Reactor Physics and Environmental Analyses Portland Oregon April May 1995 Vol 1 pp 192 201 19 RT3D v2 5 Update Document Franzen M F L J M Petersen T P Clement B S Hooker and R S Skeen 1997 Pulsing as a Strategy To Achieve Large Biologically Active Zones During In Situ Carbon Tetrachloride Remediation Computational Ge
30. osciences 1 3 4 217 288 Lu G T P Clement C Zheng and T H Wiedemeier 1999 Natural Attenuation of BTEX Compounds Model Development and Field Scale Application Ground Water 37 5 707 717 Peyton B M T P Clement and J P Connolly 2000 Modeling of Natural Remediation Contaminant Fate and Transport Chapter 5 in Natural Remediation of Environmental Contaminants M Swindoll R G Stahl and S J Ells eds Society for Environmental and Toxicology and Chemistry Pensacola Florida pp 79 120 Press W H S A Teukolsky W T Vetterling and B P Flannery 1992 Numerical Recipes in Fortran 77 The Art of Scientific Computing Cambridge University Press New York Sun Y J N Petersen T P Clement and R S Skeen 1999 Development of Analytical Solutions for Multi Species Transport with Serial and Parallel Reactions Water Resources Research 35 1 185 190 Sun Y J N Petersen J Bear T P Clement and B S Hooker 1999 Modeling Microbial Transport and Biodegradation in a Dual Porosity System Transport in Porous Media 35 1 49 65 Sun Y and T P Clement 1999 A Generalized Decomposition Method for Solving Coupled Multi Species Reactive Transport Problems Transport in Porous Media 37 3 327 346 Sun Y J N Petersen T P Clement and B S Hooker 1998 Effect of Reaction Kinetics on Predicted Concentration Profiles During Subsurface Bioremediation Journal of Contaminant Hydrology 31 359 372
31. standard in version 1 0 the following modifications to the source code must be made These modifications assume that the compiler is Compaq Visual Fortran Note that the compiler directive may take the form of IDEC or MSS the latter of which is for Microsoft Fortran Powerstation Visual Fortran appears to accept the MSS declaration although this form may not be supported in future versions e Uncomment line 53 of rt3dv25 f USE DFLIB e Uncomment line 54 of rt3dv25 f USE DFPORT e Uncomment line 1428 of rteqnv25 f DEC ATTRIBUTES DLLIMPORT rxns e Uncomment line 1445 of rteqnv25 f DEC ATTRIBUTES DLLIMPORT jrxns e Uncomment line 24 of rxns f DECS ATTRIBUTES DLLEXPORT rxns e Uncomment line 26 of jacrxns f DECS ATTRIBUTES DLLEXPORT jrxns In the DLL version the subroutines rxns f and jacrxns f should be compiled first as standalone DLLs e g the Visual Fortran command is df dll rxns f The RT3Dv2 5 executable can then be compiled and linked against the rxns lib and jacrxns lib files that were created along with the DLLs The DLL files rxns dll and jacrxns dll should be copied moved to the same directory as the RT3Dv2 5 executable or to a location in the file search path If the DLL file is not present in the same directory as the RT3Dv2 5 executable Windows will search for the file in the directories in the search path Confusion may arise if Windows finds a differ
32. t have their concentrations held constant Format Free Format The following record D8a is required only if the value of ITYPE 2 RT3D2 5 D8a Source decay 1 Source decay 2 Source decay 3 Source decay ncomp used to define ncomp number of source decay constants The values should be positive if the source is decreasing with time and negative if it is increasing with time Specify decay value as zero for the species that were assigned negative concentration values in the previous record Also zero can be used for non decaying species Format Free Format Note that constant or non decaying concentration boundary conditions can be simulated by either using the ITYPE 1 option or the ITYPE 2 option and setting the corresponding source decay constants to zero COMMENT ON USING ARTIFICIAL INJECTION WELLS TO MODEL SOURCES Injection wells are often used to simulate contaminant release mass discharge from a source into an aquifer This is a reasonable approximation for single species transport if the amount of water injected into the aquifer does not alter the flow field However for multi species simulations additional care should be exercised to maintain the injection rate at a very low level At discharge points 1 e at the hypothetical injection wells the concentrations of species other than the primary contaminant are usually assumed to be zero and an injection flow rate that is too high may introduce artifi
33. tor of weights which must always be positive and the values of rtol and atol should all be non negative Setting atol i 0 0 results in a pure relative error test on that component species Setting rtol i 0 0 results in a pure absolute error test on that component A mixed test with non zero rtol i and atol 1 corresponds roughly to a relative error test when the solution component y i is much bigger than atol and to an absolute error test when the solution component is smaller than the threshold atol For practical problems the following rules of thumb may be used to set atol and rtol values The atol i should be set to a tolerable absolute error level selected by the user around 10 to 10 Defining m as the number of significant digits required in the solution for species y i set rtol i 10 Caution Actual global errors may exceed the local tolerances so choose atol i and rtol i conservatively For ISOLVER 4 or ISOLVER 5 one global tolerance value known as the error residue eps specified as atol 1 is used to control convergence For further details about error residue RT3D v2 5 Update Document refer to Numerical Recipes in Fortran 77 Press et al 1992 p 711 For practical problems set eps atol 1 to a small value between 10 to 10 Enter E6 if IREACT gt 0 Input all the constant reaction parameter values number of entries must be equal to NCRXNDATA the values will be stored in a variable
34. with respect to each of the four species These partial derivatives are shown in Equations 11 to 14 K ex Pce 11 pd 1 1 P ae ax OPCE dt 11 RT3D v2 5 Update Document OF 0 alPCE _ PS cni dt LE OK 0 _ d PCE _ MOT aE di 59 OF 0 d PCEY aud o ari dt j on Similarly elements of the second third and fourth rows of the Jacobian matrix are the partial derivatives of the second third and fourth reaction equations respectively with respect to the four species Equations 15 to 26 SUE x x cH ES as o gt Se 5 menta l x che pie x Em B ny pito A S mu ue Beene x xe iren Om aoa 7 S S B mE Ven Pus 3 7 TE Grose 21 oe a see 122 i4 lt oe 29 o 23 EROS T TE RT ma RT3D v2 5 Update Document OF 0 dVC Yy 4 K d 4 3 4 Vc Dce Dce Bae al dt Ry 29 OF 8 d VC K d 4 4 MONS xvcl di Ry 28 The Jacobian matrix for the system of Equations 7 to 10 is then formulated as shown in Equation 27 K Pce R 0 0 0 Yu Pce K Pce I Cx K R ie 0 0 R T J 27 Yie Tce K Tce pee a K 0 R neie 0 R D 0 0 Y voie K nce Ky R V R V EXAMPLE JACRXNS F FILE A jacrxns f file that uses the Jacobian matrix for the example above Equation 27 is listed below The Jacobian matrix equations are coded in Block 6 of the subroutine c Evaluation of Jacob
35. x ce Tce 6 VC _ 9 AVC AviL VC R u ot a i Ox Oxi 1 VC Yveipee amp Deel DCE K v VC where R is the retardation factor unitless n is the aqueous phase concentration of the n specie ML Dy is the hydrodynamic dispersion coefficient L T 5 v is the pore water velocity LT 5 gs is the volumetric flux of water per unit volume of aquifer that represents sources and sinks T is the aquifer porosity unitless n is the concentration of the n specie in the sources sinks ML Yap represents the stoichiometric yield i e the amount of daughter specie d produced by degrading a unit mass of parent specie p MM and K represents the first order degradation rate constant of the n specie T Note that all reactions are assumed to occur only in the aqueous phase which is a conservative assumption Using the operator split strategy Clement et al 1998 the reaction kinetics can be separated from the transport equations and written as a set of ordinary differential equations given by Equations 7 to 10 _d PCE __ K p PCE F I di R 7 d TCE Vice Pook pel PCE K rel TCE F 8 t R d DCE Y eere K Tee TCE K bel DCE F 7 9 t R F d VC Y DeK peel DCE u Ky VC 10 dt R Elements of the first row of the Jacobian matrix represented as pd 1 1 pd 1 2 etc in the code are the partial derivatives of the first reaction equation
Download Pdf Manuals
Related Search